Pharmacological Research 103 (2016) 26–48
`
`Contents lists available at ScienceDirect
`
`Pharmacological Research
`
`j o u r n a l h o m e p a g e : w w w . e l s e v i e r . c o m / l o c a t e / y p h r s
`
`Invited review
`Classification of small molecule protein kinase inhibitors based upon
`the structures of their drug-enzyme complexes
`Robert Roskoski Jr. ∗
`
`Blue Ridge Institute for Medical Research, 3754 Brevard Road, Suite 116, Box 19, Horse Shoe, NC 28742-8814, United States
`
`a r t i c l e
`
`i n f o
`
`a b s t r a c t
`
`Article history:
`Received 25 October 2015
`Received in revised form 26 October 2015
`Accepted 27 October 2015
`Available online 31 October 2015
`
`Chemical compounds studied in this article:
`Afatinib (PubMed CID: 10184653)
`Crizotinib (PubMed CID: 11626560)
`Erlotinib (PubMed CID: 176870)
`Gefitinib (PubMed CID: 123631)
`Imatinib (PubMed CID: 5291)
`Nilotinib (PubMed CID: 644241)
`Sorafinib (PubMed CID: 216239)
`Sunitinib (PubMed CID: 5329102)
`Tofacitinib (PubMed CID: 9926791)
`Vemurafenib (PubMed CID: 42611257)
`
`Keywords:
`ATP-binding site
`Catalytic spine
`K/E/D/D
`Protein kinase structure
`Regulatory spine
`Residence time
`
`Contents
`
`Because dysregulation and mutations of protein kinases play causal roles in human disease, this family
`of enzymes has become one of the most important drug targets over the past two decades. The X-ray
`crystal structures of 21 of the 27 FDA-approved small molecule inhibitors bound to their target protein
`kinases are depicted in this paper. The structure of the enzyme-bound antagonist complex is used in
`the classification of these inhibitors. Type I inhibitors bind to the active protein kinase conformation
`␣C-helix in). Type I½ inhibitors bind to a DFG-Asp in inactive conformation while Type II
`(DFG-Asp in,
`inhibitors bind to a DFG-Asp out inactive conformation. Type I, I½, and type II inhibitors occupy part of
`the adenine binding pocket and form hydrogen bonds with the hinge region connecting the small and
`large lobes of the enzyme. Type III inhibitors bind next to the ATP-binding pocket and type IV inhibitors
`do not bind to the ATP or peptide substrate binding sites. Type III and IV inhibitors are allosteric in nature.
`Type V inhibitors bind to two different regions of the protein kinase domain and are therefore bivalent
`inhibitors. The type I–V inhibitors are reversible. In contrast, type VI inhibitors bind covalently to their
`target enzyme. Type I, I½, and II inhibitors are divided into A and B subtypes. The type A inhibitors bind
`in the front cleft, the back cleft, and near the gatekeeper residue, all of which occur within the region
`separating the small and large lobes of the protein kinase. The type B inhibitors bind in the front cleft and
`gate area but do not extend into the back cleft. An analysis of the limited available data indicates that
`type A inhibitors have a long residence time (minutes to hours) while the type B inhibitors have a short
`residence time (seconds to minutes). The catalytic spine includes residues from the small and large lobes
`and interacts with the adenine ring of ATP. Nearly all of the approved protein kinase inhibitors occupy
`the adenine-binding pocket; thus it is not surprising that these inhibitors interact with nearby catalytic
`spine (CS) residues. Moreover, a significant number of approved drugs also interact with regulatory spine
`(RS) residues.
`
`© 2015 Elsevier Ltd. All rights reserved.
`
`1.
`2.
`
`3.
`
`The protein kinase enzyme family. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .27
`Structures of active and inactive protein kinases . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27
`2.1.
`The bilobed protein kinase domain and the K/E/D/D signature motif. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .27
`2.2.
`Structures of the hydrophobic spines in active and dormant protein kinase domains . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29
`Classification of small molecule protein kinase inhibitors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31
`3.1.
`Types of inhibitor. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .31
`3.2.
`A preview of the binding properties
`
`Abbreviations: ALL, acute lymphoblastic leukemia; AS, activation segment; CDK, cyclin-dependent kinase; CML, chronic myelogenous leukemia; CS or C-spine, catalytic
`spine; EGFR or ErbB1, epidermal growth factor receptor; FGFR, fibroblast growth factor receptor; GIST, gastrointestinal stromal tumor; HER 2, human epidermal growth factor
`receptor-2 or human ErbB2; HGFR or c-Met, hepatocyte growth factor receptor; HP, hydrophobic; JAK, Janus kinase; NSCLC, non-small cell lung cancer; PDGFR, platelet-
`derived growth factor receptor; Ph+, Philadelphia chromosome positive; PKA, protein kinase A; pY, phosphotyrosine; RCC, renal cell carcinoma; RS or R-spine, regulatory
`spine; Sh, shell.
`∗ Fax: +1 828 890 8130.
`E-mail address: rrj@brimr.org
`
`http://dx.doi.org/10.1016/j.phrs.2015.10.021
`1043-6618/© 2015 Elsevier Ltd. All rights reserved.
`
`Petitioner Allgenesis Biotherapeutics Inc.
`Exhibit 1017 - Page 1 of 23
`
`

`

`R. Roskoski Jr. / Pharmacological Research 103 (2016) 26–48
`
`27
`
`6.
`
`of inhibitors to protein kinase domains. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .32
`4.
`Type I inhibitors. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .33
`5.
`Type I½ inhibitors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38
`5.1.
`Type I½A inhibitors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38
`5.2.
`Type I½B inhibitors. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .39
`Type II inhibitors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40
`6.1.
`Type IIA inhibitors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40
`6.2.
`Type IIB inhibitors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43
`Type III inhibitors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 44
`Type VI inhibitors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 44
`Drug-target residence time . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 44
`Epilogue. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .45
`10.1.
`Rationale for the current classification of small molecule protein kinase inhibitors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45
`10.2. New therapeutic indications for protein kinase inhibitors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 46
`Conflict of interest . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 46
`Acknowledgment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 46
`References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 46
`
`7.
`8.
`9.
`10.
`
`1. The protein kinase enzyme family
`
`structures and selected properties of all currently FDA-approved
`protein kinase antagonists.
`Nearly all of the approved protein kinase antagonists are steady-
`state competitive enzyme inhibitors with respect to ATP and they
`interact with the ATP-binding pocket. Targeting the ATP-binding
`site of protein kinases was not thought to be selective or effec-
`tive because of the large number of protein kinases and other
`ATP-requiring enzymes with the likelihood that these binding
`sites would be indistinguishable thereby leading to numerous side
`effects [5]. However, the approval of imatinib for the effective treat-
`ment of chronic myelogenous leukemia dispelled this notion. As
`discussed later, structural studies have exploited differences within
`the ATP-binding site and contiguous regions that can be used to
`provide specificity for protein kinase inhibitors. As with the case of
`G-protein coupled receptors, the lesson to be relearned from these
`observations is that drugs can be tailored to bind specifically to
`targets that exhibit only subtle differences.
`
`2. Structures of active and inactive protein kinases
`
`2.1. The bilobed protein kinase domain and the K/E/D/D signature
`motif
`
`We consider first the conformation of the active EGFR pro-
`tein kinase domain as a prototype for all protein kinases. These
`domains have a small N-terminal lobe and a large C-terminal lobe
`␣-helices and
`␤-strands (Fig. 1), first
`that contain several conserved
`described by Knighton and co-workers for PKA [6,7]. The small lobe
`
`␤-sheet (␤1–␤ 5) and
`is dominated by a five-stranded antiparallel
`␣C-helix that contains a glutamate that makes a salt bridge with
`an
`␤3-sheet in the active conformation [8]. The small
`a lysine in the
`lobe contains a conserved glycine-rich (GxGxxG) ATP-phosphate-
`␤1- and
`␤2-strands. The
`binding loop that occurs between the
`␤- and
`␥-phosphates of ATP for catal-
`G-rich loop helps position the
`␤1- and
`␤2-strands harbor the adenine component of
`ysis. The
`ATP.
`The glycine-rich loop is followed by a conserved valine within
`␤2-strand that makes a hydrophobic contact with the adenine
`the
`␤3-strand typically contains an Ala-Xxx-Lys sequence,
`of ATP. The
`␣- and
`␤-phosphates of ATP to the
`the lysine of which couples the
`␣C-helix. A conserved glutamate occurs near the center of the
`␣C-
`helix within the small lobe of protein kinases. The presence of a
`␤3-lysine and the
`␣C-glutamate is a pre-
`salt-bridge between the
`requisite for the formation of the active state and usually, but not
`always, corresponds to the “␣C-in” conformation; by contrast the
`␤3-lysine and the
`␣C-glutamate of the dormant form of EGFR fail
`to make contact in the “␣C-out” conformation (Fig. 1C). The
`␣C-in
`
`Protein kinases are enzymes that play key regulatory roles
`in nearly every aspect of cell biology [1]. These enzymes par-
`ticipate in signal transduction modules that regulate apoptosis,
`cell cycle progression, cytoskeletal rearrangement, differentiation,
`development, the immune response, nervous system function, and
`transcription. Protein kinases represent attractive drug targets
`because their dysregulation occurs in a variety of illnesses including
`cancer, diabetes, and autoimmune, cardiovascular, inflammatory,
`and nervous disorders. Both academic and commercial enterprises
`have expended considerable effort to determine the physiologi-
`cal and pathological functions of protein-kinase signal transduction
`pathways during the past 45 years.
`Protein kinases catalyze the following reaction:
`2– +
`MgATP1– +
`Based upon the nature of the phosphorylated OH group, these
`enzymes are classified as protein-serine/threonine kinases and
`protein-tyrosine kinases. Manning et al. identified 478 typical and
`40 atypical human protein kinase genes (total 518) that corre-
`spond to nearly 2% of all human genes [1]. Moreover, there are
`106 protein kinase pseudogenes. The protein kinase family includes
`385 serine/threonine kinases, 90 protein-tyrosine kinases, and 43
`tyrosine-kinase like proteins. Of the 90 protein-tyrosine kinases, 58
`are receptors with an extracellular, transmembrane, and intracel-
`lular domain and 32 are non-receptors occurring intracellularly.
`There is a small group of enzymes, which includes MEK1 and
`MEK2, that catalyze the phosphorylation of both threonine and
`tyrosine on target proteins. These intracellular proteins, which
`closely resemble and are classified as serine/threonine kinases, are
`called dual specificity kinases. Families of protein phosphatases
`catalyze the dephosphorylation of proteins [2,3] thus making
`phosphorylation–dephosphorylation an overall reversible process.
`Because dysregulation and mutations of protein kinases play
`causal roles in human illnesses, this family of enzymes has become
`one of the most important drug targets over the past two decades
`[4], perhaps accounting for a quarter of all current drug discov-
`ery research and development efforts. Imatinib was the first small
`molecule targeted protein kinase inhibitor that was FDA-approved
`for the treatment of chronic myelogenous leukemia (CML) in 2001.
`The current list of FDA-approved drugs includes 27 orally effec-
`tive direct protein kinase inhibitors that target a limited number of
`enzymes (Table 1). Most of these drugs are used for the treatment of
`malignancies except for ruxolitinib, which is used for the treatment
`of myelofibrosis, and tofacitinib, which is used for the treatment
`of rheumatoid arthritis. See www.brimr.org/PKI/PKIs.htm for the
`
`+
`
`H
`
`+
`
`protein O : PO3
`
` MgADP
`
`→
`
` protein O : H
`
`Petitioner Allgenesis Biotherapeutics Inc.
`Exhibit 1017 - Page 2 of 23
`
`

`

`28
`
`R. Roskoski Jr. / Pharmacological Research 103 (2016) 26–48
`
`Table 1
`Selected drug targets of FDA-approved protein kinase inhibitors.a
`
`Drug target
`
`ALK
`BCR–Abl
`EGFR family
`PDGFR␣/␤
`VEGFRfamily
`c-Met
`RET
`BTK
`JAK family
`Src family
`CDK family
`B-Raf
`MEK1/2
`
`Protein substrate
`
`Receptor
`
`Drug
`
`Tyrosine
`Tyrosine
`Tyrosine
`Tyrosine
`Tyrosine
`Tyrosine
`Tyrosine
`Tyrosine
`Tyrosine
`Tyrosine
`Serine/threonine
`Serine/threonine
`Dual specificity
`
`Yes
`No
`Yes
`Yes
`Yes
`Yes
`Yes
`No
`No
`No
`No
`No
`No
`
`Crizotinib, ceritinib
`Bosutinib, dasatinib, imatinib, nilotinib, ponatinib
`Gefitinib, erlotinib, lapatinib, vandetanib, afatinib
`Axitinib, gefitinib, imatinib, lenvatinib, nintedanib, pazopanib, regorafenib, sorafenib, sunitinib
`Axitinib, lenvatinib, nintedanib, regorafenib, pazopanib, sorafenib, sunitinib
`Crizotinib, cabozantinib
`Vandetanib
`Ibrutinib
`Ruxolitinib, tofacitinib
`Bosutinib, dasatinib, ponatinib, vandetanib
`Palbociclib, sorafenib
`Vemurafenib, dabrafenib
`Trametinib
`
`a Adapted from www.brimr.org/PKI/PKIs.htm.
`
`conformation is necessary, but not sufficient, for the expression of
`full kinase activity.
`The large lobe of protein kinase domains including that of EGFR
`␣-helical with six conserved segments (␣D-␣I) [8]. The
`is mainly
`␤-strands (␤6–␤9)
`large lobe also contains four short conserved
`that contain most of the catalytic residues associated with the phos-
`␣E-helix is followed
`phoryl transfer from ATP to its substrates. The
`␤6-stand, the catalytic loop, the
`␤7- and
`␤8-strands, and
`by the
`␤9-strand. The activa-
`the activation segment, which contains the
`tion segment in the active state forms an open structure extending
`away from the catalytic loop that allows protein/peptide substrate
`binding. The dormant and the active protein kinase domains con-
`␣EF-helix near the end of the activation segment
`tain an additional
`(Fig. 1A).
`Hanks et al. aligned the catalytic domains of 65 protein kinases
`and identified 11 subdomains (I–XI) within the catalytic core of
`these enzymes with conserved amino-acid-residue signatures [9].
`We selected a K/E/D/D (Lys/Glu/Asp/Asp) signature to exemplify
`the catalytic properties of protein kinases. As noted above, an
`␤3-strand lysine (the K of K/E/D/D) forms salt bridges
`invariant
`␣C-glutamate (the E of K/E/D/D). The catalytic loops
`with the
`surrounding the actual site of phosphoryl transfer differ in protein-
`serine/threonine and protein-tyrosine kinases. This loop is made
`up of a YRDLKPEN in the protein kinase AGC family, HRDLKPQN
`for other protein-serine/threonine kinases, HRDLAARN in receptor
`protein-tyrosine kinases, and HRDLRAAN for non-receptor protein-
`tyrosine kinases [9].
`The catalytic loop aspartate, which is the first D of K/E/D/D),
`serves as a base that abstracts a proton from the protein OH group
`of the substrate thereby facilitating the nucleophilic attack of the
`␥-phosphorous atom of ATP. The second
`hydroxyl group onto the
`aspartate of K/E/D/D constitutes the first residue of the so-called
`activation segment. In the majority of protein kinases, this segment
`begins with DFG (Asp-Phe-Gly) and ends with APE (Ala-Pro-Glu).
`Although the activation segment of EGFR begins with DFG, it ends
`with ALE (Ala-Leu-Glu). The difference between active and dormant
`conformations has received considerable emphasis in the char-
`acterization of protein kinases and the DFG-Asp in configuration
`represents the active conformation. However, dormant EGFR also
`occurs with the DFG-Asp in configuration. The DFG-Asp binds Mg2+,
`␣-
`␤- and
`␥-phosphates of ATP. The
`which in turn coordinates the
`␤7-
`primary structure of the catalytic HRD loop occurs before the
`␤8-strands, which are followed by the activation segment. The
`and
`large lobe characteristically binds the peptide/protein substrates at
`a binding loop within the activation segment (Fig. 1A).
`The configuration of the activation segment controls both sub-
`strate binding and catalysis [10]. Moreover, the initial five residues
`of the activation segment make up the magnesium-positioning
`loop. The activation segment in EGFR contains a phosphorylatable
`
`Fig. 1. (A) Active conformation of EGFR with DFG-D pointing inward toward the
`␣C-helix directed inward toward both the
`␤3-K and N-terminal
`active site and the
`region of the activation segment. (B) Inactive DFG-D out conformation. (C) Inactive
`␣C-helix out conformation. (D) Superposition of active
`␣C-helix in
`DFG-D in and
`␣C-helix out (4HJO) conformations. (E) Active EGFR with DFG-D
`(1M17) and inactive
`directed inward toward the active site. (F) Inactive Abl with DFG-D directed outward
`from the active site. The human enzyme and corresponding PDB ID are listed. AS,
`activation segment; CL, catalytic loop. Figs. 1, 2, 5, and 7 were prepared using the
`PyMOL Molecular Graphics System Version 1.5.0.4 Schrödinger, LLC.
`
`Petitioner Allgenesis Biotherapeutics Inc.
`Exhibit 1017 - Page 3 of 23
`
`

`

`R. Roskoski Jr. / Pharmacological Research 103 (2016) 26–48
`
`29
`
`tyrosine. The magnesium-positioning loop, the amino-terminus of
`␣C-helix, and the conserved HRD component of the catalytic
`the
`loop are close together in the tertiary structure. In the analysis of
`the structures of some two dozen active protein kinases, Nolan et al.
`reported that phosphorylation of one or more residues within the
`activation segment is required for activation [11]. However, such
`phosphorylation is not required for the activation of EGFR [12,13].
`Although the tertiary structure of catalytically active protein
`kinase domains are similar, Huse and Kuriyan noted that the crystal
`structures of dormant enzymes reveal distinct inactive conforma-
`tions [14]. One of the most common inactive enzyme forms is the
`DFG-Asp out conformation. When this aspartate is directed out-
`ward, the DFG-F is directed into the active site (Fig. 1B). Fig 1E
`depicts the active DFG-D in configuration and Fig. 1F shows the
`inactive DFG-D out conformation. Another commonly occurring
`␣C-helix out state (Fig. 1C) [10]. The
`inactive conformation is the
`␣C-helix out conforma-
`superposition of an active enzyme and the
`tion illustrates the differences in these two states (Fig. 1D). Also
`␣C-out conformation is in
`note that the activation segment in the
`an inactive closed state. To summarize, the three main regulatory
`␣C-
`elements within the kinase domain include the N-terminal lobe
`helix (␣C-in, active;
`␣C-out, inactive), the C-terminal lobe DFG-Asp
`(DFG-Asp in, active; DFG-Asp out, inactive), and the C-terminal lobe
`activation segment (AS open, active; AS closed, inactive). We also
`consider the structure of the regulatory spine, which is considered
`in the next section. It has a near linear structure in the active state
`and a distorted conformation in the inactive state.
`
`2.2. Structures of the hydrophobic spines in active and dormant
`protein kinase domains
`
`Kornev and co-workers compared the structures of the active
`and inactive conformations of 23 protein kinases and determined
`functionally important residues by a local spatial pattern align-
`ment algorithm [15,16]. Their analysis led to the description of the
`structural skeletons of eight hydrophobic residues that constitute a
`catalytic or C-spine and four hydrophobic residues that constitute a
`regulatory or R-spine. Each spine consists of amino acids occurring
`in both the small amino-terminal lobe and the large carboxytermi-
`␣C-helix
`nal lobe. The regulatory spine contains residues from the
`and the activation segment, both of which are important in charac-
`terizing active and dormant states. The adenine base of ATP is one
`component of the catalytic spine. The regulatory spine positions
`the protein substrate and the catalytic spine positions ATP thereby
`enabling catalysis. The structure of the spines differs between the
`active and dormant enzyme states and their correct alignment is
`necessary for the assembly of an active protein kinase domain.
`The protein kinase regulatory spine consists of a residue from
`␤4-strand, from the C-terminal end of the
`the beginning of the
`␣C-helix, the phenylalanine of the activation segment DFG, along
`with the H/YRD-His/Tyr of the catalytic loop. The spinal component
`␣C-helix is four residues C-terminal to the conserved
`␣C-
`from the
`glutamate. The backbone of the catalytic loop histidine/tyrosine is
`␣F-helix by a hydrogen bond to a conserved aspar-
`anchored to the
`␣F-helix. Going from the aspartate within
`tate residue within the
`␣F-helix up to the top residue within the
`␤4-strand, Meharena
`the
`et al. labeled the residues RS0, RS1, RS2, RS3, and RS4 (Figs. 2 and 3)
`[10].
`The regulatory spine of active protein kinase domains is nearly
`linear (Fig. 2A) while that of the dormant enzymes possesses vari-
`ous distortions. In the inactive DFG-Asp out form, the DFG-F residue
`(RS2) is displaced into the active site and separated from RS3/4; this
`␣C-helix out confor-
`form of the spine is broken (Fig. 2B). In the
`mation, RS3 is displaced away from the active site along with the
`␣C-helix (Fig. 2C). Besides these inactive forms, Meharena et al.
`described two less common inactive forms [10]. In the inactive
`
`␣C-helix in active confor-
`Fig. 2. (A) Frontal view of EGFR with its DFG-D in and
`mation. The space filling models on the left depict the C-spine and those on the
`␣C-helix out.
`right depict the R-spine. (B) Inactive I: DFG-Asp out. (C) Inactive II:
`(D) Inactive III: HRD-His out. (E) Inactive IV: twisted lobe conformation. (F) Rear
`␣C-␤4 or back loop residues that
`view of active EGFR to indicate the location of the
`interact with many small molecule protein kinase antagonists. AS, activation seg-
`ment; GF, gefitinib; RS, regulatory spine. All structures are human proteins except
`for D, which is from yeast. Classification of active and inactive forms adapted from
`Meharena et al. [10].
`
`Petitioner Allgenesis Biotherapeutics Inc.
`Exhibit 1017 - Page 4 of 23
`
`

`

`30
`
`R. Roskoski Jr. / Pharmacological Research 103 (2016) 26–48
`
`eAdapted from Lamba and Ghosh [22].
`dAdapted from Gavrin and Saiah [21].
`cAdapted from Zucotto et al. [20].
`bA, drug extends into back cleft; B, drug does not extend into back cleft.
`aAdapted from Dar and Shokat [18].
`
`Usually not
`No
`Variable
`Variable
`Variable
`Variable
`Variable
`
`Yes
`Variable
`Variable
`Variable
`Variable
`Variable
`Variable
`
`Yes
`No
`Variable
`Variable
`Variable
`Variable
`No
`
`Yes
`No
`Distorted
`Out
`Variable
`Variable
`Yes
`
`Covalent inhibitor
`
`spanning two regions
`Bivalent inhibitor
`
`ATP-site
`bound next to the
`Allosteric inhibitor not
`
`ATP-site
`bound next to the
`Allosteric inhibitor
`
`Yes
`Yes
`Usually distorted
`Variable
`Variable
`Out
`(A) Yes/(B) no
`out conformation
`the inactive DFG-Asp
`ATP-binding pocket of
`Binds in the
`
`Yes
`Yes
`Distorted
`Variable
`Variable
`In
`(A) Yes/(B) no
`conformation
`an inactive DFG-Asp in
`ATP-binding pocket of
`Binds in the
`
`Yes
`Yes
`Linear
`In
`Out
`In
`No
`conformation
`the active
`ATP-binding pocket of
`Binds in the
`
`Reversible
`ATP-competitive
`Spine
`␣C
`Activation segment
`DFG-Asp
`Extends into back cleft
`
`Type VI
`
`Type Ve
`
`Type IVd
`
`Type IIIa
`
`Type IIa(A/B)b
`
`Type I1/2c(A/B)b
`
`Type Ia
`
`Properties
`
`Classification of small molecule protein kinase inhibitors.
`Table 2
`
`Fig. 3. Frontal view of the C- and R-spines of PKA. Note that RS1 in most protein
`kinases is histidine and not tyrosine [9]. The dashed line represents a hydrogen bond.
`Modeled after PDB ID: 1ATP. CS, C-spine; GK, gatekeeper; RS, R-spine; Sh, shell.
`
`HRD-His out structure, the catalytic loop histidine (RS1) is sep-
`␣F-helix (RS0) (Fig. 2D). In the
`arated from the aspartate of the
`inactive twisted lobe structure, the N-lobe is rotated with an addi-
`tional separation of the N-lobe (RS3) from the C-lobe (RS2) (Fig. 2E).
`The catalytic spine of protein kinases consists of residues from
`the small amino-terminal and large carboxyterminal lobes that are
`completed by the adenine of ATP (Fig. 3) [10,16]. This spine medi-
`ates catalysis by facilitating ATP binding thereby accounting for
`the term catalytic. The two residues of the small lobe of protein
`kinase domains that bind to the adenine component of ATP include
`␤3-strand (CS8)
`the alanine from the conserved Ala-Xxx-Lys of the
`␤2-
`and a hydrophobic valine residue from the beginning of the
`strand (CS7) (Figs. 2A and 3). Furthermore, a hydrophobic residue
`␤7-strand (CS6) binds to the adenine base
`from the middle of the
`in the active enzyme. This residue is flanked by two hydrophobic
`residues (CS4 and CS5) that bind to a residue near the beginning of
`␣D-helix (CS3). CS3 and CS4 interact with two residues of the
`the
`␣F-helix (CS1 and CS2) to complete the C-spine (Fig. 3). Using site-
`directed mutagenesis, Meharena et al. identified three residues in
`murine PKA that stabilize the R-spine that they labeled Sh1, Sh2,
`and Sh3, where Sh refers to shell [10]. The Sh2 residue corresponds
`to the gatekeeper residue. The name gatekeeper signifies the role
`that this residue plays in controlling access to the back cleft, which
`is described in Section 3.1. The back cleft is sometimes called the
`back pocket or hydrophobic pocket II (HPII).
`␣F-helix, which is entirely within the protein
`Note that the
`kinase domain, anchors both the R-spine and C-spine. Moreover,
`the spines in turn position the protein kinase catalytic residues.
`The residues that constitute the spines were identified by a com-
`parison of the tertiary structure of 23 protein kinases in their active
`and inactive states based upon their X-ray crystallographic struc-
`tures [15,16]. This contrasts with the identification of the DFG, APE,
`or HRD amino acid signatures based upon their primary structures
`[9].
`Besides the hydrophobic interactions with the adenine group,
`the exocyclic 6-amino nitrogen of ATP characteristically forms
`a hydrogen bond with the carbonyl backbone residue of the
`first hinge residue that connects the small and large lobes of
`the protein kinase domain and the N1 nitrogen of the ade-
`nine base forms a second hydrogen bond with the N H group
`
`Petitioner Allgenesis Biotherapeutics Inc.
`Exhibit 1017 - Page 5 of 23
`
`

`

`R. Roskoski Jr. / Pharmacological Research 103 (2016) 26–48
`
`31
`
`Fig. 4. Schematic overview of the binding pockets of the DFG-Asp in and DFG-Asp out protein kinase conformations. AP, adenine pocket; BP, back pocket; FP, front pocket;
`GK, gatekeeper; Hn, hinge; HP, hydrophobic. Adapted from van Linden et al. [24].
`
`of third hinge residue. As noted
`later, most small-molecule
`inhibitors of protein kinases that are steady-state ATP compet-
`itive inhibitors also make hydrogen bonds with the backbone
`residues of the connecting hinge [17]. These antagonists also inter-
`act with residues that make up the C-spine, the R-spine, or both,
`and wi