`
`
`
`
`
`
`
`
`Paper No. _
` Filed: March 30, 2021
`
`
`
`UNITED STATES PATENT AND TRADEMARK OFFICE
`_______________
`
`BEFORE THE PATENT TRIAL AND APPEAL BOARD
`_______________
`
` ILLUMINA, INC.,
`Petitioner,
`
`v.
`
` THE TRUSTEES OF COLUMBIA UNIVERSITY
`IN THE CITY OF NEW YORK
`Patent Owner.
`_______________
`
`Case IPR2020-00988 (U.S. Patent 10,407,458)
`Case IPR2020-01065 (U.S. Patent 10,407,459)
`Case IPR2020-01177 (U.S. Patent 10,435,742)
` Case IPR2020-01125 (U.S. Patent 10,457,984)1
`
`
`
`PATENT OWNER’S RESPONSE
`
`
`
`1 An identical Paper is being entered into each listed proceeding.
`
`
`
`
`
`
`
`
`
`
`
`Table of Contents
`
`
`Introduction ...................................................................................................... 1
`
`I.
`
`II. Level of Ordinary Skill in the Art ................................................................... 5
`
`III. Claim Construction .......................................................................................... 5
`
`IV. Illumina’s Ground 1 and 2 Challenges Fail ..................................................... 6
`
`A. There Was No Expectation Of Success In Achieving The Claimed
`Invention .......................................................................................................... 8
`
`1. Illumina’s arguments are contradicted by Illumina’s positions in
`the European Opposition ............................................................................ 9
`
`2. Polymerase incorporation of 3’-O-capped nucleotides was believed
`to be rare and unpredictable .....................................................................13
`
`3. Hovinen and Kwiatkowski provided no expectation of
`incorporation for the MOM embodiment .................................................22
`
`4. There was no expectation of incorporation for the claimed adenine
`nucleotides ................................................................................................32
`
`B. There Was No Motivation To Arrive At The Claimed Invention .................33
`
`1. There was no interest in the MOM capping group for SBS .....................34
`
`2. Illumina’s Hiatt theory fails......................................................................37
`
`3. A POSA would not have been motivated to use the MOM capping
`group because there was no expectation that it would be efficient
`enough to sequence twenty base pairs ......................................................57
`
`V.
`
`Illumina’s Ground 3 and 4 Challenges Fail ...................................................70
`
`VI. Conclusion .....................................................................................................71
`
`i
`
`
`
`
`
`CASES
`
`Table of Authorities
`
`Cuozzo Speed Techs., LLC v. Lee,
`136 S. Ct. 2131 (2016)..........................................................................................64
`
`
`Intelligent Bio-Sys., Inc. v. Illumina Cambridge Ltd.,
`821 F.3d 1359 (Fed. Cir. 2016) ............................................................................64
`
`
`Otsuka Pharm. Co. v. Sandoz, Inc.,
`678 F.3d 1280 (Fed. Cir. 2012) ..................................................................... 20, 44
`
`OTHER AUTHORITIES
`
`35 U.S.C. § 311(b) ...................................................................................................70
`
`
`
`ii
`
`
`
`Case IPR2020-00988, -01065, -01125, -01177
`U.S. Patent Nos. 10,407,458; 10,407,459; 10,435,742; 10,457,984
`
`I.
`
`INTRODUCTION
`
`In the present proceedings, the claims of the patents-at-issue are narrowly
`
`tailored to a small genus of modified nucleotides. Illumina’s Grounds 1 and 2 allege
`
`that a single embodiment of the genus is invalid for obviousness, specifically, the
`
`“MOM embodiment.” The MOM embodiment is a nucleotide modified to have a
`
`MOM capping group attached to the 3’ oxygen of the nucleotide sugar and a
`
`linker/label attached to the nucleotide base:
`
`Ex. 1001 at claim 1 (relevant modifications highlighted).
`
`
`
`Notably, the claims of the patents-at-issue do not cover the allyl embodiment
`
`that was previously the focus of the Allyl Claim IPRs.2 In the Allyl Claim IPRs, two
`
`
`2 Columbia’s SBS patents with the allyl embodiment are U.S. Patent Nos.
`
`9,718,852; 9,719,139; 9,708,358; 9,725,480; and 9,868,985. They were challenged
`
`in IPR2018-00291, IPR2018-00318, IPR2018-00322, IPR2018-00385, and
`
`IPR2018-00797, respectively (collectively, “the Allyl Claim IPRs”).
`
`
`
`1
`
`
`
`Case IPR2020-00988, -01065, -01125, -01177
`U.S. Patent Nos. 10,407,458; 10,407,459; 10,435,742; 10,457,984
`
`findings led a 2-1 majority to opine that the allyl embodiment would have been
`
`obvious to a POSA:
`
`1. Metzker provided experimental data demonstrating some polymerase
`incorporation of the allyl capping group; and
`
`2.
`
`Tsien prominently disclosed using the allyl capping group for
`Sequencing by Synthesis (“SBS”).
`
`Here, no prior art demonstrated that the MOM capping group could be incorporated
`
`by any polymerase and no prior art suggested the use of the MOM capping group
`
`for SBS. Indeed, it is undisputed that the Columbia inventors were the first to
`
`disclose using the MOM capping group for SBS, an insight they arrived at after
`
`discovering the particular chemical and structural features that dictate whether a
`
`capping group could be useful for SBS, and which the MOM capping group satisfies.
`
`Just last week, in a European proceeding involving a corresponding European
`
`patent, Illumina admitted that a POSA would not have reasonably expected that the
`
`MOM embodiment could be successfully used in SBS, concluding that “[i]t was not
`
`plausible at the effective date [of October 6, 2000] that [Columbia’s SBS] invention
`
`could be put into practice with MOM as a protective group based on the common
`
`general knowledge in the prior art.” As detailed herein, that admission is one of
`
`Illumina’s central positions in the European proceeding, where Illumina alleges
`
`invalidity of Columbia’s European patent based on the contention that a POSA
`
`would not have been able to use the MOM capping group to practice SBS as of the
`
`
`
`2
`
`
`
`Case IPR2020-00988, -01065, -01125, -01177
`U.S. Patent Nos. 10,407,458; 10,407,459; 10,435,742; 10,457,984
`
`priority date. While Illumina’s argument in the European proceeding pertains to the
`
`equivalent of enablement, the argument directly conflicts with Illumina’s position in
`
`this case concerning a POSA’s understanding prior to Columbia’s filing date.3
`
`Illumina’s European positions confirm the untenability of Illumina’s Petitions
`
`here for at least two reasons. First, Columbia’s claims require that the claimed
`
`nucleotides be capable of being “incorporated” into a growing DNA strand by a
`
`polymerase, and Illumina argues here that a POSA would have expected success
`
`with MOM—the exact opposite of Illumina’s position in Europe. Before
`
`Columbia’s discovery of the requirements for a useful capping group, a POSA would
`
`have known that there was a high failure rate for achieving polymerase incorporation
`
`with 3’-O-capped nucleotides. In a seminal study, Metzker in 1994 showed an 85%
`
`failure rate, and no more than a few 3’-O-capped nucleotides were believed to be
`
`incorporated, including the 2-nitrobenzyl capping group. Given this high failure rate
`
`and the complete absence of any prior art data demonstrating MOM could be
`
`incorporated by a polymerase, a POSA prior to Columbia’s invention would have
`
`had no reason to expect success with a MOM capped nucleotide.
`
`
`3 Columbia’s position in the European proceeding and here is that the Columbia
`
`patents provide the necessary guidance on the chemical and structural requirements
`
`needed for a capping group to be efficiently incorporated and useful for SBS.
`
`
`
`3
`
`
`
`Case IPR2020-00988, -01065, -01125, -01177
`U.S. Patent Nos. 10,407,458; 10,407,459; 10,435,742; 10,457,984
`
`Illumina theorizes that a POSA would have extrapolated from data available
`
`for different modified nucleotides, specifically those disclosed in Hovinen and
`
`Kwiatkowski, to form an expectation of success with a MOM capped nucleotide.
`
`The problem with Illumina’s theory is that Metzker and Canard resoundingly reject
`
`it by demonstrating that such extrapolations are invalid. Illumina’s reliance on
`
`speculative theories that are contradicted by actual data reveals the weakness of
`
`Illumina’s hindsight-driven position. Illumina has failed to meet its burden of
`
`demonstrating that a POSA would have expected that the MOM capped nucleotide
`
`could be incorporated by a polymerase.
`
`Illumina’s Petitions fail for a second reason. Illumina’s Petitions hinge on the
`
`contention that a POSA would have been motivated to use the MOM capping group
`
`in Tsien’s or Dower’s SBS technology. To support its contention, Illumina relies on
`
`Hiatt’s non-SBS disclosure of a MOM capping group, but MOM is one of hundreds
`
`of capping groups disclosed in Hiatt, and a POSA would have had no reason to select
`
`MOM from that universe of possibilities, particularly given that Hiatt has no data
`
`showing that a MOM capping group could be incorporated by a polymerase.
`
`Additionally, Hiatt’s technology requires using template-independent, zero fidelity
`
`polymerases, which cannot be used for SBS or even for Sanger sequencing. SBS
`
`requires using template-dependent, high fidelity polymerases, which interact with
`
`DNA and nucleotide substrates differently than those used in Hiatt. A POSA would
`
`
`
`4
`
`
`
`Case IPR2020-00988, -01065, -01125, -01177
`U.S. Patent Nos. 10,407,458; 10,407,459; 10,435,742; 10,457,984
`
`not have been motivated to select any of Hiatt’s modified nucleotides for SBS given
`
`the lack of relatedness of these different technologies.
`
`A POSA would neither reasonably expect the MOM embodiment to be
`
`successfully incorporated by a polymerase, nor would he or she be motivated by
`
`Illumina’s prior art to select the MOM capping group for use in Tsien’s or Dower’s
`
`SBS methods. Therefore, Illumina’s Grounds of challenge fail and Columbia’s
`
`claims should be upheld.
`
`II. LEVEL OF ORDINARY SKILL IN THE ART
`
`For the purposes of these IPR proceedings, Columbia does not dispute
`
`Illumina’s criteria for defining a POSA.4
`
`III. CLAIM CONSTRUCTION
`
`The term “small” refers to the ability of the capping group to fit into the active
`
`site of the polymerase whose three-dimensional structure is shown in Figure 1 of the
`
`patents-at-issue. More specifically, “small” means the group has a diameter less
`
`than 3.7Å. This construction is based on the specification of the patents-at-issue.
`
`Exs. 1001-1004, each at 2:63-3:54, 5:52-59, Fig. 1, 7:51-8:28. As explained during
`
`
`4 Columbia’s position herein relates to a POSA prior to Columbia’s invention, i.e.,
`
`a POSA without the benefit of the Columbia inventors’ insights set forth in the
`
`specification of the Columbia patents-at-issue.
`
`
`
`5
`
`
`
`Case IPR2020-00988, -01065, -01125, -01177
`U.S. Patent Nos. 10,407,458; 10,407,459; 10,435,742; 10,457,984
`
`prosecution of Columbia’s patents, “[a]s of October 6, 2000, the [POSA] reading the
`
`specification would have understood that ‘small’ referred to the ability to fit into the
`
`active site of the polymerase defined by reference to the three-dimensional structure
`
`shown in FIG. 1,” and therefore capping group R must have “a diameter less than
`
`3.7Å so that it would fit into the active site of the polymerase.” Ex. 1036 at 4-5.
`
`The Examiner adopted this meaning. See, e.g., Ex. 2026 at 12 (“The declaration of
`
`Jingyue Ju . . . is sufficient to explain what is meant by ‘small.’”).
`
`IV. ILLUMINA’S GROUND 1 AND 2 CHALLENGES FAIL
`
`Columbia’s claims cover a limited genus of modified guanine, adenine,
`
`cytosine, and thymine nucleotides. See Ex. 1001 at claim 1 (guanine version of
`
`Columbia’s nucleotides); Ex. 1002 at claim 1 (adenine version); Ex. 1003 at claim 1
`
`(cytosine version); Ex. 1004 at claim 1 (thymine version). The relevant
`
`modifications to the claimed nucleotides are highlighted below:
`
`
`Moreover, the claims recite structural, chemical, and functional limitations for the R
`
`group, such as “a small, chemically cleavable, chemical group,” “does not interfere
`
`with recognition of the analogue as a substrate by a DNA polymerase,” “is stable
`
`
`
`6
`
`
`
`Case IPR2020-00988, -01065, -01125, -01177
`U.S. Patent Nos. 10,407,458; 10,407,459; 10,435,742; 10,457,984
`
`during a DNA polymerase reaction,” “does not contain a ketone group,” “is not a —
`
`CH2CH═CH2 group” (i.e., the allyl group) and “wherein OR is not a methoxy group
`
`or an ester group.” See Ex. 1001 at claim 1.5
`
`In addition, Columbia’s claims cover methods of using these same modified
`
`nucleotides to practice SBS. Specifically, the claims of U.S. Patent No. 10,428,380
`
`cover “method[s] for sequencing a nucleic acid[.]” Ex. 1005 at claims 1-4. The ’380
`
`patent is the subject of IPR2020-01323, for which Columbia’s Patent Owner’s
`
`Response is being concurrently filed with this Patent Owner’s Response.
`
`Illumina’s Grounds 1 and 2 rely on a single R group to meet all of these
`
`limitations, arguing that it would have been obvious to make an embodiment of
`
`Columbia’s claims wherein R is a methoxymethyl (“MOM” or -CH2OCH3) capping
`
`group, as shown below:
`
`
`5 In the Allyl Claim IPRs, the Board opined on the meaning of “Y” in the claims.
`
`See, e.g., Ex. 1024 at 53 n.33. Columbia agrees with the Board’s conclusion that
`
`the claims do not “exclude[] a linker attached to a propargyl amine[.]”
`
`
`
`7
`
`
`
`Case IPR2020-00988, -01065, -01125, -01177
`U.S. Patent Nos. 10,407,458; 10,407,459; 10,435,742; 10,457,984
`
`
`
`To arrive at the MOM embodiment, Illumina combines Hiatt (which discloses
`
`using MOM as a capping group in a non-SBS context) with Tsien or Dower (which
`
`allegedly disclose the remaining features of the claim, including Y and Tag). As
`
`explained below, Grounds 1 and 2 fail because (A) there was no expectation of
`
`success in achieving the claimed invention, and (B) there was no motivation to make
`
`the claimed invention.
`
`A.
`
`There Was No Expectation Of Success In Achieving The
`Claimed Invention
`The claims require that the modified nucleotides be capable of being
`
`“incorporated” by a polymerase into a growing strand of DNA in a reaction that
`
`requires a DNA primer strand and a separate template strand as well as a nucleotide
`
`analogue.
`
`
`
`See
`
`Ex.
`
`1001-1004
`
`at
`
`claim
`
`1
`
`(“wherein
`
`the
`
`[guanine/adenine/thymine/cytosine] deoxyribonucleotide analogue:
`
`.
`
`.
`
`.
`
`is
`
`incorporated at the end of a growing strand of DNA during a DNA polymerase
`
`reaction”).
`
`
`
`8
`
`
`
`Case IPR2020-00988, -01065, -01125, -01177
`U.S. Patent Nos. 10,407,458; 10,407,459; 10,435,742; 10,457,984
`
`Thus, it was Illumina’s burden to demonstrate that a POSA would have
`
`reasonably expected that the MOM embodiment could be incorporated at the end of
`
`a growing strand by a polymerase. See, e.g., Ex. 1024 at 55-57 (Board in Allyl Claim
`
`IPRs assessing whether a POSA would have had “a reasonable expectation of
`
`success in reaching the ‘incorporated at the end of a growing strand of DNA during
`
`a DNA polymerase reaction’ limitation”). While Illumina contends that a POSA
`
`would have “reasonably expected the claimed nucleotide analogue to meet each of
`
`the functional limitations recited” in the claims, Illumina has not and cannot meet its
`
`burden because a POSA would have had no such expectation.
`
`1.
`
`Illumina’s arguments are contradicted by Illumina’s
`positions in the European Opposition
`
`In a related European proceeding, Illumina has admitted a POSA would not
`
`have reasonably expected that the MOM embodiment could be successfully used in
`
`SBS. Columbia owns EP3034627, which has priority of the same application as the
`
`patents-at-issue here (i.e., 09/684,670, filed October 6, 2000), and claims SBS
`
`methods using, inter alia, the MOM capping group. See Ex. 2074. Illumina filed
`
`its Opposition against Columbia’s European patent on October 29, 2019 (Ex. 2075),
`
`and on March 25, 2021, filed a Response (Ex. 2076) to a preliminary opinion of the
`
`opposition division favorable to Columbia (Ex. 2079).
`
`In its Response, Illumina has taken the position that Columbia’s EP’627
`
`patent should be revoked on the basis that a POSA would have been unable to
`
`
`
`9
`
`
`
`Case IPR2020-00988, -01065, -01125, -01177
`U.S. Patent Nos. 10,407,458; 10,407,459; 10,435,742; 10,457,984
`
`practice the claimed SBS methods using the MOM capping group. See, e.g., Ex.
`
`2076 at 20-23, 37-38. Illumina’s Petitions here directly conflict with positions that
`
`Illumina is taking in the ongoing European proceeding. According to Illumina,
`
`nothing in the field (or in Columbia’s patent) would have provided a POSA with
`
`information needed to successfully pursue SBS using the MOM capping group. For
`
`example, Illumina just argued:
`
`[I]t was thus not plausible at the effective date of EP'627
`
`that the alleged invention could be put into practice with
`
`MOM as protective group based on the common general
`
`knowledge in the prior art.
`
`Ex. 2076 at 38. Illumina similarly argued:
`
`In this section we set forth that an SBS method cannot be
`
`put into practice when using allyl or MOM as 3'-OH
`
`protective groups by following the instructions of the
`
`contested patent in view of
`
`the common general
`
`knowledge.
`
`Ex. 2076 at 20 (emphasis in original).
`
`Illumina’s Petitions here should be rejected because Illumina’s argument in
`
`Europe is correct in one important respect: that, before the guidance in Columbia’s
`
`
`
`10
`
`
`
`Case IPR2020-00988, -01065, -01125, -01177
`U.S. Patent Nos. 10,407,458; 10,407,459; 10,435,742; 10,457,984
`
`patents, a POSA would not have expected from the teachings in the prior art that the
`
`MOM capping group could be used for SBS.6
`
`The credibility of Illumina’s present positions is further eroded by Illumina’s
`
`positions regarding the allyl embodiment. In 2019, Illumina prevailed in the Allyl
`
`Claim IPRs by convincing the Board that SBS-compatible cleavage conditions for
`
`the allyl were known in the prior art. See, e.g., Ex. 2014 (IPR2018-00291
`
`Petitioner’s Reply) at 20 (“The prior art discloses SBS-compatible cleavage
`
`conditions [for allyl]”), 21 (“[B]y at least the mid-1990’s, the state of the art included
`
`appropriate allyl cleavage conditions.”). In its present Petitions, Illumina’s Grounds
`
`3 and 4 hinge on that same argument, i.e., Illumina continues to argue that “DNA-
`
`compatible cleavage conditions for the allyl group were known.” IPR2020-00988
`
`Petition (Paper 1) at 68, 69. Currently, in the European proceedings, Illumina is
`
`taking the exact opposite position:
`
`SBS-compatible cleavage conditions for allyl were not
`
`part of the common general knowledge.
`
`
`6 Illumina’s European position is incorrect that Columbia’s specification does not
`
`teach that the MOM capping group could be used for SBS. As explained herein,
`
`the teachings in the patent specification enlighten a POSA as to the fact that the
`
`MOM embodiment could be used to practice SBS without undue experimentation,
`
`and why.
`
`
`
`11
`
`
`
`Case IPR2020-00988, -01065, -01125, -01177
`U.S. Patent Nos. 10,407,458; 10,407,459; 10,435,742; 10,457,984
`Ex. 2076 at 32.7 And:
`
`Cleavage conditions for allyl which would be compatible
`
`with SBS were also not part of the common general
`
`knowledge.
`
`Ex. 2076 at 11. And:
`
`At the filing date of the patent, a skilled person could
`
`thus not practice the claimed invention without an undue
`
`burden at least as far as allyl is concerned . . . because
`
`suitable cleavage conditions were missing.
`
`Ex. 2076 at 11. And:
`
`We will further show that the skilled person could not
`
`rely on the common general knowledge to identify
`
`suitable cleavage conditions when using allyl as the 3’-
`
`OH protective group.
`
`Ex. 2076 at 21.
`
`Having argued successfully to the Board in the Allyl Claim IPRs that the prior
`
`art disclosed SBS-compatible cleavage conditions for the allyl, and continuing to
`
`rely on that same argument in Grounds 3 and 4 of its present Petitions, it is
`
`concerning that Illumina is now taking the opposite position in the European
`
`proceeding without disclosing to the Board it is doing so.
`
`
`7 Emphases added unless otherwise indicated.
`
`
`
`12
`
`
`
`Case IPR2020-00988, -01065, -01125, -01177
`U.S. Patent Nos. 10,407,458; 10,407,459; 10,435,742; 10,457,984
`
`2.
`
`Polymerase incorporation of 3’-O-capped nucleotides was
`believed to be rare and unpredictable
`
`In trying to identify capping groups that might work for SBS, the prior art
`
`suggested
`
`that achieving any polymerase
`
`incorporation with 3’-O-capped
`
`nucleotides was unpredictable and that there was a high failure rate. This is not
`
`surprising given that DNA polymerases evolved precise mechanisms for identifying
`
`and incorporating natural nucleotides into strands of DNA, meaning that they have
`
`evolved to reject non-natural (i.e., modified) nucleotides such as the MOM
`
`embodiment. Ex. 2071 at ¶¶ 29-30. As of the priority date, only a few 3’-O-capped
`
`nucleotides had been reported to be incorporated to any degree, e.g., the 2-
`
`nitrobenzyl and methyl capping groups. Ex. 2071 at ¶¶ 29-30; see Ex. 1039 at 4263
`
`(Table 2) (reporting incorporation of 2-nitrobenzyl and methyl capping groups).
`
`In a study examining incorporation of 3’-O-capped nucleotides by DNA
`
`polymerases for SBS, Metzker 1994 conducted sixty-four experiments to determine
`
`whether eight polymerases could incorporate any of eight different 3’-O-capped
`
`nucleotides.
`
` Ex. 1039 at 4263
`
`(Table 2).
`
` Out of
`
`the sixty-four
`
`polymerase/nucleotide combinations, there was some level of incorporation
`
`(reported as Termination or Termination*) for only ten, and incorporation was
`
`reported to be efficient (reported as Termination) for only six of the sixty-four:
`
`
`
`13
`
`
`
`Case IPR2020-00988, -01065, -01125, -01177
`U.S. Patent Nos. 10,407,458; 10,407,459; 10,435,742; 10,457,984
`
`
`Ex. 1039 at 4263 (Table 2) (highlighting added, red indicates no incorporation, green
`
`indicates successful incorporation, yellow indicates incomplete incorporation); Ex.
`
`2071 at ¶¶ 31-35; Ex. 2072 at 231:6-232:15 (“Inhibition” indicates no
`
`incorporation). Stated as percentages, Metzker reported that only 15% of the
`
`polymerase/nucleotide combinations achieved any level of incorporation. Ex. 2071
`
`at ¶¶ 31-35. That amounts to an 85% failure rate. Id. Researchers in the SBS field
`
`cited Metzker as evidence that 3’-O-capped nucleotides are poorly incorporated by
`
`polymerase. E.g., Ex. 2022 at 363.
`
`Metzker’s results also suggest unpredictability. Ex. 2071 at ¶¶ 36-38.
`
`Metzker explicitly taught that “no consistent activity pattern between the
`
`polymerases and the analogs could be discerned.” Ex. 1039 at 4265. As one
`
`example, whereas AMV-RT polymerase incorporated an adenine nucleotide with a
`
`
`
`14
`
`
`
`Case IPR2020-00988, -01065, -01125, -01177
`U.S. Patent Nos. 10,407,458; 10,407,459; 10,435,742; 10,457,984
`
`methyl capping group (i.e., a 3’-O-methyl adenine nucleotide), it was incapable of
`
`incorporating a thymine nucleotide with the same exact capping group:
`
`
`Ex. 1039 at 4263 (Table 2); Ex. 2071 at ¶¶ 36-38. And the opposite was true for the
`
`Bst and AmpliTaq polymerases, which incorporated thymine nucleotides with a
`
`methyl capping group, but were incapable of incorporating adenine nucleotides with
`
`the same capping group:
`
`
`
`15
`
`
`
`Case IPR2020-00988, -01065, -01125, -01177
`U.S. Patent Nos. 10,407,458; 10,407,459; 10,435,742; 10,457,984
`
`Ex. 1039 at 4263 (Table 2); Ex. 2071 at ¶¶ 36-38.
`
`
`
`Consistent with this unpredictability, Dower repeatedly disclosed the need to
`
`conduct “screening” assays to determine whether a polymerase could incorporate
`
`any given 3’-O-capped nucleotides. Ex. 1030 at 18:21-32 (“Screening will be
`
`performed
`
`to determine appropriate polymerase and monomer analog
`
`combinations”), 17:43-48 (“Each of these analogs can be tested for compatibility
`
`with a particular polymerase by testing whether such polymerase is capable of
`
`incorporating the labeled analog”), 19:15-20 (“A test using a primer with template
`
`which directs the addition of the nucleotide analog to be incorporated will determine
`
`whether the combination is workable.”). Prior to Columbia’s invention, such assays
`
`were needed precisely because the prior art suggested a high failure rate and
`
`unpredictability. Ex. 2071 at ¶ 38. Indeed, Illumina’s expert Dr. Romesberg
`
`testified that a POSA would have to perform experimental testing to determine
`
`
`
`16
`
`
`
`Case IPR2020-00988, -01065, -01125, -01177
`U.S. Patent Nos. 10,407,458; 10,407,459; 10,435,742; 10,457,984
`
`whether a 3’-O-capped nucleotide is incorporated sufficiently for use in SBS. Ex.
`
`2072 at 20:13-23:17.
`
`Dr. Romesberg also testified as to the difficulty of achieving polymerase
`
`incorporation of 3’-O-capped nucleotides. Ex. 2013 at 66:9-67:21. Specifically, in
`
`IPR2013-00517, a company called IBS challenged Illumina’s patents covering
`
`nucleotides modified to have a 3’-O-azidomethyl capping group, which has a similar
`
`small structure to the MOM capping group (-CH2N3 and -CH2OCH3, respectively).
`
`IBS sought to establish that a POSA in 2002 would have expected success in
`
`achieving polymerase incorporation with a nucleotide having a 3’-O-azidomethyl
`
`capping group. To rebut that argument, Dr. Romesberg testified that achieving
`
`polymerase incorporation with 3’-O-capped nucleotides was “difficult,” “not
`
`something that’s common,” that there were “many, many examples of failures,” and
`
`that he is “amazed” that any such nucleotides are capable of incorporation. Ex. 2013
`
`at 66:9-67:21. Specifically:
`
`Q. Are you aware if some researchers were able
`
`to – to get 3’ oxygen protecting group nucleotide
`
`analogues with 3’ oxygen protecting groups, they
`
`were able to get incorporation and termination of
`
`a – of a template growing strand using – using a
`
`particular polymerase? …
`
`THE WITNESS [Dr. Romesberg]: I’m aware that
`
`people were looking for that. I’m aware that a lot
`
`
`
`17
`
`
`
`Case IPR2020-00988, -01065, -01125, -01177
`U.S. Patent Nos. 10,407,458; 10,407,459; 10,435,742; 10,457,984
`
`of [capping] groups were tried. I think the O-
`
`Allyl was amongst – was maybe uniquely
`
`successful, an O-Allyl protecting group, and even
`
`that was limited, meaning a very broad panel of
`
`polymerases were looked at and it was the only
`
`one that showed any promise, in my opinion; and
`
`that, again, just testifies to how difficult I think
`
`the problem was.
`
`Look, the 3’ hydroxyl group [of the nucleotide]
`
`mechanistically needs to bind the magnesium ion
`
`on the active site [of the polymerase]. It needs to
`
`be activated for attack on the next incoming
`
`triphosphate. And part – sorry. Part of its
`
`binding mechanism involves chelating, binding
`
`to a metal ion in the active site, if you’re going to
`
`go putting groups on here.
`
`Frankly, my intuition is I’m amazed any of them
`
`work, because it’s such a mechanistically central
`
`part of the enzyme.
`
`Q. But you would agree that there were some
`
`researchers that had some success prior to August
`
`2002?
`
`A. Sure. But the question I’m answering is
`
`whether it was considered straightforward or
`
`whether it was considered – you know, whether it
`
`was considered sort of – it was certainly not
`
`considered easy, and the very few examples –
`
`
`
`18
`
`
`
`Case IPR2020-00988, -01065, -01125, -01177
`U.S. Patent Nos. 10,407,458; 10,407,459; 10,435,742; 10,457,984
`
`maybe only the O-Allyl, partially – and the
`
`many, many examples of failures I think support
`
`– that’s what – that’s what I’m saying, is that
`
`amongst the natural polymerases, these – this is
`
`simply not something that’s common.
`
`Ex. 2013 at 66:9-67:21.
`
`Given the field’s high failure rate and absence of data for the MOM
`
`embodiment, absent the guidance in Columbia’s patent, a POSA would have had no
`
`reason
`
`to expect
`
`that the MOM embodiment could achieve polymerase
`
`incorporation. That remained true until the Columbia inventors disclosed their
`
`insights into the requirements needed for 3’-O-capped nucleotides to be useful for
`
`SBS. With the benefit of those insights, for the first time it became possible for a
`
`POSA to arrive at the expectation that the MOM embodiment would be capable of
`
`achieving polymerase incorporation. While Illumina derides Columbia’s patent for
`
`containing no “data,” that argument misses the point. See IPR2020-00988 Petition
`
`(Paper 1) at 19, 33, 35, 37. The Columbia patents disclose novel chemical and
`
`structural features that enable the claimed inventions and that would have provided
`
`a POSA with an expectation of success where none existed previously. No other
`
`“data” beyond those novel insights is needed to enable the claimed inventions.
`
`Specifically, the Columbia inventors were the first group to recognize the
`
`particular chemical and structural features that dictated whether a capping group
`
`
`
`19
`
`
`
`Case IPR2020-00988, -01065, -01125, -01177
`U.S. Patent Nos. 10,407,458; 10,407,459; 10,435,742; 10,457,984
`
`could be useful for SBS. First, they discovered that there are chemical constraints
`
`that limit which capping groups can work for SBS. Specifically, capping groups
`
`cannot work if they are methoxy or if they contain a ketone or ester. See Ex. 1001
`
`at 3:18-54, claim 1. Illumina contends that Columbia’s specification “admits” that
`
`these provisos would have been obvious to a POSA, but that is wrong. See IPR2020-
`
`00988 Petition (Paper 1) at 48-50. The specification discloses the inventors’
`
`insights, and it is axiomatic that “[t]he inventor’s own path itself never leads to a
`
`conclusion of obviousness; that is hindsight.” Otsuka Pharm. Co. v. Sandoz, Inc.,
`
`678 F.3d 1280, 1296 (Fed. Cir. 2012).
`
`Notably, Illumina has not disputed that the inventors’ discovery ran contrary
`
`to the prior art’s teachings to use capping groups that failed to satisfy these chemical
`
`constraints. For example, Tsien discloses that “preferred embodiments focus on the
`
`ester blocking groups such as lower (1-4 carbon) alkanoic acid and substituted lower
`
`alkanoic acid esters[.]” Ex. 1031 at 21. And Welch in 1999 tested nucleotides with
`
`esters, belying Illumina’s contention that Canard had previously taught a POSA not
`
`to use esters. Ex. 1033 at 952, 954. Illumina’s own patent applications, filed after
`
`Columbia’s priority date, teach the use of capping groups with esters, contrary to the
`
`Columbia inventors’ insight. Ex. 2077 at Fig. 3.
`
`The Columbia inventors also discovered that there are structural constraints
`
`that limit which capping groups can work for SBS. Specifically, the Columbia
`
`
`
`20
`
`
`
`Case IPR2020-00988, -01065, -01125, -01177
`U.S. Patent Nos. 10,407,458; 10,407,459; 10,435,742; 10,457,984
`
`inventors discovered that SBS requires the use of particularly small capping groups,
`
`i.e., less than 3.7Å in diameter, such that it could fit within the active site of a
`
`polymerase. Like Columbia’s discovery of chemical requirements, its discovery of
`
`the structural requirements ran contrary to the prior art, which (wrongly) suggested
`
`that larger, bulky capping groups could be used for SBS. Tsien disclosed both small
`
`and large capping groups, including the large 2-nitrobenzyl capping group, which
`
`Tsien touted as preferred. Ex. 1031 at 21; Ex. 2071 at ¶¶ 79-80. Following Tsien,
`
`Metzker 1994 likewise pursued small and large capping groups, and reported that a
`
`nucleotide with the large 2-nitrobenzyl capping group achieved complete
`
`termination in his assays. Ex. 1039 at 4263 (Table 2); Ex. 2071 at ¶ 81. Following
`
`Metzker 1994, Metzker 1998, Welch 1999, Anazawa, and Stemple similarly pursued
`
`the large 2-nitrobenzyl capping group and modifications thereof. Ex. 2071 at ¶ 81;
`
`Ex. 1033 at 951-956; Ex. 1073 at 7; Ex. 2002 at 814-817; Ex. 2003 at 13. By the
`
`priority date, a POSA would have understood that 2-nitrobenzyl was being widely
`
`pursued due to the belief that it achieved complete termination in Metzker’s assays.
`
`Ex. 2071 at ¶ 81. Based on actual data provided in the art, a POSA had no reason to
`
`conclude that capping groups smaller than the large 2-nitrobenzyl would be
`
`preferred for SBS. Ex. 2071 at ¶¶ 79-89. In fact, even after the priority date (and
`
`shortly following the first publication of a Columbia application), Illu