PGR2025-00017
`
`Declaration of Dr. Park
`
`
`
`
`
`
`
`
`UNITED STATES PATENT AND TRADEMARK OFFICE
`
`––––––––––––––––––
`
`BEFORE THE PATENT TRIAL AND APPEAL BOARD
`
`––––––––––––––––––
`
`Merck Sharp & Dohme LLC,
`Petitioner,
`
`v.
`
`Halozyme Inc.,
`Patent Owner.
`
`––––––––––––––––––
`
`Case No. PGR2025-00017
`U.S. Patent No. 12,110,520
`
`––––––––––––––––––
`
`Declaration of Dr. Sheldon Park
`
`
`
`
`Petitioner Merck
`Ex. 1004, p. i
`
`

`

`PGR2025-00017
`
`Declaration of Dr. Park
`
`Table of Contents
`
`Introduction .................................................................................................... 1
`
`I.
`
`A.
`
`B.
`
`C.
`
`D.
`
`Background and Qualifications ............................................................. 1
`
`Compensation ........................................................................................ 2
`
`Information Considered ......................................................................... 3
`
`Person of Ordinary Skill in the Art ....................................................... 3
`
`II.
`
`Scope of My Declaration ............................................................................... 4
`
`III. Terminology Used in this Declaration ......................................................... 5
`
`IV. Overview of Methodologies Used for Identification of Tolerated Amino
`Acid Substitutions in PH201-447 ..................................................................... 6
`
`A.
`
`Identification of Non-Essential Residues From Homologous
`Sequences .............................................................................................. 9
`
`1.
`
`2.
`
`3.
`
`BLAST Results and the Multiple Sequence Alignment ........... 10
`
`Essential Residues in PH20 ...................................................... 13
`
`Non-Essential Residues in PH20 .............................................. 15
`
`B. Modeling the PH20 Structure for Visual Inspection ........................... 16
`
`1.
`
`2.
`
`3.
`
`Visualization of the Protein Structure ....................................... 16
`
`Homology Models for Proteins with Unknown Structures ...... 19
`
`Generation of the Model of PH20 Structure ............................. 21
`
`C.
`
`Evaluation of Substitutions at Non-Essential Residues ...................... 23
`
`1.
`
`Published Assessments of Single Amino Acid Substitutions ... 27
`
`2. My Assessment of Factors Influencing Single Amino Acid
`Substitutions .............................................................................. 32
`
`3.
`
`4.
`
`Analysis of Published Results from Mutations of Hyaluronidase
`Proteins ...................................................................................... 47
`
`Review of the Methodology Demonstrates Unbiased Evaluation
` ................................................................................................... 56
`
`V. Analysis of Position 324 / 359 ...................................................................... 57
`
`A. Description of the Structure Near Position 324 .................................. 57
`
`Petitioner Merck
`Ex. 1004, p. ii
`
`

`

`PGR2025-00017
`
`Declaration of Dr. Park
`
`B. Assessment of E324D Substitution ..................................................... 63
`
`C. Assessment of E324N Substitution ..................................................... 67
`
`D. Assessment of E324R Substitution ..................................................... 71
`
`E.
`
`F.
`
`Assessment of E324A Substitution ..................................................... 75
`
`Assessment of E324H Substitution ..................................................... 77
`
`G. Assessment of E324S Substitution ...................................................... 79
`
`VI. Tools Used in My Analysis .......................................................................... 81
`
`A.
`
`B.
`
`C.
`
`D.
`
`BLAST Search and Narrowing of Returned Sequences ..................... 82
`
`Clustal Omega ..................................................................................... 86
`
`SWISS-MODEL .................................................................................. 87
`
`PyMol .................................................................................................. 97
`
`VII. Determination of Numbers of Distinct Polypeptides .............................. 100
`
`
`
`Petitioner Merck
`Ex. 1004, p. iii
`
`

`

`PGR2025-00017
`
`I.
`
`Introduction
`
`Declaration of Dr. Park
`
`A. Background and Qualifications
`
`1. My educational background, career history, and other relevant
`
`qualifications are summarized below. I attach to this Declaration my curriculum
`
`vitae (Appendix B) which provides a full and accurate description of my
`
`educational background, professional experience, and qualifications.
`
`2.
`
`I received my Ph.D. in Biophysics from Harvard University in 2000.
`
`I completed my postdoctoral training at the University of Pennsylvania. I also
`
`received an M.S. in Physics from the Massachusetts Institute of Technology
`
`(“MIT”) in 1994. I received a B.S. in Physics and Math from the University of
`
`California, Berkeley, in 1991.
`
`3.
`
`I currently serve as an associate professor in the Department of
`
`Chemical and Biological Engineering and previously served as Director of
`
`Graduate Studies for the department. I am also affiliated with the University’s
`
`Genetics, Genomics and Bioinformatics graduate program. I have been a professor
`
`at the University of Buffalo since completing my postdoctoral research position at
`
`the University of Pennsylvania.
`
`4.
`
`During my career I have taught courses in, among other things,
`
`Protein Engineering, and Biotechnology Principles for Chemical Engineers. I
`
`began teaching these courses in 2007.
`
`Petitioner Merck
`Ex. 1004, p. 1
`
`

`

`PGR2025-00017
`
`Declaration of Dr. Park
`
`5.
`
`I have nearly two decades of experience in the field of protein
`
`engineering. In my lab, we use computational and experimental tools to
`
`characterize and design protein molecules with novel physical and biological
`
`properties. For example, my lab engineered the first functional monomeric
`
`streptavidin that is used in in vivo imaging and biomolecular detection. My lab has
`
`also developed and engineered many proteins based on rational design and directed
`
`evolution techniques.
`
`6.
`
`In 2011, I was named a recipient of a CAREER award from the
`
`National Science Foundation, which is an award given to early-career faculty who
`
`have the potential to serve as academic role models in research and education and
`
`to lead advances in the mission of their department or organization.
`
`7.
`
`I have authored 30 peer-reviewed publications, many of which are
`
`directed to topics in proteins, protein characterization, and protein design and
`
`engineering. Along with Jennifer Cochran, I was an editor of the textbook “Protein
`
`Engineering and Design,” which was published in 2009 by CRC Press.
`
`B. Compensation
`
`8.
`
`I am being compensated for my time at the rate of $350 per hour for
`
`my work in connection with this matter. I am being reimbursed for reasonable and
`
`customary expenses associated with my work in this investigation. This
`
`compensation is not dependent in any way on the contents of this Declaration, the
`
`Petitioner Merck
`Ex. 1004, p. 2
`
`

`

`PGR2025-00017
`
`Declaration of Dr. Park
`
`substance of any further opinions or testimony that I may provide, or the ultimate
`
`outcome of this matter.
`
`C.
`
`Information Considered
`
`9. My opinions are based on my years of education, research, and
`
`experience, as well as my investigation and study of relevant materials. I also have
`
`relied upon the materials listed in Appendix A in forming these opinions.
`
`D.
`
`Person of Ordinary Skill in the Art
`
`10.
`
`I understand that my analysis and opinions are to be provided using
`
`the perspective of a person of ordinary skill in the art in the December 2011 time
`
`frame.
`
`11.
`
`I have been informed that a “person of ordinary skill” is a hypothetical
`
`person who has certain educational qualifications and experience, and possesses an
`
`ordinary level of insights and skill.
`
`12. Counsel for Merck provided the following description of a person of
`
`ordinary skill in the art in the 2011 time frame for me to evaluate:
`
`A person of ordinary skill in the art in the 2011 time-frame
`
`would have had an undergraduate degree, a Ph.D., and post-
`
`doctoral experience in scientific fields relevant to study of
`
`protein structure and function (e.g., chemistry, biochemistry,
`
`biology, biophysics). From training and experience, the
`
`person would have been familiar with factors influencing
`
`protein structure, folding and activity, production of
`
`Petitioner Merck
`Ex. 1004, p. 3
`
`

`

`PGR2025-00017
`
`Declaration of Dr. Park
`
`modified proteins using recombinant DNA techniques, and
`
`use of biological assays to characterize protein function, as
`
`well with techniques and tools used to analyze protein
`
`structure-structure relationship (i.e., sequence searching and
`
`alignments, protein modeling software, etc.).
`
`13.
`
`I believe the description of a person of ordinary skill in ¶ 12 is
`
`accurate, and that I had those qualifications by December of 2011.
`
`14.
`
`In preparing this declaration, I have used the perspective of a person
`
`of ordinary skill in the art in the 2011 time frame as it is described in ¶ 12.
`
`II.
`
`Scope of My Declaration
`
`15.
`
`I was asked if a person of ordinary skill in the art in 2011 would have
`
`been able to identify the single amino acid substitutions within non-essential
`
`regions of PH201-447 that would be tolerated by the protein (i.e., would not cause a
`
`substantial loss of hyaluronidase activity) without making and testing each possible
`
`single-substituted PH201-447 protein.
`
`16. As I demonstrate below, a skilled person, using sequence analysis and
`
`protein modeling techniques known in 2011, would have readily identified many
`
`specific amino acid substitutions, some of which are not necessarily conservative
`
`substitutions, that would be tolerated by the PH201-447 protein structure and would
`
`therefore enable the protein to retain its hyaluronidase activity. I base this
`
`conclusion on an analysis I performed, which is described in § IV.C.
`
`Petitioner Merck
`Ex. 1004, p. 4
`
`

`

`PGR2025-00017
`
`Declaration of Dr. Park
`
`17.
`
`I identified single amino acid substitutions that would be tolerated at
`
`position 324 in PH201-447, including E324D, E324N, E324R, E324A, E324H, and
`
`E324S. I summarize this analysis in § V.
`
`18. Finally, I was asked to determine the number of distinct polypeptides
`
`that share 91% or 95% sequence identity with human PH20 sequences of varying
`
`length, assuming certain conditions. I provide my calculations below in § VII.
`
`III. Terminology Used in this Declaration
`
`19.
`
`I will use the following abbreviations in this declaration:
`
`(a)
`
`I use “PH20” to refer to the human PH20 protein. The full-
`
`length sequence of the human PH20 protein has 509 amino
`
`acids and was first published in 1993.1 The sequence is
`
`reported as SEQ ID NO:1 in U.S. Patent No. 7,767,429 (the
`
`’429 Patent), which is identical to the sequence deposited with
`
`the Universal Protein Resource website (“www.uniprot.org”)
`
`with ID P38567.
`
`(b) The full-length human wild-type PH20 sequence includes a 35
`
`amino acid signal sequence (positions 1-35 of Uniprot ID:
`
`P38567). The mature amino acid sequence PH20 is found at
`
`
`1 EX1029 (Gmachl), 546, Fig. 1.
`
`Petitioner Merck
`Ex. 1004, p. 5
`
`

`

`PGR2025-00017
`
`Declaration of Dr. Park
`
`positions 36-509 of Uniprot ID: P38567, and is 1-474 in
`
`sequences that omit the signal sequence.
`
`(c)
`
`“PH201-n” refers to the human wild-type PH20 polypeptide
`
`sequence starting at position 1 and terminating at position “n”
`
`of the mature protein (i.e., lacking the signal sequence). For
`
`example, PH201-447 means the polypeptide starting at position 1
`
`and ending at position 447 of the mature human wild-type
`
`PH20 sequence. The corresponding positions in the human
`
`wild-type PH20 sequence having the signal sequence are 36-
`
`482 (e.g., SEQ ID NO: 1 of U.S. 7,767,429, Uniprot: P38567).
`
`(d)
`
`“AxxxB” refers to an amino acid substitution at position xxx,
`
`where the wild-type residue is A and the residue after the
`
`substitution is B.
`
`(e) The use of a position number in this declaration is referring to
`
`the position in the mature form of human PH20 (i.e., omitting
`
`the 1-35 amino acid signal sequence). I also will occasionally
`
`use both the full-length and the mature numbers.
`
`IV. Overview of Methodologies Used for Identification of Tolerated Amino
`Acid Substitutions in PH201-447
`
`20. Proteins often can tolerate a single amino acid substitution in non-
`
`essential regions of the protein’s structure. I use tolerate here to mean that the
`
`Petitioner Merck
`Ex. 1004, p. 6
`
`

`

`PGR2025-00017
`
`Declaration of Dr. Park
`
`presence of a single amino acid at a particular position of the protein’s amino acid
`
`sequence that is different than the naturally occurring (“wild-type”) amino acid at
`
`that position of the protein does not materially alter the local structure around that
`
`position in the protein and thus does not meaningfully alter the biological activity
`
`of the protein. Of course, there are exceptions to this general point, and any
`
`particular substitution will need to be assessed to determine if that particular
`
`substitution will be tolerated.
`
`21. The concept of tolerance can be readily appreciated by comparing
`
`evolutionarily related proteins that share structure homology. While there are
`
`amino acids at many positions within a set of evolutionarily related proteins that
`
`are conserved (i.e., amino acids that do not vary or vary only rarely), the amino
`
`acids at many other positions can vary extensively.2 The variability in the amino
`
`acids at these non-conserved positions indicates that the protein structure(s) is not
`
`
`2 EX1014 (Brandon), 351 (“The underlying assumption is that secondary and
`
`tertiary structure has been more conserved during evolution than amino acid
`
`sequence; in other words only such changes have been retained during
`
`evolution that conserve the structure. Consequently, the pattern of residue
`
`changes within homologous proteins contains specific information about the
`
`structure.”).
`
`Petitioner Merck
`Ex. 1004, p. 7
`
`

`

`PGR2025-00017
`
`Declaration of Dr. Park
`
`dependent on the exact identity of the amino acid at those positions and is able to
`
`accommodate different amino acids without causing a loss of protein function.
`
`22. To identify single amino acid substitutions that would be tolerated in
`
`human PH201-447, I used procedures that were widely used in rational design
`
`methods of protein engineering before December 2011. Generally, these include
`
`the following steps:
`
`(a)
`
`identify homologous sequences based on sequence comparison;
`
`(b)
`
`perform a multiple sequence alignment of retrieved sequences;
`
`(c)
`
`identify the frequency of occurrence (“profile”) of amino acids
`
`at each position of the protein across the set of sequences;
`
`(d)
`
`generate a protein model using SWISS-MODEL;
`
`(e)
`
`use the model to visually assess individual substitutions within
`
`the local environment of the protein at a particular position.3
`
`23. The techniques I describe above of finding homologous sequences,
`
`aligning them, and using the sequence identity information to identify single amino
`
`
`3 As I describe in further detail, below, the model would have been reliable for
`
`evaluating certain single amino acid substitutions (depending on the position of
`
`the substitution), but not for modeling multiple concurrent substitutions, which
`
`would quickly make the model unreliable. See ¶¶ 160-162.
`
`Petitioner Merck
`Ex. 1004, p. 8
`
`

`

`PGR2025-00017
`
`Declaration of Dr. Park
`
`acid substitutions that would be tolerated were well-known and widely used by
`
`skilled artisans in 2011.4
`
`24.
`
`I believe a person of ordinary skill in the art in 2011 would have been
`
`familiar with the tools and techniques discussed in Green (EX1017).5 These
`
`include: (i) BLAST, for performing protein sequence searching, (ii) sequence
`
`alignment tools like CLUSTAL (e.g., CLUSTAL-Omega), (iii) the protein
`
`structure modeling tool SWISS-MODEL, and (iv) software to view protein
`
`structures, like PyMol. I discuss these tools further in § VI, below.
`
`A.
`
`Identification of Non-Essential Residues From Homologous
`Sequences
`
`25.
`
`In general, amino acids that do not vary at the same aligned positions
`
`within a set of structurally-related proteins are essential to the structure and
`
`functions of the protein.6 When a residue does not vary in so many naturally
`
`
`4 EX1017 (Green), 223-230, 236 (discussing rational design techniques);
`
`EX1016 (Steipe), 181-186.
`
`5 These tools and techniques are also discussed in Steipe (EX1016).
`
`6 EX1017 (Green), 224 (“By considering the common features of the sequences
`
`of these proteins, it is possible to deduce the key elements that determine
`
`protein structure and function…”).
`
`Petitioner Merck
`Ex. 1004, p. 9
`
`

`

`PGR2025-00017
`
`Declaration of Dr. Park
`
`occurring proteins, it is strong evidence that it is essential to the protein’s structure
`
`and function and needs to be preserved.7
`
`26.
`
`I investigated essential residues in hyaluronidase enzymes in two
`
`ways. First, I identified invariant residues by analyzing a multiple sequence
`
`alignment of a set of published hyaluronidase sequences that was available in
`
`December of 2011. Second, I reviewed scientific literature that identified
`
`important residues in hyaluronidase proteins or which reported experimental
`
`results showing that modifying single residues impaired or eliminated activity of
`
`the enzymes.8 These two overlapped significantly and many residues that were
`
`experimentally shown to be essential for activity proved to be highly conserved
`
`based on sequence alignment. Below I address the first investigation.
`
`1.
`
`BLAST Results and the Multiple Sequence Alignment
`
`27.
`
`I generated a dataset of sequences that were homologous to PH20 and
`
`that were publicly available by the end of 2011. To do that I performed a BLAST
`
`
`7 EX1017 (Green), 224 (“Evolution provides a tremendously useful model for
`
`protein design.”).
`
`8 I reviewed the scientific literature in detail below. See § IV.C.3.
`
`Petitioner Merck
`Ex. 1004, p. 10
`
`

`

`PGR2025-00017
`
`Declaration of Dr. Park
`
`search and culled the results to yield a set of 88 nonredundant homologous
`
`sequences that were publicly available by 2011.9
`
`28.
`
`I then aligned the 88 homologous sequences in a multiple sequence
`
`alignment (“MSA”).10 The homologous sequences “must be aligned such that
`
`conserved positions are in register with one another.”11 Since the sequences are all
`
`different, the MSA attempts to align the sequences, but this can result in gaps in
`
`the MSA. MSA algorithms seek to “optimize a global score across an alignment,”
`
`which is “typically based on sequence similarity alone and does not take structural
`
`or functional information into account.”12
`
`
`9 I explain this process in further detail below. See § VI.A.
`
`10 EX1017 (Green), 224 (“One of the most straightforward applications of primary
`
`sequence data in protein engineering is the use of multiple-sequence alignments
`
`to define consensus motifs for a particular structure or function.”); EX1016
`
`(Steipe), 184 (“multiple sequence alignments are required”).
`
`11 EX1017 (Green), 224.
`
`12 EX1016 (Steipe), 184.
`
`Petitioner Merck
`Ex. 1004, p. 11
`
`

`

`PGR2025-00017
`
`Declaration of Dr. Park
`
`29.
`
`I used the Clustal Omega tool to produce an MSA of the 88
`
`homologous sequences.13 The Clustal Omega program computes and reports
`
`whether a residue is “conserved” or “semi-conserved,” which takes into account
`
`how similar a residue is across all of the sequences. A “conserved residue” is one
`
`which has the same amino acid in all of the sequences. A “semi-conserved”
`
`residue is one where all amino acids in that position have similar properties. The
`
`multiple sequence alignment I produced is in Exhibit 1058.14 Each row shows 60
`
`positions of the aligned sequences. Below each set of 60 positions is a consensus
`
`report using 70%, 80%, 90% and 100% identity thresholds for consensus residues
`
`at that position (i.e., the amino acid that appears in 70%, 80%, 90% or 100% of the
`
`aligned sequences, if any). The consensus report gives a quick overview of where
`
`
`13 See, e.g., EX1043 (Sievers), 1-2, 4-5. I explain this process in further detail
`
`below. See § VI.B.
`
`14 The MSA depicted in Exhibit 1058 was produced by the MSA viewer available
`
`from the European Bioinformatics Institute of the European Molecular Biology
`
`Laboratory organization (EMBL-EBI)
`
`(https://www.ebi.ac.uk/jdispatcher/msa/mview?stype=protein). The text-based
`
`output of the CLUSTAL-Omega MSA is also provided in Exhibit 1057.
`
`Petitioner Merck
`Ex. 1004, p. 12
`
`

`

`PGR2025-00017
`
`Declaration of Dr. Park
`
`conserved residues appear. I performed a more detailed analysis of the MSA using
`
`custom scripts I prepared, as described below.
`
`2.
`
`Essential Residues in PH20
`
`30. Using the alignment, I identified 68 largely invariant residues that a
`
`skilled artisan would have deemed “essential residues” in PH201-447 based on the
`
`sequence alignment (table below).15 These are positions where the amino acid was
`
`conserved 95% of the time or more and non-identical amino acids appear in less
`
`than ~5% of the proteins in the data set.16
`
`
`
`
`
`
`15 The essential residues are also listed in Appendix D-3 (EX1004, 171).
`
`16 The frequency at which an amino acid appears in the MSA is compiled in
`
`Appendix D-1 (EX1004, 133).
`
`Petitioner Merck
`Ex. 1004, p. 13
`
`

`

`PGR2025-00017
`
`Declaration of Dr. Park
`
`Residue
`#
`
`Mature
`Residue #
`
`PH20
`Residue
`
`Residue
`%
`
`49
`
`53
`
`60
`
`91
`
`92
`
`97
`
`99
`
`14
`
`18
`
`25
`
`56
`
`57
`
`62
`
`64
`
`F
`
`W
`
`C
`
`F
`
`Y
`
`G
`
`Y
`
`95.5
`
`100
`
`100
`
`98.9
`
`98.9
`
`100
`
`97.7
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`Residue
`#
`
`Mature
`Residue #
`
`PH20
`Residue
`
`Residue
`%
`
`222
`
`224
`
`226
`
`234
`
`236
`
`238
`
`246
`
`249
`
`187
`
`189
`
`191
`
`199
`
`201
`
`203
`
`211
`
`214
`
`P
`
`C
`
`N
`
`Y
`
`G
`
`C
`
`N
`
`L
`
`100
`
`100
`
`98.9
`
`98.9
`
`98.9
`
`100
`
`100
`
`98.9
`
`100
`
`112
`
`113
`
`115
`
`116
`
`123
`
`141
`
`144
`
`146
`
`65
`
`77
`
`78
`
`80
`
`81
`
`88
`
`106
`
`109
`
`111
`
`P
`
`G
`
`G
`
`P
`
`Q
`
`H
`
`G
`
`V
`
`D
`
`100
`
`98.9
`
`100
`
`100
`
`100
`
`98.9
`
`98.9
`
`95.5
`
`98.9
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`251
`
`253
`
`256
`
`258
`
`259
`
`261
`
`281
`
`284
`
`216
`
`218
`
`221
`
`223
`
`224
`
`226
`
`246
`
`249
`
`W
`
`W
`
`S
`
`A
`
`L
`
`P
`
`R
`
`E
`
`R
`
`98.9
`
`98.9
`
`100
`
`97.7
`
`100
`
`98.9
`
`97.7
`
`100
`
`98.9
`
`147
`
`148
`
`150
`
`152
`
`154
`
`157
`
`158
`
`164
`
`168
`
`112
`
`113
`
`115
`
`117
`
`119
`
`122
`
`123
`
`129
`
`133
`
`W
`
`E
`
`W
`
`P
`
`W
`
`N
`
`W
`
`Y
`
`S
`
`100
`
`100
`
`100
`
`100
`
`97.7
`
`98.9
`
`96.6
`
`100
`
`100
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`287
`
`299
`
`316
`
`321
`
`326
`
`332
`
`335
`
`339
`
`351
`
`252
`
`264
`
`281
`
`286
`
`291
`
`297
`
`300
`
`304
`
`316
`
`P
`
`L
`
`L
`
`G
`
`G
`
`G
`
`W
`
`C
`
`98.9
`
`98.9
`
`97.7
`
`98.9
`
`98.9
`
`98.9
`
`98.9
`
`100
`
`96.6
`
`188
`
`192
`
`203
`
`211
`
`212
`
`215
`
`216
`
`217
`
`219
`
`153
`
`157
`
`168
`
`176
`
`177
`
`180
`
`181
`
`182
`
`184
`
`A
`
`F
`
`T
`
`R
`
`P
`
`L
`
`W
`
`G
`
`Y
`
`100
`
`100
`
`98.9
`
`95.5
`
`97.7
`
`95.5
`
`100
`
`100
`
`100
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`362
`
`368
`
`369
`
`376
`
`377
`
`381
`
`385
`
`387
`
`398
`
`327
`
`333
`
`334
`
`341
`
`342
`
`346
`
`350
`
`352
`
`363
`
`L
`
`N
`
`V
`
`C
`
`S
`
`C
`
`G
`
`C
`
`L
`
`97.7
`
`96.6
`
`100
`
`96.6
`
`100
`
`100
`
`100
`
`96.6
`
`Petitioner Merck
`Ex. 1004, p. 14
`
`

`

`PGR2025-00017
`
`Declaration of Dr. Park
`
`3.
`
`Non-Essential Residues in PH20
`
`31. The 379 positions in PH201-447 other than the 68 essential residues are
`
`positions at which there is evolutionary variation among the 88 protein sequences
`
`that I analyzed. The existence of evolutionary variation at these positions indicates
`
`that the homologous proteins have tolerated different amino acids at those
`
`positions. Amino acids at these positions in the different hyaluronidase proteins
`
`would be considered “non-essential” residues because the proteins presumably still
`
`exhibit hyaluronidase activity when they are mutated. The 379 non-essential
`
`positions in PH201-447 are compiled in Appendix D-2.
`
`32.
`
`I also reviewed observations in the ’429 Patent concerning making
`
`modifications to PH20 proteins. As it explains:
`
`Suitable conservative substitutions of amino acids are
`
`known to those of skill in this art and can be made generally
`
`without altering the biological activity, for example
`
`enzymatic activity, of the resulting molecule. Those of skill
`
`in this art recognize that, in general, single amino acid
`
`substitutions in non-essential regions of a polypeptide do not
`
`substantially alter biological activity …17
`
`
`17 EX1005 (’429 Patent), 16:14-22; also 9:47-49.
`
`Petitioner Merck
`Ex. 1004, p. 15
`
`

`

`PGR2025-00017
`
`Declaration of Dr. Park
`
`I believe the positions that I identified in Appendix D-2, including position 324,
`
`align with what I consider to be the “non-essential regions” referred to by the ’429
`
`Patent. An illustration of some non-essential regions in PH20 is provided below.18
`
`
`
`B. Modeling the PH20 Structure for Visual Inspection
`
`1.
`
`Visualization of the Protein Structure
`
`33.
`
`In 2011, to assess whether a potential amino acid substitution at a
`
`particular position in a protein would be tolerated, a person of ordinary skill would
`
`have visualized the mutation by building a structural model of the protein.
`
`Visualization allows one to assess the interactions between the wild-type amino
`
`acid being changed and its neighboring amino acids at that position, and thereby
`
`compare the interactions of different amino acids at that position with neighboring
`
`
`18 I annotated Figure 3 in Chao (EX1006) to illustrate some of the non-essential
`
`regions I identified in my analysis.
`
`Petitioner Merck
`Ex. 1004, p. 16
`
`

`

`PGR2025-00017
`
`Declaration of Dr. Park
`
`amino acids. Comparing the set of interactions seen with the wild-type residue to
`
`those observed for the substituted amino acid at that position will allow prediction
`
`of whether the substitution will be favorable, neutral, or unfavorable to the
`
`protein’s structure.
`
`34. Assessing amino acid substitutions by visual inspection of the site of
`
`the substitution in a protein structural model was a commonly used technique in
`
`2011. For example, as one group reported:
`
`Many protein-engineering applications involve the creation
`
`of a small number of mutations to a naturally occurring
`
`protein so as to enhance its function in a well-defined
`
`manner. In these cases, a structural biologist’s intuition is
`
`often an important tool in the design of the desired variants,
`
`an approach that may be termed structure-based protein
`
`design to borrow a term from the drug design field.
`
`Visualization of the known reference structure is a key
`
`component of this. For example, visualization can identify
`
`unsatisfied hydrogen bond donors or acceptors that may be
`
`mutated to increase stability or affinity. Similarly,
`
`visualizing steric interactions can help engineer interactions
`
`to discriminate among several potential binding targets.19
`
`
`19 EX1017 (Green), 228-29.
`
`Petitioner Merck
`Ex. 1004, p. 17
`
`

`

`PGR2025-00017
`
`Declaration of Dr. Park
`
`35. Another example is work done by a group led by Dr. Moult at UMBI.
`
`His group performed visual inspections of protein models to assess the impact of
`
`single amino acid substitutions in proteins that were caused by single nucleotide
`
`polymorphisms. The results of those visual inspections were published in peer
`
`review journals.20 I view that as a validation of this technique of visually assessing
`
`the effects of an amino acid substitution that I used in my analysis here.
`
`36. The local environment of the protein where an amino acid substitution
`
`is being made can be visualized using a structural model of the protein. In
`
`December of 2011 (and even today), the structure of human PH20 was not solved.
`
`The absence of an experimentally determined structure for a protein does not
`
`preclude the use of structure-based modeling methods.21 That is because the
`
`
`20 EX1031 (Yue)‚ 459 (reporting effects of single amino acid substitutions “relied
`
`primarily on visual inspection of an amino acid substitution on protein structure
`
`and function.”); EX1032 (Wang), 265-266 (analyzing homology models for
`
`single amino acid substitutions to study mechanisms of hereditary disease).
`
`21 EX1017 (Green), 229 (“In many cases, protein engineering targets a protein
`
`whose structure has not been solved”), (“This does not preclude the use of
`
`structure-based methods, as the known structures of related proteins can be used
`
`to create model structures through the process of homology modeling”).
`
`Petitioner Merck
`Ex. 1004, p. 18
`
`

`

`PGR2025-00017
`
`Declaration of Dr. Park
`
`structure of a protein can be modeled in a process called “homology modeling.”22
`
`Homology modeling was routinely used by 2011 and provided an “accurate
`
`computational method to generate reliable structural models.”23
`
`2. Homology Models for Proteins with Unknown Structures
`
`37. Generally, a protein homology model is built using an experimentally
`
`determined reference structure with a highly homologous sequence.24 That is
`
`
`22 EX1017 (Green), 229-230 (“This does not preclude the use of structure-based
`
`methods, as the known structures of related proteins can be used to create
`
`model structures through the process of homology modeling”), (“Homology
`
`building can allow a model structure to be built from the structure of a related
`
`sequence”).
`
`23 EX1012 (Bordoli), 1 (“Homology modeling is currently the most accurate
`
`computational method to generate reliable structural models and is routinely
`
`used in many biological applications”), (Homology modeling is “the method of
`
`choice to build reliable” models); EX1014 (Brandon), 348 (“This model can
`
`serve as an excellent basis for identifying amino acid residues involved in the
`
`active site…”).
`
`24 EX1017 (Green), 229; EX1012 (Bordoli), 1 (“Homology modeling aims to
`
`build three-dimensional protein structure models using experimentally
`
`Petitioner Merck
`Ex. 1004, p. 19
`
`

`

`PGR2025-00017
`
`Declaration of Dr. Park
`
`possible because “[h]omologous proteins have similar three-dimensional
`
`structures.”25 To build a homology model, the “backbone of homologous residues
`
`from a protein of unknown structure” is mapped “onto a known structure.”26 In
`
`many cases, however, “there are regions of nonhomologous sequence, even in
`
`highly homologous proteins” in which a “new backbone…must be constructed for
`
`these regions.”27
`
`38. Numerous homology modeling programs were available by 2011,
`
`including the SWISS-MODEL program.28 This program searches a library of
`
`experimentally determined protein structures to identify suitable templates to use
`
`
`determined structures of related family members as templates”); EX1014
`
`(Brandon), 348 (“If significant amino acid sequence identity is found with a
`
`protein of known crystal structure, a three-dimensional model of the novel
`
`protein can be constructed, using computer modeling, on the basis of the
`
`sequence alignment and the known three-dimensional structure”).
`
`25 EX1014 (Brandon), 370.
`
`26 EX1017 (Green), 229.
`
`27 EX1017 (Green), 229.
`
`28 EX1017 (Green), 229, Table 10.2; also EX1012 (Bordoli), 2. I explain this in
`
`further detail below. See § VI.C.
`
`Petitioner Merck
`Ex. 1004, p. 20
`
`

`

`PGR2025-00017
`
`Declaration of Dr. Park
`
`for the protein of interest, and then generates a model for that protein based on the
`
`sequence alignment of the target protein and the template structure.29
`
`3. Generation of the Model of PH20 Structure
`
`39.
`
`I used the SWISS-MODE