7
`
`Chromatographic Techniques for
`the Characterization of Proteins
`
`Joost J M Holthuis
`and Reinoud J Driebergen
`
`1 INTRODUCTION
`
`and peptides
`The improvements in the chromatographic
`analysis of proteins
`significantly to the development of recombi
`during the last decade contributed
`nant pharmaceutical proteins The availability of these analytical methods for the
`and purity determination of proteins enabled the improvement of
`characterization
`the overall manufacturing
`process of pharmaceutical
`proteins
`As with all other pharmaceutical preparations recombinant pharmaceutical
`protein products must be thoroughly characterized Protein drugs must meet the
`same standards of safety purity and potency as conventional drugs The charac
`terization of the structure of a protein is complex because of the presence of a
`tertiary and quarternary structure Small structural
`primary secondary
`changes
`in a protein can influence the physicochemical
`as well as the activity
`properties
`and potency of
`the protein Different purification steps are used during the
`characteristics of the protein Such
`manufacturing that may change the structural
`changes may range from minor modifications eg deamidation of an Asp Patel
`and Borchardt 1990 to unfolding of
`the protein chain yielding biologically
`
`Joost J M Holthuis
`OctoPlus bv PO Box 722 2300 AS Leiden The Nether
`Reinoud J Driebergen
`Ares Serono 15bis Chemin des Mines CH1211
`Geneva Switzerland
`
`lands
`
`J N Herron et al eds Physical Methods to Characterize Pharmaceutical Proteins
`0 Springer Science+Business Media New York 1995
`
`243
`
`Abraxis EX2023
`Actavis LLC v Abraxis Bioscience LLC
`1PR201701101 1PR201701103 1PR201701104
`
`

`

`244
`
`Joost J M Holthuis and Reinoud J Driebergen
`
`inactive molecules Unfolding may result in the decomposition
`such as oxydation
`and fragmentation of the protein due to the fact that certain parts of the molecule
`normally part of the inner side of the tertiary structure are exposed to chemicals
`water and so forth
`For the determination of the purity of proteins the detection of impurities is of
`related Garnick et al
`importance The impurities are process host or product
`1988 Product related impurities can be the result of posttranslational changes or
`products that are formed during
`variability but are in most cases decomposition
`the purification process or upon storage Table I presents an overview of possible
`processes that can change the structure of a protein Geigert 1989 Recently an
`extended overview was published on the inactivation of proteins during chromato
`graphic analysis Sadana 1992
`The protein can also be changed intentionally With the introduction of large
`amounts of recombinant proteins more and more time is spent on the develop
`ment of chemical modifications of the proteins in order to change their disposition
`in vivo Covalent attachment of eg polyethylene glycol PEG chains shows
`promise for increasing the stability and protein solubility The resulting product
`contains a mixture of species characterized by a distribution in both the number
`and position of attachment of the PEG molecules The situation is even more
`that PEG is inherently polydispersal Characterization of
`complicated by the fact
`such a pegylated protein is important
`in order
`to be able to assure product
`
`consistency
`analysis of a pharmaceutical protein and its
`The qualitative and quantitative
`analytical methods
`relies on the use of sophisticated
`related compounds
`for
`identity and homogeneity of
`to be
`demonstrating the structural
`the proteins
`separated The application
`of multiple methods is imperative to ensure that a
`and its purity accurately as
`protein is thoroughly identified and characterized
`sessed For the analysis of pharmaceutical
`proteins the method must be able to
`
`Table I Possible Modifications of Proteins
`
`Mutation
`
`process
`
`conjugation
`
`Posttranslational
`
`Chemical
`
`Site directed mutagenesis
`
`Glycosylation
`
`Pegylation
`
`Formylation
`
`Acylation
`
`Phosphorylation
`
`Sulfatation
`
`Degradation
`
`pathways
`
`Fragmentation
`
`Deamidation
`
`Oxidation
`
`Disulfide scrambling
`
`Oligomerization
`
`Aggregation
`
`Crosslinking
`
`Proteolysis
`
`Denaturation
`
`

`

`Chromatographic Techniques
`
`245
`
`separate compounds that may only differ by one or a few amino acids or functional
`groups Considering the low diffusion coefficients of large biopolymers it
`is a real
`challenge to develop analytical chromatographic methods that enable the required
`
`affect
`
`noninteractive
`
`interactive
`
`separation
`Proteins are ampholytes and therefore the pH and the ionic strength of the
`behavior In addi
`mobile phase play an important role in their chromatographic
`tion temperature protein concentration and adsorbent surface characteristics
`in all modes of
`as well as in
`chromatography
`the analysis
`sizeexclusion chromatography These parameters are important
`structure of the protein may change when the
`because
`the threedimensional
`protein is in a nonphysiological environment and may be sensitive to slight
`changes in this environment
`these parameters also can be used to manipulate the
`protein structure in order to achieve separation
`an overview of the analytical chromatographic
`This chapter presents
`methods used for the characterization of proteins with an emphasis on pharmaceu
`tical proteins This chapter does not pretend to give a detailed and complete
`of proteins and peptides but rather presents an
`overview on the chromatography
`overview of approaches used to characterize proteins including the significant
`problems that can arise and relevant examples General
`information on the chro
`matography of peptides and proteins including the column materials used can be
`found elsewhere Mant and Hodges 1991
`
`2 REVERSED PHASE CHROMATOGRAPHY
`
`21 General
`
`Reversed phase chromatography RPC is the most frequently applied chro
`technique for the analysis especially for the determination of the
`matographic
`purity of peptides and intact proteins
`as well as for the analysis of peptide
`fragments obtained after enzymatic or chemical cleavage of the protein see also
`section 243 In RPC the protein generally loses its native structure and denatures
`Lau et al 1984 Denaturation
`and coulombic
`occurs when the hydrophobic
`interactions with the stationary andor mobile phase are stronger
`are maintaining the protein native structure The level of denaturation
`often
`depends on the time the sample resides on the column The elution of a protein is
`determined by the presence of hydrophobic amino acids as well as by the exposure
`regions to the column surface determined by a sequence of
`of hydrophobic
`amino acids The retention of the intact protein is based on the
`hydrophobic
`number of hydrophobic
`column sup
`residues that
`interact with the hydrophobic
`port and therefore is dependent on whether
`these hydrophobic areas are present on
`
`than those that
`
`

`

`246
`
`Joust J M Holthuis and Reinoud J Driebergen
`
`the outside of the protein when dissolved in the mobile phase The less polar
`residues can be shielded by conformational effects which results in an exposition
`of the more polar residues at the proteins surface during the analysis Nugent
`et al 1988 For peptides up to 15 amino acids there is a correlation between the
`hydrophobicity and retention on a C4 C8 and C18 bonded phase With larger
`compounds also the polypeptide chain length must be taken into account because
`the length of the chain determines the spatial structure Using the overall hydro
`phobicity of proteins based on retention parameters derived from small peptides
`a correlation was found with retention behavior under denaturing conditions for
`proteins ranging in molecular weight from 35 to 32 kDa Mant et al 1989 The
`conformation of the protein determines the exposure of hydrophobic
`regions to the
`stationary phase and thus regulates the retention of the protein Resolution of the
`denatured protein in RPC is achieved on the basis of the overall hydrophobicity of
`the sample and is mainly dominated by solvent effects
`Conformational changes on reversed phase and hydrophobic
`supports can
`lead to multiple states of the protein which can result in multiple or distorted
`peaks The control of the peak shape and the number of peaks in protein chroma
`of equilibria and the kinetic processes in
`tography require the understanding
`volved One can distinguish three kinetic
`processes where protein structural
`changes can affect behavior In the first case the changes are slow relative to the
`chromatographic migration time and the process is apparently irreversible In such
`cases several chromatographic peaks representing native and partially de
`natured compounds may be observed In the second case retention and halflives
`are comparable and distorted peaks can be observed Finally rapid changes occur
`and the system may act as though it were behaving ideally while at the same time
`significant changes have taken place on the chromatographic
`surface In principle
`such changes in order to enhance the possibility of
`one would like to prevent
`analyzing the native biologically active protein Karger and Blanco 1989 but in
`RPC this is almost an impossible task Therefore in order to prevent conforma
`is important to choose the circumstances in
`tional changes during the analysis it
`such a way that the analyte is present in one single conformation where in most
`cases the denaturation of the protein is complete The unfolding of proteins in a
`system can be controlled by the hydrophobicity of the packing
`chromatographic
`material Sometimes it
`reversed
`proteins under
`is possible to chromatograph
`in the folded conformation by carefully choosing the column
`phase conditions
`material Hanson et al 1992
`Changes in the quaternary structure often result in a significant change in the
`properties of the molecule because of dissocia
`molecular weight or hydrophobic
`tion or aggregation of the molecule Oligomers are high molecular weight com
`pounds and in most cases more hydrophylic when compared with the monomeric
`proteins Oligomers are therefore easily separated from the corresponding
`monomerssubunits by RPC
`
`

`

`Chromatographic Techniques
`
`247
`
`is often used in the quality control of pharmaceutical
`Peptide mapping
`lot to lot variations Isolation of peptides
`proteins and is then used to detect
`analysis and the determination of the amino acid
`following the chromatographic
`sequence of these peptides are tools that are used to determine the full sequence
`of a protein Peptide mapping of proteins is used to generate information on the
`primary structure of proteins Peptides produced from proteins by two or more
`different chain cleavage methods and analyzed by RPC give information on the
`identity of the protein Peptide mapping consists of an enzymatic eg trypsine or
`chemical eg cyanobromide CNBr digestion often preceded by reduction and
`carboxymethylation of free sulfhydryl groups with iodoacetic acid With peptide
`is possible to detect small changes in the protein eg replacement of a
`mapping it
`single amino acid The replacement of one amino acid in a peptide has in general a
`on the retention behavior on RPC In general RPC gives a
`dramatic influence
`high resolution separation of tryptic digests A typical run takes between 30 and
`60 min in order to obtain optimum separation of the peptide mixture The use of
`short reversed phase columns packed with 2µm pellicular particles
`can reduce
`the run time of a peptide to 15 min or less at elevated temperatures as was shown
`by Kalghatgi and Horvath 1988 for recombinant
`tissue plasminogen activator
`and recombinant human growth hormone rhGH For the determination of impu
`rities the method is less suitable In general
`to identify protein
`impurities below 10 by peptide mapping using RPC with UV detection as was
`shown by Clogston et al 1992 for recombinant
`human granulocyte
`stimulating factor
`
`it
`
`is difficult
`
`colony
`
`22 Stationary Phase
`
`The stationary phase used in RPC contains a hydrophobic
`surface and often
`consists of coated silica polymeric bonded silica and cross linked polymers
`Mant and Hodges 1991 Kennedy et al 1989 The matrices are in most cases
`coated with hydrophobic
`ligands such as in increasing order of hydrophobicity
`C2 C4 C8 C18 phenyl and cyanopropyl groups In general
`the retention time
`increases with increasing hydrophobicity of the ligands Decreasing particle size
`results in an increasing surface area and thus in a higher resolution
`The most frequently used column materials are silica based with a pore size
`of 100 to 300 A These phases are only stable at pH 2 to 8 The phases consisting of
`a polymer coating on silica are more inert to chemical degradation when compared
`silica bound packings These stationary phases are stable in
`with the conventional
`the pH region of 2 to 10
`Replacement of the silica as base of the reversed phase material by a polymer
`increases the stability further An example is the C4 C8 and C18 anchored to the
`
`

`

`248
`
`Joost J M Holthuis and Reinoud J Driebergen
`
`of
`
`phobic packings
`
`polymer skeleton eg polyvinylalcohol Other examples are cross linked hydro
`like hydroxylated polyether eg TSK polystyrenedivinyl
`benzene eg PLRPS polystyrene Maa and Horvath 1988 polymethacrylate
`and polymeric fluorocarbon These stationary phases are stable in the pH region
`of 1 to 13 Kato et al 1990 The pore size 504000 A of the column material is
`and must match the molecular size of the sample in order to
`importance
`eliminate undesired exclusion Polymeric polystyrene packings with particle
`diameters of 5 pm and pores of 300 A are in general superior in terms of resolution
`and sample recovery Burton et al 1988
`Reversed phase material based on silica and polymer based material often
`show a different selectivity probably due to unwanted
`between
`interactions
`silanol groups and the protein eg for proinsulin and insulin Linde and Welinder
`1991ab
`
`23 Mobile Phase
`
`The mobile phase often consists of mixtures of organic solvents to which
`organic or anorganic acids are added Isocratic elution of the proteins is possible
`however
`retention of the protein is strongly affected by the percentage of the
`organic modifier It
`is very difficult to prepare the mobile phase to be so reproduc
`ible that consistent data are obtained Therefore a gradient elution system is
`usually used that results in higher resolution and is most effective in the control of
`the retention time Elution is started with aqueous acid solutions of eg trifluoro
`acetic acid TFA phosphoric acid perchloric acid and heptafluorobutyric acid
`such as phosphoric acid perfluorinated carboxylic acids and trial
`Compounds
`kylammonium phosphate act as ion pairing agents The pH of
`the aqueous
`solutions is in the range of 2 to 5 The use of hydrochloric acid HC1 instead of
`TFA has some advantages
`such as a greater optical clarity that enables the
`low wavelengths
`detection at
`A recent development
`is the use of formic acid instead of phosphoric acid or
`TFA in the mobile phase for the chromatography
`of proteins Formic acid enables
`low wavelengths and because of its
`the detection of nonaromatic
`peptides at
`is easily removed Detection at 214 nm is possible with good resolu
`volatility it
`tion Formic acid is a mild acid that does not damage the column Poll and
`Harding 1989 1991
`The starting eluent
`is followed by a displacing
`eluent
`is an organic
`solvent such as methanol acetonitrile 1propanol or 2propanol The mobile
`and
`denaturation by disrupting the intramolecular hydrophobic
`phase causes
`coulombic forces The organic solvents can be ranked according to decreasing
`elutropic strength as compared to water which results in the following order
`
`that
`
`

`

`Chromatographic Techniques
`
`249
`
`fact
`
`propanol > acetonitrile > methanol > water Experimental variables that affect
`in RPC are temperature pH solvents buffers additives flow
`the elution behavior
`rate gradient program and sample pretreatment Nugent et al 1988
`RPC is very suitable for the analysis of hydrophilic proteins and peptides
`proteins eg membrane proteins often give poor recovery
`More hydrophobic
`and multiple peaks and may not elute or may be considerably
`retarded Protein
`aggregates can result in plugging of the stationary phase and poor recovery The
`that many proteins are not fully recovered from the analytical column often
`makes RPC not suitable for quantitative analysis These problems are sometimes
`and stationary phase andor a
`solved by optimization of the elution conditions
`correct sample preparation Aggregation can sometimes be prevented by modify
`ing the mobile phase by the addition of sodium dodecyl sulfate SDS 313chol
`CHAPS Solubilization of
`amidopropyldimethylammonio1propanesulfonate
`the protein can be obtained by using acids detergents andor chaotropic agents
`eg SDS CHAPS guanidine Triton X100 octyl glucoside or urea Also
`sodium cholate or Emulgen 991 have been used in order to solubilize membrane
`analysis Kato et al 1987 The addition of
`proteins prior to chromatographic
`such compounds results in dissolution of the protein but very often it also results
`in loss of the secondary and tertiary structure Because of the denaturation of the
`proteins in RPC it
`is not always suitable to study small changes in the protein
`conformation
`
`24 Examples
`
`241 CONFORMATIONAL CHANGES DURING CHROMATOGRAPHY
`AND RETENTION BEHAVIOR
`
`is
`
`acid acts
`
`Thevenon and Regnier 1989 studied the influence of formic acid concentra
`tion on retention and protein structure The conformation of
`the proteins
`below 10 the
`strongly dependent on the acid concentration At concentrations
`as an amine pairing agent and the protein is partially denaturated
`Between 10 and 30 the protein is further denaturated
`and the hydrophobic
`residues become exposed to the solvent The hydrophobicity of the species is
`increased and hence the retention increases Above 30 the solution becomes so
`residues begin to migrate into the interior of some
`the hydrophobic
`polar that
`highly denatured structure and they are finally shielded from the aqueous phase
`above 30 is not only due to the
`The decrease in retention at concentrations
`higher solvophilic power of formic acid relative to water but also to various
`conformational states of the proteins Thevenon and Regnier 1989
`the retention behavior and resolution
`The temperature may also influence
`
`

`

`250
`
`Joost J M Holthuis and Reinoud J Driebergen
`
`Cohen et al 1984 The effect of the temperature can be explained by different
`levels of unfolding of the protein in the mobile phase The effect of the tempera
`experiments with ribonuclease A RNase on
`ture was shown in chromatographic
`an n butyl column RNase elutes in two peaks at 25 °C and as one peak at 37°C
`The first peak proved to be the native peak while the second peak is the denatu
`rated compound Fig 1
`behavior of RNase A on RPC is not
`ideal Lu et al
`The chromatographic
`1986 Cohen et al 1985 showed that
`the peak shape can be distorted due to
`refolding of the protein during elution The binding of the proteins to the column
`material was also determined by the components of the mobile phase This
`indicated that broad peak shapes can be due to reversible conformational changes
`during the elution of a protein Cohen et al 1985
`The choice of organic solvent can influence the conformation of the protein
`in solution and influences the binding of proteins to C8 alkyl bonded silica It was
`shown for several proteins that 1propanol
`induces conformational changes of the
`surface associated protein Katzenstein etal 1986 At low pH a lower concen
`
`10
`
`30°
`
`1137
`
`10
`
`15
`
`10
`
`1
`
`15
`
`10
`
`15
`
`Figure 1 RPLC behavior of pancreatic
`
`TIME minutes
`ribonuclease A RNase as a function of temperature at pH 22
`Conditionssolvent A 10 mM H3PO4 pH 22 solvent B 5545 vv HH3PO4 concentration
`mM gradient 585 B in 30 mm linear flow rate 1 mlmin column 10m C4 LiChrospher SI 500
`100 x 46 rmn inner diameter sample size 140 pg RNase detection 254nm 01 AUFS Cohen et al
`1984
`
`is 10
`
`

`

`Chromatographic Techniques
`
`251
`
`tration of 1propanol is required to induce conformational changes as compared to
`neutral conditions Sadler et al 1984
`Retention parameter Z slope of log capacity factor k versus log molar
`in the mobile phase and log the
`concentration of organic modifier 1propanol
`value of log k at 1 M 1propanol can be used to describe the relation between the
`retention and the state of unfolding of a protein Lin and Karger 1990 The Z and
`values increased from folded surface unfolded urea unfolded to reduced un
`folded for a series of small globular proteins These experiments proved that
`proteins with their disulfide bridges cleaved have the largest degree of unfolding
`Each protein was present in four conformational states This relation between the
`parameters and the state of unfolding opens possibilities for the measurement of
`the rate or state of unfolding The differences between strongly related compounds
`with regard to the change in conformation on the column during elution can be
`used to obtain an optimal separation as was shown by Oroszlan et al 1992 who
`studied the retention of rhGH and the retention of methioninerhGH MetrhGH
`Retention of rhGH decreases with increasing column temperature when 1pro
`is used as an organic modifier while the retention increases with tempera
`panol
`ture when acetonitrile is employed Desorption and elution with 1propanol was
`correlated with a solvent induced conformational change The retention increase
`obtained with acetonitrile correlated with the temperature rise and was caused by a
`change yielding a more hydrophobic
`In this case a
`partial structural
`species
`surface driven process was suggested
`lower protein yields from
`In general
`are obtained with acetonitrile when compared with
`reversed chromatography
`1propanol The change in the 3 sheet structure of bovine achymotrypsinogen on
`alkyl and phenyl bonded phases when eluted at different pH and concentration
`3040 of 1propanol proved to be independent of the alkyl chain length The
`structure was changed to the same extent after leaving the protein interaction with
`the column material and releasing it back into solution Drake et al 1989
`suggesting a solvent driven rather
`than a surface driven process
`Depending on the type of column used it
`is sometimes possible to perform
`RPC without an organic solvent A polymeric stationary phase divinylbenzene
`based in combination with acetic acid gradients in water without organic mod
`ifiers was used to determine rhGH in a crude acetic acid extract of Escherichia coli
`in which the protein was expressed Other proteins such as interleukin113 glu
`cagon and insulin are eluted as sharp symmetrical peaks with acetic acid gradients
`in water acetonitrile or isopropanol Welinder and Sorensen 1991
`
`242 SEPARATION OF INTACT PROTEINS
`
`For the development and quality control of pharmaceutical proteins it
`is often
`necessary to have analytical methods available that separate proteins that differ
`
`

`

`252
`
`Joost J M Holthuis and Reinoud J Driebergen
`
`it
`
`only in one or a few amino acids One example is the separation of muteins
`proteins that differ by only a few amino acids caused by a mutation in the DNA
`sequence Other examples are the separation of oxidized proteins and the deter
`mination of the number and the position of Cys residues
`Under carefully controlled conditions
`is possible to separate mixtures of
`insulins that have minimaldifferences in structure eg a single uncharged residue
`McLeod and Wood 1984 From this it
`is concluded that retention of insulin
`proteins on the reversed phase column is determined by the entire exposed surface
`of the molecule rather
`than a specific region
`RPC can also be used for the separation of related recombinant proteins in the
`fermentation broth after a selective online sample clean up Hayashi et al 1987
`Human recombinant epidermal growth factors rhEGF were determined quantita
`tively in cultured media The analytes were isolated on a C18 protein coated pre
`column and separated on an analytical TSK ODS 120T column The rhEGF153
`was separated from rhEGF1471 and 1511 The isolation of the compounds on
`the precolumn was superior when compared with offline protein precipitation
`with trichloroacetic acid TCA
`The replacement of one amino acid can change the conformation of a protein
`molecule Such conformational changes can often be monitored by RPC Brems
`et al 1988 studied the effect of replacing Lys 112 with Leu in bovine growth
`hormone bGH approx 190 amino acids on the folding process The retention
`time of the Leu mutant was increased from 165 to 25 min when compared with the
`conditions The conformation of the mutant
`wild type under denaturing
`in its
`native state was indistinguishable from that of the wild type These results indicate
`form is more sensitive for the denaturing
`agent guanidineHC1
`the mutant
`36 M than the wild type
`Kunitani et al 1986 studied the retention behavior of several recombinant
`human interleukin2 IL2 muteins These muteins differ in the presence of Ala at
`position 1 and contain different amino acids at positions 58 104 105 and 125
`Table II The replacement of one single amino acid resulted in most cases in a
`in retention time The muteins eluting as peak A contain a methionine
`sulfoxide at 104 When the muteins were reduced compounds with subscript red
`toward longer retention times was observed
`indicating that more hydro
`a shift
`
`that
`
`shift
`
`phobic regions are exposed to the stationary phase
`The Nterminal heterogeneity
`and the oxidation states of the methionines
`were also studied The species that contains the oxidized methionine group elutes
`prior to the nonoxidized species Table II Oxidation of methionine residues of
`results in more hydrophilic compounds Frelinger and Zull
`proteins in general
`1984 described the elucidation of the structure of the three oxidized forms of
`bovine parathyroid hormone which could be separated with RPC
`human granulocyte macrophagecolony
`Purified recombinant
`stimulating
`rhGMCSF was characterized with RPC Analysis of
`the compound
`
`factor
`
`

`

`Chromatographic Techniques
`
`253
`
`Table II Retention Times of IL 2 Muteins in RPHPLCa
`time mm
`Peak B Peak A
`
`Mutein
`
`Retention
`
`Peak B Peak Are
`
`AlaiCys125 parent
`
`AlalSer125
`
`desAlaiCysi25
`
`desAlalSer125
`
`desAlaiA1a125
`
`desAlaiAlai°4Ser125
`
`desAlaiSer58
`
`desAlalSerio
`
`390
`
`346
`
`388
`
`346
`
`409
`
`328
`
`389
`
`380
`
`°From Kunitani et al 1986
`
`332
`
`296
`
`330
`
`297
`
`346
`
`363
`
`319
`
`361
`
`321
`
`381
`
`364
`
`352
`
`350
`
`314
`
`349
`
`314
`
`366
`
`300
`
`199
`
`203
`
`5 is the native
`
`showed the presence of the major peak and five small peaks see Fig 2 The six
`compounds were isolated by RPC and subjected to proteolytic digest with trypsin
`and protease V8 The digests were analyzed by RPC and the separated peptides
`were subjected to fast atom bombardment mass spectrometry FABMS Frac
`tions 1 to 4 proved to be rhGMCSF derivatives of which different methionines are
`oxidized Fraction 6 proved to be rhGMCSF of which two amino acids at the N
`terminal were deleted Met Ala Ohgami et al 1989 Peak
`compound
`RPC is the method of choice for studying the position number and oxidation
`status of the cysteine residues The presence of S S bridges strongly contributes to
`the conformation and refolding of a protein Direct analysis of the disulfide
`content of proteins can be a method for monitoring the stability and refolding
`process of cysteinecontaining proteins The general approach of the determina
`tion of the number of cysteines and cystines involves
`the carboxymethylation of
`acid and radiolabeled iodoacetic
`acid The first
`the protein with iodoacetic
`carboxymethylation is performed prior to reduction of the protein and the second
`carboxymethylation is performed following reduction The samples are analyzed
`by RPC with UV and radiochemical detection
`A compound whose refolding process was studied extensively by RPC is
`rhIL2 Recombinant hIL2 is a hydrophobic protein with a limited aqueous solu
`using aqueous acetonitrile and TFA as eluent
`bility and can be chromatographed
`Due to the presence of three cysteines in endogenous rhIL2 three isomers can be
`formed each with a different disulfide bridge Weir and Sparks 1987 studied the
`refolding of rhIL2 Recombinant hIL2 expressed in E colt was isolated as
`inclusion bodies after cell breakage Recombinant hIL2 and contaminants were
`dissolved and further purified in reduced and denatured form by gel permeation
`chromatography Refolding took place by diluting the solvent Renaturationauto
`
`

`

`254
`
`Joost J M Holthuis and Reinoud J Driebergen
`
`r5
`
`r6
`
`Fr4
`Fr3
`Fr2
`Fr1
`
`nn
`
`280
`
`at
`
`Absorbance
`
`0
`
`20
`
`40
`
`Retention time min
`Figure 2 HPLC profile on a reversed phase column of renatured rhGMCSF Ohgami et al 1989
`
`the
`
`oxidation was initiated and two peaks were observed on RPC C18 The first peak
`to elute was oxidized folded IL2 and the second peak was the less bioactive
`reduced IL2 When the renaturation
`time was increased peak 1 grew at
`expense of peak 2
`Browning and coworkers 1986 also used RPC C4 to clarify the structure
`rhIL2 In the presence of a chaotrope
`and disulfide scramble of nonglycosylated
`the proper disulfide bond is scrambled resulting in a mixture of the three disulfide
`isomers The identity of each of the three forms was determined by labeling of the
`acid and subsequent peptide mapping The
`free cysteines with 311iodoacetic
`nonnative isomers showed a less hydrophobic
`behavior when compared with the
`native form The fully reduced rhIL2 is more hydrophobic
`than the native form
`From this it was concluded that
`the nonnative isomers retain some folded struc
`ture to mask a hydrophobic
`region of the molecule
`Kunitani and coworkers 1986 studied variants of rh1L2 Endogenous IL2 has
`only one disulfide bridge between amino acid 58 and 105 but it contains three cys
`teines at positions 58105 and 125 respectively Recombinant desAlaSer125IL2
`is a protein that can only form one cystine disulfide The cys125 is replaced by a
`serine by sitespecific mutagenesis which resulted in an improved stability The
`and the complete reduced intact native
`of this compound
`complete reduction
`retention Table II The complete reduction
`molecule result in an increase of
`regions to the stationary phase
`probably results in an exposure of the hydrophobic
`The compounds with intact cystine bridges are less hydrophobic
`
`because
`
`the
`
`

`

`Chromatographic Techniques
`
`255
`
`intact
`
`regions are covered in the tertiary structure The other muteins
`hydrophobic
`exhibit the same behavior Reduction results in an increase of retention
`OKeefe and coworkers 1992 determined
`the number of cysteines in
`aeruginosa exotoxin A40 TGFa
`transforming growth factoraPseudomonas
`PE40 a fusion protein expressed in E colt The number of cysteines was
`determined by derivatization of the cysteine moieties in reduced protein with
`can be selectively
`monobromobimane The resulting derivative
`detected with
`
`fluorescence detection It not possible to separate PE40 M = 40000 from the
`fusion protein TGFaPE40 M = 44960 However with derivatization
`of the cysteines in TGFaPE40 prior to RPC it was possible to discriminate the
`cysteine free PE40 from TGFetPE40 OKeefe et al 1992
`Chromatography of pituitary derived hGH PhGH and biosynthetic hGH
`BhGH was described Christensen et al 1990 Small peaks were visible in the
`in addition to the dominating hGH peak
`chromatograms of both compounds
`BhGH consistently elutes as a slightly sharper peak when compared with PhGH
`isocratic conditions This is caused by the absence of a 201cDa
`under various
`fraction in the BhGH The 201Da compound is a decomposition
`product of hGH
`obtained after deletion of amino acid residues 3246 from the 22K hGH Al
`though both compounds differ significantly in molecular weight separation could
`not be obtained with RPC
`human interferons rhINF were analyzed on
`Different forms of recombinant
`a reversed system Felix et al 1985 The method was able to separate rhIFNetA
`rhIFNaD and the hybrid rhIFNaAD Recombinant hIFNaA and rhIFNaD
`differ about 15 in their amino acids In addition it was possible to separate the
`rhIFNaA fast migrating monomer CyslCys98 and Cys29Cys138 from the
`slow migrating monomer Cys29Cys138 and from the oligomers When the
`CyslCys98 bridge was replaced only the slow migrating monomer was found
`
`243 PEPTIDE MAPPING
`
`Recombinant hIL2 expressed in E coli was identified by Nterminal se
`quence and by peptide mapping Lahm and Stein 1985 Disulfides were deter
`mined by using a double label Scarboxymethylation procedure with tritiated
`iodoacetic acid and cold iodoacetic acid The labeled compound was digested with
`The retention of labeled peptides
`trypsin and the peptides were chromatographed
`was detected and indicated the presence of cysteine residue A disulfide linkage
`between cysteine residues at positions 58 and 105 was confirmed About 90 of
`the isolated rhIL2 contained a methionine residue at the N terminus
`Renlund et al 1990 described the routine peptide mapping of recombinant
`human immunodeficiency virus HIV proteins p24 core an