Misbehaving Proteins
`Protein (Mis)Folding, Aggregation,
`and Stability
`
`Exhibit 2072
`
`Edited by
`
`Regina M. Murphy
`University of Wisconsin
`Madison, Wisconsin
`
`Amos M. Tsai
`Human Genome Sciences
`
`Rockville, Maryland
`
`Q) Springer
`
`Mylan v. Regeneron
`IPR2021-00880
`U.S. Pat. 9,669,069
`
`Exhibit 2072
`Page 01 of 15
`
`

`

`Regina M. Murphy
`Department of Chemical and Biological
`Engineering
`University of Wisconsin
`15 Engineering Drive
`Madison, Wisconsin 53706-1691
`USA
`
`regina @engr.wisc.edu
`
`AmosM. Tsai
`Human GenomeSciences
`14200 Shady Grove Road
`Rockville, Maryland 20850
`USA
`
`amos-_tsai@hgsi.com
`
`Steenbock Memorial Library
`University of Wisconsin - Madison
`550 Babcock Drive
`Madison, WI 53706-1293
`
`Library of Congress Control Number: 2005938663
`
`ISBN-10: 0-387-30508-4
`ISBN-13: 978-0387-30508-0
`
`Printed on acid-free paper.
`
`© 2006 Springer Science+Business Media, LLC.
`Allrights reserved. This work maynotbe translated or copied in wholeorin part without the written permission
`of the publisher (Springer Science+Business Media, LLC, 233 Spring Street, New York, NY 10013, USA),
`except for brief excerpts in connection with reviews or scholarly analysis. Use in connection with any form
`of information storage andretrieval, electronic adaptation, computer software, or by similar or dissimilar
`methodology now knownor hereafter developedis forbidden.
`The use in this publication of trade names, trademarks, service marks, and similar terms, even if they are
`notidentified as such, is not to be taken as an expression of opinion as to, whetheror not they are subject to
`proprietary rights.
`
`Printed in the United States of America.
`
`(TB/M¥V)
`
`10987654321
`springer.com
`
`General Library System
`University of Wisconsin - Madison
`728 State Street
`Madison, WI 53706-1494
`U.S.A.
`
`Exhibit 2072
`
`Page 02 of 15
`
`Exhibit 2072
`Page 02 of 15
`
`

`

`Contents
`
`Part I. Introduction
`
`Protein Folding, Misfolding, Stability, and Aggregation: An Overview........
`Regina M. Murphy and Amos M.Tsai
`
`Part II. Mathematical Models and Computational Methods
`
`Nonnative Protein Aggregation: Pathways, Kinetics, and Stability
`PrSdiGtiONn ss seas oececsasews on ag we ee eleOA Oe ot tn veel eantate
`Christopher J. Roberts
`
`Simulations of Protein Aggregation: A Review ...........0. eee er sree eee e eee
`Carol K. Hall, Hung D. Nguyen, Alexander J. Marchut,
`and Victoria Wagoner
`
`Part III. Experimental Methods
`
`Elucidating Structure, Stability, and Conformational Distributions during
`Protein Aggregation with Hydrogen Exchange and Mass Spectrometry .....
`Erik J. Fernandez and Scott A. Tobler
`
`Application of Spectroscopic and Calorimetric Techniques in Protein
`Formulation Development... «0... 200.0600 cs ce ba ee eee eee ne eee cn ne ee an
`Angela Wilcox and Rajesh Krishnamurthy
`Small-Angle Neutron Scattering as a Probe for Protein Aggregation
`al. Many Lenst SCales:. 0 cco eacsccrsiieieeaed ub bi oo oe HOR RSRONET ER OS HE EG BB
`Susan Krueger, Derek Ho, and Amos Tsai
`
`Laser Light Scattering as an Indispensable Toolfor
`Probing Protein: ApPresation woccc cuca us a ee eanaenneasenes ae os oa cena
`Regina M. Murphy and Christine C. Lee
`
`X-Ray Diffraction for Characterizing Structure in Protein Aggregates.........
`Hideyo Inouye, Deepak Sharma, and Daniel A, Kirschner
`
`Glass Dynamics and the Preservation of Proteins ............-.....-2-..0055
`Christopher L. Soles, Amos M. Tsai, and Marcus T. Cicerone
`
`3
`
`17
`
`47
`
`81
`
`99
`
`125
`
`147
`
`167
`
`193
`
`vii
`
`Exhibit 2072
`
`Page 03 of 15
`
`Exhibit 2072
`Page 03 of 15
`
`

`

`CONTENTS
`
`Part IV. Fundamental Studies in Model Systems
`
`Folding and Misfolding as a Function of Polypeptide Chain Elongation:
`Conformational Trends and Implications for Intracellular Events........
`Silvia Cavagnero and Nese Kurt
`
`et
`
`F207
`
`Determinantsof Protein Folding and Aggregation in P22 Tailspike Protein ....
`Matthew J. Gage, Brian G. Lefebvre, and Anne S. Robinson
`
`247
`
`Factors Affecting the Fibrillation of a-Synuclein, a Natively
`Wintolded Proteiiicrsatae: a eas i er lc ae veseniomnnrrensonons dt em
`Anthony L. Fink
`
`view
`
`§=265
`
`Molten Globule-Lipid Bilayer Interactions and Their Implications
`for Protein Transport and Aggregation..............0.0ccecceceeeeeee
`sam«=eee
`Lisa A. Kueltzo and C. Russell Middaugh
`
`Part V. Protein Product Development
`
`Self-Association of Therapeutic Proteins: Implications for
`Product: Developments oa ss 6: is vewvaseavenes da os ey 4 A EEL dw
`a»_31S
`Mary E. M. Cromweil, Chantel Felten, Heather Flores, Jun Liu,
`and Steven J. Shire
`
`Mutational Approach to Improve Physical Stability of Protein
`Therapeutics Susceptible to Aggregation: Role of Altered
`Conformationin Irreversible Precipitation ...............0..0.c0eee0es
`Margaret Speed Ricci, Monica M. Pailitto, Linda Owens Narhi,
`Thomas Boone, and David N. Brems
`
`5 a
`
`331
`
`TNOOX: we wayyy crae es eA Bo WU Gencteecacosncecannomencaceve tee moe een dev wininvaumcaramecaraiacauesacanee «
`
`soe
`
`OO
`
`Exhibit 2072
`
`Page 04 of 15
`
`Exhibit 2072
`Page 04 of 15
`
`

`

`This material may be protected by Copyright law (Title 17 U.S. Code)
`
`Exhibit 2072
`Page 05 of 15
`
`

`

`4
`
`R. M. MURPHY AND A. M. TSAI
`
`antibodies are often purified using protein A chromatography that requires acid pH for
`elution. Viral inactivation required for mammalian cell-derived products is carried out
`in even harsher acidic conditions. Acid denaturation of the protein product as well as
`other host cell biomolecules leads to aggregation and loss of product. Depthfiltration,
`diafiltration, and other similar processes can cause aggregation due to shear-induced
`protein denaturation. A protein product that survives purification processes then must
`be formulated for a useful shelf life, to maintain a drug’s stability until the point of
`administration. Cosolutes andother excipients often are recruited to improve a product's
`stability. As a greater diversity of proteins reach the market, the industry must incorpo-
`rate new understanding of mechanisms of protein misfolding and aggregation in order to
`develop robust manufacturing and formulation processesthatresult in stable, correctly
`folded, and active products.
`
`1.2. Protein Misfolding Diseases
`
`The problemofprotein structuralinstability, misfolding, and aggregationis notlim-
`ited to manufacturing. The “protein misfolding” diseases constitute a newly recognized
`group of diseases with a diverse array of symptoms. Some of the most commonprotein
`misfolding diseases are neurodegenerative, including Alzheimer’s disease, Parkinson's
`disease, Huntington’s disease, and the prion diseases. These diseases share a common
`feature: the deposition of insoluble, usually fibrillar, B-sheet-rich protein aggregates.
`The source and nature of the aggregating protein, the location of the deposit, and the
`biological consequencesdiffer from disease to disease (Table 1). The factors that trig-
`ger formation of aggregates and the mechanisms by which aggregation leadsto disease
`are poorly understood. Developmentof effective treatment and prevention therapies for
`these diseases requires elucidation of the molecular basis for protein misfolding and
`aggregation.
`
`2. PROTEIN STRUCTURE AND PROTEIN FOLDING
`
`2.1. Amino Acids
`
`Twenty amino acids makeupthe library from which naturalproteins are synthesized
`(Table 2). Within their side chains is contained a wide diversity of chemical function.
`Table 3 summarizes important physicochemical properties of these amino acids. At a
`fundamental level, these properties direct the folding, misfolding, and aggregation of
`proteins. It should be noted that many different scales have been proposed to measure
`the relative hydrophobicity of the side chains.
`Covalent modifications of these side chains by reactions such as phosphorylation,
`glycosylation, oxidation, or deamidation introduced by design or by accident further in-
`crease chemicaldiversity and affect protein structure and stability. In a manufacturing en-
`vironment,the ability to control, minimize, or completely eliminate these modifications
`directly contributes to the quality attributes of a product. Forinstance, the glycosylation
`pattern of a monoclonal antibody can be part of the release specifications and one of
`
`Exhibit 2072
`
`Page 06 of 15
`
`Exhibit 2072
`Page 06 of 15
`
`

`

`PROTEIN MISFOLDING AND AGGREGATION
`
`5
`
`TABLE|. Protein misfolding diseases!?
`
`Majoraffected
`Aggregating
`
`Disease Nature of deposits—_brain regionsprotein Characteristics of protein
`
`
`
`
`
`Alzheimer
`
`Beta-amyloid
`
`Huntington
`
`Huntingtin
`
`4-kDa peptide cleaved from Extracellular
`the membrane-bound
`amyloid fibrils
`precursor protein APP
`
`350-kDaprotein, with an
`expanded (>35)
`polyglutamine domain in
`the disease state; number
`of glutamines correlates
`inversely with age of
`onset
`
`Intranuclear
`inclusions,
`cytoplasmic
`aggregates
`
`Hippocampus,
`cortex
`
`Striatum, basal
`ganglia
`
`Parkinson
`
`Alpha-
`synuclein
`
`14-kDanatively unfolded
`protein
`
`Substantia nigra
`
`Lewy bodies, a
`cytoplasmic
`inclusion body
`often localized
`near the nucleus
`
`
`
`Prion diseases—_Prion protein ~34-kDa glycoprotein that Extracellular and Cortex, thalamus,
`
`(kuru, CJD*,
`is normal cell-surface
`intracellular
`brain stem,
`
`others) cerebellum component of neurons amyloid deposits
`
`
`
`“Creutzfeldt-Jakob disease
`
`the quality attributes that defines lot-to-lot consistency. Oxidation or deamidation are
`considered product-related impurities raising regulatory concerns because such product
`variants could affect a drug’s potency, modeof action, and immunogenicity. Control of
`the level of covalent modifications consistently is required for productrelease. In protein-
`folding diseases, covalent modifications are known to affect the aggregation propensity
`and aggregate morphologyof relevant proteins.
`Unnatural aminoacids are readily incorporated into peptides and short proteins by
`solid-phase chemical synthesis. Biosynthetic routes for incorporation of a few analogs
`of natural amino acids recently have been developed.® These methods providea tech-
`nology base that could lead to development ofa great variety of designer proteins with
`highly tunable physicochemical properties. Controlling the folding of these designer pro-
`teins and preventing aggregationor facilitating self-assembly (depending on the specific
`application) present formidable challenges.
`
`2.2. Forces Driving Folding, Misfolding, and Aggregation
`
`Correct folding of a polypeptide chain, containing perhaps ~ 100 aminoacids,into a
`compactstructure is truly a remarkable accomplishmentof nature. The possible configu-
`rationsthat a polypeptide can sample during folding are enormous. Yet a typical folding
`process is extremely rapid, taking place in milliseconds to seconds, and a folded struc-
`ture uniqueto that particular chain is reliably generated. This suggests that folding of a
`denatured chain proceeds through multiple pathways and still moves toward the same
`
`Exhibit 2072
`
`Page 07 of 15
`
`Exhibit 2072
`Page 07 of 15
`
`

`

`R. M. MURPHY ANDA. M.TSAI
`
`TABLE2. Chemical structures of common 20 amino acids
`
`
`Glycine, Gly, G|Alanine, Ala, A|Valine, Val, V|Leucine, Leu, L Isoleucine,Ile, |
`°
`
`H,N—CH—C—OH
`H,N—CH-C—OH
`sdima
`es
`CH,
`ot
`CH,
`GiGi,
`CH,
`CH,
`
`
`Aspartic acid,|Glutamic acid,|Asparagine, Glutamine, Gin,| Lysine, Lys, K] Arginine, Arg,
`oO
`Asp, D
`Glu, E
`Asn, N
`Q
`R
`I
`iiean
`H,
`—7
`OH
`
`ps
`cH,
`|
`ve
`NH
`—NH
`
`|N
`
`H,N—CH—C— OH
`
`:
`
`|
`CH,
`vi
`oa
`OH
`
`|
`HN—GH-C—OH
`vig
`oa
`NH,
`
`Ms
`oe
`i
`NH,
`
`“Me
`ie
`Ghe
`NH,
`
`
`
`
`
`
`
`
`
`
`H,
`
`
`
`Serine, Ser, S Threonine, Thr,|Cysteine, Cys,|Methionine, Proline, Pro, P
`1
`T
`Cc
`Met, M
`i
`o
`1
`H,N—CH—C—OH
`G—OH
`H3N
`H—C—OH llai H,N——GH—C—OH
`om
`on
`H—OH
`He
`i
`CH,
`SH
`rtes
`
`HN
`
`|C
`
`H
`
`
`| Histidine, His,H]Phenylalanine,|Tyrosine, Tyr,.|Tryptophan,
`o
`Phe, F
`Y
`Trp, W
`°
`oO
`H,N—CH—C—OH
`CH,
`
`° I
`
`HN
`
`H—C —OH ee H,N—CH—C—OH
`CH.
`CH.
`CH.
`2
`2
`2
`
`a
`Hi
`
`“ALi
`
`L
`
`final intended structure determined by the amino acid sequence. This view assumesthat
`the proper foldedstructure has the lowestfree energy and thatthefree energy of folding
`acts to guide the chain along different pathways that lead to the final structure.’ It is
`important to recognize that a foldedprotein is a collection of closely related structures
`in equilibrium with each other. The conformational distribution of all the states in the
`ensemble allowsa protein to perform its function, but at the same time makesit prone to
`denaturation. The forces that contribute toward the overall folding free energy include
`hydrogen bonding, hydrophobic interactions, electrostatics, and conformational or en-
`tropic forces. Correct folding of a protein requires a delicate balance of forces between
`different parts of the polypetide chain and between the polypeptide and surrounding
`water molecules. Interestingly enough, the sameforces that drive protein folding also
`
`Exhibit 2072
`
`Page 08 of 15
`
`Exhibit 2072
`Page 08 of 15
`
`

`

`PROTEIN MISFOLDING AND AGGREGATION
`
`7
`
`~12
`
`3.9-4.8
`8.8-9.5
`
`3.9-4.8
`
`TABLE3. Key physicochemical properties of amino acid side chains?~>
`Aminoacid
`Molecular
`pKa ofside chain
`Accessible surface
`Hydrophobicity of side
`residue
`weight (Da)
`(in polypeptides)
`area (A?)
`chain analogs (kJ/mole)
`
`Alanine
`71
`113
`—3.65
`Arginine
`156
`241
`66.61
`Asparagine
`114
`158
`21.83
`Aspartic acid
`115
`151
`40.57
`Cysteine
`103
`140
`—1.42
`Glutamine
`128
`189
`27.21
`Glutamic acid
`129
`183
`32:55
`Glycine
`57
`85
`0
`Histidine
`137
`194
`23.52
`Isoleucine
`113
`182
`—16.71
`Leucine
`113
`180
`—16.71
`Lysine
`128
`211
`27.25
`Methionine
`131
`204
`—§5.92
`Phenylalanine
`147
`218
`—8.57
`Proline
`97
`143
`18.23
`Serine
`87
`122
`14.74
`Threonine
`101
`146
`—5.84
`Tryptophan
`186
`259
`4.54
`9.4—10.8
`Tyrosine
`163
`229
`
`99 160Valine —13.02
`
`
`
`6-7.5
`
`9.8—11.1
`
`drive protein misfolding and aggregation. Indeed, perhaps the relevant question is not
`why someproteins misfold and aggregate, but why mostdo not!
`2.2.1. Hydrogen bonding Hydrogen bonds between carbonyl and amide groups along
`the polypeptide backbonestabilize the basic structural elements of folded proteins, Al-
`though formation of a hydrogen bond between two moieties on the backbone means the
`loss of hydrogen bonds between peptide amides and water, the multivalency possible
`in long helices and B-sheets stabilizes these structuresrelative to the unfolded protein.
`Urea (H2N-CO-NH2) and the guanidinium ion (H2N-CNH; -NH2) denature proteins in
`part by competing for hydrogen bondswith the polypeptide backbone.
`Hydrogen bonding,althoughit can explain the stability of folded proteins, cannot
`explain by itself why one folded structureis native and anotherfolded structureis not.
`Intermolecular hydrogen bondingis importantfor stabilizing aggregated proteins with
`defined structural elements such as B-sheet-rich amyloidfibrils.
`Several aminoacid side chainsalso are capable of participating in hydrogen bonds.
`Of particular note are the amide side chains glutamine and asparagine. Expanded
`polyglutamine domainsare involved in the abnormal protein aggregates occurring in
`Huntington’s disease as well as several less commondiseases.It is believed that hydrogen
`bonding between glutamine side chain and backbone amidesstabilize these aggregates.
`2.2.2. The hydrophobic effect Burial of hydrophobic side chains in the protein core
`(hydrophobic collapse)is essential for the development of tertiary structure; indeed,col-
`lapse to a molten globule state may sometimes precede formation of secondary structural
`elements such as helices. The driving force for burial can be considered most simply as
`
`Exhibit 2072
`
`Page 09 of 15
`
`Exhibit 2072
`Page 09 of 15
`
`

`

`8
`
`R. M. MURPHY AND A. M. TSAI
`
`the difference betweenthe weak attractive van der Waals interactions between nonpolar
`groupsand the weak attractive van der Waalsinteractions of nonpolar groups with water.
`The hydrophobic effectis unusualin thatit increases with temperature, whereasthe other
`forces favoring protein structure decrease with temperature. This changein the relative
`importanceofforces often results in misfolding and aggregationat higher temperatures.
`2.2.3. Coulombic interactions The basic aminoacidslysine andarginine andthe acidic
`aminoacids glutamate and aspartate carry charge at neutral pH.Histidine also may be
`charged at neutral or slightly acidic pH. Charged residues most often are found on
`the exterior of a correctly folded protein, but ionic pairing between a positively and
`negatively charged residue may occur in the interior of a protein; such an interaction
`may be unusually strong because ofthe very low dielectric constantin the interior of a
`folded protein.
`Perhaps of more importance,althoughless easily understood,is the role of dipole-ion
`and dipole-dipole interactionsin stabilizing protein structure. The peptide bonditself as
`well as manyside chainsfunction as dipoles becauseofthedifferent electronegativities of
`the atoms. For example,interactions betweenthe tyrosine dipole and charged aminoacids
`(e.g., glutamate, aspartate) stabilize somefolded protein structures.® Protein aggregation
`and insolubility with pH adjusted to at or near the isoelectric point is a well-known
`phenomenon,causedbythe loss of repulsive electrostatic interactions.
`2.2.4. Disulfide bondformation Disulfide bonds are formed during folding when two
`cysteine side chains are brought into close contact under oxidizing conditions. These
`covalent bondscross-link tether two sections of a polypeptide chain andstabilize folded
`structure. For polypeptides with multiple cysteines, incorrect disulfide bond formation
`maylock the protein in a misfolded conformation and could lead to further aggregation.
`
`2.3. Role of Cosolutes
`Proteins and peptides operate in a complex environmentof salts, other proteins,
`carbohydrates,lipids, and other solutes. Eachofthese cosolutesinfluencesthe folding and
`aggregation propertiesofa protein. Cosolutes maybe classified as kosmotropes(structure
`stabilizers) or chaotropes (structure destablizers). Both correctly folded proteins and
`misfolded structured protein aggregates are more structured than unfolded polypeptide
`chains; thus, addition of kosmotropes may enhance aggregation as much,if not more,
`than it enhancescorrect folding.
`The Hofmeisterseries is a useful tool for correlating the chaotropic or kosmotropic
`nature of cosolutes (Figure 1). In general, kosmotropesact by a preferential exclusion
`(also called preferential hydration) mechanism.Essentially, kosmotropes are preferen-
`tially excluded from the protein-solventinterface. As a result, the water concentration
`near the interface is higher than in the bulk solvent and there is an increased driving
`force for burial of hydrophobic residues. This stabilizes folded (or misfolded) protein
`structures. Stabilization of protein structure correlates well with the kosmotrope’s ability
`to increase the surface tension of water.’ Chaotropic action generally is believed to be
`mediated by preferential binding. The chaotrope preferentially associates with specific
`chemical moieties on the side chain or backbone, thus favoring increased protein-solvent
`interfacial area and therefore unfolding. The behaviorof some cosolutes suchasarginine
`
`Exhibit 2072
`
`Page 10 of 15
`
`Exhibit 2072
`Page 10 of 15
`
`

`

`PROTEIN MISFOLDING AND AGGREGATION
`
`9
`
`KOSMOTROPES
`
`CHAOTROPES
`
`F-
`
`PO,
`
`S0,2- CH,;COO-
`
`(CH3),N*
`
`(CH))NH,” NH,*
`
`K*
`
`cr Be
`
`Na* Cs”
`
`Fr
`
`Lit
`
`CNS
`
`Mg** Ca?
`
`betaine
`
`arginine
`
`ied g
`a
`‘NH —C—HN—(CHs)—cH-L
`o-
`>
`
`lysine
`
`NH;* o
`*NH,—(CH3)4—CH
`oO
`i
`
`oO
`
`+
`
`oO
`CH;
`I
`_C
`CH;—N*
`| Cy ™\o-
`CH;
`urea
`glutamate
`sarcosine i
`guanidinivm:
`oO
`4 9 NH;
`\
`|
`\
`pe.
`H,C—Nicy—
`C=O *NH;—C
`[oo Se of
`'E
`o
`|
`H
`Hy
`NHy
`NH,
`
`FIGURE 1. Common kosmotropes and chaotropesofinterest in protein-folding studies.
`
`is complex: argininespecifically interacts with and therefore destabilizes many proteins;
`however,arginine increases water surface tension and can alternatively behave as a kos-
`motrope. Furthermore, the protein stabilizing or destabilizing activity of a particular
`cosolute may be a function of the cosolute’s concentration.
`
`2.4. Predicting Folded Structure and Aggregation Propensity
`
`Several programs, available free of charge to academic users, are of inter-
`est. The Biology Workbench, operated under the San Diego Supercomputer Center
`(http://workbench.sdsc.edu), provides a slate of tools that, for example, allow the user to
`search databases for known sequences,predict secondary structure, plot the hydrophobic
`profile of a sequence, estimate isoelectric points, or align multiple sequences. The Euro-
`pean Molecular Biology Laboratory (EMBL)has developed two programsofparticular
`interest. FoldX (http://foldx.embl.de) provides a quantitative analysis of the effects of
`mutations on protein stability. Tango (http://tango.embl.de) predicts regions of unfolded
`polypeptides that are mostlikely to initiate aggregation.
`
`3. KINETICS AND THERMODYNAMICSOF PROTEIN
`FOLDING AND AGGREGATION
`
`Protein folding from the denatured state often is modeled as a simple two-state
`transition between the unfolded U and the native N states:
`
`USN.
`
`(1)
`
`The folded protein is only marginally thermodynamically stable relative to the un-
`folded protein. The Gibbs energy ofstabilization (net Gibbs energy difference between
`
`Exhibit 2072
`
`Page 11 of 15
`
`Exhibit 2072
`Page 11 of 15
`
`

`

`10
`
`R. M. MURPHY AND A. M. TSAI
`
`
` Activation
`
`energy
`
`
`
`Gibbsenergy
`
`Reaction coordinate
`
`FIGURE 2. Energetic relationship between unfolded protein U, natively folded protein N, intermediate J, and
`aggregate A. In this example,the partially refolded I can follow two pathways. One path, with a loweractivation
`energy, producesthe natively folded protein NV. The other path, shown as a dashedline, has a higheractivation
`energy but produces the lower-energy aggregate A. Changes in pH, temperature, or cosolute concentration
`may affect both the activation energies and the energy of the metastablestates.
`
`folded protein and unfolded polypeptide chain) is generally —20 to —60 kJ/mol.!° Since
`K= 107 =o (—Oi}
`(2)
`_IN]__
`(-aG
`at physiological temperatures, ~0.00001% to 0.01% of polypeptide chains are unfolded.
`Therelationship betweenprotein folding, misfolding, and aggregation can be under-
`stood through a reaction kinetic framework. This framework treats the folding process
`as a unimolecular equilibrium betweenthe folded and the unfolded states (Eq. 1). In the
`forward andreversereactions, the folding energetics can be incorporated in the appro-
`priate form in the reaction constants. The two-state model has been used successfully
`to explain protein-folding data captured from thermal melting, chemical denaturation,
`hydrogen exchange,fluorescence spectroscopy, and similar experimental studies.
`Proteins that tend to aggregate, however, may notbe accurately modeled using a
`simple two-state model. Rather, formation of one (or more) metastable, partially folded
`aggregation-prone intermediate / often is postulated (Figure 2). Several examples of
`these multistate transitions are given in later chaptersin this volume. Another plausible
`model for aggregating proteins postulates the formation of two alternate conformers
`(J, and Jy) from the unfolded polypeptide chain, one of which leads to aggregates and
`the other to correctly folded protein. Perhaps proteins with regions of “conformational
`confusion.” where alternate folded structures have similar thermodynamicstability, are
`mostlikely to be aggregation-prone.
`
`Exhibit 2072
`
`Page 12 of 15
`
`Exhibit 2072
`Page 12 of 15
`
`

`

`PROTEIN MISFOLDING AND AGGREGATION
`
`ll
`
`amorphous
`aggregate
`
`partially-structured
`aggregate
`
`Ea ties ore Mead
`
`simeges
`
`(ED
`
`unfolded
`polypeptide
`
`partially-folded
`intermediate
`
`partially-folded
`intermediate
`
`natively-folded
`protein
`
`prefibrillar
`oligomer
`
`fibrillar
`aggregate
`
`FIGURE 3. Kinetic pathways leading from unfoldedprotein to aggregates with various morphologies.In this
`sketch formationof large aggregated species is shownasirreversible; in some casesthese steps may bepartially
`reversible. Adapted from Foguel andSilva.!!
`
`In favorable cases, small proteins refold rapidly and correctly within seconds. In
`these cases, protein-refolding kinetics often can be modeled using a single forward and
`reverse rate constant. Slow steps include cis-trans isomeration ofpeptide bonds preceding
`prolines and disulfide bond formation. If metastable intermediates form, however, com-
`plex kinetic expressions may be required to model protein refolding. Refolding kinetics
`are generally first-order, while aggregation kinetics may be first-order or higher-order,
`depending on the exact mechanism andrate-limiting steps (Figure 3).
`At one pointit was thought that misfolded aggregated protein was simply an amor-
`phous mass, lacking any kind of organized structure. This turns out to be untrue; many
`aggregates that appear grossly amorphous retain some evidence (by FTIR or other
`means) of secondary structural elements and some aggregates such as amyloidfibrils are
`highly structured, althoughnotcrystalline. Micelles, globular assemblies,fibrils, ribbons,
`sheets, and other morphologies have been reported. There are examples of polypeptides
`
`Exhibit 2072
`
`Page 13 of 15
`
`Exhibit 2072
`Page 13 of 15
`
`

`

`12
`
`R. M. MURPHY ANDA. M. TSAI
`
`forming alternatively folded conformationalstates, one of which leads to amorphous ag-
`gregates and the other leading to fibrils.!* There is continuing discussion as to whether
`these various aggregate morphologies are stuck in kinetic traps (local minima in the
`energetic landscape) or are globally equilibrated.
`
`4. OUTLINE OF CHAPTERS
`
`This book is organized into four themes: theory (Chapters 2 and 3), experimen-
`tal techniques (Chapters 4-9), mechanisms and model systems (Chapters 10-13) and
`industrial applications (Chapter 14-15).
`In Chapter 2, Chris Roberts extensively discusses a reaction kinetics framework to
`model the aggregation pathways and provides guidelines for experimental design and
`data interpretation. Carol Hall and colleagues, in Chapter 3, give us a glimpse ofthe
`future in computer simulation. While full atomic scale modeling of protein aggregation
`processesis still far into the future, the approachespresentedin this chapter provide
`detailed and novelinsights into aggregation phenomena.
`The secondsection ofthis volume introducesus to techniques of special interest to
`the study of protein aggregation. Fernandez and Toblerdiscuss the use of hydrogen ex-
`change coupled with mass spectrometry to identify regions within a molecule that are in-
`volvedin aggregation. Three experimentaltools widely used for protein folding studies—
`circular dichroism, fluorescence spectroscopy, and calorimetry—can be adapted for
`investigations of aggregating systems, as discussed by Wilcox and Krishnamurthy in
`Chapter 5. In separate chapters, Krueger and colleagues. and Murphyand Lee, describe
`how neutron andlight scattering, respectively, afford us the ability to probe structure
`over a wide range of length scales. By using appropriate mathematical models to ana-
`lyze scattering data, one can gain additionalinsightinto the geometry of the aggregates
`and the kinetics of growth. Daniel Kirschner and colleagues describe X-ray diffraction
`techniques and theory and demonstrate how this method is employed to identify the
`existence of different morphologies along the aggregation pathway. Lastly, Soles and
`colleagues. explain how the dynamicsof a protein at the atomic level can be directly
`measured and how molecular motionis intimately related to the macroscopic stabil-
`ity of a protein, giving us a glimpse ofthe earliest events that may eventually lead to
`aggregation.
`Chapters 10-13 describe the mechanisms through which folding, misfolding, and
`self-assembly or aggregation may occur in model systems. Cavagnero and Kurt demon-
`strate that cotranslational folding, leading to early formation of nativelike secondary
`andtertiary structures, often make a protein aggregation-prone. Anne Robinson and
`co-workers show that aggregation is species-specific, i.e., protein aggregationis a self-
`assembly processevenin the midstof otherproteins that can be recruited into the process.
`Using a natively unfoldedprotein, Fink uses chemicals identified as amyloid diseaserisk
`factors to study the mechanism offibril formationin vitro. Kueltzo and Middaughdiscuss
`how formationofthe structurally labile molten globule state facilitates protein transport
`across membranes. Read together, these chapters offer us a comprehensive insight on
`
`Exhibit 2072
`
`Page 14 of 15
`
`Exhibit 2072
`Page 14 of 15
`
`

`

`PROTEIN MISFOLDING AND AGGREGATION
`
`13
`
`howprotein structure, dynamics,andstability contribute to the formationofeither highly
`structured or amorphous aggregates.
`The last section of the volume contains two case studies from the biotechnology
`industry. Mary Cromwell's chapter discusses the unique challenges presented during
`protein product developmentin trying to monitor andcontrolnative protein aggregation.
`Ricci and co-workersdescribe in detail an exhaustive search for mutations in a protein
`that lead to improved stability and reduced aggregation. These two case studiesillustrate
`beautifully the fundamental themesand approachesdiscussedin earlier chapters, thereby
`confirming their universal appeal.
`In this volume, ourintent is to show that protein misfolding and aggregationis
`best tackled by breaking downthe probleminto its fundamental components. We hope
`to reveal the connections between the molecular and the macroscopic, the monomeric
`and the multimeric, the theoretical and the practical. The many techniques reviewed
`here should assist both novice and experienced investigators to formulate their studies
`and to expand beyondtraditional approachesin their search for additional insight. The
`examples includedin this volume, from both academic andindustrial laboratories, should
`give hope that the apparently unwieldy problem ofprotein misfolding and aggregation
`will succumbto soundscientific approaches.
`
`REFERENCES
`
`Ne
`
`By
`
`5.
`
`1. C.A. Ross and M. A.Poirier, Protein aggregation and neurodegenerative disease, Nature Med. 10,
`$10-S17 (2004).
`R. M. Murphy, Peptide aggregation in neurodegenerative disease, Anau. Rev. Biomed. Eng. 4,
`155-174 (2002).
`N. J. Darby and T. E. Creighton, Protein Structure (Oxford University Press, New York, 1993).
`T. E. Creighton, Proteins: Structures and Molecular Properties (New York: W. H. Freeman and
`Co., 1984).
`J.D. Rawn, Biochemistry (New York: Harper and Row, 1983).
`T. L. Hendrickson, V. de Crecy-Lagard, and P. Schimmel, Incorporation of nonnatural amino acids
`into proteins, Annu. Rev. Biochem. 73, 147-176 (2004).
`7, K. A. Dill, Polymerprinciples and protein folding, Protein Sci. 8, 1166-1180 (1999),
`8.
`T, Cserhati and M. Szogyi, Role of hydrophobic and hydrophilic forces in peptide-protein interac-
`tion: new advances, Peptides 16, 165-173 (1995).
`9. M.G. Cacace, E. M. Landau, and J. J Ramsden, The Hofmeister series: salt and solventeffects on
`interfacial phenomena, Q. Rev. Biophys. 30, 241-277 (1997).
`R. Jaenicki, Protein stability and protein folding, Ciba Foundation Symposium 161, 206-221 (1991).
`D, Foguel and J. L. Silva, New insights into the mechanismsofprotein misfolding and aggregation
`in amyloidogenicdiseases derived from pressure studies, Biochemistry 43, 11361—11370 (2004).
`R. Khurana,J. R. Gillespie, A. Talaptra, L. J. Minert, C. lonescu-Zanetti, I. Millett, and A. L. Fink,
`Partially folded intermediates as critical precursors of light chain amyloid fibrils and amorphous
`aggregates, Biochemistry 40, 3525-3535 (2001).
`
`10,
`11.
`
`12.
`
`Exhibit 2072
`
`Page 15 of 15
`
`Exhibit 2072
`Page 15 of 15
`
`