AFFIDAVIT OF JERRY L. ATWOOD, Ph.D.
`
`I, the undersigned, Jerry L. Atwood, US Passport No. 028224440 with a business address
`
`of Department of Chemistry, 601 S. College Avenue, University of Missouri-Columbia,
`
`Columbia, MO 65211, after having been warned that I must state the truth and if I fail to
`
`do so I will be liable to penalties prescribed by law, hereby declare in writing as follows:
`
`1. I reside at 5704 Short Line Dr., Columbia, Missouri 65203. I hold a B.S. degree in
`
`Chemistry and Mathematics from Southwest Missouri State University (1964) and a
`
`Ph.D. in Chemistry from the University of Illinois (1968).
`
`2. Since 1994, I have been employed as Professor and Chairman of the Department of
`
`Chemistry at the University of Missouri-Columbia. From 1968 to 1994, I was
`
`employed by the University of Alabama, where I successively held the titles of
`
`Assistant Professor, Associate Professor, Professor, and University Research Professor.
`
`In 1999 I became Curators’ Professor at the University of Missouri-Columbia.
`
`3. From 1985 to 1998, I was Editor of the Journal of Chemical Crystallography. In 1999 I
`
`was named Consulting Editor for the Journal of Chemical Crystallography. I edited the
`
`Journal of Supramolecular Chemistry from 2000 to 2003, and I was Associate Editor
`
`of Chemical Communications from 1996 to 2005. From 1992 until 2000, I was Editor
`
`of Supramolecular Chemistry. From 1985 to 1993, I was Regional Editor for the
`
`Journal of Coordination Chemistry. I have been Co-Editor-in-Chief of the New Journal
`
`of Chemistry since 2005. I am Co-Editor of the Inclusion Compounds book series (five
`
`volumes), Comprehensive Supramolecular Chemistry (ten volumes), and the
`
`Encyclopedia of Supramolecular Chemistry (2 volumes). I currently serve on the
`
`Editorial Boards of Crystal Growth & Design, Crystal Engineering, the Journal of
`
`Coordination Chemistry, the New Journal of Chemistry, and Supramolecular
`
`Chemistry. I have published more than 650 articles in refereed journals. I have authored
`
`twelve patents. I have taught more than 10,000 students in undergraduate University
`
`chemistry courses and I have taught and supervised graduate students (at both the
`
`Merck Exhibit 2221, Page 1
`Mylan Pharmaceuticals Inc. v. Merck Sharp & Dohme Corp.
`IPR2020-00040
`
`

`

`Masters and Ph.D. level) with a primary emphasis on organic synthesis and
`
`crystallization. I am an expert in the fields of organic and inorganic chemistry and
`
`crystal growth and engineering. Over the years, I have consulted numerous innovative
`
`and generic pharmaceutical companies with regard to issues pertaining to synthesis,
`
`processes and pre-formulation including crystallization, salt formation and
`
`polymorphism. I have consulted widely for industry, particularly in the fields of
`
`pharmaceutical chemistry and polymer chemistry. A copy of my curriculum vitae is
`
`attached hereto as Appendix "A".
`
`4. I was requested by Applicant to review Prof. Serajuddin's declaration and IL 172,563
`
`and respond to Prof. Serajuddin's assertions relating to the formation of phosphoric acid
`
`salts of 4-oxo-4-[3-(trifluoromethyl)-5,6-dihydro[1,2,4]triazolo[4,3-a]pyrazin-7(8H)-
`
`yl]-1-(2,4,5-trifluorophenyl)butan-2-amine (sitagliptin).
`
`5. Prof. Serajuddin essentially argues that a skilled person would have expected that the
`
`only possible stable phosphoric acid salt form of sitagliptin would be the
`
`dihydrogenphosphate salt and that other phosphoric acid salt forms of sitagliptin are not
`
`chemically feasible. Briefly, Prof. Serajuddin argues that the pKa difference between
`
`phosphoric acid and sitagliptin is not sufficient to allow a stable phosphoric acid salt
`
`other than the dihydrogenphosphate salt.
`
`6. As detailed below, Prof. Serajuddin's assertions are without any scientific merit and
`
`contradict first principles of chemistry. Prof. Serajuddin's assertions are also
`
`inconsistent with his own publications. As I elaborate below, there is every expectation
`
`that sitagliptin can form various phosphoric acid salts. Moreover, had Prof. Serajuddin
`
`carried out some experiments, he would have immediately found out that his entire
`
`assertions fly in the face of reality.
`
`Phosphoric acids
`
`7. Phosphoric acid (H3PO4) is a triprotic compound, i.e. it has three acidic protons which
`
`can be donated to an appropriate base to form salts. Therefore, phosphoric acid has the
`
`2
`
`
`
`Merck Exhibit 2221, Page 2
`Mylan Pharmaceuticals Inc. v. Merck Sharp & Dohme Corp.
`IPR2020-00040
`
`

`

`-), the
`potential for forming the dihydrogenphosphate salt (H2PO4
`-2) and the phosphate salt (PO4
`
`monohydrogenphosphate salt (HPO4
`
`-3). For instance, the
`
`known stable ammonium phosphates are (NH4)H2PO4 (ammonium acid
`
`dihydrogenphosphate; CAS 7722-76-1), (NH4)2HPO4 (diammonium acid phosphate;
`
`CAS 7783-28-0) and (NH4)3PO4 (triammonium phosphate; CAS 10361-65-6). As
`
`phosphoric acid has three donatable protons, it is sometimes used with another cation
`
`such as potassium, sodium or ammonium. All these phosphoric acids are well known
`
`stable commercial products.
`
`Sitagliptin base molecule and simple acid-base chemistry
`
`8. The chemical structure of sitagliptin is as follows:
`
`3
`
`
`
`

`

`11. pKa is the negative log of the ionization constant for an acid or a base in aqueous
`
`solution. It is used as a measurement of the strength of the base or the acid in an
`
`aqueous medium (McMurry, pp. 52-53, Appendix "B").
`
`12. The first pKa value of phosphoric acid is 2.12; the second is 7.21 and the third is 12.67.
`
`The pKa value of sitagliptin, as reported by Applicant during prosecution of IL
`
`172,563, is 7.7. According to Prof. Serajuddin, in order to obtain a stable salt of a basic
`
`drug, "the pKa of the acid used should be at least 2-3 units below that of the drug".
`
`Therefore, Prof. Serajuddin concludes that phosphoric acid and sitagliptin can form a
`
`dihydrogenphosphate salt because the pKa difference between the first pKa of
`
`phosphoric acid (2.12) and sitagliptin (7.7) is above 3 units. However, the
`
`monohydrogenphosphate salt of sitagliptin cannot be formed because the pKa
`
`difference between the second pKa of phosphoric acid (7.21) and sitagliptin (7.7) is
`
`only 0.5.
`
`13. As detailed below, Prof. Serajuddin's arguments are incorrect speculations.
`
`pKa values in water are different from pKa values in other solvent systems
`
`14. Prof. Serajuddin's starting point is totally flawed. He bases his entire argument on the
`
`pKa values of phosphoric acid and of sitagliptin in aqueous medium. These pKa values
`
`are an indication of ionization in water. However, salts are commonly formed in non-
`
`aqueous solutions. The pKa values in water are not a measure of the acid and base
`
`strength in a non-aqueous medium or in mixed organic–aqueous solvent systems and
`
`can vary significantly (F.G. Bordwell, "Equilibrium Acidities in Dimethyl Sulfoxide
`
`Solution", Acc. Chem. Res. 21 (1988), 456-463, Appendix "C"). The difference in pKa
`
`value of the same compound in aqueous medium vs. organic medium can be in many
`
`orders of magnitude.
`
`15. For instance, A. Bhattacharyya et al., "Conductometric Studies on the Dissociation
`
`Constants of Phosphoric Acid in Methanol-Water Mixture", Electrochimica Acta 25
`
`(1980), 559-561 (Appendix "D") teaches that in solutions of water and methanol in
`
`4
`
`
`
`Merck Exhibit 2221, Page 4
`Mylan Pharmaceuticals Inc. v. Merck Sharp & Dohme Corp.
`IPR2020-00040
`
`

`

`varying concentrations the first pKa value of phosphoric acid can vary from 2.1 to 4.9.
`
`Similarly, H. Yüksek et al., "Synthesis and Determination of pKa Values of Some New
`
`3,4-Disubstituted-4,5-Dihydro-1H-1,2,4-triazol-5-one Derivatives in Non-aqueous
`
`Solvents", Molecules 9 (2004), 232-240 (Appendix "E") also teaches that different
`
`medium can substantially impact the pKa value.
`
`16. Therefore, even if Prof. Serajuddin's argument had any substance at all, it would have
`
`been relevant to the prediction of the making of sitagliptin phosphate salt only in water.
`
`It is not predictive for other commonly used solvents or solvent systems which are
`
`routinely used for salt formation. This is basic chemistry. It is also acknowledged by
`
`Prof. Serajuddin on page 159 of Handbook of Pharmaceutical Salts (Appendix 5 to
`
`Prof. Serajuddin's declaration). I am therefore surprised that Prof. Serajuddin did not
`
`mention this fact in his declaration.
`
`pKa of sitagliptin was not known at the priority date of IL 172,563
`
`17. It is my understanding that the pKa of sitagliptin was not known at the priority date of
`
`IL 172,563. Prof. Serajuddin also mentions that he relies on the pKa value of sitagliptin
`
`as reported by Applicant during prosecution of IL 172,563, well after the priority date
`
`of the application. Without the pKa value of sitagliptin, Prof. Serajuddin's arguments
`
`collapse as the skilled artisan had no reference value for sitagliptin.
`
`18. Moreover, calculation or measurement of pKa value is an intricate exercise for a
`
`molecule of the chemical complexity of sitagliptin. Not only is the measurement itself
`
`difficult, but the stability of the molecule must also be ascertained throughout the entire
`
`measurement in which the molecule is exposed to highly acidic conditions. This is
`
`particularly so for a sensitive compound such as the free base of sitagliptin. The
`
`multiple ionization sites further add to the complexity of the exercise.
`
`19. The pKa value of sitagliptin is the key data of Prof. Serajuddin's analysis. As indicated,
`
`the data were only known to Applicant and were not available at the priority date to the
`
`skilled person. Hence, in order to challenge the validity of the application, Prof.
`
`5
`
`
`
`Merck Exhibit 2221, Page 5
`Mylan Pharmaceuticals Inc. v. Merck Sharp & Dohme Corp.
`IPR2020-00040
`
`

`

`Serajuddin relies on data that only became available much later than the priority date.
`
`This is pure hindsight.
`
`Prof. Serajuddin ignores that sitagliptin has numerous protonation sites
`
`20. Prof. Serajuddin discusses and dismisses the possibility of a phosphoric acid salt with
`
`more than one substrate base of sitagliptin. He however disregards the possibility that
`
`more than one site of the sitagliptin molecule will be protonated, thus forming a salt of
`
`sitagliptin cation with two or more dihydrogenphosphate anions (i.e. instead of a salt
`
`consisting of 1 phosphoric acid with 1, 2 or 3 molecules of base, a salt consisting of 1
`
`molecule of base with 2 or more phosphoric acid molecules). According to Prof.
`
`Serajuddin's arguments, in order to predict whether or not such salts can be formed, one
`
`needs to know all the pKa values of sitagliptin and not only the first pKa. These pKa
`
`values were not known at the priority date of the application. I am not aware that they
`
`have been reported even today. Hence, the lack of pKa information for sitagliptin
`
`renders Prof. Serajuddin's analysis useless.
`
`Salt formation is a trial and error endeavor
`
`21. In any event, Prof. Serajuddin grossly exaggerates the significance of the ∆pKa
`
`between conjugate base and acid. This is not a scientific principle as Prof. Serajuddin
`
`seems to incorrectly imply. Even Prof. Serajuddin himself stated in his publications
`
`that, in any case, notwithstanding the pKa differences, the reality of salt formation
`
`remains "a trial and error endeavor" (C.G. Smith et al., The Process of New Drug
`
`Discovery and Development (editors), 2006, at p. 26, Appendix "F"). Moreover, in
`
`section 2.1 of Handbook of Pharmaceutical Salts on "feasibility assessment for salt
`
`formation" (Appendix 4 to Prof. Serajuddin's declaration), Prof. Serajuddin states that
`
`"no predictive procedure to determine whether a particular acidic or basic drug would
`
`form a salt with a particular counter-ion has been reported in the literature".
`
`22. In addition, to the extent that the pKa difference can provide any guidance, it is that the
`
`pKa value of the acid should be equal to or lower than the pKa of the conjugate base in
`
`6
`
`
`
`Merck Exhibit 2221, Page 6
`Mylan Pharmaceuticals Inc. v. Merck Sharp & Dohme Corp.
`IPR2020-00040
`
`

`

`order to create a salt (i.e. a 2-3 unit difference is not necessary). For instance, Gould
`
`states that "to form a salt the pKa of the conjugate acid has to be less than or equal to
`
`the pKa of a basic centre of the drug" (P.L. Gould, Salt Selection of Basic Drugs,
`
`International Journal of Pharmaceutics, 33 (1986) 201, at p. 202, Appendix "G"). As
`
`discussed below, salts may be formed even in cases where the pKa value of the acid is
`
`higher than the pKa value of the base. Prof. Serajuddin himself states that "although it
`
`is generally agreed that a successful salt formation requires the pKa of a conjugate acid
`
`to be less than the pKa of the conjugate base to ensure sufficient proton transfer from
`
`the acidic to the basic species, the salt formation still remains a trial and error
`
`endeavor" (Smith, page 26, Appendix "F").
`
`23. In any event, if one were to rely on the pKa difference, any pKa difference would be
`
`considered sufficient. This is exactly the case for sitagliptin phosphate. The first pKa of
`
`sitagliptin is 7.7 and the second pKa of phosphoric acid is 7.21. Accordingly, contrary
`
`to Prof. Serajuddin's assertions, sufficient pKa differences exist for more than one
`
`ionization to occur. The pKa difference does not support any expectation that only the
`
`dihydrogenphosphate can be formed. Just the contrary.
`
`Complete ionization in solution is not necessary to form a salt
`
`24. Prof. Serajuddin further asserts that in order to form a stable salt, the acid must be
`
`completely ionized. According to Prof. Serajuddin's explanations, complete ionization
`
`of the second hydrogen atom of phosphoric acid occurs at a pH of 9.21 (section 49).
`
`Unless a pH of 9.21 is reached, it will be "extremely difficult" to obtain a stable
`
`monohydrogenphosphate salt. This is a total misconception.
`
`25. The aim is to obtain a solid salt, not a salt in solution. A stable solid will be obtained by
`
`precipitation when the solution is super-saturated. When the solution is in equilibrium
`
`at pH 7.21, the concentration of dihydrogenphosphate salt equals the concentration of
`
`monohydrogenphosphate salt. If, under those conditions, the monohydrogen salt
`
`becomes super-saturated and precipitates, it will drive the equilibrium to form more
`
`7
`
`
`
`Merck Exhibit 2221, Page 7
`Mylan Pharmaceuticals Inc. v. Merck Sharp & Dohme Corp.
`IPR2020-00040
`
`

`

`monohydrogenphosphate salt in solution which will further precipitate and so on. It is
`
`therefore not at all necessary to obtain complete ionization in solution and there is no
`
`need to reach a pH of 9.21.
`
`Ammonium phosphate experiments are perfectly valid examples showing that the
`
`pKa of the conjugate acid does not have to be lower than the pKa of the base
`
`26. As indicated above, ammonium phosphate experiments demonstrate that phosphoric
`
`acid has three acidic protons, which can be donated to an appropriate base to form
`
`dihydrogenphosphate, monohydrogenphosphate and phosphate salts. Prof. Serajuddin
`
`argues that this example is "inappropriate and irrelevant" and presents a long list of
`
`irrelevant properties of ammonia as a liquid and as a gas, all intended to demonstrate
`
`that ammonia is not useful in a pharmaceutical environment. Prof. Serajuddin totally
`
`misses the point. Ammonia is able to remove all three protons from phosphoric acid to
`
`make ammonium phosphate. The pKa of aqueous ammonia is 9.21 (see
`
`http://chemweb.unp.ac.za/chemistry/Physical_Data/pKa_compilation.pdf) whereas the
`
`third pKa of phosphoric acid is much higher (12.67). Therefore, the ammonium
`
`phosphate experiments are perfectly valid examples to demonstrate that in order to
`
`create a salt, the pKa of the conjugate acid does not have to be lower than the pKa of
`
`the base.
`
`27. There are abundant additional examples in the literature to the same effect, including
`
`phosphate salts specifically. For instance, bis(4-amino-trans-azobenzene) hydrogen
`
`phosphate salt is reported in the literature (I. Halasz et al, "Hydrogen Phosphate and
`
`Dihydrogen Phosphate Salts of 4-aminoazobenzene", Acta Cryst C63 (2007), o61-o64,
`
`Appendix "H"). The pKa difference between the base and the first pKa of phosphoric
`
`acid (2.12) is 0.70. The pKa difference between the base and the second pKa of
`
`phosphoric acid (7.21) is -4.39 (the pKa value of the amino group of 4-
`
`aminoazobenzene is reported in K.N. Bascombe et al., "Acidity Functions of Some
`
`Aqueous Acids", J. Chem. Soc. (1959), 1096-1104, Appendix "I"). 4-aminoazobenzene
`
`also forms the dihydrogen phosphate phosphoric acid solvate salt (see, I. Halasz et al,
`
`8
`
`
`
`Merck Exhibit 2221, Page 8
`Mylan Pharmaceuticals Inc. v. Merck Sharp & Dohme Corp.
`IPR2020-00040
`
`

`

`Appendix "H"). This salt is formed by the protonation of the azo group of the 4-
`
`aminoazobenzene molecule. The pKa value of the azo group of this molecule is less
`
`basic than the pKa of the amino group (see, G. Cilento, "Resolution of the Over-all
`
`Basicity of the Carcinogenic and Noncarcinogenic Derivatives of 4-
`
`Aminoazobenzene", Cancer Res 20 (1960), 120-124, Appendix "J"). Thus the pKa
`
`difference between the base and the first pKa of phosphoric acid (2.12) is even less
`
`than 0.70.
`
`28. Other examples are guanine phosphate salts. Guaninium dihydrogenphosphate salt and
`
`bis(guaninium) hydrogen phosphate salt are reported in E. Bendeif et al., "Tautomerism
`
`and Hydrogen Bonding in Guaninium Phosphite and Guaninium Phosphate Salt", Acta
`
`Cryst. B63 (2007), 448-458 (Appendix "K") and J.N. Low et al., "Structure of
`
`Bis(guaninium) Hydrogenphosphate 2.5 hydrate", Acta Cryst. C42 (1986), 1045-1047
`
`(Appendix "L"). The pKa differences are 1.18 and -3.91, respectively (the pKa value of
`
`guanine is reported in V. Verdolino et al., "Calculation of pKa Values of Nucleobases
`
`and the Guanine Oxidation Products Guanidinohydantoin and Spiroiminodihydantoin
`
`Using Density Functional Theory and a Polarizable Continuum Model", J. Phys. Chem,
`
`112 (2008), 16860-16873, Appendix "M"). As can be seen, salts are formed not only
`
`when the pKa difference is below 2 or 3 units, but also when the pKa difference is
`
`negative.
`
`29. To sum up, Prof. Serajuddin's arguments that sitagliptin and phosphoric acid can only
`
`form a dihydrogenphosphate salt and that this was the "expectation" based on the pKa
`
`values of the acid and the base are artificial and totally unfounded.
`
`30. In any event, in order to completely rebut Prof. Serajuddin's assertions, I ran some
`
`experiments demonstrating that sitagliptin and phosphoric acids form additional salts.
`
`In all experiments, sitagliptin free base obtained from Merck was used (Batch no.
`
`0000013867). Certificate of Analysis, XRPD and DSC of Batch no. 0000013867 of the
`
`free base of sitagliptin are attached hereto as Appendix "N". The Laboratory Notebook
`
`pages of the experiments are attached hereto as Appendix "O".
`
`9
`
`
`
`Merck Exhibit 2221, Page 9
`Mylan Pharmaceuticals Inc. v. Merck Sharp & Dohme Corp.
`IPR2020-00040
`
`

`

`Bis(sitagliptin) phosphoric acid salt
`
`31. Sitagliptin free base (1.50 g) (0.00368 moles) was combined with isopropanol (3.2 mL)
`
`and distilled water (1.4 mL). The mixture was stirred for 5-10 minutes to form a clear
`
`solution. Phosphoric acid (85% w/w) 0.215 g (0.00186 moles) was added with stirring.
`
`The solution was heated with stirring to 70ºC for 15 minutes, cooled to room
`
`temperature and left stirring overnight. The solution solidified. The solid was dried for
`
`~ 6h at room temperature under vacuum and was analyzed by XRPD (file
`
`0431_sample1_2.xrdml), TGA (file 0431_Sample_1_2.tai) and DSC (file
`
`0431_Sample_1_2.002), Appendix "P". In addition, a sample was analyzed by
`
`elemental analysis (Appendix "Q").
`
`
`
`C
`
`H
`
`N
`
`P
`
`Observed
`
`39.36% 3.83% 14.30% 3.02%
`
`Calculated (sitagliptin)2(H3PO4)(H2O)3
`
`39.76% 4.07% 14.49% 3.20%
`
`32. As can be determined from the analytical results, the compound produced is
`
`bis(sitagliptin)H3PO4 trihydrate, i.e., phosphoric acid can donate two protons to
`
`sitagliptin.
`
`33. This conclusion is further supported by the agreement between the calculated and
`
`experimental values for the nitrogen and phosphorous content that corresponds to the
`
`stoichiometry of a bis(sitagliptin) phosphoric acid ((sitagliptin)2(H3PO4)) salt.
`
`34. To further demonstrate that phosphoric acid can donate two protons to sitagliptin, I also
`
`ran the following similar procedure, this time using methanol as a solvent:
`
`Bis(sitagliptin) phosphoric acid salt – Procedure in Methanol
`
`35. Sitagliptin free base (1.50 g) (0.00368 moles) was combined with methanol (4.6 mL).
`
`Phosphoric acid (85% w/w) (0.21 g) (0.0018 moles) was added to the solution. The
`
`10
`
`
`
`Merck Exhibit 2221, Page 10
`Mylan Pharmaceuticals Inc. v. Merck Sharp & Dohme Corp.
`IPR2020-00040
`
`

`

`solution was heated with stirring to 70ºC for 15 minutes and cooled to room
`
`temperature. A white crystalline powder was formed. The solid was dried overnight at
`
`room temperature under vacuum and was analyzed by XRPD (file
`
`0431_sample8.xrdml), TGA (file 0431_Sample_8.tai) and DSC (file
`
`0431_SAMPLE8.008), Appendix "R". In addition, a sample was analyzed by elemental
`
`analysis (Appendix "S").
`
`
`
`C
`
`H
`
`N
`
`P
`
`Observed
`
`40.81% 3.54% 14.82% 3.39%
`
`Calculated (sitagliptin)2(H3PO4)(H2O)
`
`41.30% 3.79% 15.05% 3.33%
`
`36. As can be determined from the analytical results, the compound produced is
`
`bis(sitagliptin)H3PO4 monohydrate, again demonstrating that phosphoric acid can
`
`donate two protons to sitagliptin.
`
`37. As elaborated above, this conclusion is further supported by the agreement between the
`
`calculated and experimental values for the nitrogen and phosphorous content that
`
`corresponds to the stoichiometry of a bis(sitagliptin) phosphoric acid
`
`((sitagliptin)2(H3PO4)) salt.
`
`Sitagliptin ammonia phosphoric acid salt
`
`38. As another demonstration that the second proton of phosphoric acid can be donated to
`
`sitagliptin, I reacted sitagliptin with ammonia phosphoric acid (NH4H2PO4).
`
`39. Sitagliptin free base (1.50 g) (0.00368 moles) was combined with isopropanol (3.2 mL)
`
`and distilled water (1.4 mL). The mixture was stirred for 5-10 minutes to form a
`
`solution. Ammonia phosphoric acid (NH4H2PO4) (0.42 g) (0.00365 moles) was added
`
`with stirring. The mixture was heated to 70ºC with stirring for 15 minutes and then
`
`cooled to room temperature to yield a white crystalline powder. The solid was dried
`
`overnight at room temperature under vacuum and was analyzed by XRPD (file
`
`11
`
`
`
`Merck Exhibit 2221, Page 11
`Mylan Pharmaceuticals Inc. v. Merck Sharp & Dohme Corp.
`IPR2020-00040
`
`

`

`0431_sample3.xrdml), TGA (file 0431_Sample_3.tai) and DSC (file
`
`0431_SAMPLE3.004), Appendix "T". In addition, the sample was analyzed by
`
`elemental analysis (Appendix "U").
`
`
`
`Observed
`
`C
`
`H
`
`N
`
`P
`
`34.36% 4.21% 14.67%
`
`5.50%
`
`Calculated (sitagliptin) (NH4H2PO4)
`
`33.87% 4.62% 14.81%
`
`5.46%
`
`40. As can be determined from the analytical results, the compound produced is (sitagliptin
`
`NH4H2PO4)·2.5 H2O, i.e., phosphoric acid can donate its second proton to sitagliptin.
`
`This conclusion is further supported by the agreement between the calculated and
`
`experimental values for the nitrogen and phosphorous content that corresponds to the
`
`stoichiometry of sitagliptin ammonia phosphoric acid salt.
`
` Sitagliptin bis(phosphoric acid) salt
`
`41. As a demonstration that sitagliptin can form salt with phosphoric acids by protonation
`
`of two sites, I carried out the following experiment.
`
`42. Sitagliptin free base (0.75 g) (0.00184 moles) was combined with acetone (4.0
`
`mL). The mixture was stirred for 5-10 minutes to form a solution. Anhydrous,
`
`crystalline H3PO4 (0.36 g) (0.00368 moles) was added with stirring. The mixture was
`
`heated to 50°C with stirring for 15 minutes and the clear solution was then cooled to
`
`room temperature. The acetone was removed under vacuum at room temperature. The
`
`white product was crushed with a spatula and then placed under vacuum overnight and
`
`was analyzed by XRPD, TGA and DSC (Appendix "V"). In addition, a sample was
`
`analyzed by elemental analysis (Appendix "W").
`
`
`
`C
`
`H
`
`N
`
`P
`
`Observed
`
`31.80% 3.54% 10.71% 10.10%
`
`12
`
`
`
`Merck Exhibit 2221, Page 12
`Mylan Pharmaceuticals Inc. v. Merck Sharp & Dohme Corp.
`IPR2020-00040
`
`

`

`Calculated (Sitagliptin) (H3PO4)2
`
`32.14% 3.72% 11.26% 9.96%
`
`(with 2% acetone and 1% water residues)
`
`43. As can be determined from the analytical results, the compound produced is
`
`amorphous (Sitagliptin)(H3PO4)2, i.e., sitagliptin forms a salt with 2 molecules of
`
`phosphoric acid by protonation of two of its sites. This conclusion is further supported
`
`by the agreement between the calculated and experimental values for the nitrogen and
`
`phosphorous content.
`
`44. The experiments reported above clearly demonstrate that contrary to Prof. Serajuddin's
`
`unfounded assumptions, the dihydrogenphosphate salt is by no means the inevitable
`
`result of reacting sitagliptin with phosphoric acids. As clearly exemplified in these
`
`experiments, reacting sitagliptin with phosphoric acids result in the formation of
`
`various salts. Therefore, Prof. Serajuddin's unfounded assumptions do not stand up to
`
`scientific reality.
`
`45. In addition to the experiments described above, I ran additional experiments as
`
`described in the attached LNB's. These experiments would require additional work to
`
`isolate or obtain pure compounds. As the experiments reported above unequivocally
`
`rebut Prof. Serajuddin's assertions, I did not find it necessary to fine tune every
`
`experiment that I ran. However, there is no doubt in my mind that, in addition to the
`
`salt claimed in IL 172,563 and the salts reported above, additional phosphoric acid salts
`
`of sitagliptin can be made.
`
`WO '498
`
`46. In sections 44-47 and 54-55 of his declaration, Prof. Serajuddin comments on the WO
`
`'498 patent application. I reviewed WO '498 and, in particular, the sections on pages 9
`
`line 27 – page 10, line 15, on which Prof. Serajuddin relies. It is my clear opinion that
`
`Prof. Serajuddin attempts to incorrectly interpret WO '498 based on hindsight and with
`
`the benefit of Applicant's subsequent research.
`
`13
`
`
`
`Merck Exhibit 2221, Page 13
`Mylan Pharmaceuticals Inc. v. Merck Sharp & Dohme Corp.
`IPR2020-00040
`
`

`

`47. WO '498 describes a large class of beta-aminotetrahydrotriazolo[4,3-a]pyrazine
`
`compounds which are optionally in the form of pharmaceutically acceptable salts. Lists
`
`of potentially suitable salt-forming acids and bases are provided. WO '498 exemplifies
`
`33 compounds belonging to this class, including sitagliptin. All 33 beta
`
`aminotetrahydrotriazolo[4,3-a]pyrazine compounds are exemplified as the
`
`hydrochloride salts. It is therefore incorrect to say that WO '498 directs the skilled
`
`person to sitagliptin, let alone to a phosphate salt of sitagliptin, and not to one of the
`
`many other salt formers listed in WO '498.
`
`48. Even if the skilled person would have opted to make a phosphate salt, numerous
`
`phosphate salts are possible as I already explained above. A phosphate salt is not
`
`synonymous to a dihydrogenphosphate salt. Prof. Serajuddin's assertion that only one
`
`single species of phosphate salt of sitagliptin is possible or expected is erroneous as I
`
`demonstrated theoretically and empirically above. Therefore, any attempt to assert that
`
`WO '498 in any manner renders Applicant's invention "obvious" is totally misleading.
`
`Declaration of Dr. Leonard Chyall
`
`49. I am advised that on October 17, 2010, Opposer submitted an additional declaration of
`
`Dr. Leonard Chyall. At the request of Applicant, I also reviewed Dr. Chyall's
`
`declaration.
`
`50. In his declaration, Dr. Chyall asserts that he ran salt formation experiments for the
`
`purpose of obtaining a phosphate salt other than a 1:1 adduct of sitagliptin and
`
`phosphoric acid. He further asserts that he varied common parameters used in the
`
`experiments in a "deliberate attempt" to obtain different possible salts and that his
`
`experiments "encompassed a wide range of experimental conditions in deliberate
`
`attempts to prepare a phosphate salt other than a 1:1 adduct of sitagliptin and
`
`phosphoric acid" (see, for instance, sections 23 and 72 of Dr. Chyall's declaration). Dr.
`
`Chyall nevertheless reports that he failed to obtain any salt other than the
`
`dihydrogenphosphate salt of sitagliptin. He concludes that his experiments "indicate"
`
`14
`
`
`
`Merck Exhibit 2221, Page 14
`Mylan Pharmaceuticals Inc. v. Merck Sharp & Dohme Corp.
`IPR2020-00040
`
`

`

`that there is "only one possible molecular ratio, a 1:1 ratio", namely,
`
`dihydrogenphosphate salt of sitagliptin.
`
`51. I was surprised to see that Dr. Chyall did not attach to his declaration the laboratory
`
`notebooks describing the experiments which he conducted. This is highly uncustomary
`
`and does not allow a thorough analysis of the experiments. I am advised that
`
`Applicant's counsel requested Opposer to provide the notebooks and that a disc
`
`containing the notebooks was thereafter received on October 28, 2010. A printout of
`
`the laboratory notebooks as received is attached hereto as Appendices "X1", "X2",
`
`"X3" and "X4". Upon review of the notebooks, it transpired that numerous analytical
`
`results or pages were missing, partial or unclear. I am advised that on January 4, 2011,
`
`January 11, 2011 and January 19, 2011, Applicant's counsel requested Opposer to
`
`provide the missing material and information. On February 24, 2011 Opposer provided
`
`another disc with some additional laboratory notebooks and analytical results. A
`
`printout of the additional material is attached as Appendix "Y". Also attached hereto is
`
`the correspondence between the parties' counsel with respect to the missing laboratory
`
`notebooks and data (Appendices "Z1" and "Z2", correspondence in Hebrew and
`
`translations to English).
`
`Dr. Chyall's failed experiments
`
`52. The experiments I ran unequivocally rebut Dr. Chyall's assertions and render Dr.
`
`Chyall's failed attempts to obtain other salts irrelevant. In any event, the procedures
`
`reported by Dr. Chyall do not represent attempts to deliberately obtain different
`
`possible molecular combinations of the sitagliptin and phosphoric acid. As detailed
`
`below, they rather appear as experiments that were designed to fail. Among others, Dr.
`
`Chyall conducted the experiments in temperatures expected to suppress the reaction,
`
`added one of the reagents (phosphoric acid) drop-wise, causing the reaction to start
`
`without the required excess molar amount, used excess solvent causing the salt to
`
`disproportionate, etc.
`
`15
`
`
`
`Merck Exhibit 2221, Page 15
`Mylan Pharmaceuticals Inc. v. Merck Sharp & Dohme Corp.
`IPR2020-00040
`
`

`

`53. Moreover, contrary to Dr. Chyall's assertions, his experiments did not encompass a
`
`wide range of experimental conditions. Just the opposite. As detailed below, the
`
`experimental parameters used by Dr. Chyall were in fact extremely narrow. Among
`
`others, Dr. Chyall conducted his "salt screening" experiments in one single solvent
`
`(methanol), did not try any other solvent system and went to great lengths to avoid the
`
`presence of water in the reactions which he conducted.
`
`54. I will now discuss in detail Dr. Chyall's experiments:
`
`Two experiments are irrelevant
`
`55. Overall, Dr. Chyall reported that he performed 12 experiments. In two of the
`
`experiments, Dr. Chyall used a molar ratio of 1:1.05 between the sitagliptin base and
`
`the acid (Nos. 4063-02-01 and 4063-34-01). In order to obtain salts of different
`
`molecular combinations, one needs to use an excess molar ratio of either the base or the
`
`acid, not a molar ratio which is essentially a 1:1 ratio. Thus, these experiments are
`
`irrelevant.
`
`Five experiments with excess sitagliptin base
`
`56. In five of the experiments (4063-18-01, 4063-04-01, 4063-35-01, 4063-51-01, 4063-
`
`57-01), Dr. Chyall used a molar excess of the sitagliptin base.
`
`Temperature of the reaction
`
`57. It is common to heat organic reactions including salt formation experiments. The
`
`heating is generally at temperatures just below the boiling point of the solvent (in the
`
`case of methanol, 65º C). Among others, the solubility of the compound greatly
`
`increase