Hydrophobic Interaction Manual - AMERSHAM

7/22/2019 Hydrophobic Interaction Manual - AMERSHAM

1/1041

Hydrophobic InteractionChromatography

18-1020-90

PRINCIPLESAN D M ETHODS

Edition AB


2/104

to order:

Asia PacificTel: +852 2811 8693 Fax: +852 2811 5251 AustralasiaTel: +61 2 9894 5152 Fax: +61 2 9899 7511AustriaTel: 01 576 0616 20 Fax: 01 576 0616

2 7

BelgiumTel: 0800 73 888 Fax: 03 272 1637 CanadaTel: 1 800 463 5800 Fax: 1 800 567 1008 Central, East, South East EuropeTel: +43 1 982 3826 Fax: +43 1

985 8327 DenmarkTel: 45 16 2400 Fax: 45 16 2424FinlandTel: 09 512 3940 Fax: 09 512 1710FranceTel: 0169 35 67 00 Fax: 0169 41 9677GermanyTel: 0761

4903 401

Fax: 0761 4903 405 ItalyTel: 02 27322 1 Fax: 02 27302 212JapanTel: 81 3 5331 9336 Fax: 81 3 5331 9370 Latin AmericaTel: +55 11 3667 5700 Fax: +55

11 3667 87 99 Middle East and AfricaTel: +30 (1) 96 00 687 Fax: +30 (1) 96 00 693 NetherlandsTel: 0165 580 410 Fax: 0165 580 401 NorwayTel: 2318 5800

Fax: 2318 6800

PortugalTel: 01 417 7035 Fax: 01 417 3184 Russian FederationTel: +7 (095) 232 0250,956 1137 Fax: +7 (095) 230 6377 South East AsiaTel: 60 3 724 2080

Fax: 60 3 724 2090 SpainTel: 93 594 49 50 Fax: 93 594 49 55 SwedenTel: 018 612 19 00 Fax: 018 612 19 10 SwitzerlandTel: 01 802 81 50 Fax: 01 802 81

5 1

UKTel: 0800 616 928 Fax: 0800 616 927 USATel: +1 800 526 3593 Fax: +1 877 295 8102

Sephadex, Sepharose, Superose, BioProcess, FPLC, SMART, Mono Q, SOURCE, HiTrap, HiLoad, BPG, Superloop, Superdex, Ultroscan, PhastSystem, and

RESOURCE are trademarks of Amersham Pharmacia Biotech Limited or its subsidiaries. Amersham is a trademark of Nycomed Amersham plc. Pharmacia and

Drop Design are trademarks of Pharmacia & Upjohn. AmershamPharmacia Biotech ABBjrkgatan 30, SE-751 84 Uppsala, Sweden. AmershamPharmacia Biotech

UK LimitedAmersham Place, Little Chalfont, Buckinghamshire HP7 9NA, England. AmershamPharmacia Biotech Inc800 Centennial Avenue, PO Box 1327,

Piscataway, NJ 08855 USA. AmershamPharmacia Biotech Europe GmbHMunzinger Strasse 9, D-79111 Freiburg, Germany. AmershamPharmacia Biotech K.K.

Sanken Building, 3-25-1, Hyakunincho, Shinjuku-ku, Tokyo 169-0073, J apan. All goods and services are sold subject to the terms and conditions of sale of the

company within the Amersham Pharmacia Biotech group that supplies them. A copy of these terms and conditions is available on request. Amersham

Pharmacia Biotech AB 2000 All rights reserved.


3/104

PRINCIPLESANDM ETHODS

Hydrophobic Interaction

Chromatography


4/104

ISBN 91-970490-4-2


5/104

Many biotechnologists began their careers in chromatography reading Gel Filt-

rat ion: Theory and Practice.First published in 1966, this monograph has had over250,000 copies printed in five languages. It was soon followed by another helpfulmonograph from Amersham Pharmacia B iotech on ion exchange. About 15 years ago,Affi nity Chromatography: Principles and M ethods was published describing theemergence of t his pow erful separa ting method for ma cromolecules. In some w ays thismonograph series has defined the critical methods in the field a t the time of publicationand ha s been bo th go od business and a public service for over 25 years.

With the rise of the modern biotechnology industry and its requirement for highlypurified pharmaceutical proteins, a further emphasis has been placed on entire

processes w ith respect to their economy , capacity and resultant product quality . O ftenthe extent of separation pow er req uired is defined by the need to resolve the productnot o nly from the ba ckground impurities derived from the fermentat ion but a lso fromdegrada tion products and analogues of the drug itself. For many cases, hydrophobicinteraction chromatography (H IC) is an ideal separa tion method.

In my experience, H IC is finding dra matically increased use bo th in laboratory andproduction processes. Since the molecular mechanism of HIC relies on uniquestructural features, it serves as an ortho gona l method to ion exchange, gel filtra tion andaffinity chroma tography. It is very generic, yet capable of pow erful resolution. Usually

media have high capacity and are economical and stable. Adsorption takes place inhigh salt a nd desorption in low salt concentra tions. These special properties make H ICvery useful in whole processes for bridging or transitioning between other steps inad dition to the separa tion w hich is effected.

This book can serve as an excellent intro duction to the subject of H IC for those newto this method of separation. More experienced chromatographers can also benefitfrom the useful review . Topics include the molecular mechanism of separation by H ICin contrast to reversed phase chroma tography, a helpful section on stra tegies for rapidmethod development, as well as a w ide selection of examples. Practical aspects such

as packing, use and sanitization of columns are discussed. There are many tricks,techniques and insights to b e gained in a complete reading.

I recommend it be read a nd kept handy on your persona l book shelf and I predictthat you will find HIC a surprisingly helpful technique both alone and especially incombination w ith other modes of separa tion.

Stuart E. Bui lder

So. San FranciscoJanuary 15, 1993

Foreword


6/104

1. Introduction to HIC ................................................... 9

2. Principles of HIC.................................................... 11

Theory.............................................................................11

HIC vs RPC .....................................................................12

Factors affecting HIC.......................................................13

Type of ligand.......................................................13

Degree of substitution..........................................14

Type of base matrix..............................................14

Type and concentration of salt .............................15

Effect of pH...........................................................16

Effect of temperature............................................17

Additives...............................................................18

3. Product Guide........................................................ 19

BioProcess Media ...........................................................20

Base matrices .......................................................20

Coupling ...............................................................21

Chemical stability .................................................21

Physical stability...................................................22

Binding capacity ...................................................22

Phenyl Sepharose 6 Fast Flow (low sub) andPhenyl Sepharose 6 Fast Flow (high sub) .......................23

Butyl Sepharose 4 Fast Flow...........................................24

Phenyl Sepharose High Performance .............................25

Custom Designed HIC Media ..........................................26

HIC Media Test Kit ..........................................................26

Contents


7/104

Phenyl Sepharose CL-4B and

Octyl Sepharose CL-4B ...................................................27Phenyl Superose and Alkyl Superose..............................27

4. Experimental Design............................................. 29

Hydrophobicity of proteins .............................................29

Multivariate mapping ......................................................29

Strategic considerations .................................................30

Choice of HIC media .......................................................31

General considerations .........................................31

Screening experiments .........................................32

Optimizing a HIC step .....................................................39

The solute.............................................................39

The solvent ...........................................................41

Elution ..................................................................42Sample load and flow rate....................................45

Regeneration ........................................................45

Process considerations...................................................46

Method optimization in process ...............................chromatography ...................................................46

Scaleability ...........................................................49

Regulatory considerations....................................50

5. Experimental Technique........................................ 53

Choice of column............................................................53

Column dimensions..............................................53

Packing the column ........................................................53

Packing Sepharose Fast Flow based HIC gels ......54

Packing Phenyl Sepharose High Performance .....55

Packing Sepharose CL-4B based HIC gels ...........55

Contents


8/104

Use of an adaptor .................................................55

Checking the packed bed......................................56

Prepacked HIC media ...........................................58

Sample preparation.........................................................59

Sample composition.............................................59

Sample volume.....................................................59

Sample viscosity...................................................60

Particle content.....................................................60Sample application..........................................................61

Sample reservoir ..................................................61

Sample applicators ...............................................61

Sample loops with valves LV-4 or SRV-4 .............62

Sample loops or Superloop withvalves V-7 or MV-7...............................................62

Batch separation .............................................................63Cleaning, sanitization and sterilization procedures .........63

Storage of gels and columns ..........................................65

Prevention of microbial growth ............................65

Antimicrobial agents .............................................65

Storage of unused media......................................67

Storage of used media..........................................67

Storage of packed columns ..................................67

Process considerations...................................................68

Selecting a column ...............................................68

Aspects of column design ....................................69

Packing large scale columns ................................71

Scale-up ...............................................................74

Contents


9/104

Contents

6. Applications ...........................................................77

Preparative and analytical HICapplications in the research laboratory ...........................77

HIC in combination with ammonium sulphateprecipitation..........................................................77

HIC in combination with ion exchangechromatography ...................................................78

HIC in combination with gel filtration ...................80

HIC as a single step purification technique........81Analysis of conformational changes with HIC.......84

Other HIC application areas in the researchlaboratory .............................................................84

Preparative, large scale applications ...............................85

Purification of a monoclonal antibody forclinical studies of passive immuno-

therapy of HIV-1 ...................................................85Purification of recombinant humanEpidermal Growth Factor (h-EGF) from yeast .......87

Purification of a monoclonal antibody forin vitro diagnostic use ..........................................90

Purification of a recombinant Pseudomonasaeruginosa exotoxin, produced in E. Coli ..............92

7. References............................................................ 97

Order from ................................................................ 102


10/10410


11/10411

Introductionto HIC

In a classical paper published in 1948 and entitled: Adsorption Separation bySalting Out, Tiselius [1] laid dow n the foundation for a separation method w hich isnow popularly know n as hydrophobic interaction chromatogra phy (H IC). He notedtha t, ...proteins and other substances w hich a re precipitated a t high concentra tionsof neutra l sa lts (salting out), often are adsorbed q uite strongly a lready in salt solutionsof lower concentration than is required for their precipitation, and that someadsorbents w hich in salt-free solutions show no or only slight af finity fo r proteins, a tmoderately high sa lt concentra tions become excellent a dsorbents . Since then, greatstrides have been ma de in developing a lmost ideal stat ionary phases for chroma togra phy(such as cellulose, cross-linked dextra n (Sephadex ), cross-linked agarose (SepharoseCL, Sepharose High Performance and Sepharose Fast Flow), and in developingcoupling methods for immobilizing ligands of choice [2,3] to such ma trices. It w as a

combina tion o f these tw o events w hich, in the beginning of 1970's, led to the synthesisof a variety of hydrophobic adsorbents for biopolymer separations based on thispreviously rarely exploited principle.

The first at tempt a t synthesizing such adsorbents w as made by Yo n [4] follow ed byEr-el et al. [5], Hofstee [6] and Shaltiel & Er-el [7]. Characteristically, these earlyadsorbents show ed a mixed ionic-hydrophob ic chara cter [8]. Despite this, H alperinet al. [9] claimed that protein binding to such adsorbents was predominantly of ahydrophobic character. Porath et al. [10] and Hjertn et al. [11] later synthesizedcharge-free hydrophobic adsorbents and demonstrated that the binding of proteins

was enhanced by high concentrations of neutral salts, as previously observed byTiselius [1], and that elution of the bound proteins w as achieved simply by w ashing thecolumn with salt-free buffer or by decreasing the polarity of the eluent [6, 10, 11].Amersham Pharmacia Biotech was first in producing commercial HIC adsorbents(Phenyl and O ctyl Sepharose CL-4B [12]) of the charge-free type and ha s continuouslyfollow ed this up with new developments in aga rose matrix design by introducing newstable HIC media ba sed on Superose , Sepharose Fast Flow and Sepharose H ighPerforma nce, meeting various demands on chroma tographic productivity , selectivityand efficiency.

1


12/10412

The commercial availability of well-characterized HIC adsorbents opened newpossibilities for purifying a variety of biomolecules such as serum proteins [12, 13],membra ne-bo und proteins [14], nuclear proteins [15], receptors [16], cells [17], andrecombinant proteins [18, 19] in research a nd industrial laboratories. These adsorbentsw ere a lso used for the reversible immobilizat ion o f enzymes [20] and liposomes [21].

The principle for protein adsorption to HIC media is complementary to ion

exchange chromatography and gel filtration. HIC is even sensitive enough to beinfluenced by non-polar groups normally buried within the tertiary structure ofproteins but exposed if the polypeptide chain is incorrectly folded or da maged (e.g. byproteases). This sensitivity can be useful for separating the pure native protein fromother forms.

Altogether this makes HIC a versatile liquid chromatography technique, being alogical part of any rational purification strategy, often in combination with ionexchange chromatogra phy and gel filtration. H IC has also found use as an a nalyticaltoo l to detect protein conforma tional changes.

H IC req uires a minimum of sample pre-treatment and ca n thus be used effectivelyin combination w ith tra ditional protein precipitation techniques.

Protein binding to H IC ad sorbents is promoted by moderately high concentra tionsof anti-chaotropic salts, which also have a stabilizing influence on protein structure.Elution is achieved by a linear or stepw ise decrease in the concentra tion o f sa lt in theadsorption buffer. Recoveries are often very satisfactory.

A number of mechanisms have been proposed for HIC over the years and factorsthat affect the binding of proteins to such adsorbents have been investigated. Theseaspects will be briefly out lined in this handbook. G reater emphasis has been given topractical considerat ions on how to make optimal use of Amersham Pharmacia B iotechrange of HIC products.


13/10413

Principlesof HIC

Theory

The discussions tha t follow in this chapter w ill be limited to the non-charged t ypeof H IC a dsorbents.

The many t heories tha t ha ve been proposed fo r H IC are essentia lly based upon thosederived fo r interactions betw een hydrophob ic solutes and w ater (22,23), but no ne ofthem has enjoyed universal accepta nce. What is common to a ll is the centra l role playedby the structure-forming salts and the effects they exert on the individua l components(i.e., solute, solvent and adsorbent) of the chromatographic system to bring about thebinding of solute to adsorbent. In view of this, Porath (24) proposed salt-promotedadsorption as a general concept for HIC and other types of solute-adsorbent

interactions occuring in the presence of moderately high concentrations of neutralsalts.

Hofstee (6) and later Shaltiel (7) proposed hydrophobic chromatography withthe implicit assumption that the mode of interaction between proteins and theimmobilized hydrophobic ligands is similar to the self association of small aliphaticorganic molecules in water. Porath et al. (10) suggested a salting-out effect inhydrophobic adsorption, thus extending the earlier observations of Tiselius (1). Theyalso suggested that . . .the driving force is the entropy gain arising from structurechanges in the water surrounding the interacting hydrophob ic groups. This concept

was later extended and formalized by Hjertn (25) who based his theory on the wellknow n thermodyna mic relationship:

G =

H - T

S. He proposed tha t the displace-

ment of the ordered water molecules surrounding the hydrophobic ligands and theproteins leads to an increase in entropy ( S) resulting in a negative value for the changein free energy (

G ) of the system. This implies tha t the hydrophobic ligand-protein

interaction is thermodynamically favourable, as is illustrated in Fig. 1.

An alternat ive theory is ba sed on the parallelism betw een the effect of neutra l sa ltsin salting out (precipitation) and H IC (26,27). According to M elander and H orvath(27), hydrophobic interaction is accounted for by increase in the surface tension of

w ater arising from the structure forming salts dissolved in it. In fact, a combina tionof these tw o mechanisms seems to be an obvious extension and ha s been exploited long

2


14/10414

Fig. 1.

Close to the surface of the

hydrophobic ligand and solute

(L and H), the water molecules

are more highly ordered than

in the bulk water and appear

to shield off the

hydrophobic ligand and solute

molecules. Added saltinteracts strongly with the

water molecules leaving less

water available for the

shielding off effect, which is

the driving force for Land H to

interact with each other.

before H IC adsorbents w ere synthesized (28). Finally, Srinivasan and Ruckenstein (29)

have proposed that H IC is due to van der Waa ls a ttra ction forces betw een proteins andimmobilized ligands. The ba sis for this theory is tha t the van der Waals a tt raction forcesbetw een protein and ligand increase as the ordered structure of w ater increases in thepresence of salting o ut salts.

HIC vs RPC

In theory, HIC and reverse-phase chromatography (RPC) are closely related LC

techniques. Both are based upon interactions between solvent-accessible non-polargroups (hydrophobic patches) on the surface of biomolecules and the hydrophobicligands (alkyl or aryl groups) covalently attached to the gel matrix. In practice,however, they are different. Adsorbents for RPC are more highly substituted withhydrophobic ligands than HIC adsorbents. The degree of substitution of HICadsorbents is usually in the range of 1050

moles/ml gel of C

2C

8a lkyl or simple aryl

ligands, compared with severa l hundred

moles/ml gel of C4C

18a lkyl ligands usually

used for R PC adsorbents. Consequently, protein binding to RPC adsorbents is usuallyvery strong, which requires the use of non-polar solvents for their elution. RPC hasfound extensive applications in analytical and preparative separations of mainlypeptides and low molecular weight proteins that are stable in aqueous-organicsolvents.

In summary, H IC is an alternative w ay of exploiting the hydrophobic properties ofproteins, working in a more polar and less denaturing environment.

C ompared w ith RPC , the polarity of the complete system of H IC is increased bydecreased ligand density on the sta tiona ry phase and by a dding sa lt to the mobile phase.

L + H S L

H S +

W

P=Polymer matrixS=Solute moleculeL=Ligand attached to polymer matrixH=Hydrophobic patch on surface of solute moleculeW=Water molecules in the bulk solution

P


15/10415

Factors affecting HIC

The main parameters to consider when selecting HIC media and optimizingseparation processes on HIC media are:

Ligand type and degree of substitution

Type of base matrix

Type and concentration of salt

pH

Temperature

Additives

Type of ligand

The type of immobilized ligand (alkyl or aryl) determines primarily the proteinadsorption selectivity of the HIC adsorbent (6,7,30). In general, straight chain alkyl(hydrocarbon) ligands show pure hydrophobic character w hile aryl ligands show amixed mode behaviour where both aromatic and hydrophobic interactions arepossible (30). It is also esta blished tha t, a t a constant degree of substitution, the proteinbinding capacities of HIC adsorbents increase with increased alkyl chain length(Fig. 2A) (30,31). The charged type HIC adsorbents (6,7) show an additiona l mode ofinteraction, w hich will not be discussed here. The choice betw een a lkyl or a ryl ligands

is empirical and must be established by screening experiments for each individualseparation problem.

Fig. 2.

The effect of alkyl

chain length and

degree of

substitution on

binding capacity

in HIC. In Fig. 2A

it is assumed that

the degree of

substitution is the

same for each

alkyl chain length

shown.

A B

C C C 10 20 30

n-Alkyl chain length Degree of substitution

(mol ligand/ml gel)

Bindingcapacity

(mgpr

otein/mlgel)

4 6 8


16/10416

H IC media show n in Fig. 3 are all ba sed on the glycidyl ether coupling procedure,w hich produces gels that are charge free and tha t should thus only have hydrophobicinteractions w ith proteins. The phenyl group show n in Fig. 3-C a lso ha s a potential for- interactions. The glycidyl-ether coupling technique will introduce a short spacerbut the effect of this will be very limited since the short hydrophobic chain isneutralized w ith the hydro philic O H -group.

Degree of substitution

The protein binding capa cities of H IC adsorbents increase w ith increased degree ofsubstitut ion of immobilized ligand. At a sufficiently high degree of ligand substitut ion,the apparent binding capacity of the adsorbent remains constant (plateau is reached)but the strength o f the interact ion increases (3133, 35) (Fig. 2B). Solutes bound undersuch circumsta nces are difficult to elute due to multi-point a tt achment (34).

Type of base matrix

It is important not to overlook the contribution of the base matrix. The tw o mostw idely used types of support are strongly hydrophilic carbohydra tes, e.g. cross-linkedagarose, or synthetic copolymer materials. The selectivity o f a copolymer support w illnot be exactly the same as for an agarose ba sed support substituted w ith the same typeof ligand.

To a chieve the same type of results on an agarose-ba sed ma trix a s on a copolymersupport, it may be necessary to modify adsorption and elution conditions.

OCH CHCH O(CH ) CH2 2 2 3 3

OH

OCH CHCH O(CH ) CH2 2 2 7 3

OH

Butyl Sepharose 4 Fast Flow

Octyl Sepharose CL-4B

OCH CHCH O2 2

OH

Phenyl SuperosePhenyl Sepharose High PerformancePhenyl Sepharose CL-4BPhenyl Sepharose 6 Fast Flow (low sub)Phenyl Sepharose 6 Fast Flow (high sub)

OCH CHCH OCH C(CH )2 2 2 3

OH

Alkyl Superose3

A

B

C

D

Fig. 3.

Different hydrophobic

ligands coupled to

cross-linked agarose

matrices.


17/10417

Type and concentration of salt

The addition of various structure-forming (sa lting out ) sa lts to the eq uilibra tionbuffer and sample solution promotes ligand-protein interactions in H IC (10, 12, 36,65, 66). As the concentration o f such sa lts is increased, the amount of proteins boundalso increases almost linearly up to a specific salt concentration and continues toincrease in an exponential ma nner at still higher concentra tions.

This lat ter phenomenon is demonstra ted in Fig. 4 w here to ta l binding capacity ofPhenyl Sepharose H igh Performance for

-chymotrypsinogen and R NAse w as exa mi-

ned at gradually increasing salt concentrations.

1

20

3

40

60

80

2 4

Proteincapacitymg/m

lpackedbed

Initial salt concentration M (NH ) SO

-chymotrypsinogen

RNA se

4 2 4

In this experiment, the column was first equilibra ted w ith buffer conta ining varyingconcentra tions of sa lt a s indica ted in the Figure. The sample w as dissolved in buffer

including this initial salt concentration prior to application to the column. However,in those experiments w here the protein begins to precipita te a t high sa lt concentrat ion(1.3 M and 2.3 M ammonium sulphate for -chymotrypsinogen and RNAse respectively)the sample was dissolved at a slightly lower salt concentration.

The samples w ere loaded on the column until breakthrough could be observed a tthe column outlet. Then start buffer w ith initial salt concentrat ion w as run through thecolumn until UV-absorption in the eluent returned to the ba seline. Finally, the boundproteins were eluted with a decreasing salt gradient.

A significant increase in adsorption capacity can be seen w hen the salt concentra tionis increased a bove the precipita tion po int.

Fig. 4.

Protein binding capacity on

Phenyl Sepharose High

Performance as a function of

salt concentration in the

column equilibration buffer(Work from Amersham

Pharmacia Biotech, Uppsala,

Sweden).


18/10418

This phenomenon is proba bly due to the precipita tion o f proteins on the column.It ha s a concomita nt negative effect on the selectivity of the H IC ad sorbent.

The effects of sa lts in H IC can be accounted for by reference to the Hofmeister seriesfor the precipita tion of proteins or fo r their positive influence in increasing the molalsurface tension of water (for extensive review, see refs. 27,29). These effects aresummarized in Tables 1 and 2.

In both insta nces, sodium, potassium or ammonium sulphates produce rela tivelyhigher salting-out (precipita tion) or mo lal surface tension increment effects. It is alsothese salts that effectively promote ligand-protein interactions in HIC. Most of thebound proteins are effectively desorbed by simply w ashing the HIC adsorbent w ith

w at er or d ilute buffer solutions at near neutral pH .

Effect of pH

The effect o f pH in H IC is a lso no t stra ightforw ard . In general, a n increase in pHw eakens hydrophob ic interactions (10,41), proba bly as a result of increased titra tionof cha rged groups, thereby leading to an increase in the hydrophilicity of the proteins.On the other hand, a decrease in pH results in an apparent increase in hydrophobic

interactions. Thus, proteins which do not bind to a H IC adsorbent a t neutra l pH b indat a cidic pH (9). H jertn et al.(42) found tha t the retention o f proteins changed moredra stically a t pH values above 8.5 and/or below 5 tha n in the range pH 58.5 (Fig 5).

These findings suggest that pH is an important separation parameter in theoptimizat ion of hydrophobic interaction chromatogra phy and it is advisable to checkthe applicability of these observations to the particular separation problem at hand.

Increasing precipitation (salting -out) effect

Anions: PO43, SO

42, CH

3COO, Cl, Br, NO

3, CLO

4, I, SCN

Cations: NH4+, Rb+, K+, Na+, Cs+, Li+, Mg2+, Ca2+, Ba2+

Increasing chaotropic (salting-in) effect

Table 1.

The Hofmeister

series on the effect

of some anions and

cations in

precipitating

proteins.

Table 2.Relative effects of

some salts on the

molal surface

tension of water.

Na2SO

4>K

2SO

4>(NH

4)

2SO

4>Na

2HPO

4>NaCl>LiCl. . . >KSCN


19/10419

Fig. 5.

The pH dependence of the interaction between proteins and an octyl agarose gel

expressed as Ve/V

T(V

eis the elution volume of the different proteins and V

Tis the

elution volume of a non-retarded solute). Elution was by a negative linear gradient ofsalt. The model proteins used were STI=soy trypsin inhibitor, A=human serum

albumin, L=lysozyme, T=transferrin, E=enolase, O=ovalbumin, R=ribonuclease,

ETI=egg trypsin inhibitor and C=cytochrome c. (Reproduced with permission, from

ref. 42).

Effect of temperature

Based on theories developed for the interaction of hydrophobic solutes in water(22,37), Hjertn (38) proposed that the binding of proteins to HIC adsorbents is

entropy driven [ G = ( H -T S) ~ -T S], w hich implies that the interaction increaseswith an increase in temperature. Experimental evidence to this effect has beenpresented by H jertn (25) and Jennissen (34). It is interesting to no te tha t the van derWaals a ttraction forces, which operate in hydro phobic interactions (29), a lso increasew ith increase in tempera ture (39). H ow ever, an o pposite effect w as reported by Visser& Strat ing (40) indicat ing tha t the role of t empera ture in H IC is of a complex na ture.This apparent discrepancy is probably due to the differential effects exerted bytempera ture on the conforma tiona l sta te of different proteins and their solubilities inaqueous solutions.

In practical terms, one should thus be aw are tha t a dow nstream purification processdeveloped at room temperature might not be reproduced in the cold room, or viceversa.


20/10420

Viscosity Dielectric Surface tensionSolvent (centipoise) constant (dynes/cm)

Water 0.89 78.3 72.00Ethylene glycol 16.90 40.7 46.70Dimethyl Sulphoxide 1.96 46.7 43.54Dimethyl Formamide 0.796 36.71 36.76n-propanol 2.00 20.33 23.71

Table 3.

Physical properties of some

solvents used in HIC (data at 25 oC).

Additives

Low concentrations of water-miscible alcohols, detergents and aqueous solutionsof chaotropic (sa lting-in) salts result in a w eakening of the protein-ligand interactionsin HIC leading to the desorption of the bound solutes. The non-polar part s of a lcoholsand detergents compete effectively w ith the bound proteins for the adsorption sites onthe HIC media resulting in the displacement of the latter. Chaotropic salts affect theordered structure of w ater and/or tha t of the bound proteins. Both types of a dditivesalso decrease the surface tension o f w ater (see Table 3) thus w eakening the hydrophobicinteractions to give a subsequent dissocia tion o f the ligand-solute complex.

Although additives can be used in the elution buffer to affect selectivity duringdesorption, there is a risk tha t prot eins could be denatured or inactivated by exposureto high concentra tions of such chemicals. H ow ever, additives can be very effective incleaning up HIC columns that have strongly hydrophobic proteins bound to the gelmedium.


21/10421

BioPro

cessM

edia

Quality

anperfo

rmance

from

andreg

ulatory

suppor

tservic

es

fromPh

armacia

.

produc

tion.Ba

ckedb

yfullte

chnical

resear

chthro

ughsca

le-upto

BioPro

cessMe

dia

Quality

anperfo

rmance

from

andreg

ulatory

suppor

tservic

es

fromPh

armacia

.

produc

tion.Ba

ckedby

fulltec

hnical

research

throug

hscale-

upto

3

Product Guide

Amersham Pharma cia Biotech manufa ctures a w ide range of H IC media suitable foranalytical, small scale preparative and process scale applications. The HIC productrange is summarized in Table 4.

* Octyl Sepharose 4 Fast Flow is currently (December 1992) only availa ble as a C DM product(see p. 17), but will later be available as a standard catalogue product.

Phenyl Sepharose 6 Fast Flow (low sub) Suitable for all initial and

Phenyl Sepharose 6 Fast Flow (high sub) intermediate step purifications.

Butyl Sepharose 4 Fast Flow Available in laboratory pack sizes and

Octyl Sepharose 4 Fast Flow* bulk quantities.

Phenyl Sepharose High Performance Suitable for all high resolution

purifications.

Available in laboratory pack sizes,bulk quantities and as prepacked

columns.

Phenyl Sepharose CL-4B Traditional medium for all applications.

Octyl Sepharose CL-4B Available in laboratory pack sizes and

bulk quantities.

Alkyl Superose and For analytical and small scale

Phenyl Superose preparative applications.

Available as prepacked columns.

HIC Media Test Kit For screening different types of ligands

and for method development work at

small scale.

Five different HIC media as prepacked

1 ml columns.

Table 4.

HIC

products

available

from

Amersham

Pharmacia

Biotech.


22/10422

Fig. 6.

Structure of cross-linked agarose gels.

BioProcess M edia form a full range of separa tionmedia especially designed to meet the demands of toda ysindustrial production of biomolecules.

Productive: High flow rates, high capacity and high recovery lead to goodprocess economy.

Validated: Manufactured according to fully validated process with strictq uality standa rds and complete documenta tion.

Scaleable: Work equally w ell in labo ratory a nd pilot production systems aswell as in industrial operation.

Cleanable: Very high chemical stability enables thorough cleaning and

sanitiza tion treatments that reduce the risk of conta mination ofthe end product and increase the media lifetime.

Documented: Regulatory Support Files give full details of approval supportdata such as performance, stability (including leakage data),extractable compounds and analytical methods. A RegulatorySupport File is an invaluable starting point, especially forpharmaceutical process validations.

Guaranteed supply: Large production capacity and guara nteed future supply.

BioProcess Media

Base matricesThe BioProcess HIC media ra nge is ba sed o n the highly cross-linked beaded agarose

matrices Sepharose Fast Flow and Sepharo se H igh Performance. Their ma crostructuresconta ining polysaccharide chains arra nged in bundles (Fig. 6) are further strengthenedby dif ferent degrees of inter-chain cross-linking. The resulting ma cropo rous structurescombine good capacities for molecules up to 4x106(6% agarose) and 2.7x107(4%agarose) in molecular mass with excellent flow properties and high physical andchemical stability.

BioPro

cessM

edia

Quality

anperfo

rmance

from

andreg

ulatory

suppor

tservic

es

fromPh

armacia

.

produc

tion.Ba

ckedby

fulltec

hnical

resear

chthro

ughsca

le-upto


23/10423

All Sepharose based matrices have virtually no non-specific adsorption propertiesand a re a lso resista nt to microbial degrada tion due to t he presence of the unusual sugar3,6-anhydro-L-galactose.

Coupling

The HIC ligands are coupled to the monosaccharide units by stable ether linkages.The structures of the coupled ligands are shown in Fig. 3.

Chemical stabilityBioPro cess H IC M edia are stable in all commo nly used aq ueous buffers and solvents

in the pH range 2-14. When these media w ere challenged by storage for 7 da ys at 40oCin the solutions listed in Table 5, no significant change in chromatographic functionw as seen.

O f special interest is their sta bility in a lkaline solutions, as cleaning and sanitiza tionwith NaOH solutions are preferred in process applications. The functional stabilityand recommended pH ranges are summarized in Table 6.

The ligand leaka ge of B ioProcess H IC M edia a t different pH values has been testedand generally found to be extremely low (43). The pH range 214 can be used forcleaning-in-place (CIP) and sanitiza tion-in-place (SIP), see C leaning, sanitiza tion andsterilization procedures, page 63. BioProcess HIC Media are stable at high tempe-ratures and can be sterilized by autoclaving at 120 oC for 20 min.

Long term stability and recommended working pH range: 313

Short term stability and recommended CIP and SIP pH range: 214

Recommended long term storage: 0.01 MNaOH or20%ethanol.

Table 6.

Stability and recommendedpH ranges for BioProcess

HIC Media.

Tested media Test solutions

1 M acetic 3 M 70% 30%1 M NaOH acid 1 mM HCL (NH

4)2SO

4ethanol isopropanol 6 M GuHCl 8 M Urea

Phenyl Sepharose6 Fast Flow (low sub) X (n. t.) (n. t.) X X X X X

Phenyl Sepharose6 Fast Flow (high sub) X (n. t.) (n. t.) X X X X X

Butyl Sepharose4 Fast Flow X (n. t.) X (n. t.) X X X (n. t.)

Phenyl SepharoseHigh Performance X X (n. t.) (n. t.) X X X X

X = Functionally stable when tested for 7 days at + 40 C(n. t.) = Not tested

Table 5. Chemical stability test of BioProcess HIC Media.


24/10424

Physical stability

The highly cross-linked structures of Sepharose Fast Flow and Sepharose HighPerforma nce matrices are physically stable resulting in very good flow properties. Thisis illustrated by the pressure-flow rate curves for Phenyl Sepharose 6 Fast Flow show nin Fig. 7. In columns w ith 5 cm inner diameter and a bed height o f 15 cm, flow ratesup to 500 cm/h can be used w ithout exceeding a ba ck pressure of 1 ba r. The optima l

w orking flow ra te during elution is norma lly 50150 cm/h but during equilibrat ion,regeneration, and also often during sample application, higher flow rates of 200300 cm/h can be used. These higher flow ra tes reduce cycle times.

1.21.00.80.60.40.20.0

100

200

300

400

500

600

700

high sub

low sub

Pressure (bar)

Flow rate (cm/h)

Binding capacity

O ne of the major features of BioProcess H IC M edia is the high binding capacity,which results in high throughput and productivity even at relatively low saltconcentrations. Fig. 8 shows the total dynamic binding capacities of human serumalbumin and human IgG at different concentrations of ammonium sulphate asdetermined by frontal analysis. Phenyl Sepharose 6 Fast Flow (high sub) showed the

highest capacities for both hIgG and HSA. Phenyl Sepharose High Performance hadhigher capacity for hIgG compared with HSA while Butyl Sepharose 4 Fast Flowshow ed the reverse, indicating the difference in selectivity. The protein recoveries w heneluting with low salt buffer were all 80% or more.

The dynamic binding capa city w ill decrease w ith increasing linear flow rates. Thisis especially importa nt to consider w hen optimizing initial separa tion steps where largevolumes need to be processed. Productivity may be higher at high flow rates eventhough the binding capacity is decreased.

Fig. 7.

Typical pressure/flow rate curves

for Phenyl Sepharose 6 Fast Flow

(low sub) and Phenyl Sepharose 6

Fast Flow (high sub) in an XK 50/30

Column, bed height 15 cm, mobile

phase 0.1 M NaCl. (Work fromAmersham Pharmacia Biotech,

Uppsala, Sweden).


25/10425

Fig. 8.

Total adsorption capacities of Phenyl

and Butyl Sepharose media for

human IgG and HSA as a function of

the concentration of ammonium

sulphate in the equilibration buffer.

1=Phenyl Sepharose 6 Fast Flow

(high sub), 2=Phenyl Sepharose High

Performance, 3=Phenyl Sepharose 6Fast Flow (low sub), 4=Butyl

Sepharose 4 Fast Flow. (Work from

Amersham Pharmacia Biotech,

Uppsala, Sweden).

Phenyl Sepharose 6 Fast Flow (low sub)

Phenyl Sepharose 6 Fast Flow (high sub)

Phenyl Sepharose 6 Fast Flow (low sub) and Phenyl Sepharose 6 Fast Flow (high sub)

are based on highly cross-linked 6% agarose with phenyl ligands coupled via stableether linkages. The media characteristics are summarized in Table 7.

Phenyl Sepharose 6 Fast Flow (low sub) and Phenyl Sepharose 6 Fast Flow (high sub)were initially developed and tested in cooperation with leading pharmaceuticalmanufacturers. They are ideal for initial or intermedia te step purifica tion o f proteins

30

20

10

0.39 0.45 0.57 0.68

0.6 0.9 1.2 1.5

Concn. of ammonium sulphate (M)

Adsorptioncapa

city(mghIgG/mlgel)

Concn. of ammonium sulphate (M)

30

20

10

Adsorptioncapacity(mgH

SA/mlgel)

hIgG

HSA

1

2

3

4

1

4

32

Bead structure cross-linked agarose, 6%, spherical

Mean particle size 90 m

Particle size range 45165 m

Degree of substitution approx. 20 (low sub) and 40 (high sub)mol phenyl groups/ml gel

Further information is available in Data File 2040 (Code No. 18-1020-53).

Table 7.

Characteristics of Phenyl Sepharose 6

Fast Flow (low sub) and Phenyl

Sepharose 6 Fast Flow (high sub).

BioPro

cessM

edia

Quality

anperfo

rmance

from

andreg

ulatory

suppor

tservic

es

fromPh

armacia

.

produc

tion.Ba

ckedby

fulltec

hnical

research

throug

hscale-

upto


26/10426

BioProces

sMedi

a

Quality

anperfo

rmancefro

m

andreg

ulatory

suppor

tservic

es

fromPh

armacia

.

produc

tion.Ba

ckedby

fulltec

hnical

research

throug

hscale-upto

and peptides w ith a low to medium degree of hydrophobicity. The availability o f tw odegrees of substitution increases the possibility of finding the best selectivity andcapacity f or a given applica tion.

Phenyl Sepharose 6 Fast Flow (high sub) has been used a s an effective capture stepin methods for the purification of recombinant human Epidermal Growth Factor(h-EG F) and recombinant Pseudomonas aeruginosaexotoxin. These applications a re

presented in chapter 6, pages 87 and 92 respectively.

Product availability

Phenyl Sepharose 6 Fast Flow (low sub) and Phenyl Sepharose 6 Fast Flow (highsub) are supplied a s suspensions in 20% ethanol in packs of 200 ml, 1 litre and 5 litres.


Buty l Sepharose 4 Fast Flow is based on highly cross-linked 4% aga rose with butylligands coupled via stable ether linkages. The characteristics of this medium aresummarized in Table 8.

Buty l Sepharose 4 Fast Flow w as initially developed and tested in cooperation w ithleading pharmaceutical ma nufacturers. It is intended for the initia l or intermedia te steppurificat ion of proteins and peptides w ith a low to medium degree of hydrophobicityand often w orks efficiently w ith rather low salt concentra tions.

For the buty l ligand, the mechanism of adsorption a nd desorption is different tha nfor the phenyl ligand , w hich gives a difference in selectivity. This was illustra ted in a napplicat ion w here recombinant human Annexin V, expressed in E. coli ,w as purifiedusing HIC after an initial capture step on a cation exchanger. A comparison of thechromatograms in Fig. 9 shows that the elution position of Annexin V and the mainimpurities interchanged when changing from Butyl Sepharose 4 Fast Flow to PhenylSepharose 6 Fast Flow (high sub).


Butyl Sepharose 4 Fast Flow is supplied as suspension in 20% ethanol in packs of200 ml, 500 ml and 5 litres.


Mean particle size 90 mParticle size range 45165 m

Degree of substitution approx. 50 mol butyl groups/ml gel

Further informa tion is a vailable in D at a File 3300 (C ode N o. 18-1020-70).

Table 8.

Characteristics of Butyl Sepharose 4

Fast Flow.


27/10427

Time (min)0

A280nm

Annexin V

60Time (min)0

A280nm

Annexin V

60

Medium: Butyl Sepharose 4 Fast FlowColumn: XK 16/20Buffer A: 20 mM sodium phosphate pH 7.0

+ 1.0 M ammonium sulphateBuffer B: 20 mM sodium phosphate pH 7.0Sample: Partially purified Annexin V

expressed in E. ColiSamplevolume: 5 mlFlow rate:100 cm/hGradient: 0100% B, 10 column volumes

Medium: Phenyl Sepharose 6 FastFlow (high sub.)

Column: XK 16/20Buffer A: 20 mM sodium phosphate

pH 7.0 + 1.0 M ammoniumsulphate

Buffer B: 20 mM sodium phosphatepH 7.0

Sample: Partially purified Annexin Vexpressed in E. Coli

Samplevolume: 5 mlFlow rate:100 cm/hGradient: 0100% B, 10 column

volumes

Fig. 9. Purification of Annexin V on Butyl Sepharose 4 Fast Flow and Phenyl Sepharose 6 Fast Flow

(high sub). (Work from Amersham Pharmacia Biotech, Uppsala, Sweden).

Phenyl Sepharose High Performance

Phenyl Sepharose High Performance is based on very highly crossed-linked 6%agarose w ith phenyl ligands coupled via stab le ether linkages. The characteristics ofthis medium are summarized in Table 9.

BioPro

cessM

edia

Quality

anperfo

rmance

from

andreg

ulatory

suppor

tservic

es

fromPh

armacia

.

produc

tion.Ba

ckedby

fulltec

hnical

research

throug

hscale-

upto



Particle size range 2444 m

Degree of substitution approx. 25 mol phenyl/ml gel

Further information is available in Data File 2050 (Code No. 18-1020-56).

Table 9.Characteristics of

Phenyl Sepharose High Performance.

Phenyl Sepharose High Performance is ideal for laboratory and process scaleintermediate step purifications where high resolution is needed. The separation of

slightly modified variant s, clipped forms etc., o f a recombinant protein from the nat iveprotein is a typica l application exa mple. It ha s also proven to be very efficient for thepurifica tion o f mo noclonal antibodies. Tw o large scale applications on monoclonal


28/10428

antibodies, one for the purification of anti-gp120, which is in clinical trials fortreatment of AIDS, the other for the purification of an antibody used in d iagnostic tests,are presented in Chapter 6, pages 85 and 90 respectively.


Phenyl Sepharose High Performa nce is supplied a s a suspension in 20% ethanol in

packs of 75 ml, 1 litre and 5 litres and prepacked in H iLoad 16/10 and 26/10columns.

Custom Designed HIC Media

Custom Designed Media (CDM) meet the needs of specific industrial processsepara tions where chromato graphy media from our standa rd range are not suita ble.CDM can be made to meet BioProcess Media specifications if required.

The CDM group at Amersham Pharmacia Biotech works in close collaborationwith the customer to design, manufacture, test and deliver media for specializedseparation requirements. Several CDM products are also available to the generalmarket. Some HIC media first produced as C ustom D esigned M edia have proven sosuccessful that they have subsequently been introduced as standard products, e.g.Phenyl Sepharose 6 Fast Flow (low sub), Phenyl Sepharose 6 Fast Flow (high sub) andButyl Sepharose 4 Fast Flow.


Please conta ct your loca l Amersham Pharmacia Biotech representa tive for furtherdetails of C D M products and services.

HIC Media Test Kit

H IC M edia Test Kit consists of five ready-to-use 1 ml plastic columns for screeningdifferent types of ligands and for method development work at small scale.

The kit conta ins the follow ing H IC media:

Phenyl Sepharose High Performance

Phenyl Sepharose 6 Fast Flow (low sub)

Phenyl Sepharose 6 Fast Flow (high sub)


Octyl Sepharose 4 Fast Flow


Please conta ct your loca l Amersham Pharmacia Biotech representa tive for furtherinformation.


29/10429

Phenyl Sepharose CL-4B and Octyl Sepharose CL-4B

Phenyl Sepharose CL-4B and Octyl Sepharose CL-4B are produced in largequantities w ith high and consistent quality. Their performance has been demonstra tedin hundreds of applicat ions and they have been approved by regulatory a uthorities foruse in many pharmaceutical production processes.

Phenyl Sepharose CL-4B a nd O ctyl Sepharose CL-4B a re based o n cross-linked 4%agarose matrices w ith ligands coupled via stab le ether linkages. The media chara cte-ristics are summarized in Table 10.

Phenyl Sepharose CL-4B and Octyl Sepharose CL-4B are stable in all commonlyused aqueous buffers. Long term stability and recommended working pH range is312. 1 M Na OH can be used for short term exposure in cleaning and sanitizationprocedures, see C leaning, sanitiza tion and sterilizat ion procedures, page 63. Shortterm stability and recommended CIP and SIP pH range is 214.

M aximum flow rate for a laboratory-sca le column w ith an internal diameter of upto 5 cm and a bed height of up to 15 cm is 150 cm/h.


Phenyl Sepharose CL-4B a nd O ctyl Sepharose CL-4B a re supplied a s suspensionsin 20% ethanol in packs of 50 ml, 200 ml and 10 litres.

Phenyl Superose and Alkyl Superose

Phenyl Superose and Alkyl Superose are media for high performance H IC, availablein prepacked columns for use in FPLC , SMART System (Phenyl Superose only) orH PLC systems.

Phenyl and neopentyl groups respectively are attached to the matrix via a stableether linkage. The characteristics of these media and columns are summarized in

Table 11.



Paricle size range 45-165 m

Degree of substitution approx. 40 mol phenyl or octylgroups/ml gel

Table 10.

Characteristics of Phenyl Sepharose

CL-4B and Octyl Sepharose CL-4B.


30/10430



Column sizes 5x50 mm (HR 5/5)10x100 mm (HR 10/10)1.6x50 mm (Phenyl Superose, PC 1.6/5for SMART System)

Further information is ava ilable in Dat a File for prepacked HR columns (Co de

No. 18-1009-26) and in D at a File for prepacked PC columns(Co de N o. 18-1009-02).

Table 11.

Characteristics of Phenyl Superose

and Alkyl Superose.

Phenyl Superose and Alkyl Superose are stable in all commonly used aqueousbuffers. Long term stability a nd recommended w orking pH range is 213. 1 M NaOHcan be used for cleaning and sanitizat ion, see C leaning, sanitizat ion and sterilizationprocedures, pa ge 63. Short term sta bility and recommended C IP and SIP pH range is214.

The columns are typically used in labora tory scale protein purificat ion schemes oras an ana lytica l tool, a s a complement to e.g. ion exchange chromatogra phy and gelfiltration. Examples of applications are shown in chapter 6, page 7983. Suitableprotein loads are in the mg range (HR columns) or, for micropurification, in the ng-g ra nge (Phenyl Superose, PC 1.6/5). Alkyl Superose is less hydrophobic tha n PhenylSuperose and is therefore part icularly suitable for high performance H IC w ith retainedbiological activities of la bile proteins and of proteins which bind very tight ly to mediawith higher hydrophobicities.


31/10431

ExperimentalDesign

This chapter will deal with experimental methods of HIC which are applicable inthe majority of cases. Since the facto rs w hich influence HIC are numerous, the relevantchromatographic parameters that lead to the selective purification o f the protein(s) ofinterest should be opt imized on a case to case ba sis.

Hydrophobicity of proteins

It is estimated tha t a s much as 4050% of the accessible surface area of proteins isnon-polar (44, 45). These areas are responsible for the binding of proteins to HICadsorbents in the presence of moderate to high concentra tions of salting-out salts. The

strength o f this salt-promoted interaction may be predicted from the close relat ionshipbetw een precipitation da ta for proteins and their relat ive retention on H IC adsorbents(27). Since such retention data are not readily available for the large majority ofproteins, they must be esta blished from case to case for the protein(s) of interest in abiologica l sample.

Multivariate mapping

This is a useful method for:i. C haracterizing hydrophobic media o n the basis of their selectivity (46).

ii. Choosing the most suita ble medium for the optimum resolution of tw o closelyrelated proteins.

iii. D etermining the adsorption behaviour of proteins on H IC media and therebyestablishing a practical hydrophobic scale for the proteins in question.

4


32/10432

The results obtained in our laboratories (46) suggest that:

i. The adsorption selectivity of O ctyl Sepharose CL-4B is related to the fraction ofhydrophobic amino acids in the model proteins examined.

ii. The retention of proteins on a lkyl Superose and Pyridine sulphide-Sepharose6 Fast Flow is proportional to a parameter best described as absence of surface

charge on the sample molecules.iii. The phenyl- and butyl-based media separa ted proteins according to a combination

of t he above tw o mechanisms.

iv. D ifferent hydrophobicity coefficients co-variat e with t he retention dat a establishedfor the various hydrophobic media examined.

Multivariate analysis thus opens new possibilities in the design of HIC and otherchromatography-ba sed separa tions by using a minimum number of experimenta l da ta .

Strategic considerations

O ne of the most important a spects of developing a complete purificat ion scheme isto keep the number of unit operations to a minimum. A logical approach to reach thehighest possible purity w ith the smallest number of individua l chromatographic stepsis to combine techniques based on d ifferent principles and thus exploit different surfaceproperties of the substances to be separated. However, the sequence in which thechosen techniques are used must be carefully planned. In many applications HIC is

useful especially in combinat ion w ith techniques such as ion exchange chromatographyand gel filtra tion. As an example, hydrophobic interaction chromatogra phy is a logica lchoice when the sample already has a high ionic strength. The conductivity of mostbiologica l start ing ma terials is typically in the range of 1530 mS/cm, w hich makesH IC an at tractive alternat ive to ion exchange chromatogra phy (IEX) in the first stepof a dow nstream purification scheme. H igh conductivity in the sta rting material w illreduce the binding capacity of ion exchange media and some type of conditioning suchas desalting, diafiltration or dilution has to be included before an ion exchange step.In contrast, the only conditioning needed if HIC is used, is to add enough salt to

promote the proper binding to the medium. Used in the first step HIC, like IEX andother adsorpt ion techniq ues, w ill serve as an effective means of concentra ting a dilutesample.

O ther typical points in a purification scheme w here HIC fits in natura lly a re afteran ammonium sulphate precipitation, which often comes in the beginning of adow nstream process, and after an ion exchange step w here the sample is eluted w itha ra ther high ionic strength. The further addition of salt tha t might be needed to reta rdthe components in a desired w ay on the H IC medium is thus a very simple linking step.

In a similar w ay , a sample eluted from a H IC step in a low ionic strength buffer can

often be directly applied to an ion exchange column without an extra dialysis ordesalting step.


33/10433

Choice of HIC media

The type of immobilized ligand, the degree of substitution and the type andconcentra tion of salt and pH used during the adsorption stage have a profound effecton the overall performance (i.e., selectivity and capacity) of a HIC medium [seeChapter 2]. Moreover, the type of matrix used and the coupling chemistry can also

influence to a variab le degree the binding a nd elution behaviour o f ma ny proteins. Thepractical implications of these effects are that different H IC media must be comparedmuch more rigorously than ion exchange or affinity media, especially when the HICstep is part of a downstream purification process intended for an industrial scaleoperation.

General considerations

i. The H IC medium should bind the protein of interest a t a reasonably lowconcentra tion o f salt . This is often dependent on the type of salt chosen, e.g. up tofour t imes higher concentra tion of Na Cl might be necessary to obt ain a bindingeffect comparable to tha t o bta ined w ith ammonium or sodium sulphate. The sa ltconcentration should be below the concentration that causes precipitation ofdifferent proteins in the crude feed stock. 1 M ammonium sulphate is a goodstarting point fo r screening experiments. If the substance does not bind in e.g. 1 Mammonium sulphate, then choose a more hydrophobic medium. The right choiceof a suitable HIC medium can often lead to a lower consumption of salt in thebinding buffer. This in turn ha s a direct bearing on the economic and environmentalaspects of the purification process, especially fo r large-scale HIC applicat ions.

ii. The bound protein should be eluted from the column w ith salt-free buffer and w ithhigh recovery (75% or higher). If non-pola r solvents are required fo r its elution, trya less hydrophobic medium.

iii. The pH o f the start buffer and the type of salt to use are both para meters tha t canbe exploited to maximize selectivity during the adsorption phase. This is done bychecking the adsorption properties of the media at different pH-values and withdifferent types of sa lts during the screening of different ligands.

iv. Since hydrophobic interaction is dependent on temperature, it is important thatmethod development w ork is performed at the intended fina l working tempera ture.


34/10434

Screening experiments

This section out lines a general procedure for performing H IC screening experimentsw here emphasis is laid on optimizing selectivity by proper choice of H IC medium andby roughly defining the most critical experimental pa rameters. It a lso presents sometypical elution profiles that could be obta ined in a variety of situations follow ed byrelevant discussions of the results and recommendations for further experimentalwork.

i. Pack the media in suitable columns according to our packing recommendations(a bed volume of 110 ml is adequate) or use the HIC Media Test Kit fromAmersham Pharmacia Biotech. The HIC M edia Test Kit consists of five 1 ml plasticcolumns prepacked w ith BioProcess H IC media. For more information a bout theH IC M edia Test Kit , see Chapter 3, Selection G uide.

ii. Equilibra te the column w ith 2 bed volumes of the equilibrat ion Buffer A (50 mMsodium phosphate, 1.0 M ammonium sulphate, pH 7.0). Use a consta nt flow rate

througho ut (e.g. 100 cm/h).

iii. Apply a suitab le amo unt of sample, a lso conta ining 1.0 M ammonium sulphate (pHadjusted to 7.0), to the column and wash with 23 column volumes of Buffer A,or until the UV-tra ce of the effluent returns to near baseline.

iv. Elute the bound f raction using a linear and descending salt gra dient from 0 to 100%Buffer B (50 mM sodium phosphate buffer, pH 7.0). A to ta l gradient volume of 10bed volumes is usually sufficient.


35/10435

Evaluation of results

Figs. 10 to 15 show some typical elution profiles that could be obtained fromscreening experiments. The shaded a rea show s the elution position of the protein ofinterest. Each chromat ogra m is accompanied by a genera l discussion of the results andsuggestions for further experiments to optimize the separation of the protein ofinterest.

Result: Product is eluted early in gradient. Resolution is not satisfactory.

Discussion: No t much can be gained in this situat ion by changing sa lt concentra tion.Decreasing the salt concentration will decrease the binding capacity ofthe protein of interest and might even lead to its elution together w ith theunbound fraction.

Increasing the salt concentration might lead to the co-adsorption ofunw anted impurities and thereby lead to a decrease in the selectivity o f

the adsorbent for the protein of interest.Changing the pH of the equilibration buffer might result in strongerbinding a nd higher selectivity for the protein o f interest. The effect o f pHis variab le for different proteins and usually a low ering of the pH leadsto increased binding of proteins. Increasing the pH usually leads to adecreased b inding of proteins, w hich, in this part icular case, might resultin the elution of the protein of interest together with the unboundfraction.

Next step: Repeat the experiment at a lower and a higher pH. If no improvementin selectivity is obtained TRY A MEDIUM WITH A DIFFERENTLIG AND or, if available, A MEDIUM WITH A H IG H ER D EG REE OFLIG AND SUBSTITUTION.

Elution volume

Rel.Abs

Fig. 10.


36/10436

Elution volume

Rel.Abs

Fig. 11.

Result: Product is eluted near the end o f the gradient.Resolution is not satisfactory

Discussion: A decrease of the salt concentration w ill w eaken the strength of bindingresulting in the earlier elution o f the protein of interest. It may a lso ha vea positive effect on selectivity since more of the less hydrophobicsubstances w ill be eluted together w ith the unbound f raction. H ow ever,the effect of this approach on the resolution is marginal since theconta minant s are eluted very close to t he protein of interest, both beforeand after. Changing the pH of the equilibration buffer may have apositive effect on resolution and should be tried.

Next step: Repeat the experiment a t a higher and a low er pH of the eq uilibra tionbuffer. If no improvement in resolution is obta ined TRY A M EDIUMWITH A DIFFERENT LIG AND or, if available, A MEDIUM WITH ALOWER DEGREE OF LIGAND SUBSTITUTION.


37/10437

Elution volume

Rel.Abs

Fig. 12.

Result: Product is eluted in the middle of the gradient.Resolution is not satisfactory.

Discussion: C hanging the concentrat ion of salt in the equilibration buffer w ill havea limited effect on resolution. H ow ever, a change of pH of the equilibrationbuffer (bo th low er and higher pH values) might have a f avoura ble effect.

Next step: Repeat the experiment at a higher and a lower pH value. If noimprovement in resolution is obtained TRY A MEDIUM WITH AD IFFERENT LIG AND or, if available, A MED IUM WITH A HIG H ERDEGREE OF LIGAND SUBSTITUTION.


38/10438

Elution volume

Rel.Abs

Fig. 13.

Result: Product is eluted early in gradient. Resolution is satisfactory.

Discussion: In principle, this can be a good choice of medium. H ow ever, the fact tha tthe protein of interest is eluted very early in the gradient indicates tha tthe binding capacity may be low. This might be compensated for, ifnecessary, by a moderate increase of the salt in the equilibra tion buffer.This in turn may lead to a decrease in the selectivity o f the adsorbent sincesome of the unbound proteins might be adsorbed together with theprotein of interest. Another negative effect of increased salt concent-ra tion may be a decrease in resolution caused by the increase in grad ientslope if the total gradient volume, or the cycle time, is kept constant.Increased salt concentra tion w ill a lso give increased costs w hich may beof importa nce if the HIC step is to be a part of a manufa cturing process.Finally, not much can be gained by changing the pH of the eq uilibra tionbuffer since the resolution obtained was considered to be satisfactory.

Next step: Continue with method development as outlined under Optimizing aH IC step.

If low binding capacity is a problem a nd problems w ith increased sa ltconcentration as outlined above are encountered TRY A MEDIUMWITH A DIFFERENT LIG AND or, if a vailable, A M EDIUM WITH AH IG H ER D EG REE OF LIG AND SUBSTITUTION .


39/10439

Result: Product is eluted near the end o f the gradient. Resolution is sat isfa ctory .

Discussion: This can a lso be a good choice of medium. Decreasing the concentra tionof sa lt in the eq uilibra tion buffer w ill give earlier elution o f the proteinof interest, reduced cycle time and decreased cost for salt.

A disadvantage in this situation might be tha t some of the most hydro-

phobic contaminating substances bind so strongly that some organicsolvent or chaotropic agent has to be used for their removal.

Not much can be gained by changing pH since the selectivity is alreadygood.

Next step: Continue with method development as outlined under Optimizing aHIC step. If problems with very strong binding of hydrophobiccontaminants are encountered TRY A MEDIUM WITH A DIFFE-RENT LIGAND or, if available, A MEDIUMWITH A LOWER D EG REEOF LIGAND SUBSTITUTION.

Elution volume

Rel.Abs

Fig. 14.


40/10440

Result: Product is eluted in the middle of the gradient. Resolution is satisfactory.

Discussion: The choice of ligand is very good a nd there is less risk of strong bindingof the most hydrophobic contaminants.

Next step: Continue with method development as outlined under Optimizing aH IC step.

The exa mples presented a bove do no t cover tw o extreme cases tha t ma y a rise, i.e.the situation in which the protein of interest is either not bound to the HIC mediumor that it binds so strongly that it is difficult to elute it without using denaturingsolvents. In both instances, one should t ry to use a different H IC medium or use anothermedium which opera tes on a different separation principle.

In some of the examples above it is assumed that resolution is inadequate. Thereq uirements for resolution in any part icular chromatographic step must be stipulatedon a case-by-case ba sis. What sometimes seems to be fa irly ba d resolution can of tenbe good enough if it is an initial capture step w here the main objective is reduction o fvolume, removal of critical contaminants and preparation for higher resolutionchromatography.

Elution volume

Rel.Abs

Fig. 15.


41/10441

Optimizing a HIC step

The main purpose of opt imizing a chromatographic step is to reach the pre-definedpurity level w ith highest possible recovery by choosing the most suitable combinationof the critical chroma tographic parameters. In process applicat ions there is a lso a needto reach the highest possible throughput. The screening experiments outlined previously

will mainly help in establishing the most suitable medium to use. The sections belowwill deal with some important guidelines for optimizing the critical operationalparameters which affect the maximum utilization of the HIC step. These parametersinclude: type of buffer salt, salt concentration, buffer pH, temperature, bed height,flow rate, gradient shape and gra dient slope.

The solute

As in other adsorption chromatogra phy techniques, the way H IC is used dependson the size of the solute molecule.

Small molecules such as small peptides interact with the medium by single pointattachment. Their migration velocity depends directly on the binding constant of asingle bond and can vary over a wide interval depending on the ionic strength of themobile phase. Larger molecules such as proteins and nucleic acids interact with themedium by multi-point a tt achment. Their migration velocity depends on the sum ofseveral bonds. Thus their velocity is extremely low at a ll ionic strengths over a certa invalue. The protein is more or less stuck to the column. Below this ionic strength, t he

protein is practically not retarded at all (47).The interval of eluting strength where a large molecule is partly retarded on the

column is thus much smaller tha n fo r a small molecule. This means tha t purifying la rgemolecules on H IC is a typical on-off t echnique where the difference in retention for themolecules to be separated can be substantial at any specific ionic strength. In otherwords, separation of large molecules on HIC is a high selectivity technique. Theseparation should be optimized by ma nipulating the parameters af fecting the selectivityof the system, i.e. optimizing the chemistry of the system by means of salt concent-ration, type of salt, pH, gradient slopes or stepwise elution schemes. By effecting

relatively small changes in selectivity, large changes in resolution can occur.

When purifying small molecules on the other hand, the selectivity of the system isusually much low er and the requirements for purity might no t be met by w orking onthe selectivity alone.

The efficiency parameters such as bed height, bead size, theoretical plates, linearflow rate and sample volume may a lso ha ve to be optimized.

In this handbook how ever, the focus will be on large molecules such as proteins andlarge peptides.


42/10442

In conclusion, when purifying large molecules such as proteins, relatively shortcolumns can be used if the selectivity of the adsorbent is exploited in an o ptimal w ay.The linear flow rate should, if required, be sufficiently reduced in order to opt imize thekinetics of the adsorption a nd desorpt ion process. This can a lso be further enhancedby choosing a smaller bead size. Smaller beads w ill also provide the necessary increacein efficiency w hen more difficult separation problems are encountered.

The solvent

This is one of the most important pa rameters to ha ve a significant influence on thebinding capa city and selectivity of a H IC medium. In general, the adsorpt ion processis often more selective than the desorption process and it is therefore important tooptimize the starting binding buffer conditions with respect to critical parameterssuch as pH , type of salt, concentra tion of salt and temperature. The combinat ion of saltand pH can be manipulated to give optimum selectivity during purificat ion by H IC.O ptimal conditions differ from applicat ion to applica tion a nd a re best established byrunning linear grad ients and va rying the parameters in a controlled w ay (for exa mpleby using Facto rial design). Changes of tempera ture and pH are sometimes restrictedby the stab ility o f the substa nce of interest or by system constraints etc. but may of tenbe of interest t o eva luate. The Hofmeister series (Table 1) gives importa nt guidelinesin choosing the type of salt to use. The most efficient salts are normally ammoniumsulphate and sodium sulphate but a lso w eaker sa lts such as sodium chloride shouldbe considered. In an ideal situa tion, the correct choice of sa lt and salt concentra tion w illresult in the selective binding of the protein of interest while the majority of the

impurities pass through the column unreta rded. If the protein of interest binds w eaklyto the column, an a lternative approa ch is to choo se the sta rting buffer conditions w hichw ill result in the maximum binding of a large proportion o f the conta minating proteinsbut allowing the protein of interest to pass through unretarded. An extension of thisstra tegy is to increase the sa lt concentra tion in the unbound fra ction to such an extenttha t the protein of interest binds to the same column in a second run w hile most o f theimpurities pass through the column unretarded.

The effect of varying the concentration of salt in the binding buffer on thepurification of a monoclonal antibody (IgG

1) from mouse ascites fluid is shown in

Fig. 16. The column of Alkyl Superose w as equilibra ted w ith varying concentra tionsof a mmonium sulphate (2 M to 0.8 M) and its selectivity for the IgG

1investiga ted. The

results show tha t high selectivity fo r IgG1is obta ined using 1 M a mmonium sulphate

in the binding buffer.

It should be pointed out tha t the higher the salt concentra tion in the equilibra tionbuffer, the greater the risk tha t some of the proteins in the sample w ill precipitate. Sincesuch precipitates can clog tub ings and column filters, the sample must be filtered orcentrifuged. This extra step can be avoided by equilibra ting the sample in a low er saltconcentration than is required for its precipitation and then applying it to a column

w hich is eq uilibra ted w ith a higher salt concentrat ion (48). Some of the proteins w illprecipita te on the column (zone precipita tion) but they redissolve upon reduction o fthe salt concentration during stepwise or gradient elution.


43/10443

Fig. 16. The effect of starting conditions in HIC. Sample, 100 l anti-CEA MAB (-IgG1) from mouse ascites

fluid in 0.8 M (NH4)

2SO

4(corresponding to 20 l ascites); column. Alkyl Superose HR 5/5; flow rate, 0.5 ml

min -1; buffer A, 0.1 M sodium phosphate, pH 7.0, (NH4)2SO

4). (a) Sample applied in 2 M (NH

4)

2SO

4: both

albumin and IgG are absorbed. (b) Sample applied in 1.5 M (NH4)

2SO

4: less albumin binds and IgG elutes

earlier in the gradient. (c) Sample applied in 1.0 M (NH4)2SO

4: albumin does not bind and, therefore, the

column has a greater capacity for binding IgG. (d) Sample applied in 0.8 M (NH4)

2SO

4: albumin does not

bind; IgG is retarded, but elutes in a broad peak. (Work from Amersham Pharmacia Biotech, Uppsala,

Sweden).

a) b)

c) d)

Albumin

IgG IgG

Albumin

IgG

IgG


44/10444

When sample is applied at a salt concentra tion low er tha n tha t used for equilibra tionof the column, the sample volume becomes important. This is demonstrated in Fig. 17.When a 500 l sample of a scites fluid w as applied t o a 1 ml column of Alkyl Superose,a lbumin, the w eakest interact ing substance, sta rted to elute during sample applicat ion(Fig. 17 a). Dividing the sample into portions, e.g. five 100

l samples and adding

eq uilibra tion buffer (1.3 ml) a fter each sample applicat ion to enhance the hydrophobicinteraction prevented early elution of a lbumin (Fig. 17 b).

ElutionThis can be achieved by:

i. A linear o r step-w ise decrease of the concentration of salt.

ii. Adding various proportions of o rganic solvents to the elution buffer (see Chapter 2)provided tha t the protein o f interest is stab le upon exposure to such solvents. Theseadditives decrease the polarity or surface tension of the eluent resulting in areduction in the binding strength and the elution of the bound proteins from thecolumn. Usually, 40% ethylene glycol or 30% iso-propa nol, d issolved in salt-free

buffer, is used. In some applicat ions, it can be adva nta geous to linearly increase theconcentration of such additives as the salt concentration of the elution buffer issimultaneously decreased by a linear gra dient. The latt er procedure can sometimeslead to increased resolution o f the bound pro teins.

Fig. 17. The effect of loading conditions in HIC. Column, Alkyl Superose HR 5/5; flow rate, 0.5 ml min-1;

buffer A, 0.1 M sodium phosphate, pH 7.0, 2 M (NH4)2SO

4. (a) Sample (500 l anti-CEA MAB (IgG

1) from

mouse ascites fluid in 0.9 M (NH4)

2SO

4(corresponding to 115 l ascites) applied in one injection. (b)

Sample as (a) applied in five 100 l injections with 1.3 ml 2.0 M (NH4)

2SO

4after each portion. (Work from

Amersham Pharmacia Biotech, Uppsala, Sweden).

20I I I

40 60

A280 nm

Time (min)

2.00.5

0

I

I

conc.

(NH4)2SO4

(M)

0

a

20I I I

40 60

A280 nm

Time (min)

2.00.5

0

I

conc.

(NH4)2SO4

(M)

0

b

0.25


45/10445

iii. Adding neutra l detergents (usua lly 1%) to the elution buffer. H ow ever, somedetergents are bound so strongly tha t they a re difficult to w ash out completely w ithcommon organic solvents (e.g. ethanol). In the worst case, this might lead to adecrease in the capacity of the HIC medium for subsequent applications. Theseadditives must therefore be used w ith care.

The preferred method of elution is a linear or step-wise decrease of the salt

concenta tion in the elution buffer. Some typical examples are presented below .

Gradient elution

Simple linear gradients are the first choice for screening experiments, but w hen moreexperience is at hand it might be advantageous to make a grad ient more shallow inareas where resolution is inadequa te. Consequently, a reas where resolution is good canbe covered by a steep gradient (Fig. 18). Such complex gradients offer maximum

flexibility in terms of combining resolution w ith speed during the same separation.By increasing the tota l gradient volume (i.e. decreasing gra dient slope) of a linear

gradient, resolution w ill be improved in a ll parts of the chromatogram (Fig. 19). Thisis usually no t the best a pproach in preparative mode w here the prime issue is not toresolve as many peaks as possible but to separate the compound of interest from therest of the compounds in the feed material. Increased gradient volume will also giveincreased cycle time and the separated fra ctions w ill a lso be more diluted.

Fig. 18. Effect of a complex gradient on resolution.

Elution volume

Rel.Abs

Elution volume

Rel.Abs


46/10446

Elution volume

Rel.Abs

Fig. 19. Effect of gradient slope on resolution.

Fig. 20. Switching from a continuous gradient to step-wise elution.

Elution volume

Rel.Abs

Elution volume

Rel.Abs

Elution volume

Rel.Abs


47/10447

Step-wise elution

Step-w ise elution is often preferred in large scale prepara tive applica tions since it istechnically more simple and reproducible than gradient elution.

Step-w ise elution can sometimes be advanta geous a lso in small scale applicationssince the compound of interest can be eluted in a more concentra ted form if the eluting

strength of the buffer can be kept high enough without causing co-elution of morestrongly bound compounds.

The principle of step-w ise elution is to increase resolution in the area w here the peakof int erest elutes. Fig. 20 illustra tes how a t hree step increase in eluting strength can beused to obtain maximum resolution of the fraction of interest (shaded peak).

In the first step, the strength a nd the volume of the elution buf fer is optimized toelute all compounds binding less strongly to the gel tha n the compound of interest. Theelution strength a nd vo lume of buffer should be large enough to elute these contam-inat ing w eaker binding substances, but it must not exceed tha t level w here the peak of

interest starts to co-elute with the contaminating compounds.

In the second step the elution strength is increased t o the point w here the compoundof interest elutes. The elution strength should be large enough to elute the compo undof interest w ithout excessive dilution, b ut must b e kept below the level w here the morestrongly bound contaminating compounds start to co-elute.

In the final step, the elution strength is further increased to elute all of the remainingconta minat ing compounds. This step can be a very short one w ith high elution strength.

When step-wise elution is applied, one has to keep in mind the danger of getting

artefact peaks w hen a subsequent step is administered too early af ter a ta iling peak. Forthis reason it is recommended to use continuous gradients in the initial experi-mentsto characterize the sample and its chromatographic behaviour.

Sample load and flow rate

The through-put of the method can be increased by increasing sample loa d a nd f lowrate. H ow ever, this has to be tra ded of f a ga inst decreased resolution (efficiency). The

effects of sample load and flow rate are further discussed below under Processconsiderations.

Regeneration

After each cycle, bound substa nces must be w ashed out fro m the column to restorethe original function of the medium. H IC adsorbents can norma lly be regenerated byw ashing w ith distilled w ater after each run. To prevent a slow build up of contaminants

on the column over time, more rigorous cleaning proto cols may have to be applied ona regular basis. (See page 63, C leaning, sanitiza tion a nd sterilizat ion procedures).


48/10448

Process considerations

In contra st to ana lytical chromatogra phy or small scale prepara tive chromatogra phyin research and development, process chromatography is used as part of a manufacturingprocess. M ethod development w ork ha s to focus on purifying the product o f interestto the highest yield a nd the required purity as quickly, cheaply and easily as possible,

i. e. to find the conditions that give the highest possible productivity (amount ofproduct produced per volume of media a nd unit t ime) and process economy.

Method optimization in process chromatography

Firstly, selectivity for the substance of interest is max imized by choosing the propertype of media , pH , type of sa lt, salt concentra tion and temperature, a s has already beenoutlined above.

In HIC, as for most other adsorption techniques, there are then basically twoalternative routes to follow :

i. If H IC is used in an intermediate or final step w here the need for resolution is highin order to meet purity requirements for the final product, the resolution ismaximized by w orking on the eluting conditions such as grad ient shape, grad ientslope or concentration and volume of steps in a step-wise elution procedure.Resolution should be the highest possible while still keeping separation timereasonably short and avoiding excessive dilution of eluted product. From thispoint, f low rate and sample load a re optimized to find highest possible product ivity

where resolution is still high enough to meet the predefined purity requirements.

In HIC, as in ion exchange chromat ography, sample load , flow rate and gradientvolume are interrelated. Increased f low rate w ill give a decrease in resolution, butthis decrease w ill not be very significant a t high sample loadings. This means tha tunder process conditions, where maximum sample load is applied to achievemaximum throughput, the flow rate is limited primarily by the rigidity ofchromatography media and by system constraints. The effect on resolution ofincreased gra dient volume is usua lly more significant than the effect of flow ra te.This means tha t w hen increasing gradient volume to increase resolution, flow ra tecan also be increased a ccordingly to compensa te for loss in separation speed. Theresult is an increase in resolution that may be traded off for increased sampleloading and thereby increased productivity.

In other w ords, in process chromatography the best result w ill be obta ined by usingthe maximum flow with the gradient volume

Documents

Hydrophobic Interaction Manual - AMERSHAM