UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

A new process for the separation and purification of egg yolk antibodies Charter, Edward 1993

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata


831-ubc_1993_spring_phd_charter_edward.pdf [ 5.82MB ]
JSON: 831-1.0058569.json
JSON-LD: 831-1.0058569-ld.json
RDF/XML (Pretty): 831-1.0058569-rdf.xml
RDF/JSON: 831-1.0058569-rdf.json
Turtle: 831-1.0058569-turtle.txt
N-Triples: 831-1.0058569-rdf-ntriples.txt
Original Record: 831-1.0058569-source.json
Full Text

Full Text

We accept this thesis as conformingto the required standardA NEW PROCESS FOR THE SEPARATION AND PURIFICATION OFEGG YOLK ANTIBODIESByEdward A. CharterB.A.Sc., The University of British Columbia, 1987A THESIS SUBMITTED IN PARTIAL FULFILLMENT OFTHE REQUIREMENTS FOR THE DEGREE OFDOCTOR OF PHILOSOPHYinTHE FACULTY OF GRADUATE STUDIESDEPARTMENT OF CHEMICAL ENGINEERINGTHE UNIVERSITY OF BRITISH COLUMBIAApril 1993© Edward A. Charter, 1993In presenting this thesis in partial fulfilment of the requirements for an advanced degree atthe University of British Columbia, I agree that the Library shall make it freely availablefor reference and study. I further agree that permission for extensive copying of thisthesis for scholarly purposes may be granted by the head of my department or by hisor her representatives. It is understood that copying or publication of this thesis forfinancial gain shall not be allowed without my written permission.Department of Chemical EngineeringThe University of British Columbia2075 Wesbrook PlaceVancouver, CanadaV6T 1W5Date: -111,rvi 30) /773AbstractA three stage process was developed for the separation of an antibody (IgY) from in-dustrially separated chicken egg yolk. This included aqueous extraction of the watersoluble fraction (WSF) from the yolk by dilution with distilled water and pH adjust-ment, separation of IgY from the WSF using a cation exchange column in an automatedchromatography system, and finally purification of IgY using precipitation with sodiumsulphate. The overall recovery of the process was approximately 50%, and the purity95% or greater depending on the number of precipitation steps used. An automatedliquid chromatography system was developed to allow efficient study of various chro-matographic media for the separation of the antibody. Analysis of breakthrough curvesindicates that superficial velocity is the governing parameter in the binding of IgY tothe cation exchanger. Results of a pilot scale experiment involving the application of46.5 litres of WSF containing approximately 0.8 mg/ml IgY to a 1500 nil column arepresented.11Table of ContentsAbstractList of Tables^ ivList of FiguresList of Abbreviations^ viAcknowledgement^ viii1 Introduction^ 12 Literature Review^ 52.1 Introduction to IgY  ^52.1.1 A History of the Study of IgY  ^52.1.2 Characteristics of IgY  ^82.1.3 Other Constituents of Egg Yolk ^  162.2 Potential Applications of IgY ^  172.2.1 Immunoassays ^  172.2.2 Therapeutic Applications ^  202.2.3 Infant Formula ^  202.2.4 Immunoaffinity Chromatography ^  212.3 Relative Merits of Using IgY in Lieu of other Antibodies ^ 212.3.1 Advantages ^^Disadvantages ^Methods of Isolation and Purification ^2.4.1^Extraction of the Water Soluble Fraction from Yolk ^2.4.2^Separation of IgY from WSF^222324263 Assembly of an Automated Chromatography System 323.1 Introduction ^ 323.2 Initial System Design ^ 333.3 Improved System with Process Controller ^ 333.4 Final System Design ^ 353.5 Conclusions ^ 384 Extraction of the Water Soluble Proteins 394.1 Introduction ^ 394.2 Materials and Methods ^ 394.2.1^Raw Material 394.2.2^Separation of Lipoproteins ^ 404.2.3^Analytical Procedures 404.3 Results and Discussion ^ 424.3.1^Comparison of Industrially and Manually Separated Egg Yolk 424.3.2^Separation of Lipoproteins from Egg Yolk Soluble Protein 424.4 Conclusions ^ 495 Separation of IgY 505.1 General Introduction ^ 505.2 Ultrafiltration ^ 515.2.1^Introduction 51iv5.35.2.2^Materials and Methods ^5.2.3^Results and Discussion 5.2.4^Conclusions ^Anion Exchange Chromatography ^535557575.3.1^Introduction ^ 575.3.2^Materials and Methods ^ 585.3.3^Results and Discussion 595.3 .4^Conclusions ^ 595.4 Cation Exchange Chromatography ^ 605.4.1^Introduction ^ 605.4.2^Materials and Methods ^ 605.4.3^Results and Discussion 615.4.4^Conclusions ^ 615.5 General Conclusions 646 Comparison of Two Cation Exchangers 656.1 Introduction ^ 656.2 Materials and Methods ^ 656.2.1^Cation Exchange Chromatography ^ 656.2.2^Analytical Procedures ^ 666.3 Results and Discussion ^ 666.4 Conclusions ^ 787 HPLC Analysis of IgY 797.1 Introduction ^ 797.2 Materials and Methods ^ 817.3 Results and Discussion 83V7.3.1 Molecular Weight Determination ^  837.3.2 IgY Concentration Determination  857.3.3 IgY Purity Determination ^  917.4 Conclusions ^  978 Purification of IgY^ 988.1 Introduction  988.2 Materials and Methods ^  998.2.1 Anion Exchange Chromatography ^  998.2.2 Salt Precipitation  ^998.2.3 Ultrafiltration ^  1008.2.4 Methods of Analysis ^  1008.3 Results and Discussion  1008.3.1 Anion Exchange Chromatography ^  1008.3.2 Salt Precipitation ^  1018.3.3 Ultrafiltration  1028.4 Conclusions ^  1029 Breakthrough Curve Analysis^ 1039.1 Introduction ^  1039.2 Materials and Methods ^  1069.2.1 Cation Exchange Chromatography ^  1069.2.2 Methods of Analysis ^  1099.3 Results and Discussion  1099.4 Conclusions ^  11410 Scale-up of the Separation Process^ 116vi10.1 Introduction ^  11610.2 Materials and Methods ^  11610.2.1 Flow Testing a Column of HC-2 ^  11610.2.2 Batch Separation Using HC-2  11710.2.3 Column Separation Using HC-2 ^  11810.3 Results and Discussion ^  11810.3.1 Flow Test Results  11810.3.2 Batch Separation ^  120^10.3.3 Column Separation     12010.4 Conclusions ^  12711 Conclusions^ 12912 Recommendations^ 131Bibliography^ 133Appendices^ 144A Schematic Diagram of ChemResearch/Solenoid Valve Interface^144B Sample Chromatograms of Molecular Weight Markers^145C Economics of IgY Purification^ 147viiList of Tables2.1 Milestones in the Study of IgY  ^72.2 Some Early Papers Dealing with Antibody Activity in Egg Yolks  ^92.3 Survey of Molecular Weight Values Reported for IgY Monomer ^ 112.4 Examples of Recent Papers on Applications of IgY ^ 182.5 Summary of Abbreviations for Table 2.4 ^  194.6 Effect of pH on residual lipids and IgY recovery after 24 h sedimentationusing 10x dilution ^  486.7 IgY Recovery and Mass Balance using Linear Gradient and Step-wise Elu-tion on a Column of HC-2 Cation Exchanger for the Representative Ex-periments shown in Figures 6.14 and 6.15.   707.8 References in the literature to gel filtration of chicken antibodies ^ 807.9 Summary of Molecular Weight Information ^  847.10 Comparison of HPLC and RID Results  927.11 Summary of HPLC Results for Figure 7.24 ^  968.12 Purities obtained in Salt Precipitation and Determined by HPLC Analysis 10110.13Example Protocol for IgY Purification. Extraction was using distilledwater and pH adjustment; separation using CEC; and purification usingsodium sulphate precipitation  127viiiList of Figures1.1 Schematic Diagram of a Typical Immunoglobulin Molecule. S-S representsdisulphide bonds. Fab is the antigen binding portion of the molecule. F,is the heavy chain portion of the molecule  23.2 Simple Laboratory Chromatography System ^  343.3 Improved Chromatography System ^  363.4 Computer Controlled Chromatography System ^  374.5 SDS-PAGE of Manually versus Industrially Separated Egg Yolk ^ 434.6 Effect of pH and Dilution on Residual Lipids (a) and IgY Recovery (b) ^ 454.7 Diluted Yolk after 24 hours at 4°C ^  464.8 Effect of pH and Time on Volume of Sedimented Lipoprotein and IgYRecovery^  475.9 Ultrafiltration Apparatus ^  545.10 Chromatogram from the first CM-92 Experiment ^  625.11 Chromatogram from a later CM-92 Experiment  636.12 Separation of IgY on a CM-92 Column using Linear (a) and Step (b)Elution Profiles ^  686.13 Separation of IgY on a CM-92 Column using Double Step Elution Profile ^The sample application peak has been omitted for clarity. ^ 696.14 Separation of IgY on an HC-2 Column using a Linear Elution Profile .^716.15 Separation of IgY on an HC-2 Column using a Step Elution Profile . .^72ix6.16 Non-Denaturing SDS-PAGE of Fractions from HC-2 Experiments. Thelanes represent A - WSF prepared from industrially separated yolk; B -Unbound fraction; C - Eluted protein - IgY rich fraction; D - Saltwash peak. 746.17 Purity of Individual Fractions Eluted from an HC-2 Column ^ 756.18 Purity as a Function of Recovery using Linear Gradient Elution ^ 766.19 Purity as a Function of Recovery for HC-2 Experiments ^ 777.20 HPLC System Layout ^  827.21 Relationship Between Molecular Weight and Retention Time ^ 867.22 HPLC Analysis of Pure IgY Samples. The graphs show the chromatogramobtained for a sample of IgY standard (a) and a sample prepared by pu-rification with Na2SO4 precipitation after extraction as described in Chap-ter 4, and separation as described in Chapter 6 (b) ^897.23 HPLC Standard Curve (a) and Comparison to RID (b) ^ 907.24 Analysis Carried out at a Flowrate of 0.25 ml/min of Samples with ThreeDifferent IgY Purities: (a) - filtered yolk supernatant (WSF); (b) - theIgY rich fraction from cation exchange chromatography; (c) - IgY purifiedby precipitation with sodium sulphate precipitation  947.25 Electrophoresis of Samples with Three Different IgY Purities: A - theWSF; B - the IgY rich fraction from CEC; C - IgY purified by precipitationwith sodium sulphate   959.26 Schematic Representation of Protein Binding to Adsorption Column^1059.27 A Small Column Made from a Disposable 5 cc Syringe ^ 1079.28 Apparatus used for Breakthrough Experiments ^  1089.29 Break Through Analysis using 3 ml Column at 1.0 ml/min ^ 1119.30 Comparison of Breakthrough Curves at Two Different Flowrates ^ 1129.31 Plotted Data from Breakthrough Experiments at Five Different Flowrates(the solid lines represent curves fitted to data for the highest and lowestflowrates)^  1139.32 Recovery as a Function of Loading for 3 Different Flowrates ^ 11510.33Automated Chromatography System Including Pilot Scale Column .^11910.34Pressure Drop vs. Superficial Velocity for a Pilot Scale HC-2 Column .^12110.35Breakthrough Curve for the Pilot Scale Column ^  12210.36Breakthrough Comparison for Pilot Scale and Small Scale Columns^12510.37Recovery as a Function of Superficial Velocity for Several Experiments^126xiList of AbbreviationsAEC^Anion exchange chromatographyAIDS^Acquired immune deficiency syndromeBCA^Bicinchoninic acidBSA^Bovine serum albuminCEC^Cation exchange chromatographyCV^Coefficient of varianceDEAE Diethylaminoethyl-ELISA Enzyme linked immunosorbent assayEYS^Egg yolk supernatantHPLC High performance liquid chromatographIEF^Isoelectric focussingIg^ImmunoglobulinIgG^Immunoglobulin GIgY^Immunoglobulin YLDL^Low density lipoproteinsLP S^LipopolysaccharideLSF^Lipid soluble fractionMCIC Metal chelate interaction chromatographyNMWC Nominal molecular weight cut-offPB^Phosphate buffer (sodium form)PBS^Phosphate buffered saline (sodium form)xiiPEG^Polyethylene glycolpI^Isoelectric pointR2 Coefficient of correlationRID^Radial ImmunodiffusionS^Standard deviationSDS-PAGE Sodium dodecylsulphate polyacrylamide gel electrophoresisTB^Tris bufferTBS^Tris buffered salineTMV^Tobacco mosaic virusTP^Total proteinUF^UltrafiltrationVDC^Volts direct current-1--^Mean (statistical)WSF^Water soluble fractionAcknowledgementWithout the support of several people and organizations I could never have completedthis work. I would like to thank Agriculture Canada, the Natural Science and EngineeringResearch Council and the Canadian Egg Marketing Agency for supporting the researchproject on which much of this thesis is based; both N.S.E.R.C. and the Science Councilof B.C. for the scholarships which enabled me to remain at university; the Departmentof Bio-Resource Engineering, UBC, where the majority of the experimental work wascarried out; the Department of Food Science, UBC, where much of the analytical workwas done; several individuals at Canadian Lysozyme Inc. (most notably Mr. Steve Smithand Mrs. Jennifer Power) and Vanderpol's Eggs Ltd. (especially Mr. Ian Cooke), fortheir guidance and support; fellow graduate students such as Mr. Emmanuel Akita forhis help in understanding analytical procedures, and Mr. Colin Smith for sharing some ofhis vast computer experience; my co-supervisors, Dr. R.M.R. Branion and Dr. K.V. Lo;the other members of my Ph.D. committee, Dr. S. Nakai and Dr. K. Pinder; my friendand colleague Dr. J. Fichtali who helped me to identify my priorities while carrying outthe research, and who was integrally involved in the research, especially of chapters 3to 6, and 7; and my wife, Christine, and daughter, April, who supported me through somuch and made all the trials worthwhile. Finally, I wish to thank God for giving me thestrength to finish what I started, because I could never have done it without His help.xivChapter 1IntroductionThis dissertation is principally concerned with the purification of an antibody found inchicken egg yolk. Antibodies, also termed immunoglobulins (abbreviated Ig's), are glyco-proteins (Weir, 1988) which are produced in response to immunization with antigens, andwhich specifically react with the antigen (Sell, 1987). A schematic diagram of a typicalimmunoglobulin molecule is shown in Figure 1.1. Mammals, including humans, produceseveral classes of antibodies that reside in the blood and perform various functions inprotecting the individual from pathogens.Human beings have been plagued by diseases as far back as our recorded history cantake us. In describing a plague in Athens nearly 2500 years ago, Thucidides, the famousGreek historian, noted that those who had recovered from the plague were able to tendthe sick without fear of contracting the disease again - in other words, they were immune(Steward, 1984).Today, so much is known about immunology and the function of the various immuno-logical molecules, and so much is being published about this topic that it is virtuallyimpossible to keep abreast of all of the new information. Fortunately, there are somesimilarities between the classes of antibody found in different animals and bird species,and many of the techniques for analyzing and utilizing these molecules are now welldeveloped. Antibodies in the chicken have been studied for a century, but large scaleuse of chicken antibodies has been hindered by problems associated with acquisition and1Chapter 1. Introduction^ 2Figure 1.1: Schematic Diagram of a Typical Immunoglobulin Molecule. S-S representsdisulphide bonds. Fab is the antigen binding portion of the molecule. Fe is the heavychain portion of the molecule._§ AFcFragment- s - s --- s - 9-PapainCieavage _/7Antigen Bin" dingRegionFabFragmentChapter 1. Introduction^ 3purification in large quantities. Similar problems exist in attempting to purify large quan-tities of mammalian antibodies. Antibodies are used in a host of scientific procedures,most involving some sort of assay for another protein, enzyme, virus or bacterium. Suchprocedures are termed immunoassays because they make use of the strong affinity ofthe immunoglobulin for a specific antigen. They can also be used as ligands for affinitychromatography columns, and interest in using purified antibodies in prophylaxis andtherapy is also growing. Thus the demand for purified antibodies is also growing andpresent sources are not likely to be sufficient for future demand.One of the most common mammalian antibody classes is known as 7-globulin orimmunoglobulin-G (IgG). This molecule has a molecular weight of about 150 kDa, andin a purified state from various animal sources it is used extensively in the applications de-scribed above. A non-mammalian but roughly equivalent antibody with similar functionis chicken 7-globulin, known as immunoglobulin-Y (IgY) to emphasize certain differencesbetween it and its mammalian counterpart. IgY is a slightly larger molecule and hascertain other charateristics that set it apart from IgG, but has been used successfullyin many of the same applications. What is of especial interest about this molecule isthat it is efficiently tranferred from chicken serum to egg yolk in order to provide passiveimmunity to the chick. Many researchers have considered chicken egg yolk as a potentialsource of antibodies. Much of the impetus for such research comes from the advantagesof using egg yolk as a source of antibodies instead of serum from either birds or mammals.This dissertation describes efforts to develop a practical purification process beginningwith industrially separated egg yolk as raw material, and ending with a purified productthat retains its biological activity. The emphasis has been on simplicity in order todevelop a process that is not so complicated that it is difficult to scale up. Labour andcapital intensive methods have been avoided where possible so as to develop a processChapter 1. Introduction^ 4that is also economically viable.The process consists of 3 basic stages. The first stage, extraction of IgY, involves theseparation of the lipid soluble fraction (LSF) of egg yolk from the water soluble fraction(WSF), which contains IgY. The second stage, separation of IgY, involves the removal ofthe majority of the contaminants from the WSF. An automated chromatography systemdevised by the author is described. The third and final stage, purification of IgY, involvesthe isolation of the antibody to purities well in excess of 90%. An HPLC technique formolecular weight determination and quantification of IgY in aqueous solution is describedand compared to radial immunodiffusion (RID). Following the process development, astudy of factors affecting the scale-up of the chromatographic step is carried out, and theresults of a pilot-scale experiment presented.This work represents a pioneering effort in the use of cation exchange chromatography(CEC) to separate IgY from the WSF of ckicken egg yolk. As such, it is not meant to bethe final word on the subject, but rather to demonstrate CEC as a potentially practicalindustrial method.An attempt has been made to avoid the verbose style often used in large documentsof this nature, and to this end gobbledygook has been avoidedl. However, for sectionswhere this was not achieved, the author sincerely apologizes.1Gobbledygook has been defined as writing that suffers from a swelling of its parts, see Meyer, 1977,Reports full of gobbledygook, J. of Irreproducible Results, 22(4), 12.Chapter 2Literature Review2.1 Introduction to IgYIgY is a molecule which has received considerable attention in recent years, and yet stillis not fully understood from an immunological perspective. In this first section of thereview, a brief history of the study and use of the molecule is presented, along withsome of its biochemical and physical characteristics, and a discussion of the other majorconstituents of egg yolk.2.1.1 A History of the Study of IgYAccording to Brambell (1970), the earliest published paper dealing with the transfer ofpassive immunity from mother to offspring in mammals was by Ehrlich (1892). A yearlater, Klemperer (1893) discussed the transfer of passive immunity in the fowl. Thus ithas been only in the last 100 years that the modern scientific approach has been used tostudy the presence of immunological factors in the egg and young chick.At the turn of the century, Osbourne and Campbell (1900) isolated what they termed"a large amount of protein" from egg yolk, and identified it as vitellinl. Plimmer (1908),upon repeating their work, was also able to isolate a water soluble protein in the yolk andnamed it "livetin", an anagram formed from the letters of vitellin (Burley and Vadehra,1989). He also quoted Gross (1899) as having noted the presence of this water soluble1The vitellin of egg yolk is today termed "lipovitellin" due to its lipoprotein structure.5Chapter 2. Literature Review^ 6protein.Meanwhile, the ability of chicken serum to precipitate blood components was receivingsome attention from medical researchers. Ewing (1903) first used fowl "precipitins"in blood identification, while Sutherland (1910) used this technique on a large scale.Hektoen (1918) reported on the immunization of chickens with human blood and serum,and determined that the maximum concentration of "anti-human precipitin" occurred 9to 12 days after injection.Kay and Marshall (1928) isolated the livetin fraction in a relatively pure form, andin 1932, Jukes and Kay determined that IgY and chicken IgG are either closely relatedor identical. The livetin fraction was separated into 3 proteins using electrophoresis byShephard and Hottle (1949), and later by Martin et al. (1957), who named the three a-,3-, and ry-livetin.Since the early 1960's, many papers have been published dealing with methods ofisolation and purification of IgY, techniques for obtaining high antibody titers in serumand yolk, and new applications of the purified antibody. Perhaps the most notablepurification procedure developed is that of Poison et al. (1980a), in which polyethyleneg-lycol (PEG) is used to precipitate IgY. Although this method, as well as two revisedprocedures (Poison et al., 1985; Poison, 1990) provide IgY of relatively high purity forlaboratory scale use, they are not as appealing for large scale purification of IgY due totheir labour-intensive steps, and the use of PEG, which may prevent the use of the prod-uct in food applications. For these reasons, numerous other procedures for IgY isolationand purification have been studied, and will be discussed in more detail later.Table 2.1 presents some of the important milestones in IgY research over the lastcentury.Chapter 2. Literature Review^ 7Table 2.1: Milestones in the Study of IgYYear^Author(s)^ Event1892 Ehrlich1893 Klemperer1899 Gross1900 Osbourne & Campbell1903 Ewing1908 Plimmer1910 Sutherland1918 Hektoen1928 Kay Sz Marshall1932 Jukes 86 Kay1949 Shepard & Hottle1950 Shmittle1957 Martin et al.1980 Poison et al.studied transfer of passive immunity frommother to offspringdiscussed transfer of passive immunityin the fowlnoted presence of a water soluble proteinin yolkisolated "a large amount of protein" fromyolk and identified it as vitellinfirst used fowl precipitins in bloodidentificationisolated water soluble protein in yolkand named it livetinused fowl precipitins on a large scalein blood identificationstudied anti-human precipitins in chickenserumisolated livetin fraction in a relativelypure form, but noted signs of heterogeneitydetermined that IgY and chicken IgG areeither closely related or identicalisolated 3 proteins from livetinfraction using electrophoresis.detected antibodies to Newcastle disease inegg yolkisolated the 3 livetins and named thema-, p- and -y-livetin.developed the polyethylene glycol methodfor IgY purificationChapter 2. Literature Review^ 82.1.2 Characteristics of IgYIdentification as an ImmunoglobulinAlthough they were not aware of the divesity of proteins in the yolk livetin fraction,Jukes and Kay (1932) were convinced that livetin and serum globulin were one and thesame. Nace (1953) noted that livetin absorbs antibodies from rabbit antisera to fowlserum albumin and fowl -y-globulin and therefore contains proteins that are similar tothese.Knight and Schechtman (1954) decided to study the ease with which proteins aretransferred from the hen's serum to the yolk. They injected hens with rat serum, bovinealbumin and 7-globulin, and lobster serum. They were able to detect all of these antigensin the yolk a few days after injection by precipitin reactions with rabbit antisera againstthe corresponding antigens. Bovine albumin and IgG appeared to be transferred insubstantially unmodified form from the hen's blood into the ovarian egg.Using immunoelectrophoresis and starch gel electrophoresis, Williams (1962) identi-fied 7-livetin as 7-globulin. In 1969, Leslie and Clem suggested that fowl IgG be calledIgY, because in their words "it does not align itself very well with any of the mammalianimmunoglobulins". Leslie (1975) stated that IgY is without question the predominantimmunoglobufin of yolk, and that most of it is transferred to embryonic circulation in thelast 5 to 6 days of embryonation. Wang et al. (1986) found a mannose-binding subclassof chicken IgY that was identical in serum and yolk when compared by electrophoresis,gel filtration, amino acid composition and several other physical and chemical properties.Based on sodium dodecylsulphate - polyacrylamide gel electrophoresis (SDS-PAGE)and isoelectric focussing (IEF), Loeken and Roth (1983) determined that ovarian IgYreceptors in the fowl selectively transport all subpopulations of maternal IgY, but notChapter 2. Literature Review^ 9Table 2.2: Some Early Papers Dealing with Antibody Activity in Egg YolksSource Year Author(s) Antigenchicken 1893 Klemperer tetanus,, 1901 Dzierzgowski diphtheria,, 1928a/b Ramon tetanus1934 Jukes et al. diphtheriaduck 1934 Fraser et al. diphtheriachicken 1946 Brandly et al. Newcastle disease1948 Briles hemolytic activity77 1952 Bornstein et al. Newcastle disease1951/52 Buxton Salmonellaother immunoglobulins.Thus the evidence is substantial to support the identification of yolk immunoglobulinsas chicken IgY, originating from maternal serum.Antibody ActivityNumerous authors have reported the presence of antibody activity in egg yolks appearingseveral days after immunization of hens with various antigens.Schmittle (1950) tested 136 hens that had been exposed to Tortor furens (Newcastledisease virus) and found that the serums and egg yolk of 96.3% of the hens were positive(ie: precipitated the virus). He believed that contamination of the yolk with albumencaused some nonspecific hemagglutination inhibition. Table 2.2 lists a few more specificexamples.It seems that the antibody activity of the yolk is well maintained for at least 6 monthsin eggs stored at cooler temperatures (Brandly et al., 1946). However, the half-life ofchicken IgG has been reported as about 35 h in an adult hen, and about 72 h in a newlyChapter 2. Literature Review^ 10hatched chick (Patterson et al., 1962b).Heller (1975) monitored the antibody production in hen serum and yolk after injectionof the hens with a pathogenic E. coli strain. Yolk antibody titres were slightly lower thanthose found in the serum of the hens.Otani et al. (1991) determined that IgY antibodies lost little of their antibody activityby incubation for 10 mmn. at pH 7.2 below 60°C, or by incubation for 10 min at 40°Cabove pH 4.0. However, above 65°C or below pH 4.0 their antibody activity was greatlydiminished.Molecular WeightMany authors have reported values for the molecular weight of the IgY monomer. Mosthave used methods such as sedimentation equilibrium or approach to equilibrium, al-though a few have reported using other methods such as gel filtration or SDS-PAGE. Asummary of literature values is given in Table 2.3.From the literature, it is apparent that the IgY monomer is larger than mammalianIgG, with a mass close to 170 kDa.Two authors have reported estimations of the molecular weights of IgY aggregates.Hersh and Benedict (1966) reported a value of 550 kDa, and Van Orden and Treffers(1968) a value of 540 kDa. Thus, it is apparent that IgY has a tendency to form a trimer,since the expected molecular weight of the trimer, based on the size of the monomer,would be 510 kDa.Concentration in Blood and YolkThe amount of IgY in yolk relative to the other livetin proteins is reported to be quitevariable, depending on the level of immunization, and the time since injection. MartinChapter 2. Literature Review^ 11Table 2.3: Survey of Molecular Weight Values Reported for IgY MonomerSource Reference Method MWkDayolk Martin & Cook (1958) Sedimentation Equilibrium 150serum Tenenhouse Si Deutsch (1966) Sedimentation Equilibrium 206serum Leslie Sz Clem (1969) Sedimentation Equilibrium 170yolk Wang et al. (1986) SD S -PAGE 169yolk Shirman et al. (1988) Gel filtration 160yolk Hassl & Aspock (1988) Gel filtration 175yolk Polson et al. (1980) Various methods 165yolk/serum Loeken Si Roth (1983) SD S -PAGE 180yolk/serum Rose et al. (1974) Sedimentation Equilibrium 174serum Van Orden & Treffers (1968) Gel filtration 150yolk Akita & Nakai (1992) SD S -PAGE 180yolk Fichtali et al. (1992) SD S -PAGE 175X = 171.2S= 15.0CV = 8.8%Chapter 2. Literature Review^ 12and Cook (1958) were unable, at the time of publication of their work, to estimate theamount of IgY present in egg yolk, but suggested that it may constitute nearly half of thelivetin based on the relative areas in electrophoretic and sedimentation diagrams Poisonet al. (1980a) determined that IgY comprised more than 70% of the total livetin proteinsin yolk, seemingly in contradiction to the results of Shephard and Hottle (1949), whofound the 3 livetin proteins to be present in relatively the same proportion. Poison etal. attributed this descrepancy to the fact that they used eggs of hyperimmunized hens,and suggested that the increase in the IgY component is due to an increased amount ofantibody against the virus.In general, the literature seems to indicate that the concentration of antibodies in eggyolk from immunized hens tends to be slightly lower than the corresponding concentrationin hen's serum (Orlans, 1967), although the occasional paper will indicate a higher yolkthan serum IgY concentration (Rose et al., 1974; Wallman et al., 1990). Since bothserum and yolk levels of IgY first increase and later decrease after immunization of ahen, the concentration is usually in a state of flux. For example, Patterson et al. (1962a)compared serum and yolk antibody levels for hens immunized with a single injection ofinfluenza virus. They found the peak serum level was nearly double that of the yolk, butyolk antibody levels remained near peak levels for a longer period of time (almost 6 dayscompared to 1 to 2 days for serum). Gardner and Kaye (1982) report that quantities ofIgY corresponding to almost 500 ml of antiserum can be recovered from a hen's eggs inone month.In terms of absolute concentration, Wang et al. (1986) reported yolk to have an IgYconcentration of at least 9 mg/ml based on a semiquantitative double immunodiffusionmethod. Levels as high as 25 mg/ml (Rose et al., 1974) have been reported for the yolksof hyper-immunized chickens.Chapter 2. Literature Review^ 13The amount of antibody produced is also a function of the type of antigen used to elicitthe response. Poison et al. (1980b) immunized hens against several proteins and naturalmixtures of proteins and found that hens produced significant quantities of antibodiesto the high molecular weight antigens (>150,000 daltons), but did not react as stronglyto the low molecular weight antigens (<30,000 daltons). "It is feasible that more avidantibodies to low molecular weight antigens could be elicited in hens by cross-linkingthem to bacteria or giant molecules such as haemocyanin " (Poison et al., 1980b).Another factor that influences the amount of antibody produced is the physical stateof the hen. Serum antibody levels have a direct influence on the resulting yolk levels,since it has been determined that IgY subpopulations are transported in proportion totheir concentration in maternal serum (Loeken and Roth, 1983). It appears that theserum of a laying hen has a higher IgY concentration than the serum of a hen before itsreproductive period, or of a rooster (Poison, 1980a).Heterogeneity (subclasses)A significant amount of evidence exists that supports the idea that IgY represents aheterogeneous population of molecules made up of several subclasses. Wilkinson andFrench (1968) used immunoelectrophoresis and radioimmunoelectrophoresis to identify 2distinct heavy-chain subtypes. Using a radiochromatographic assay, Howell et al. (1973)also concluded that at least 2 major subpopulations exist2. Other authors have confirmedthe existence of at least 2 major subclasses (Rose et al., 1974; Loeken and Roth, 1983),and others possibly even more'.2A specific IgG (anti-2,4- dinitrophenyl 7S antibody) contained 5.5% carbohydrate by weight com-pared to 6.0% for normal chicken 7S, indicating that at least 2 subclasses exist.3Using IgY as an antigen to elicit antibodies in rabbit, and then carrying out 2 dimensional Lau-ren electrophoresis, Poison et al. (1980b) discovered that five antigenic components were recognized,supporting the idea that IgY has several subclasses.Chapter 2. Literature Review^ 14The diffuse bands obtained in electrophoresis (McIndoe and Culbert, 1979), and iso-electric focussing of IgY (Awdeh et al., 1968; Rokhlin et al., 1975) are another indicationof the heterogeneity of this molecule. Recently, Ternes (1989) carried out isoelectric fo-cussing of the livetins and found 25 bands occurring in the pH range 4.3-7.6. He foundthe alpha and P-livetins concentrated in the ranges 4.3-4.6, 5.3 and 5.5, while the gammalivetin bands were centered around pH 4.5, 4.7 and 5.0. McCannel and Nakai (1990) pre-sented evidence of subclasses based on variations in binding patterns on a copper loadedmetal chelate interaction chromatography columnSolubility and Aggregative PropertiesAlthough soluble in aqueous solution, the solubility of IgY is highly dependent on ionicstrength and changes in temperature.While attempting to purify IgY, Martin et al. (1957) found it to be less stable thanthe other livetins. They found that once most of the other proteins had been removed,the solubility of IgY in water was reduced and an ionic strength of about 0.1 M wasrequired to dissolve it. They noted that a marked loss in solubility occurred at eachpurification step, on freeze drying, and even on standing. In 1958, Martin and Cookreported that on purification, -y-livetin has the solubility behaviour of a euglobulin butloses its solubility irreversibly on prolonged dialysis against water or freeze-drying. Thesefindings have been reiterated by numerous authors throughout the literature, who havenoted the tendency of IgY to form aggregates and even precipitate under conditions ofboth very low and very high ionic strengths.Most of the literature dealing with the aggregative properties of IgY has been focussedon antibody purified from chicken serum. As far back as 1918, Hektoen found thatchicken antihuman serum had a tendency to precipitate if stored in 0.9% NaCl solution,Chapter 2. Literature Review^ 15and especially on rapid transfer from lower to higher temperature. He found that thisprecipitation could be avoided if 1.9% NaCl solutions were used instead.At still higher salt concentrations, IgY has a tendency to aggregate. The salt inducedaggregation of IgY into trimers or tetramers has been studied by Hersh and Benedict(1966) and Van Orden and Treffers (1968a). In using gel diffusion techniques, Poison etal. (1980b) found that IgY formed dimers when the concentration of NaCl was 1.5 M,and remained as monomers in gels containing 0.15 M NaCl. They further state that"Although it is unknown which chains of the IgY are involved in the mutual linkage, atleast two of the four antigen binding sites of the dimer remain reactive, and are thus ableto bind antigens to form a lattice".Cser et al. (1982) used a low IgY concentration of 0.62% w/v in X-ray small-anglescattering experiments to avoid aggregation. They still had to use Sephadex G-200 toremove IgY aggregates from solution' just prior to their experiments. Thus the strongtendency of IgY to form aggregates and come out of solution means care must be takenin its handling.Miscellaneous PropertiesIgY possesses considerable acid and temperature resistance (LOsch et al., 1986). Theisoelectric point of IgY has been variously determined to be: 4.8-5.0 (Kay and Marshall,1928, for livetins); 5.2 (Tenenhouse and Deutsch, 1966); 6.5-9.0 (Loeken and Roth, 1983);and 4.5-5.0 (Ternes, 1989).The extinction coefficient has been determined to be 13.5 (Tenenhouse and Deutsch,1966) and 12.74-14.42 (Leslie and Clem, 1969) depending on the buffer. Tenenhouse'The solution used was PBS - pH 7.4, 3.186 g Na2HPO4.H20, 0.1533 g K112PO4 and 8.765 g NaC1per 1000 ml of 1120Chapter 2. Literature Review^ 16and Deutsch (1966) also determined the partial specific volume to be 0.718 cm3/g, thenitrogen content to be 14.8%, and the diffusion coefficient, D20,tv (X10-7 CM2), to be3.21-3.24 for the monomer. The diffusion coefficient has also been reported as 2.2-2.4 in8% NaC1, where aggregates are certain to be present (Van Orden and Treffers, 1968a).Cser et al. (1982) determined the radius of gyration to be 6.13 nm, somewhat lowerthan rabbit (6.68 nm), pig (6.78 nm) or human (7.4 nm)Finally, Wang et al. (1986) have determined that the optimal pH for IgY binding ispH 5 to 9 (based on a mannose binding subclass of IgY).2.1.3 Other Constituents of Egg YolkIgY is only one protein in a complex mixture of proteins, lipids and lipoproteins. A reviewby Powrie and Nakai (1985) provides a thorough discussion of the major components.Fresh yolk has a solids content of 52 to 53%, of which the principal constituents areproteins and lipids. The lipid fraction includes triacylglycerol (66%), phospholipid (28%),cholesterol (5%), as well as minor lipids. The proteins (and lipoproteins) can be dividedbetween those found in the granules (phosvitin, lipovitellins, low density lipoproteins(LDL), and myelin figures), and those present in the plasma (ie: the WSF). The WSFproteins include a-, /3-, and 7-livetin. The molecular weight of a-livetin is about 80,000Da, and that of P-livetin approximately 45,000 Da. Two components of low-densitylipoprotein are also present in the WSF, with molecular weights of several million Daltons.There has not been as much interest in studying a- and 0-livetin and not too much isknown about their characteristics compared to IgY. However, Martin et al. (1957) deter-0mined the extinction coefficients of a- and 13-livetin to be E0,1% at A=2800 A (280 nm)= 7.38 and 9.44 respectively.Chapter 2. Literature Review^ 172.2 Potential Applications of IgYIn the last decade there has been a flurry of publications dealing with specific uses of IgY.Table 2.4 lists several papers published since 1980 dealing with some of these applications.2.2.1 ImmunoassaysBy far the most prevalent use of the purified antibody is in immunoassays of various kindsincluding RID (radial immunodiffusion), various ELISA's (enzyme linked immunosor-bent assays), radioimmunoassays, latex agglutination, hemagglutination inhibition tests,immunoblotting and various forms of immunoelectrophoresis, to name just a few. En-teroviruses such as poliomyelitis, Coxsackie, hepatitis A and rotaviruses have a significantrole in the development of many human and animal diseases (Shirman et al., 1988). Tra-ditional methods of detection based on tissue cultures are time consuming and not alwayseffective for detecting viruses, such as hepatitis A and rotavirus, with weak cytopathiceffect. ELISA's are becoming a common method of detection because of their simplicity,high sensitivity and specificity.Immunoassays are also rapidly becoming important for the detection and quantifi-cation of various components in food (McCannel and Nakai, 1990). Not only are foodimmunoassays easy to perform, sensitive, specific and relatively inexpensive, but thereare many potential applications in the determination of food additives, meat species,fungal and bacterial contamination, antinutritional factors, pesticide residues and hor-mones (Allen and Smith, 1987). Immunotechniques have also been used to trace proteinsthrough the technological processes of brewing, and to detect contaminants, yeast pro-teins and additives in beer (Vaag and Munck, 1987).The potential for the use of IgY in immunoassays is therefore quite significant, butChapter 2. Literature Review^ 18Table 2.4: Examples of Recent Papers on Applications of IgYAntigencitrus tristeza virustobacco mosaic viruspregnancy associated murineprotein-2 and human pregnancy-specific /31-glycoproteinhuman parathyroid hormoneCucumovirusesvarious influenzae, adenovirusgroup antigen, rotavirusCalf Thymus RNA Polymerase IIhuman parathyroid hormoneNewcastle disease, Infectiousbronchitis, Mycoplasma gallisepticummonoclonal antibodiesserum IgG & IgMHuman plasma kallikreina-subunit of insulin receptor1a,25-dihydroxyvitamin DRuminal bacteriaEnterovirusesrot avirusesprocine enteropathogenic E. coliRheumatoid factor positive sera1a,25-dihydroxyvitamin DOrganophosphorous compoundsCanine distemper virusBovine lactoferrinchlamydiaedental cariesporcine-enterotoxic E. coliMicrocystis aeruginosainorganic pyrophosphatasehuman spermhuman type 0 red blood cellsnative & recombinanta-amidating enzymeReference^ ApplicationBar-Joseph & Malikson (1980) ELISAHau et al. (1980)^IEFertel et al. (1981)^RIADevergne et al. (1981) ELISAGardner & Kaye (1982)^IACarroll & Stollar (1983)^ELISAVieira et al. (1984)^RIAPiela et al. (1984) ELISA/HIAl Moudallal et al. (1984)^ELISAAltschuh et al. (1984)^RIEBurger et al. (1985) RIDSong et al. (1985)^AR/DIPBauwens et al. (1987)^RIARicke et al. (1988) ELISAShirman et al. (1988)^ETAYolken et al. (1988) PIKuhlmann et al. (1988)^PILarsson & SjOnquist (1988)^LA/ELISABauwens et al. (1988)^RIASchmidt et al. (1988) CIEIASchmidt et al. (1989)^IHCMeisel (1990)^ELISAKunz et al. (1991)^IHCOtake et al. (1991) PIWiedemann et al. (1991)^PIKang & Ho (1991)^TBPoison (1991) ZEYazawa et al. (1991)^HISturmer et al. (1992) ELISAChapter 2. Literature Review^ 19Table 2.5: Summary of Abbreviations for Table 2.4Abbreviation Full NameAR^AutoradiographyCIEIA^Competitive Inhibition Enzyme ImmunoassayDIP Double ImmunoprecipitationELIS A^Enzyme Linked Immunosorbent AssayHI Hemagglutination InhibitionIA^ImmunoassayIB ImmunoblottingIE^ImmunoelectrophoresisIHC ImmunohistochemistryLA^Latex AgglutinationPI Passive ImmunityRIA^RadioimmunoassayRID Radial Immuno diffusionRIE^Rocket ImmunoelectrophoresisZE Zone ElectrophoresisChapter 2. Literature Review^ 20requires careful control of the purity and specificity of the antibody.2.2.2 Therapeutic ApplicationsTherapeutic applications, such as passive immunization are also being explored andpresent a potentially large-scale application of IgY (Losch et al., 1986). Chicken anti-bodies are acid- and heat-resistant and might be used orally to prevent or cure infectiousintestinal diseases in young animals or humans (Gassmann et al., 1990). For example,Wiedemann et al. (1991) concluded that IgY was a good alternative to antibiotic therapyin pigs for enterotoxic E. coli. Hatta et al. (1990) stated that the effects of oral adminis-tration of antigen-specific IgY have been reported as a promising method for preventionof gastro-intestinal or dental infections.Kuhlmann et al. (1988) mentioned a variety of specific uses for IgY such as thetreatment of infectious diarrhea in foals, puppies, lambs and zoo animals; protectionagainst epidemics (typhoid, cholera) after natural catastrophes; protection of low birthweight or imrnunodeficient infants; and the treatment of diarrhea of AIDS patients orwomen during pregnancy.2.2.3 Infant FormulaAnother potential use of IgY is as a supplement for infant formula to prevent intestinalinfections (Shimizu et al., 1988). It has been shown that breast fed infants have feweroccurrences of infection than those that are bottle fed (Ballabriga, 1982; Friend et al.,1983), and the effectiveness of oral administration of bovine immunoglobulins in reducinggastrointestinal infections in infants has also been demonstrated by Ballabriga (1982).The only serious concern is the possibility of allergenicity. Some cases of hypersensitivityto bird egg yolk livetins has been reported (Mandallaz et al., 1988; de Maat-Bleeker et al.,Chapter 2. Literature Review^ 211988). This concern would require further study before large-scale use of IgY in infantformula would be possible.2.2.4 Immunoaffinity ChromatographyOne potential application of interest to the biotechnology community is the use of IgYas a ligan.d for immunoaffinity chromatography. It is likely that IgY will be used in thefuture for the separation of bioactive compounds from fermentation culture media as wellas natural resources due to the simplicity of such separation procedures (Nakai et al.,1992). Key concerns in developing such applications are the cost of both purified IgYand the supporting gel materials on which to immobilize the antibody.2.3 Relative Merits of Using IgY in Lieu of other Antibodies2.3.1 AdvantagesThe numerous advantages of using IgY in lieu of serum IgG's from various animals arewhat have spurred on research in this area. The major advantages include:• Convenience: The ease with which eggs can be collected compared to bleeding oflaboratory animals (Eichler and Rubach, 1986; Shirman et al., 1988; Poison et al.,1980a) means that IgY can be harvested and processed in large batches (Gardnerand Kaye, 1982). The animals used most commonly for producing immune sera fordiagnostic purposes are rabbits and guinea pigs. Hens require less attention thanthese animals (Poison et al., 1980a).• The relatively high concentration of IgY present in the yolk and its low cost: Chick-ens are more efficient antibody producers, since via the yolk they produce approxi-mately 20 times more antibody per kg of bodyweight than a cow does in colostrumChapter 2. Literature Review^ 22(Kuhlmann et al., 1988). Several authors have noted that a heightened level ofIgY activity can be found in the yolk', and that this level remains high long afterimmunization has ceased (Poison et al., 1980a; Shirman et al., 1988).• Compatibility with animal protection regulations: Several authors have noted thatcollecting eggs is obviously much less of a strain on the birds than bleeding (Jense-nius et al., 1981; Poison et al., 1980; Rose et al., 1974).• Health of the flock: Hens are less susceptible to diseases than laboratory animals(Poison et al., 1980a).• IgY does not bind strongly to protein A or Fc receptors: This means that fewer non-specific effects should be encountered (Gardner and Kaye, 1982). Also, Shirman etal. (1988) noted that the use of IgY in the indirect sandwich variant used in enzymeimmunoassays is especially convenient because, according to Al Moudallal et al.(1984), chicken antibodies do not give cross-serological reactions with mammalianimmunoglobulins.• Immunization is efficient: The quantity of antigen required for an immune responseis only 20 to 30 fig (Gassmann et al., 1990).2.3.2 DisadvantagesThere are, of course, some disadvantages to using IgY. Chicken antibodies do not fix com-plement and therefore cannot be used in tests employing complement fixation (Gardner5Shirman et al. (1988), using a variation of the method of Poison et al. (1980a), were able to obtain50 to 100 mg of IgY from one egg of an immunized hen. Thus they claim that in one month they couldobtain the same amount of immunoglobulins from one laying hen as could be obtained from 300 ml ofserum.Chapter 2. Literature Review^ 23and Kaye, 1982). As well, problems of isolation and purification have hindered wider use-age of IgY reagents. Although many different isolation procedures have been described,none is as easy as the purification of most mammalian IgG antibodies using affinity chro-matography with Protein A (Hassl and Aspock, 1988). Although several methods existfor the isolation and purification of IgY, most of these methods are either tedious anddifficult to scale-up, or do not lend themselves to food applications (Fichtali et al., 1992).Thus the development of a process to isolate and purify IgY from egg yolk that canbe easily scaled-up to an industrial scale would be of great benefit.2.4 Methods of Isolation and PurificationSeveral methods have been assessed for the isolation and purification of IgY from eggyolk. These have been divided into three basic parts for review. The first deals with theinitial extraction of the water soluble protein fraction (WSF) of yolk, which is generallyconsidered to be necessary step before various separation methods can be applied. Thesecond concerns the separation of IgY from the WSF. Finally, methods for purificationof IgY to relatively high levels of purity (ie: over 90%) are considered.Very few authors report attempting direct separation of IgY from egg yolk withoutthe preceding WSF extraction step. McBee and Cotterill (1979) used DEAE celluloseto fractionate egg yolk proteins into 18 bands, and stated that removal of lipids was notnecessary prior to chromatography. Later Woodward and Cotterill (1983) used DEAE-Sephacel to obtain 18 peaks, but many showed multiple protein bands in polyacrylamidegel electrophoresis (PAGE) indicating incomplete resolution. IgY was apparently not wellseparated from the other proteins. As well, the difficulties involved in using raw yolk,such as its viscosity and tendency to rapidly solidify, have deterred most researchers fromattempting to apply yolk directly to a chromatography column such as can be done moreChapter 2. Literature Review^ 24readily with egg albumen in lysozyme purification.2.4.1 Extraction of the Water Soluble Fraction from YolkThe water soluble fraction can be isolated from the lipid soluble fraction (LSF) by eitherextracting the WSF in aqueous solution, or extracting the LSF using organic solvents.Aqueous ExtractionNumerous authors have used dilution with aqueous solution to precipitate lipids andlipoproteins from yolk. The lipid granules have a strong tendency to aggregate andsettle in aqueous solution, and this effect is promoted by large dilution factors. Varioussalts, natural gums, and other additives have been used to enhance the aggregationphenomenon. Some examples are given below.Prior to gel filtration chromatography, Burley and Vadehra (1979) used dilution ofyolk 1:1 with 0.16 M NaC1 followed by centrifugation at 100,000 x g and 10°C for 30minutes. They then resuspended the sedimented granules in twice their volume of saltsolution and centrifuged again. The supernatants, containing the WSF, were then com-bined for further processing.Jensenius et al. (1981) compared two methods of aqueous extraction involving dilu-tion with Tris buffered saline (TBS), and 10 times dilution with water and pH adjustmentto 7 with 0.1 N NaOH.Hatta et al. (1988) used the simplex-centroid method (Nakai, 1988) to optimize theseparation of egg yolk into the LSF and WSF by varying pH, concentration of sodiumalginate and NaCl. The optimum conditions obtained were: sodium alginate - 0.1%; pH5.8 - 6.4; no NaCl. Hatta et al. (1990) separated the WSF in yolk using dilution withwater and precipitation of the lipoproteins using A-carrageenan. Delipidation efficiencyChapter 2. Literature Review^ 25is reported as 99.6% for this method.Kwan et al. (1991) developed a method for fractionation of both water-soluble andwater-insoluble components from egg yolk with maximum retention of biological andfunctional properties. The water soluble proteins were separated from the lipoproteinsby simple dilution with distilled water, and were thus concentrated in the supernatant.The yolk pellet was found to be useable in food applications (for instance mayonnaisepreparation) and was further fractionated to recover other valuable components. Theynoted that the efficiency of separation of lipids and phopholipids from the WSF usingsimple dilution of egg yolk with distilled water increased with the age of the egg.Simple dilution with distilled water and pH adjustment has also been used by Akitaand Nakai (1992) with optimal recovery of IgY in the WSF (and minimal residual lipids)reported to occur with 10x dilution and pH 5.2.Extraction of Lipid Soluble Fraction (LSF) with Organic SolventsA method developed by Martin et al. (1957) used ethyl ether and carbon tetrachlorideto extract the LSF.Polson et al. (1980a) used polyethylene glycol (PEG) to precipitate IgY from adiluted egg yolk mixture. Poison (1985) developed an improved PEG procedure in whichcontaminating PEG in the final product can be removed by diluting in 25% ethanol at -20°C. This yield was almost double the former technique. Traces of ethanol were removedby evaporation or by dialysis of IgY solutions against an appropriate buffer. Poison (1990)developed a new improved method of IgY isolation in which yolk homogenate was mixedwith an equal volume of chloroform and then spun at 1000 x g for 30 minutes. The IgYconcentrated in a watery phase on the top of the centrifuge tube, and was decanted priorto the use of 12% PEG Mr 6000 to precipitate IgY. He claimed that the yield of IgYChapter 2. Literature Review^ 26increased by a factor of 2.57 compared to the previous method.Bade and Stegemann (1984) used 4 extractions with pre-cooled -20°C propane-2-ol,followed by 2 extractions with pre-cooled acetone to remove the LSF, resulting in a finalprecipitate containing IgY. This precipitate was then dissolved in 0.01 M PB (pH 7.5,0.1 M NaC1, 0.01% NaN3) and centrifuged at 25,000 x g, 10°C, for 15 minutes.2.4.2 Separation of IgY from WSFSalt PrecipitationAlthough a labour intensive, often repetitive process, salt precipitation has been usedsuccessfully to obtain IgY of high purity. The reported methods generally employedprecipitation using ammonium sulphate (Martin and Cook, 1958; Wallmann et al., 1990),sodium sulphate (Kekwick, 1940; Hatta et al., 1988), or dextran sulphate (Jensenius etal., 1981). This technique is more practical, both from an economical and technical pointof view, to use as a final polishing step in industrial applications since towards the endof a purification process the product volume is generally smaller and the concentrationof the target protein much higher than in the initial stages.UltrafiltrationAkita and Nakai (1992) were able to improve IgY purity from 30% to above 93% us-ing a hollow fibre ultrafiltration cartridge with 100 kDa nominal MWC. However it wasnecessary to first separate crude IgY by ammonium sulphate precipitation. Otherwise,clogging of the membrane with lipid-containing mucous materials would prevent efficientseparation. Recently (Nakai et al., 1992) were successful in applying ultrafiltration di-rectly to the WSF by adjusting the pH to 9.0. The resulting recovery and purity were95% and 80% respectively.Chapter 2. Literature Review^ 27Protein-A ColumnsIt is not possible to use protein-A columns to purify IgY, as is done for mammalian IgG.Ansari and Chang (1983), using a protein-A Sepharose column, found that IgY doesnot bind to protein-A molecules. They concluded that the reasons for the non-bindingproperties of chicken IgY may be:• the presence of protein which binds to the protein-A molecule and inhibits thebinding of protein-A to IgY• the amino acid configuration of the F, region of chicken IgY is different from mam-malian IgG and has different binding propertiesPrecipitation with AntigenPoison et al. (1980a) raised antibodies to tobacco mosaic virus (TMV) in hens andpurified this specific IgY using absorption to the virus and removal of the nonspecificIgY by centrifugation and discarding of the supernatant. They determined that 15-18%of the total IgY was directed to TMV.Schram et al. (1971) immunized hens against bovine albumin and recovered IgY fromthe other water soluble proteins by precipitation with BSA. The IgY was then purifiedusing a Sephadex G-200 gel filtration procedure.ChromatographyDue to recent innovations in the development of chromatographic media, industrial grademedia are now available with greatly improved capacities and flow properties, and atreasonable prices. Several forms of chromatography have been used in IgY separationand purification.Chapter 2. Literature Review^ 28Anion Exchange To date, anion exchange chromatography (AEC) has been the pre-ferred chromatographic method for purification of IgY. In 1960, Mandeles used DEAE-cellulose with an acidic gradient elution profile to separate the WSF of the yolk. Theresulting chromatogram consisted of 10 peaks, a larger number of components than usu-ally obtained using electrophoresis. Although many researchers have made use of AEC,not all agree on its effectiveness. For instance, Higgins (1976) used DEAE Sephadex A-50 with a linear gradient to fractionate chicken serum. He concluded that ion-exchangechromatography has little application in the separation of fowl immunoglobulins since allthree Ig classes (IgG, IgM and IgA) eluted throughout most of the linear salt gradient.Using DEAE-Sephacel anion exchange media, McCannel and Nakai (1990) were ableto obtain a purity of only about 36% for a recovery of 60%. The highest purity obtainedwas about 60%, but for a fraction representing only 15.8% of the IgY applied. However,the first eluted fraction, although containing only about 20% of the IgY applied, rep-resented the majority of the specific IgY based on ELISA analysis. It follows that thismethod could be used to selectively purify IgY that is specific for the antigens used inthe study, namely E. Coli LPS (lipopolysaccharide) and 0-lactoglobulin.McCannel and Nakai (1990) state that "Separation of immunoglobulins by ion- ex-change chromatography may be due to amino-acid sequence differences occurring in thevariable region of the heavy chain, which dictates the specificity of antibodies." Thus asa separation step, early in the process of IgY purification, AEC holds the potential fordifferentiating IgY subclasses.Gel Filtration A table is given in the introduction of Chapter 7 which lists many ofthe published papers in which gel filtration of IgY has been reported. Gel filtration isgenerally used in the final stages of IgY purification. It has been used as a polishingChapter 2. Literature Review^ 29step, as well as a method for estimating the molecular weight and concentration of themolecule.High Performance Liquid Chromatography One paper only (Burley and Back,1987) has been found dealing with the use of an HPLC column to fractionate yolk pro-teins. However, this was done only on an analytical basis.Hydrophobic Interaction Chromatography Hassl and Aspock (1988) developeda 2 step chromatographic procedure using hydrophobic interaction chromatography fol-lowed by gel filtration. The major advantage of the technique is that a relatively highpurity is obtained with only one initial batch precipitation step, followed by 2 chromato-graphic steps. Therefore the process is rapid, there are potential time and labour savingssince it could easily be automated, and antibody activity is preserved. Their procedure,however, has a relatively low yield, less than the yield for the method of Poison et al.(1985) with a purity of 85% or greater.Immunoaffinity Chromatography Otani et al. (1991) used immunoaffinity chro-matography to purify IgY produced in the yolk of hens hyper-immunized against a-s1-casein. The casein was first bound to a column of Sepharose 4B, the anti-a-s1-casein eggyolk IgY6 applied to the column, and the IgY eluted with 0.5M glycine-HC1 buffer, pH2.3, and then 4M guanadine-HC1, pH 7.0, in a cold room.Metal Chelate Interaction Chromatography Copper- loaded MCIC has been usedfor the separation of bovine IgG from blood and cheese whey (Lee et al., 1988). McCanneland Nakai (1989) used this technique to separate IgY from the other livetin proteins, and'Previously purified according to the method of Poison et al. (1985), followed by DEAE SephacelchromatographyChapter 2. Literature Review^ 30found that the majority of IgY eluted in the unbound fraction. For moderate loading ofthe gel (60 mg of protein applied per ml of MCIC gel), it was possible to pool the unboundfraction and obtain a recovery of 74%, but with a purity of only about 50%. A purityof about 67% was possible, but with a much reduced recovery of only 36%. Consideringthe rather high cost of this gel in comparison with ion exchangers, it is unlikely that thisprocess would prove as economical.Cation Exchange Very few authors report the use of cation exchangers to separateIgY from other yolk proteins. Most of the published work involves whole egg or egg yolk.Dreesman and Benedict (1965) used carboxymethyl-cellulose (CMC) to fractionatepapain digested chicken antisera. "Immunoelectrophoresis of digested antibovine serumalbumin IgG revealed 2 antigenically distinct fragments" that they termed electrophore-tically-slow and fast. They concluded that the fragments separated poorly on CMC atpH 5.4, and a better separation could be achieved by chromatography on DEAE celluloseand by starch block electrophoresis.Seideman and Cotterill (1965) fractionated yolk proteins on CMC and noted that thefractions obtained did not completely correspond with proteins prepared by other inves-tigators. They suggested that these fractions may represent combinations or complexesof proteins or lipoproteins.Parkinson (1967) applied whole egg, egg white and egg yolk to DEAE-cellulose (DE52)and whole egg to CMC (CM52) in order to fractionate the proteins. He concluded thatDEAE-cellulose was more suitable than CMC, but that even using the anion exchanger,most of the fractions contained more than one protein, and there was considerable tailingin some of the fractions.Seideman et al. (1969) investigated the use of ion exchange cellulose for the separationChapter 2. Literature Review^ 31of egg yolk proteins and lipoproteins. They determined that DEAE cellulose showedlittle promise for separating yolk proteins and lipoproteins, and likewise for CM celluloseusing acetate buffers. However, they were able to obtain a separation using a citrate-phosphate buffer system. Using a column equilibrated to between pH 4.5 and 5.0, andthen stored in a starting buffer of pH 5.4, application of native chicken egg yolk apparentlyresulted in the a-, #-, and -y-livetin passing directly through the column, along with lipidcontamination of about 4.2% (a free lipid or lipoprotein). The other yolk proteins andlipoproteins were eluted by increasing the pH.Parkinson (1972) used CMC to fractionate yolk proteins into a large number of peaksfollowing elution with an NaC1 gradient and other salt solutions.Thus papers dealing with cation exchange chromatography of undiluted egg yolk andliquid whole egg are concerned more with identification of various fractions rather thanpractical purification of any of the proteins.Chapter 3Assembly of an Automated Chromatography System3.1 IntroductionSince the goal of this work was to determine the potential for large scale purification ofIgY using some form of chromatography, it was necessary to consider what type of equip-ment would be most beneficial in carrying out first small scale tests at the laboratory orbench scale, and later pilot scale experiments. Because it was considered likely that nu-merous experiments at the lab scale would be required in order to test different types andbrands of chromatographic media, and to investigate the effect of changes in various pro-cess parameters such as flowrate, pH, ionic strength, etc., an automated system seemedessential. In order to simplify the work of scaling up the process to pilot scale, it was alsoconsidered advantageous if the same system could be adapted to run the larger experi-mental trials as well. Although many companies carry automated systems for laboratoryscale chromatography, it is not as easy to find fully automated process scale equipmentwith a variety of interchangeable components (Johansson et al., 1988). Since differenttypes of chromatographic media might require the monitoring of different parameters,the preparation and delivery of buffers of various elution profiles (linear vs step gradientsin pH, ionic strength, etc.), or widely differing flowrates, it was necessary that types ofvarious pumps, valves, and other equipment be easily interchangeable. Thus the designof a highly flexible chromatography system is an important step for purification processdevelopment.32Chapter 3. Assembly of an Automated Chromatography System^333.2 Initial System DesignThe author, having had no prior experience in operating chromatography equipment,sought the advice of those with practical experience in operating such equipment, aswell as information available in scientific catalogues, manuals and textbooks on practicalbiochemistry. The idea was to determine a simple but proven starting point for thedesign of the system. Using primarily equipment already available in the laboratory,an automated system was pieced together as shown in Figure 3.2. As can be seen, aperistaltic pump is used to maintain a constant flow of sample or buffer through thechromatography column Upon exiting from the column, the solution flows through anin-line UV monitor, and finally to a fraction collector. The chart recorder produceshardcopy of the chromatograms which must then be digitized for computer analysis, orintegrated using other methods.The system is crude and unsophisticated, but it gets the job done - as long as there arenot too many buffers involved, no elution gradients (other than steps) are required, andthe operator has the time to hang around and look after switching buffers and turningequipment on and off. In many laboratories, and for many small experiments, this typeof system is quite adequate. However, since an anion exchanger requiring linear gradientelution was one of the first media tested for purifying IgY, a slightly more sophisticatedsystem was quickly assembled.3.3 Improved System with Process ControllerThe next permutation included a second peristaltic pump and mixing chamber (formedby placing a flask on a stirring hot plate) to allow the formation of linear salt or pHgradients, several valves and buffer tanks, and a Pharmacia C3 dedicated controller.Chapter 3. Assembly of an Automated Chromatography System^34Figure 3.2: Simple Laboratory Chromatography SystemChapter 3. Assembly of an Automated Chromatography System^35The control unit provided timed on/off control of the solenoid valves, and on/off powercontrol for the pumps, mixing chamber and fraction collector. The chromatograms werestill registered by the chart recorder. A schematic diagram of this system is shown inFigure 3.3.This system was used for the first series of experiments in which an anion exchangecolumn was used to separate IgY from the other water soluble proteins in diluted eggyolk supernatant (to be described in Chapter 4). Although the full potential of theC3 controller was not used', it was still limited in its ability to collect and analyzechromatograms Therefore, a microcomputer based controller was sought which could beinterfaced easily with the equipment already available and provide rapid data acquisitionand analysis.3.4 Final System DesignFigure 3.4 is a schematic diagram of the automated chromatography system which wasfirst designed for bench-scale separation. The system includes a 2.5 cm inside diametercolumn, buffer and sample tanks, two remote control Ismatec peristaltic pumps (Cole-Parmer, Chicago, USA), an in-line Pharmacia UV detector and monitor (Pharmacia-LKB, Upsalla, Sweden), an ISCO Retriever II fraction collector (ISCO Inc., Lincoln,Nebraska), several solenoid valves (Burkert, Germany), an IBM compatible personalcomputer and monitor, as well as tubing and fittings.An ISCO ChemResearch data management/system controller specifically designed forHPLC was adapted for use with the low pressure liquid chromatography system used inthis work. The software is menu driven, and includes routines to allow calibration and'The C3 is capable of controlling processes involving more complex elution gradients, but requiresa gradient former. As well, the C3 can accept input from an in-line flow meter, and can trigger thecollection of individual peaks based on sudden changes in the monitored baseline.Chapter 3. Assembly of an Automated Chromatography System^36Figure 3.3: Improved Chromatography SystemSolution ASaltwashSampleWashingBufferEquilibrationButlerSolution BComputer withinterface moduleChromatographyColumnPUMP Ao-MixingValve xl^PUMP B^oTM,Waste FractionCollectorChapter 3. Assembly of an Automated Chromatography System^37Figure 3.4: Computer Controlled Chromatography SystemChapter 3. Assembly of an Automated Chromatography System^ 38control of devices (i.e. pumps, valves), data acquisition, analysis and graphic display. Anexternal distribution interface module into which externally controlled instruments andaccessories are connected is included in the system. However, it was necessary to usea custom designed and built control interface unit to actuate the process scale solenoidvalves. The unit amplifies the low/high (0/5VDC) signal from the ISCO interface moduleby using a Darlington transistor arrangement powered by a 24 VDC power supply2.Tygon food grade tubing (B-44-4X, 6.4 mm outside diameter, 3 2 mm inside diameter)was connected to the valves by nylon fittings.3.5 ConclusionsThe automated system developed offers several advantages for carrying out repetitivechromatography experiments. The data acquisition software allows for routine orga-nization and saving of equipment settings and parameter values, and rapid analysis ofchromatograms. It is relatively easy to transfer the data to a graphing program in order toproduce publication quality graphs. Time savings are substantial since many operationsare accurately controlled and carried out. It is possible to run experiments overnight,which in turn allows more efficient use of laboratory equipment. Finally, experimentscan be accurately controlled and easily reproduced.'See Appendix A for a schematic diagram of the interface.Chapter 4Extraction of the Water Soluble Proteins4.1 IntroductionAs mentioned in the literature review, there are two basic approaches used to separatethe water and lipid soluble fractions: aqueous extraction of the WSF or the use of organicsolvents to extract the LSF. Since other authors have reported good separation of theWSF from manually separated yolk using only dilution with H 20 and pH adjustmentwith diluted HC1 or NaOH (Kwan et al., 1991), it was decided to study the aqueousextraction of industrially separated yolk using different dilution ratios at various valuesof pH.Such a procedure meets the goal of developing a process yielding purified IgY of foodquality. This would result in a fairly simple extraction that meets with the requirementsof developing a simple process for future scale-up.4.2 Materials and Methods4.2.1 Raw MaterialIndustrially separated egg yolk obtained from a local egg breaking plant (VanderpolsEggs Ltd., Abbotsford, B.C.) was stored at 4°C with 0.02% sodium azide until use. Eggswere also obtained from a local market and the yolk separated manually for comparison.39Chapter 4. Extraction of the Water Soluble Proteins^ 404.2.2 Separation of LipoproteinsSimple water dilution, followed by sedimentation, was used for lipoprotein separationfrom egg yolk water soluble proteins. Industrially separated fresh egg yolk was diluted10x with distilled water, the pH adjusted with 0.1N HC1/0.1N NaOH, and the solutionfurther mixed with magnetic stirrer for about 10 min The effect of pH on residual lipidsin the supernatant and IgY recovery after 24 h sedimentation of lipoproteins was studiedfor pH ranging from 5.0 to 7.0. Three pH values (5.0, 5.25, and 5.5) were considered instudying the effect of pH and time on lipoprotein sedimentation and IgY recovery. Thesolutions prepared under the specified conditions were allowed to settle at 4°C in one litreImhoff cylinders, and the precipitate volume and IgY concentration in the supernatantmeasured as a function of time up to 7 days. The effect of a second dilution/extractionstep on the IgY recovery after 24 hours sedimentation, using 10x dilution and pH 5.5, wasalso investigated. All solutions were prepared in duplicate, and the average experimentalresults reported for each treatment.4.2.3 Analytical ProceduresNon-denaturing sodium dodecylsulphate polyacrylamide gel electrophoresis (SDS-PAGE)was carried out on a Pharmacia Phast System using a 10-15% gradient PhastGel andCoomassie brilliant blue stain according to the manufacturer's recommendations (Pharmacia-LKB, Upsalla, Sweden). Molecular weight standards (BioRad, Redmond, CA) were usedto estimate the molecular weights of proteins in the WSF.Total protein was determined using the BCA method (Pierce, Rockford, Ii.). In thisprocedure, 0.1 nil of unknown protein sample is mixed with 2.0 ml of working reagentcontaining bicinchoninic acid and Cu2+ ions. The protein present in the sample reducesChapter 4. Extraction of the Water Soluble Proteins^ 41Cu2+ to Cui+ (the biuret reaction) which then reacts with BCA to produce a purple re-action product. Spectrophotometric quantitation of the protein is possible by measuringthe absorbance of the solution at 562 nm after incubating the sample for 30 minutes at37°C.The total lipids content of the supernatant samples was determined using solventextraction (see Hatta et al., 1990).Radial immunodiffusion (Kwan et al, 1991) using chicken serum IgG (ICN, Cleveland,OH) as a standard was used for quantitative analysis of IgY in aqueous solution. Rabbitanti-chicken antibody (0.15 ml) was added to 1.85 ml phosphate buffer (0.05 M, pH7.5), and the test tube then warmed in a 56°C water bath. Agarose (0.07 g) (Sigma,St. Louis) was added to 4.6 ml PB and 0.4 ml 0.35% sodium azide. This mixture washeated in a boiling water bath until the agarose had dissolved, and was then placed inthe 56°C bath. After about 10 minutes, the two solutions were mixed and poured intoradial immunodiffusion plates. After approximately 15 minutes at room temperature,the plates were transferred to a sealed container to which moist paper towel had beenadded. The container was then stored at 4°C until use. Wells were cut in the gels priorto use, and 4 pl samples applied to these wells. After storage for 3 days in the moistchamber at 4°C, measurements of ring diameters were taken using a digital micrometer.IgY purity was determined by dividing IgY concentration determined using RID by thetotal protein of the sample, as determined by the BCA method.Chapter 4. Extraction of the Water Soluble Proteins^ 424.3 Results and Discussion4.3.1 Comparison of Industrially and Manually Separated Egg YolkFigure 4.5 shows the SDS-PAGE of supernatants obtained from the industrially separatedegg yolk used in our study (column A) and from manually separated egg yolk (column B).BioRad molecular weight standards were used to determine the molecular weight of thebands obtained in non- denaturing SDS-PAGE of the WSF from manually vs. industriallyseparated yolk. The slowest moving band was IgY with an estimated molecular weightof 175 kDa. Next came 2 bands at 78 and 57 kDa, which have been identified by otherauthors as a- and 13-livetin, respectively. Both samples contained two smaller bands at35 and 38 kDa which have not been identified. One major additional band appears inthe supernatant obtained from the industrial egg yolk. Since the industrial separationof the yolk from the white is not complete (the separated yolk typically containing 22%egg white on wet basis or 5.5% on dry basis), and since the molecular weight of theadditional band is in the vicinity of 42 kD, it would appear that this additional protein isovalbumin. Electrophoresis of purified ovalbumin standard revealed a band in a similarposition on the gel. Thus the purification of IgY from industrial egg yolk is somewhatmore difficult than from manually separated yolk.4.3.2 Separation of Lipoproteins from Egg Yolk Soluble ProteinThe first step in the purification process is the separation of the granular lipoproteinsfrom the soluble fraction consisting mainly of livetins (Hatta et al., 1990). The effectof pH and dilution on lipid removal and IgY recovery has already been studied usinglaboratory separated egg yolk as the raw material and using centrifugation for separation(Akita and Nakai, 1992). As the objective was to develop a process useful for large scaleChapter 4. Extraction of the Water Soluble Proteins^ 43Figure 4.5: SDS-PAGE of Manually versus Industrially Separated Egg YolkM.W. (kD)• 175• 78574•00. •^NT-- 42• 3835.0111.10.Chapter 4. Extraction of the Water Soluble Proteins^ 44purification of IgY, industrial egg yolk was used instead, and sedimentation rather thancentrifugation was considered in order to decrease the overall cost of the process.Figure 4.6 shows the effect of pH and dilution on residual lipids (Figure 4.6a) andIgY recovery (Figure 4.6b). For the case of 5x dilution, the IgY recovery at pH 5 and 5.5has not been reported as the separation was poor and often inconsistent, perhaps dueto the low dilution ratio. At pH 5.0 the effect of dilution on residual lipids is negligible,but at higher pH the residual lipids drops significantly with increasing dilution. At pH6.0, the maximum recovery of IgY occurs for all dilutions (around 80%), but this alsocorresponds to the highest lipid residual for the conditions studied. Only at pH 5.0 forall dilutions, and pH 5.5 for 10 times dilution, is the residual lipids below 10%.Under these conditions, where the residual lipid content is less than 10%, the bestrecovery was obtained for 10x dilution at pH 5.5 (recovery of 53%). Having settled on a10x dilution ratio, two additional pH's were tested (pH 5.25 and 5.75) and the residuallipids and IgY recovery after 24 h presented in Table 4.6. This table shows that residuallipids increases sharply when the pH is greater than 5.5. The experiment was carried outonce, and the residual lipids and IgY concentration determined in duplicate.Figure 4.7 shows a 1 litre Imhoff cylinder of 10x diluted yolk adjusted to pH 5.5 andallowed to sit at 4°C overnight. A relatively clear supernatant layer (50 - 60% of thetotal volume) forms above a fluffy yellow precipitate. A thin lipid layer also forms overthe top of the aqueous supernatant.In order to follow the sedimentation of the LSF over a period of time, 10x dilutedyolk adjusted to pH 5.0, 5.25, and 5.5 was allowed to settle in 1 litre Imhoff cylindersfor several days at 4°C. Figure 4.8 shows the sedimentation curves (Fig. 4.8a) and IgYrecovery (Fig. 4.8b) as a function of time with the experimental sedimentation data fittedto the following model:Chapter 4. Extraction of the Water Soluble Proteins^ 45Figure 4.6: Effect of pH and Dilution on Residual Lipids (a) and IgY Recovery (b)Chapter 4. Extraction of the Water Soluble Proteins^ 46Figure 4.7: Diluted Yolk after 24 hours at 4°CChapter 4. Extraction of the Water Soluble Proteins^ 47Figure 4.8: Effect of pH and Time on Volume of Sedimented Lipoprotein and IgY Re-coveryapH5.005.255.50b80 -pH5.505.255.0010^40^80^120^160^200^240Time (h)Chapter 4. Extraction of the Water Soluble Proteins^ 48Table 4.6: Effect of pH on residual lipids and IgY recovery after 24 h sedimentation using10x dilutionpH Residual lipids (%) IgY recovery (%)5.0 5 225.25 5 435.5 6 535.75 16 616.0 41 826.5 28 817.0 41 77s = a/(t + b) + cand IgY recovery data fitted to the following model:r = d(1 _ e-ft) + ollhwhere s is the percent volume of sedimented lipoproteins; t is the time; r is the percentIgY recovery; and a, b, c, d, f, g and h are empirical parameters determined to best fiteach model. These models are strictly7 empirical best fits to the data determined bytrial and error and a standard curve fitting routine.The yield under the chosen conditions (10x dilution, pH 5.5) is 53% after 24 h sedimen-tation, and slightly higher (57%) after 48 h sedimentation (Figure 4.8b). An additionaldilution/extraction step under the same conditions improved the recovery from 53% to79% after 24 h sedimentation. However, this recovery is still lower than what could beobtained using centrifugation (>90% using 10x dilution and pH 5.2) according to Akitaand Nakai (1992).Chapter 4. Extraction of the Water Soluble Proteins^ 494.4 ConclusionsAlthough the use of additives such as sodium alginate (Hatta et al, 1988b) and A-carageenan (Hatta et al., 1990), centrifugation of the settled yolk mixture (Ibid., Akitaand Nakai, 1992), or filtration (Akita and Nakai, 1992) can result in significantly higherrecoveries and/or better lipids removal in the first step, simple water dilution and pHadjustment is still a valuable low cost technique for recovery of significant quantities ofthe WSF. In addition, the remaining water-insoluble yolk fraction could still be used infood applications or for the separation of other biologically active components (Kwan etal., 1991). The specific economics of a particular large-scale operation would dictate ifthe added benefits are worth the additional cost of materials and equipment (such as thegums and a continuous centrifuge).In the experiments described in the following chapters, unless otherwise stated, thesupernatant containing the WSF was prepared by 10 times dilution with distilled water,pH adjustment to 5.5, and settling for 24 h at 4°C.Chapter 5Separation of IgY5.1 General IntroductionAlthough many techniques have been developed to isolate and purify IgY from the WSF ofegg yolk (see literature review in Chapter 2), the vast majority have only been attemptedat the laboratory scale, and using yolk carefully separated by hand from the yolk sac soas to minimize albumen contamination. For large-scale isolation of IgY, a technique isrequired which would be practical to use with the large volume of dilute protein solutionobtained after aqueous extraction.Ultrafiltration is a method that permits both concentration and fractionation of pro-teins in aqueous solution'. It is also easily scaled-up, although the energy costs can besignificant compared to other methods (Kroner et al., 1984).Anion exchange chromatography has been used by many authors for IgY separationusing laboratory exchangers such as DEAE-Sephacel (Pharmacia, Uppsala, Sweden),and since industrial grade exchangers are readily available it was considered a viablepossibility.Finally, although no previous publications mentioned the use of cation exchange chro-matography (CEC) to separate IgY from the WSF, CEC has been used to fractionateliquid whole egg (Parkinson, 1967), yolk (Seideman and Cotterill, 1965; Parkinson, 1967;Seidem.an et al., 1969; Parkinson, 1972), and the papain digested fragments of the livetin'Provided the difference in mass of the molecules to be separated is sufficiently large.50Chapter 5. Separation of IgY^ 51fraction (Dreesman and Benedict, 1965).Ion exchange chromatography is used economically at the large scale for other chemi-cal and food industry separations (Streat, 1988; Barker and Ganetsos, 1988) and so bothanion and cation exchange chromatography were tested for applicability to separate IgYfrom industrial yolk.5.2 Ultrafiltration5.2.1 IntroductionThe technology of cross-flow filtration dates from the early 1960's, when the porousmembranes used were initially developed for water desalination. Cross-flow filtration canbe used to fractionate liquid streams from three digit molecular weight to submicronlevels (Davis, 1987). There are four basic classes of membrane separation used today:microfiltration, ultrafiltration, reverse osmosis and nanofiltration. Ultrafiltration canachieve simultaneous fractionation and concentration of solute in the range 0.002 to 0.2microns, corresponding to a molecular weight range of 500 - 300,000 Daltons (Dziezak,1990).The use of ultrafiltration in the food and pharmaceutical industries to concentratedilute protein solutions and to separate proteins from solutes or cells is now quite com-mon. Its use for fractionating proteins according to size is not yet as common due toseveral problems (Ingham et al., 1980; Dziezak, 1990):• membrane characteristics - non-uniform pore size and compaction problems• solute/membrane interactions - concentration polarization and fouling• protein-protein interactions - hetero- and self- association of proteinsChapter 5. Separation of IgY^ 52Over the last decade, major improvements have been made by the manufacturersin membrane quality, uniformity of pore size and in the development of new membranematerials (Abelson, 1989; Kroner et al., 1984).Concentration polarization refers to the increase in concentration of rejected specieswith decreasing distance from the membrane. At high enough concentrations, a gel layerforms which can lead to fouling, the build-up of proteins, fats and suspended solids thatcauses hydrodynamic resistance and interferes with the flux. The challenge for membranedesigners is to develop a membrane that is hydrophilic, but has very little net charge.Concentration polarization, however, occurs regardless of whether or not there is anyprotein adsorption to the membrane (Ingham et al., 1980).Protein-protein interactions such as hetero-interactions2 can significantly reduce theefficiency of fractionation. Ingham et al. (1980) carried out ultrafiltration of a binarymixture of lysozyme (molecular weight approx. 14,000 Da) and albumin (molecularweight approx. 69,000 Da) using an Amicon PM-30 UF membrane in an Amicon TCF-10 spiral channel unit in diaflltration mode3 In the absence of albumin and salt, lysozymepassed through the membrane and virtually all resided in the permeate after three vol-ume changes. When excess albumin (50 mg/ml) was added to the initial solution, lessthan 10% of the lysozyme passed through the membrane due to hetero-association withalbumin. The effect was strongly reversed by the presence of 0.25 M KC1, which partiallyinhibits complex formation.Likewise, many proteins self-associate (Ingham et al., 1980) and may not pass througha filter even though the nominal molecular weight cut-off value of the filter is much higherthan that of the protein in its monomeric form.2Interactions between different molecular species.3In diafiltration mode, the sample volume remains constant. Used in this way, the ultrafiltrationapparatus is effectively replacing dialysis, with the added benefit of protein fractionation.Chapter 5. Separation of IgY^ 535.2.2 Materials and MethodsExperimental MethodsInitial experiments involving the concentration of diluted yolk supernatant were carriedout using a Romicon PM-100 (nominal molecular weight cut-off (NMWC) 100,000) hol-low fibre ultrafiltration cartridge (Romicon Inc., Woburn, Mass.), with tygon tubing andMASTERFLEX peristaltic pump (Cole Parmer, Chicago, IL). Yolk supernatant was pre-pared by diluting industrially separated yolk ten times with distilled water. The mixture(pH 6.2 to 6.4) was allowed to sit a minimum of 24 hours at 4°C before ultrafiltration.Later experiments involved the use of a lab-scale hollow fibre ultrafiltration cartridge(AG Technology Corporation, Needham, MA) of the same NMWC as above. Yolk wasdiluted ten times with distilled water and the pH adjusted to 5.0 with 0.1 N HC1. Settlingwas allowed to occur for one week at 4°C, and then the WSF filtered4 and stored untiluse. One day prior to ultrafiltration, the pH was adjusted to 9.0 with 0.1 N NaOH, andthe solution allowed to sit overnight at 4°C. After 0.45 m filtration, the solution was di-afiltered, the permeate volume being replaced occasionally with distilled water previouslyadjusted with 0.1 N NaOH to pH 9.0. The appartus used is shown in Figure 5.9.Analytical MethodsElectrophoresis (SDS-PAGE) was carried out as described in Chapter 4, and HPLCanalysis followed essentially the same procedure as that described in Chapter 7 exceptthat the running buffer was PBS (0.01 M phosphate buffer pH 7.0 with 0.14 M NaC1 and0.02% NaN3).4See filtration of WSF in Chapter 4.Chapter 5. Separation of IgY^ 54Figure 5.9: Ultrafiltration ApparatusChapter 5. Separation of IgY^ 555.2.3 Results and DiscussionInitial Ultrafiltration ExperimentsEven before the optimal conditions for the aqueous extraction step had been determined,ultrafiltration was considered as a possible separation step to follow. As mentionedpreviously, at that time the supernatant was being prepared at ten times dilution butwithout pH adjustment. The resulting pH was generally slightly acidic, pH 6.2 to 6.4,and as shown in Figure 4.6, the lipid residual was 30-40% under these conditions.Since SDS-PAGE of the supernatant indicated the presence of 5 major contaminatingproteins (of which alpha-livetin at a molecular weight of 80,000 is the largest), the idea wasto retain only IgY and allow all 5 contaminants to pass through the filter by selecting anappropriate nominal molecular weight cut-off. Since a Romicon PM-100 (nominal MWC100,000) was readily available', ultrafiltration was attempted with this cartridge.SDS-PAGE results indicate that the apparatus functioned well in concentrating thesupernatant, but no significant reduction in contaminating proteins occurred. This ratherdisconcerting result can be attributed to several possible causes:• molecular weight cut-off too close to a-livetin• fouling with lipids/lipoproteins• protein-protein interactionsSince the molecular weight cut-off of the UF cartridge was close to the molecularweight of the largest contaminating protein, a-livetin, it is likely that a substantialamount of this protein would remain in the retentate. However, since all of the a-livetinappears to be retained, it is unlikely that this accounts solely for the lack of separation.5 Courtesy of Canadian Lysozyme Incorporated, Abbotsford, BC.Chapter 5. Separation of IgY^ 56A significant amount of lipid is present in the supernatant, so fouling is definitelya possibility. Based on the data available in this study, it was impossible to determinewhether or not protein-protein interactions were wholly or partially responsible for theretention of a-livetin. However, ovalbumin contamination of the industrially separatedyolk is significant, and as noted previously, albumins (such as BSA) are known to formcomplexes with other proteins and hinder their passage across UF membranes. Thus thispossibility also exists, and further studies are needed to elucidate the problem.Later Ultrafiltration ExperimentsLater in the work, another attempt was made to use ultrafiltration to separate IgYfrom other proteins in yolk supernatant. Mr. Emmanuel Akita of the Department ofFood Science, UBC, reported obtaining excellent separation using a 100 kDa NMWCmembrane (Akita, 1992, personal communication). Briefly, his starting material wasmanually separated yolk, and after 10 times dilution of the yolk and pH adjustment topH 5.0-5.2, the mixture was allowed to sit for at least two hours. It was then eithercentrifuged at 10,000 x g for one hour at 4°C, or filtered through Whatman No.1 filterpaper in the cold. After adjustment to pH 9.0, the supernatant was concentrated 5 timesby ultrafiltration and then diafiltered using distilled water adjusted to pH 9.0.The results obtained following Akita's method were only marginally better than thoseobtained without pH adjustment of the supernatant. An HPLC system was used to studythe permeate and retentate. Using a gel filtration column under the conditions describedin Chapter 7, it was determined that the two smalles contaminating peaks were almostcompletely removed after three volume changes.Once again, a-livetin remained in the retentate. The two major differences betweenthis experiment and those of Akita are that he used manually separated yolk free fromChapter 5. Separation of IgY^ 57albumen contamination (compared to the albumen contaminated industrially separatedyolk used here), and centrifuged the yolk mixture instead of using sedimentation at4°C. The difference in results obtained might again be explained by lipid fouling of themembrane or complex formation caused by the presence of ovalbumin or other egg whiteproteins.5.2.4 ConclusionsUltrafiltration of the supernatant prepared by diluting egg yolk and allowing the mix-ture to settle overnight at 4°C failed to achieve significant separation of IgY from othercontaminating proteins. Further work is necessary to investigate the possibility of usingthis technique for separation of IgY from industrially separated yolk.5.3 Anion Exchange Chromatography5.3.1 IntroductionAnion exchange chromatography is the chromatographic technique used most commonlyfor IgY separation and purification. Pioneering work was carried out in the 1960's and70's by several authors (Mandeles, 1960; Benedict, 1967; Parkinson, 1967 and 1972; Bur-ley and Vadehra, 1979) involving laboratory scale apparatus and most emphasizing theuse of a linear salt gradient to elute a rather large number of peaks. Papers have contin-ued to appear describing several variations of these protocols (Higgins, 1976; Gassmannet al., 1990; McCannel and Nakai, 1990; Otani et al., 1991).In order to test the basic anion exchange protocol for its applicability to the separationof IgY from the WSF of industrially separated yolk, a number of small scale experimentswere carried out.Chapter 5. Separation of IgY^ 585.3.2 Materials and MethodsA 50 ml column (i.d. 2.5 cm) was packed with DEAE-Sephacel (Pharmacia, Uppsala,Sweden) and used with the second generation semi-automated chromatography systemdescribed in Chapter 3 (shown in Figure 3.3). Two sets of buffer systems were utilized.The first set of experiments involved the use of Tris-buffer (TB). The column was equi-librated with at least 6 bed volumes' of 0.5 M TB (pH 7.5), and then washed with 4bed volumes of 0.03 M TB, pH 7.5 (washing buffer). Sample was buffered to pH 7.5with 0.03 M TB before application to the column. The column was then washed withwashing buffer, and a multiple step or linear salt gradient formed using either a valvingsystem or mixing chamber and 2 peristaltic pumps. A conductivity meter (Cole Parmer,Chicago, Illinois) was used to verify the ionic strength of the mixing chamber before andafter each run, from which the NaC1 molarity could be estimated for comparison withpredicted values. A salt wash of 3 bed volumes of Tris buffered saline (TBS) pH 7.5,0.5 M NaC1, was used to elute any unbound proteins and was followed by 1 to 2 bedvolumes of washing buffer.In later experiments, a phosphate buffer (PB) system was used, since this buffer isacceptable for food applications. The same procedure as described above was followed.Equilibration buffer was 0.2 M PB pH 8.0, washing buffer 0.02 M PB, pH 8.0, andsaltwash 0.02 M PBS pH 8.0, 0.5 M NaCl.'One bed volume is generally defined as the volume occupied by the exchanger in the column, in-cluding all pore spaces, ie: the total volume of the exchanger.Chapter 5. Separation of IgY^ 595.3.3 Results and DiscussionInitial experiments were carried out to study the extent to which IgY would bind tothe DEAE-Sephacel column. A linear salt gradient was used for elution because one-step elution resulted in virtually all of the bound protein eluting from the column ina single peak. Using a linear gradient from 0 to 0.25 M NaC1, the majority of IgY(according to SDS-PAGE) was present in several peaks eluted in the range 0.1 to 0.15 MNaCl. However, several contaminating proteins were also present with IgY in these peaks.Significant "tailing" of peaks occurred in most but not all cases, indicating that a largenumber of proteins and possibly subgroups, with different isoelectric points are present.The buffer system was then changed to phosphate buffer in order to ensure compat-ibility with food safety regulations. For most experiments, a linear elution profile from0.02 to 0.18 M NaCl was used. The largest peak eluted when the salt concentration ofthe mobile phase was between 0.12 and 0.15 M, but again, IgY was spread across severalpeaks. The majority of adsorbed protein eluted near the beginning of the gradient as theionic strength approached 0.10 M.5.3.4 ConclusionsAlthough anion exchange chromatography has been used extensively by numerous re-searchers for laboratory scale purification of IgY, it appears to hold little hope for use ona large scale using industrially separated egg yolk as a starting material. This is not dueto a lack of good industrial grade exchangers, but rather a lack of differentiation betweenIgY and other livetin proteins under the basic conditions of anion exchange.Chapter 5. Separation of IgY^ 605.4 Cation Exchange Chromatography5.4.1 IntroductionDespite the extensive use of anion exchange chromatography to separate IgY, no attemptto use a weak cation exchanger to separate IgY from other yolk water soluble proteinsat a preparative scale has yet been reported.5.4.2 Materials and MethodsChromatography ExperimentsThe automated system shown in Figure 3.4 was used with a column (2 cm inside diameter)packed with about 35 ml of CM-92 (Whatman Biosystems Inc., USA) cation exchangemedia. The column was equilibrated with 0.2 M Phosphate Buffer (PB) at pH 5.0and then washed with 0.01 M PB at pH 5.5. Approximately 15 ml of filtered egg yolksupernatant (EYS) was applied to the column, the column washed again with 45 mlof 0.01 M PB and the bound proteins eluted using a linear gradient for 60 min withincreasing phosphate buffer molarity from 0.01 to 0.2 M, followed by continued elutionfor 60 min with 0.2 M PB. After elution, the column was washed with 30 ml of 0.01 MPB/0.5M NaC1 solution followed by 60 ml of 0.01 M PB. The flow rate was maintainedat approximately 1.0 ml/min.In later experiments, 50 ml of CM-92 was packed into a 2.5 cm i.d. column and theflowrate for washing, elution and saltwash was increased to approximately 2 ml/minMethods of AnalysisSDS-PAGE using a Phast System (Parmacia, Upsalla, Sweden) was used to obtain aqualitative determination of the IgY content of each peak.Chapter 5. Separation of IgY^ 615.4.3 Results and DiscussionThe first attempt to bind IgY to a cation exchanger was carried out at pH 5.5 and witha washing buffer of 0.02 M PB pH 5.5. Unfortunately, even this low ionic strength bufferproved to be too strong to allow binding to occur. The washing buffer was then reducedin ionic strength to just 0.01 M PB, and an experiment carried out as described aboveusing a 35 ml column of CM-92. Figure 5.10 shows the chromatogram obtained from thefirst successful experiment involving the use of CM-92.As with the anion exchange experiments, a significant amount of protein is not boundand elutes with the mobile phase. The protein that is bound is held much more weaklythan in the case of anion exchange since the majority of the bound proteins eluted earlyin the 0.01 to 0.2 M PB elution gradient. SDS-PAGE of the eluted peak showed that itwas largely IgY, a result not entirely expected since the isoelectric points of the otherlivetin proteins are known to be in the same range as IgY. Thus this first experiment inthe use of a cation exchanger to separate IgY from other water soluble proteins in eggyolk supernatant appeared to be very encouraging.Figure 5.11 shows another cation exchange experiment carried out using Whatm.anCM-92 exchanger. In this case a 50 ml column was used and only 30 ml of EYS applied.This chromatogram includes the saltwash peak which indicates that under the conditionsused, certain proteins are bound quite strongly to the column and must be eluted withrelatively high ionic strength buffers (in this case 0.01 M PB/0.5 M NaC1).5.4.4 ConclusionsThe data obtained in this study indicate that cation exchange chromatography can po-tentially be used for the separation of egg yolk immunoglobulins. Under the conditionsutilized, the majority of IgY is initially bound to the column, but can easily be eluted_Chapter 5. Separation of IgY^ 62Figure 5.10: Chromatogram from the first CM-92 Experiment100 ^,..,_—_,..... ^0^30Sample Application Peak::.1. Elution../ ProfileIgY Rich Peak60^90^120^150^180Time (min)1 1tt0 50 1 00 1 50 250200100IgY Rich PeakElutionGradientBeginsSalt Wash PeakSaltwashBeginsSample ApplicationChapter 5. Separation of IgY^ 63Figure 5.11: Chromatogram from a later CM-92 ExperimentTime (min)Chapter 5. Separation of IgY^ 64with a relatively low ionic strength buffer.5.5 General ConclusionsBoth ultrafiltration and anion exchange chromatography failed to significantly improvethe IgY purity in the WSF. Of the 3 techniques tested, cation exchange chromatographyappears to hold the greatest promise for separating IgY from the WSF of supernatantprepared from industrially separated egg yolk. Under the conditions utilized in thisstudy, it would appear that all of the ovaibumin (the principal egg white contaminant ofindustrial yolk) and the majority of the other contaminating proteins have been removedfrom the purified fraction, with the exception of a portion of the a-livetin. Further studyof this technique with a view to incorporating it into a pilot-scale purification process iswarranted.Chapter 6Comparison of Two Cation Exchangers6.1 IntroductionBased on the conclusions of the work described in Chapter 5, the use of cation exchangechromatography for IgY separation was further investigated. Besides carrying out a morein depth study of the use of the CM-92 exchanger, an industrial grade exchanger (HC-2)was also tested and compared. It was decided that with a heavier sample loadingl theeffect of linear gradient could be compared to that of a single step elution profile for theelution of IgY from the column This would indicate whether or not IgY elutes as a singlepeak, and if so, if it would be possible to separate it more efficiently from a-livetin bycontrolling the shape of the elution profile.6.2 Materials and Methods6.2.1 Cation Exchange ChromatographyThe experimental equipment and buffers were as described in Section 5.4. The cation ex-change media used were the weakly acidic carboxymethyl (CM) cellulose cation exchang-ers CM-92 (Whatman BIosystems Inc., USA) and HC-2 (Gibco CEL, New Zealand).CM-92 is a fibrous exchanger, which is supposed to allow faster flow rates. HC-2 is an'The loading used in the preliminary CM-92 experiments was on the order of about 1 mg of IgYper ml of exchanger, of which about i bound. Considering that today's ion- exchangers are generallycapable of binding 50 to 100 mg of protein per ml, this is an fairly light loading, but sufficient for testingthe exchanger.65Chapter 6. Comparison of Two Cation Exchangers^ 66industrial grade exchanger and contains particles with a range of diameters from 150 to250 ym. It is based on a cellulose matrix.Each exchanger was packed into a 2.5 cm i.d. column to a total bed volume of50 ml. After equilibration with 3 to 4 bed volumes of equilibration buffer, approximately3 bed volumes ( 140 nil) of yolk supernatant prepared from industrially separated yolkaccording to the protocol recommended in Chapter 4 was applied to the column Aftera wash of 1 to 2 bed volumes, the bound proteins were eluted using either a lineargradient of PB with increasing molarity from 0.01 to 0.2 M, or a step change from 0.01to 0.2 M PB. A saltwash followed elution. The flow rate was maintained at 1.95 ml/minwith the exception of sample application where the flow rate was reduced to 0.93 ml/minFractions were collected for analysis every 7.5 min. during elution and salt washing, whilethe total volume was collected and measured after both sample application and washing,and representative samples kept for analysis. The collected samples were analyzed fortotal protein and IgY concentration, and the purity and recovery of each fraction wereestimated.6.2.2 Analytical ProceduresTotal protein determination was carried out using the BCA method, non-denaturingelectrophoresis (SDS-PAGE) with a Pharamacia Phast System using 10-15% gradientPhastGels and IgY concentration determination was by radial immunodiffusion. All ofthese techniques were applied as described previously.Chapter 6. Comparison of Two Cation Exchangers^ 676.3 Results and DiscussionFigures 6.12a and 6.12b are typical chromatograms obtained with the small columnpacked with CM-92 cation exchanger, using linear and step gradient elution, respec-tively. These experiments were repeated once for a total of 2 replicates. The resultingchromatograms were very repeatable. The first peak in each chromatogram occurs duringsample application and washing, which represents the majority of impurities and about30% of the IgY not initially bound to the column. The majority of the remaining IgY issplit into two peaks using a linear gradient, while most is eluted in a single peak usingstep-wise elution.The separation of the IgY rich peak in Figure 6.12a did not improve significantlyeither recovery or purity as the IgY concentration was high in both portions of the peak.Therefore, similar recoveries (about 60%) and purities (67-69%) were obtained for thetotal IgY peak using either linear or step-wise gradient.Figure 6.13 shows the resulting elution chromatogram when a double step elutionprofile (steps of 0.1 M PB and 0.2 M PB) is used. In this case, despite the fact thattwo well separated peaks result, the IgY purity of each one is very similar. This alsodemonstrates the existence of different IgY subclasses, as previously noted. This tech-nique might be useful if the separation of IgY into its subclasses is required. However,from the point of view of purifying the whole IgY population , there does not appear tobe any great advantage in using linear or multi-step gradients compared to a single stepelution profile. In all three cases, the saltwash peak that occurs at the end of the runcontained very little IgY.Figures 6.14 and 6.15 are typical chromatograms obtained using a colurrm packedwith HC-2 and using linear and step-wise elution profiles, respectively. Each experimentChapter 6. Comparison of Two Cation Exchangers^ 68Figure 6.12: Separation of IgY on a CM-92 Column using Linear (a) and Step (b) ElutionProfiles1 00Sample^ aApplicationPeakElutedPeaksSaltwashPeak40 -20 -^Or^0^60^120^180^240^300^360^420Time (min)10080608060 -SampleApplicationPeakElutedPeakb40 -20 -SaltwashPeak00^60^120^180^240^300^360^420Time (min)Chapter 6. Comparison of Two Cation Exchangers^ 69Figure 6.13: Separation of IgY on a CM-92 Column using Double Step Elution Profile.The sample application peak has been omitted for clarity.100 -ElutionProfile-806040SaltwashPeak20 -0^r—^1240^300 360^420Time (min)Chapter 6. Comparison of Two Cation Exchangers^ 70Table 6.7: IgY Recovery and Mass Balance using Linear Gradient and Step-wise Elutionon a Column of HC-2 Cation Exchanger for the Representative Experiments shown inFigures 6.14 and 6.15.Linear gradient Step-wise gradientStep IgY IgY IgY IgY(mg) (%) (mg) (%)Sample application 22.5 22.2 20.1 20.3Washing 8.8 8.7 9.8 9.9Elution 64.8 63.9 67.4 68.0Salt washing 6.7 6.6 3.6 3.6Total 102.8 101.4 100.9 101.7Total applied 101.4 100.0 99.1 100.0was repeated twice for a total of 3 replicates In addition to showing the absorbance at280 nm as a solid line, and the elution profile as a dashed line, these graphs also includehistograms indicating the IgY concentration of individual fractions collected during thecourse of the experiments. The fractions with the highest IgY concentration occur in theIgY rich peak in both cases. Similar IgY concentration profiles were obtained for fractionsfrom the two CM-92 experiments of Figures 6.12a and 6.12b. In the case of step elution,the HC-2 and CM-92 chromatograms are almost identical. In linear gradient elution, theHC-2 chromatogram contains only one major eluted peak instead of the two occurringwith CM-92. The recoveries and mass balance calculations for HC-2 experiments areshown in Table 6.7. The close agreement in the mass balances obtained by comparingthe sum of individual fractions to the total applied shows that all of the IgY is beingaccounted for and is not denatured throughout the course of the separation.SDS-PAGE of a representative IgY rich peak is shown in Figure 6.16. Column Ais the WSF from industrially separated yolk. Column B, the sample application andChapter 6. Comparison of Two Cation Exchangers^ 71Figure 6.14: Separation of IgY on an HC-2 Column using a Linear Elution ProfileIgY Concentration= Sample100 —^application=I Washing1^I Elution andsalt washing- 2.00Elution profile - 1.7580 - Absorbanceat 280 nm- 1.50/60 -r \tui- 1.25 Eo0....,oi- 1.00^$_,q.)o0o- 0.75 0>-.ela-40 --I - 0.5020 -_^- 0.25rev,,f^A ,\N I 0.000^60^120 180 240 300 360 420Time (min)Chapter 6. Comparison of Two Cation Exchangers^ 72Figure 6.15: Separation of IgY on an HC-2 Column using a Step Elution ProfileIgY ConcentrationE2j SampleapplicationWashing100 - - 2.00Elution andsalt washingI- 1.75Absorbanceat 280 nmElutionProfile- 1.5080 -60 -40 -- 1.25- 1.004a1a)c.)- 0.75>-■- 0.5020 -0, 0.0060^120 180 240 300 360 420Time (min)II- 0.25Chapter 6. Comparison of Two Cation Exchangers^ 73washing peak, contains most of the 0-livetin and smaller proteins. Column C is theIgY rich peak showing the strong IgY band which makes up almost of the fraction.The major contaminant, a-livetin, is also quite evident along with some lower molecularweight proteins. Column D, the saltwash peak, contains a small amount of IgY and therest of the a-livetin and other proteins.The IgY purity was also determined by dividing the IgY concentration by the totalprotein for each fraction. Figure 6.17 shows the resulting histograms for fractions elutedfrom the HC-2 column using linear gradient (Figure 6.17a) and step-wise (Figure 6.17b)elution profiles. Slightly higher purities were obtained for the corresponding CM-92experiments. What is particularly interesting in these graphs is the fact that the fractionsof highest purity are also the fractions of greatest total protein concentration. This isvery helpful when sacrificing recovery in order to obtain higher purity.A graph comparing the maximum purity obtainable as a function of recovery is shownin Figure 6.18. The data are based on the fractions collected from the CM-92 and HC-2columns operating under identical conditions and using linear elution gradients. Thecurves are obtained by starting in each case with the fraction of greatest purity, andadding fractions of successively lower purity. Overall, the use of the CM-92 exchangerresults in a somewhat higher purity for the same recovery. However, the price of HC-2is considerably lower than that of CM-92 ($60 vs. $250 US/kg), and a large amount (10kg) of HC-2 exchanger was already available. As well, the pressure drop across the HC-2column was found to be quite low, even at relatively high flow rates. It was decided,therefore, to carry out further work with HC-2.Figure 6.19 shows a similar plot of purity as a function of recovery for the lineargradient and step elution profile experiments with the HC-2 column. There appears tobe little or no difference between results obtained with the different elution profiles. ItChapter 6. Comparison of Two Cation Exchangers^ 74Figure 6.16: Non-Denaturing SDS-PAGE of Fractions from HC-2 Experiments. The lanesrepresent A - WSF prepared from industrially separated yolk; B - Unbound fraction; C- Eluted protein - IgY rich fraction; D - Saltwash peak.- 80- 60- 40- 20Chapter 6. Comparison of Two Cation Exchangers^ 75Figure 6.17: Purity of Individual Fractions Eluted from an HC-2 Column100'100a- 80— Absorbance at 280 nm80 -60 -40I.^1 IgY Purity-- 60- 4020 -..../".........-.,.......o ^100 -b80 -60 -40- 20^ o- 1000240 280 320 360 400Time (min)0 CM-92V HC-2VChapter 6. Comparison of Two Cation Exchangers^ 76Figure 6.18: Purity as a Function of Recovery using Linear Gradient Elution9080 -70;.460 -1^i^130^40^50^60Recovery (%)5020 70Chapter 6. Comparison of Two Cation Exchangers^ 77is much simpler to use a step-wise elution profile in practice.6.4 ConclusionsIt does not appear that the shape of the elution profile has any significant effect on theoverall IgY recovery and purity that can be achieved using cation exchange chromatog-raphy. From analysis of individual fractions in the IgY rich peak, it was determined thatIgY is not separated from a-livetin, but appears to elute over the entire peak.Of the two exchangers tested, Whatman CM-92 consistently bound a larger amountof IgY per ml of exchanger. However, considering the exchanger cost, the difference inbinding ability does not appear to be significant enough to warrant its use on a large-scalein place of the industrial cation exchanger HC-2.Chapter 6. Comparison of Two Cation Exchangers^ 78Figure 6.19: Purity as a Function of Recovery for HC-2 Experiments905010^20^30^40^50^60^70^80Recovery (%)Chapter 7HPLC Analysis of IgY7.1 IntroductionNumerous authors have made use of gel filtration media either as a means of identifyingor purifying chicken antibodies. Gel filtration has been used to purify chicken serumIgY, and similar methods have been used to fractionate chicken yolk IgY. A summary ofliterature references pertaining to the use of gel filtration in the study of chicken serum oregg yolk IgY is given in Table 7.8. This list is by no means comprehensive, but is a goodselection of many papers that mention gel filtration. In almost all of these papers, gelfiltration was used either to fractionate IgY from other water soluble proteins, or as finalpolishing step in a purification scheme. In three of the more recent papers (Yamamotoet al., 1975; Hassl et al., 1987; and Shirman et al., 1988) it was used to quantify the IgYpresent in solution. Van Orden and Trefrers (1968) used gel filtration on Biogel P-200 todetermine the molecular weight of serum IgY. Cser et al. (1982) used gel filtration onSephadex G-200 as a final polishing step to remove IgY aggregates from solution priorto X-ray analysis of the monomer. The reported methods, however, have all made use oflow pressure gel filtration columns packed with media such as Sephadex G-100 or G-200,Sephacryl S-200 or S-300 Superfine, or Bio-Gel P-200. No-one, as yet, has reported theuse of an HPLC gel filtration column to either purify or quantify IgY. One paper only hasbeen encountered, that of Burley and Back (1987), which refers to the use of an HPLCcolumn in the study of the livetin fraction of egg yolk. The column used was an Ultropac79Chapter 7. HPLC Analysis of IgY^ 80Table 7.8: References in the literature to gel filtration of chicken antibodiesReference Year Media SourceFlodin and Killander 1962 Sephadex G-200 serumDreesman and Benedict 1965 Sephadex G-200, G-50 serumHersh and Benedict 1966 Sephadex G-200 serumTenenhouse and Deutsch 1966 Sephadex G-200, G-100 serumVan Orden and Treffers 1968 Biogel P-200 serumLeslie and Clem 1969 Sephadex G-200 serumWilkinson and French 1969 Sephadex G-200, G-75 yolkLeslie and Clem 1970 Sephadex G-200 serumOrlans and Rose 1972 Sephadex G-200 yolkRose et al. 1974 Sephadex G-200 yolkYamamoto et al. 1975 Sephadex G-200 yolkHiggins 1976 Sephadex G-200 yolkBurley and Vadehra 1979 Ultragel AcA 34 yolkCser et al. 1982 Sephadex G-200 yolkWang et al. 1986 Fractogel TSK HW-55 (F) yolk77 Bio-Gel A-1.5mHassl et al. 1987 Sephacryl 5-300 yolkHassl and AspOck 1988 Sephacryl S-300 Superfine yolkShirman et al. 1988 Sephacryl S-300 yolkAkita and Nakai 1992 Sephacryl S-200 Superfine yolkDEAE-5PW, 10 eum column (7.5 mm x 7.5 cm).Since the goal of this work was to study IgY purification with a view to developing aprocess viable for industrial scale production of the antibody, HPLC gel filtration was notconsidered as part of the separation process. Even today's so-called large-scale HPLC isstill far from being large-scale when it comes to producing significant quantities of productfor food use. However, HPLC appeared to be an interesting method to consider foranalyzing and quantifying purified samples of IgY, and potentially for on-line monitoringof a separation process.Chapter 7. HPLC Analysis of IgY^ 81One of the drawbacks of the RID method used for IgY concentration determinationis the delay of 3 days before obtaining the results of the analysis. As well, RID is subjectto significant errors due to errors in pipetting, distortion of precipitin rings due to edgeeffects and uneven gel thickness, and errors in reading the precipitin ring diameters.HPLC analysis, on the other hand, allows for a much more rapid analysis time usinga larger sample volume (20 ftl applied by HPLC syringe compared to 4 id for RID),decreasing the influence of pipetting errors. In this work, it was carried out using analmost fully automated system with computerized data acquisition, and once equlibratedwith running buffer, the operator need only inject the sample and turn a valve in orderto start the analysis. In other words, the equipment is easy to operate, ideal for qualitycontrol situations where technicians with minimal training on HPLC equipment coulduse the technique with confidence.7.2 Materials and MethodsA 30 cm long TSK-G4000SW gel filtration HPLC column (Tosoh Corporation, Japan)was used with 0.1 M PB (pH 5.4 with 0.05% sodium azide as preservative) as runningbuffer. The column was connected to a Hewlett Packard HPLC system including a Series1050 solvent cabinet with injection valve, quaternary pump and multiple wavelengthdetector. A Hewlett Packard ChemStation running HPLC software was used to acquireand analyze the chromatograms. The running buffer was degassed with helium prior toand during all runs, and samples were filtered prior to injection with a 13 mm diameter0.45 Am (cellulose acetate membrane) disposable sterile syringe filter (Corning, NY). Aschematic diagram of the HPLC system is presented in Figure 7.20.The multiple wavelength detector has a diode array type light source that allowsthe user to program the wavelength and bandwidth of the source, as well as a referenceChapter 7. HPLC Analysis of IgY^ 82Figure 7.20: HPLC System LayoutChapter 7. HPLC Analysis of IgY^ 83wavelength and bandwidth. The wavelength used for all HPLC analysis was 280 nmand a narrow bandwidth of 4 nm, with a reference wavelength of 450 nm and 80 nmbandwidth.In the case of purity determination, a flowrate of 0.25 ml/min was used resulting in60 minute runs. For molecular weight and concentration determination, a flowrate of0.5 ml/min was used resulting in 30 minute runs.7.3 Results and Discussion7.3.1 Molecular Weight DeterminationGel filtration has been used to determine the molecular weight of IgY. Using SephacrylS-300 Superfine, Hassl and Aspock (1987) obtained a value of 175 kDa, and Shirman etal. (1988) a value of 160 kDa.In order to obtain an estimate of the molecular weight of the IgY being purified inthe separation process, proteins of various molecular weights were applied to the TSK-G4000SW column under the same conditions to be used for studying the purified samples.A total of 8 different protein standards were used, plus the dimers of two of the proteins,in order to obtain a curve relating the molecular weight of each protein to its retentiontime on the column A summary of the molecular weights and corresponding retentiontimes of the protein standards is given in Table 7.9. Sample chromatograms of some ofthe standards can be found in Appendix B. Figure 7.21a illustrates that as expected, anexponential relationship exists between the molecular weight of a protein and its retentiontime. Each point in the figure represents a single determination of the retention time fora purified sample. However, some of the samples were retested and the retention timewas always within 0.5% of the previously determined value. Figure 7.21b is a plot ofChapter 7. HPLC Analysis of IgY^ 84Table 7.9: Summary of Molecular Weight InformationData Protein Retention Molecular ReferencePoint (mm) Weight(kDa)1 Phosphorylase-a 17.37 370,000 a2 Bovine IgG dimer 17.45 300,000 b3 Bovine IgG monomer 19.32 150,000 b4 BSA dimer 18.95 132,534 b5 Conalbumin 20.56 76,000 c6 BSA monomer 20.37 66,267 b7 Ovalbumin 21.14 44,500 c8 P-lactoglobulin 21.37 35,000 a9 Carbonic anhydrase 22.19 30,000 a10 Lysozyme 23.77 14,300 cReference^ Sourcea^Fasman, G.D., ed. 1989. CRC Practical Handbook ofBiochemistry and Molecular Biology. CRC Press Inc.,Boca Raton, Florida.b^Fox, P.F., ed. 1989. Developments in Dairy Chemistry 4.Elsevier Science Publishers Ltd, London, UK.C^Powrie, W. and S. Nakai. 1985. In: Food Chemistry, 2nded., Fennema, O.R., ed. Marcel Dekker, Inc., NewYork, NY, pp 829-855.Chapter 7. HPLC Analysis of IgY^ 85the natural logarithm of the molecular weight vs. retention time, and demonstrates thata strong relationship (R2 = 0.991) results when a linear regression is performed on thetransformed data.Using the resulting regression equation and the retention time for IgY monomer ob-tained from numerous experiments (approximately 18.87 minutes on average), the molec-ular weight of the purified IgY is determined to be in the vicinity of 156 kDa. This com-pares with the value of approximately 175 kDa obtained previously using electrophoresis.Bovine IgY, known to have a molecular weight of 150 kDa, is estimated by this proce-dure to be only 124 kDa. Thus it is clear that the error involved in this estimation issignificant, and this is further supported by the range of values obtained for the molec-ular weight of IgY by other researchers' The estimation obtained nonetheless providesone means of verification when identifying the IgY peak in purified samples. The equa-tion obtained also allows for the estimation of the molecular weights of impurities andantibody aggregates in these samples.7.3.2 IgY Concentration DeterminationThe second use to which the gel column was put involved the determination of the IgYconcentration of partially and highly purified samples. Unfortunately, proteins in themolecular weight range 30,000 to about 100,000 Daltons did not separate well from oneanother under various conditions of pH and PB concentration (pH 5 to 7, 0.05 to 0.2 MPB, 0 to 140 mM NaC1). Since there are at least three major impurities present in thelOne explanation for the large difference in estimated values is that gel filtreation separates on thebasis of molecular size rather than molecular weight (hence the more correct, but not yet universallyused term "size exclusion chromatography"). Although there is a direct relationship between these twocharacteristics of a molecule, the correlation is not perfect since some molecules have a more tightlypacked structure than others, and this can also be affected by the molecule's micro-environment. Sinceit is known that the radius of gyration of IgY is somewhat smaller than expected based on it's molecularweight (Cser et al., 1982), it is not surprising the HPLC estimate is lower than expected.Chapter 7. HPLC Analysis of IgY^ 86Figure 7.21: Relationship Between Molecular Weight and Retention Time400 a300 -200 -100 -01012 -2R = 0.99111 -yi= -0.508x0-21.54310 -915^20^ 25Retention Time (min)Chapter 7. HPLC Analysis of IgY^ 87supernatant obtained from industrially separated yolk (a- and /3-livetins, and ovalbumin)that fall within this range (80,000, 48,000 and 44,500 respectively), a very large peak witha retention time of around 21.0 minutes was obtained from supernatant samples appliedto the gel filtration column. Since the retention time of ovalbumin was determined aboveto be 21.14 minutes, and since electrophoresis indicates the presence in yolk supernatantof a large proportion of ovalbumin, the large peak is apparently predominantly due tothis protein. In samples of supernatant, this large impurity peak often distorts theIgY monomer peak (causing an increase in its calculated retention time) and seems tocause a reduction in the expected area of the IgY peak, perhaps due to matrix effects.Consequently, any attempt to use the TSK-G4000SW gel filtration to develop an IgYstandard curve for concentration determination in impure samples appeared futile.However, in relatively pure samples (IgY purities around 50% or higher), the effectsdiscussed above do not appear to be significant. The only anomaly noted is the presenceof what was at first assumed to be a dimer peak, present in both pure IgY standard(minimum 98% IgY according to Sigma) and the purified IgY. It is a well recognizedfact that IgY has a tendency to form aggregates of 2, 3 or more monomer units2 Ifthe standard obtained from Sigma is as pure as the company claims, then this highermolecular weight peak must be primarily IgY.Figure 7.22 shows the chromatograms obtained when samples of the standard (7.22a)2llassl et al. (1987) used gel filtration analysis to estimate the IgY concentration partially purifiedsamples. Using a 60 x 1.6 cm column operating at room temperature and PBS as the running buffer,k, the extinction coefficient, was measured at 280 nm and the IgY content of samples determined byintegration of peak areas. They compared IgY purification using 3 methods: precipitation with propanolas described by Bade and Stegemann (1984); precipitation with polyethylene glycol (PEG 6000) followingthe method of Poison et al. (1980); and an improved PEG method developed by Poison et al. (1985).In their gel filtration analysis of the IgY content of the purified fractions, significant amounts of lowermolecular weight impurities (in the weight range of a-liveltin) were present in purified fractions for thefirst two methods, while a small amount of higher molecular weight impurity (a shoulder peak) waspresent in purified fractions using the third method. They suggested that this "shoulder" peak mayrepresent IgY aggregates.Chapter 7. HPLC Analysis of IgY^ 88and purified IgY (7.22h) are analyzed. Using the regression equation obtained above todetermine the molecular weight of the smaller "shoulder" ahead of the monomer peak,the molecular weight appears to be about 440 kDa for the IgY standard "shoulder" and500 kDa for the IgY sample "shoulder". Thus it is apparent that the peak formerlythought to be a dimer is actually more likely a trimer of IgY. In order to integratethe area for concentration determination, it was decided to integrate both peaks in allanalyses.A stock solution of chicken IgY standard (Sigma Chemical Co., St. Louis) at concen-tration of 2.0 mg/ml was prepared in 0.1 M PB, pH 5.4, 0.05% NaN3 and then passedthrough a 0.45 pm syringe filter. A series of standards were prepared and applied tothe HPLC column at a flowrate of 0.5 ml/min. The known concentration was plottedagainst the area obtained on the chromatogram for each standard in Figure 7.23a, andthe resulting regression equation found to be:C = 0.00083(Area) — 0.03574where C is the IgY concentration in mg/ml and Area is the area of the IgY peak onthe chromatogram. A very high regression coefficient (R2 = 0.9993) was obtained. Thisequation was then used to determine IgY concentration in partially purified samples.Since the standard method that had been used thus far in the work for IgY concen-tration determination was RID, a comparison was made of the two methods. A sampleof purified IgY with a concentration in the vicinity of 1.2 mg/nil was diluted to obtain5 "unknowns", which were then applied to both HPLC and RID using the chicken IgYstandards prepared earlier. As well, a concentrated sample of partially purified IgY,eluted after the loading of a 15 cm HC-2 column with about 40 bed volumes of su-pernatant, was diluted 10x and both methods applied. Figure 7.23b demonstrates the42000^5^10^15^20^25^301aChapter 7. HPLC Analysis of IgY^ 89Figure 7.22: HPLC Analysis of Pure IgY Samples. The graphs show the chromatogramobtained for a sample of IgY standard (a) and a sample prepared by purification withNa2SO4 precipitation after extraction as described in Chapter 4, and separation as de-scribed in Chapter 6 (b).Retention Time (min)I^1^I^10.2^0.4^0.6^0.8IgY Concentration from HPLC (mg/ml)0.00.0 1.0Chapter 7. HPLC Analysis of IgY^ 90Figure 7.23: HPLC Standard Curve (a) and Comparison to RID (b)^1.2 ^1.0 -0.8 -0.6 -0.4 -0.2 -0.0 ^0^200 400 600 800 1000 1200 1400Area on HPLC ChromatogramChapter 7. HPLC Analysis of IgY^ 91resulting correlation between RID and HPLC results. Note that one of the unknownswas omitted in determining the regression due to the erroneously low RID result, leavingthe four unknowns and the 10x diluted, partially purified IgY sample.It is not clear from Figure 7.23b, but the variance in the results obtained by RIDanalysis is much greater than that for the HPLC results. Table 7.10 presents the concen-trations determined by each method for each of the five "unknowns" tested (excludingthe partially purified sample), as well as their dilution factors, calculated concentrationof the original solution, and statistical parameters. From this data one can see that thecoefficient of variance for RID was 4 times that of the HPLC technique. If one rejectscompletely the 3 times diluted sample in the RID portion, then the average estimate forthe concentration of the original solution becomes identical to that estimated by HPLC,but with a coefficient of variance still more than twice that of the HPLC determined value(CV of 6.1% vs 2.7%). Numerous other analyses of IgY samples carried out with HPLChave shown a similar superior reproducibility in the results. Clearly then, for purifiedsamples of IgY the HPLC method developed here is more rapid and reliable than RID.7.3.3 IgY Purity DeterminationA third potential use for HPLC gel filtration is in the determination of IgY purity. Theprimary method used previously involved RID. For solutions with relatively low antibodyconcentration, the purity is calculated by first determining the IgY concentration by RID,and then dividing this value by the total protein (TP) concentration of the sample, asdetermined by the BCA method. Using HPLC, the purity can be estimated by integratingthe area of the IgY monomer and trimer peaks and comparing to the integrated area forthe entire chromatogram. Although an error in the estimation will occur due to theinherent difference in extinction coefficient at 280 nm from protein to protein, this errorChapter 7. HPLC Analysis of IgY92Table 7.10: Comparison of HPLC and RID ResultsHPLC ResultsHPLC(mg/ml)DilutionFactorConcentration(mg/ml)0.867 1.33 1.150.777 1.5 1.170.583 2 1.170.398 3 1.190.277 4 1.11x= 1.157s = 0.031CV(%) . 2.71RID Results including erroneous 3x dilution valueRID(mg/ml)DilutionFactorConcentration(mg/mi.)0.941 1.33 1.250.779 1.5 1.170.556 2 1.110.300 3 0.900.274 4 1.10x= 1.106s = 0.130CV(%) . 11.76RID Results excluding erroneous 3x dilution valueRID^Dilution^Concentration(mg/ml)^Factor (mg/ml)0.941 1.33 1.250.779 1.5 1.170.556 2 1.110.274 4 1.101.1570.0706.07Chapter 7. HPLC Analysis of IgY^ 93will be greatly reduced as the sample is purified to a greater degree. Even for sampleswith relatively low IgY purity the method can provide a reliable estimate for purposesof comparison.A flowrate of 0.25 ml/min was found to provide the best resolution of the peaks.Figure 7.24 shows the resulting chromatograms and Figure 7.25 the electrophoretogramobtained when samples containing IgY of 3 very different purities were applied to theHPLC column Figure 7.24a is the chromatogram for a sample of egg yolk supernatantwith IgY monomer eluting at 37.94 mm The integrated areas of the individual peaks areshown in Table 7.11. Based on the integrated areas, this sample is approximately 19%IgY (combining monomer and trimer peaks). This actually agrees well with the purityof IgY determined by the RID/TP technique, generally 20-22% for yolk supernatant.A significant error may occur in the integration of the peaks. In the case of samplesin the range of purity of 50 - 90 %, the size of peaks relative to baseline shift is generallylarge enough to keep this error fairly small. However, for very pure samples (especially>95%), it was necessary to set fixed limits for integration.Figure 7.24b represents a sample of the IgY rich fraction eluted from an HC-2 cationexchange column The purity as determined by HPLC is roughly 63%, which againcompares favourably to the RID/TP estimate of 62%.Finally, Figure 7.24c shows the chromatogram of a highly purified sample of IgY,following salt precipitation. It is in this case that the HPLC technique is of especialvalue, since the purity cannot be accurately estimated using RID/TP (even a 2% errorin either IgY concentration or total protein can result in meaningless values for purity).The purity of this sample as estimated by HPLC is greater than 98%. Thus, the methoddeveloped provides an accurate means of determining IgY purity, especially for very puresamples.60 -40 -20 -80 -60 -40 -20 -4-;Chapter 7. HPLC Analysis of IgY^ 94Figure 7.24: Analysis Carried out at a Flowrate of 0.25 ml/min of Samples with ThreeDifferent IgY Purities: (a) - filtered yolk supernatant (WSF); (b) - the IgY rich fractionfrom cation exchange chromatography; (c) - IgY purified by precipitation with sodiumsulphate precipitation.1 00 ^80 - a60 -40 -"C.180 -I^I^I^i^10^10^20^30^40^50^60Retention Time (min)20-0 ^COC\20Chapter 7. HPLC Analysis of IgY^ 95Figure 7.25: Electrophoresis of Samples with Three Different IgY Purities: A - the WSF;B - the IgY rich fraction from CEC; C - IgY purified by precipitation with sodiumsulphate.ABCDirection ofbandmigration4111.111AIgYChapter 7. HPLC Analysis of IgY^ 96Table 7.11: Summary of HPLC Results for Figure 7.24Sample: Egg Yolk SupernatantPeakNumberRetentionTimeArea %1 19.38 2.942 33.38 1.123 37.94 17.344 42.33 63.905 47.86 4.036 49.98 3.917 53.82 6.76Sample: Eluate from Cation Exchange StepPeak^Retention Area %Number Time1 33.34 10.642 37.91 52.173 41.31 29.914 43.37 6.865 51.97 0.43Sample: Purified IgY after salt precipitationPeak^Retention Area %Number Time1 33.29 11.252 37.88 83.453 50.58 5.29Chapter 7. HPLC Analysis of IgY^ 977.4 ConclusionsHPLC gel filtration is a convenient method for rapidly determining IgY concentrationor purity in partially or very pure samples. Although analysis time for concentrationand purity determination was 30 and 60 minutes respectively, this time could be cutin half for pure samples without significant loss of accuracy. This technique could beespecially useful where rapid analysis of a small number of samples is required, such asin monitoring a purification process.Chapter 8Purification of IgY8.1 IntroductionOf the techniques used to obtain very pure preparations of IgY, the most successfulwould appear to be gel filtration, salt precipitation and anion exchange chromatography.Gel filtration, although very effective in producing small quantities of very pure IgY, isat present generally considered impractical for large-scale purification of protein. Sincethe purification of IgY, which is the third step in the process being presented here,involves a smaller volume of solution of 60-70% purity, the use of a labour intensive batchtechnique such as salt precipitation was considered an economically feasible possibilityfor obtaining IgY of high purity for applications that would demand it. Likewise, anionexchange chromatography, having been used in previous publications as a polishing step,was considered a potentially feasible final purification step.In addition to these two methods, and despite the failure to obtain good separationof IgY at the second stage of the process, ultrafiltration was considered as a potentialmethod for purifying the IgY rich eluate from the cation exchange step. Thus, each ofthese methods was tested to further purify the eluate.98Chapter 8. Purification of IgY^ 998.2 Materials and Methods8.2.1 Anion Exchange ChromatographyA 50 ml column (2.5 cm diameter) was packed with HA-2 (Gibco CEL, New Zealand)anion exchanger Like HC-2, HA-2 possesses excellent flow properties and is based on thesame cellulose matrix. The column was calibrated with 0.2 M PB at pH 8.0 accordingto Hatta et al., (1990) and washed with 0.01 M PB at the same pH. A sample of 85 mlof the IgY rich peak collected in the first chromatographic step was adjusted to pH 8.0using NaOH and applied to the column After washing extensively with the 0.01 M PB(7 bed volumes), the IgY adsorbed was eluted with 0.2 M PB/pH 8.0 using step gradient.After elution, the column was washed with 0.01M PB/0.5 M NaC1 solution followed by0.01 M PB. The superficial velocity was maintained at 36 cm/h (flowrate of 3 ml/min)with the exception of sample application where the superficial velocity was reduced tohalf.8.2.2 Salt PrecipitationThe IgY rich peak fraction obtained from the cation exchange chromatography run wasused for the salting out experiment according to the method described by Hatta et al.,(1990). Aliquots were mixed with sodium sulfate at a concentration of 15% (w/v) and20°C for 30 min After 15 mm centrifugation at 10,000 x g, superanatants were discardedand precipitates dissolved in PB (0.1 M, pH 8.0). This salting-out procedure was repeatedtwice. Two different starting pH's were investigated: pH 5.0 (no adjustment) and pH 8.0(adjustment with NaOH).Chapter 8. Purification of IgY^ 1008.2.3 UltrafiltrationThe same basic procedure as described in Chapter 5 was used, except that the pH of thesample eluate was not adjusted to Methods of AnalysisRID and SDS-PAGE were as described in Chapter 4. HPLC analysis was carried outaccording to the method developed in Chapter 7.8.3 Results and Discussion8.3.1 Anion Exchange ChromatographyThe 60% pure IgY solution from the first chromatographic step was applied to a 50 mlcolumn packed with HA-2 anion exchanger. Only a slight improvement in purity from60% to 66% was achieved and SDS-PAGE revealed no major difference between thesample obtained after HC-2 chromatography and the one obtained after the additionalstep of HA-2 chromatography. Hatta et al., (1990) used DEAE-Sephacel chromatographyafter lipoprotein separation resulting in an increase in purity from 19% to 46%, while theoverall recovery dropped from 86% to 76%. It seems that anion exchange chromatographywould be more useful in the initial stages of purification. However, DEAE-Sephacel wasalso investigated as a final purification step by Akita and Nakai (1992) where the puritywas improved from 93% to 99%.HA-2 anion exchanger is much less expensive than DEAE-Sephacel and has superiorflow properties. It is therefore generally more practical for large-scale purification pro-cesses. With the minor improvement in purity obtained here, however, it does not appearto be worth considering as part of the process being developed.Chapter 8. Purification of IgY^ 101Table 8.12: Purities obtained in Salt Precipitation and Determined by HPLC AnalysisSalt precipitation^Purityt^PurityIstep 1step 2step 394% 96%98% 98%99% 99%t starting pH = 5.01 starting pH = Salt PrecipitationSalt precipitation was considered in an attempt to produce high purity IgY which mightbe needed for specific uses. Samples purified after one, two, and three steps of saltingout were analyzed using RID, HPLC and electrophoresis. All the methods agreed onthe high purity of the samples as shown in Table 8.12. Results obtained by HPLC weremore consistent but slightly lower than RID data. Salt precipitation improved the IgYpurity significantly from 60% to 94% and 96% at a starting pH of 5 and 8, respectively.A recovery of 99% is possible if care is taken in decanting the supernatant. Additionalsalt precipitation steps increased the purity to approximately 99%. A sample obtainedfrom a batch separation (to be discussed in Chapter 10) was also used for the salting outexperiment, and the purity after one step was 95%.With heavier loading of the cation exchange column in the second stage of the process,IgY can be concentrated in the eluate. This in turn results in lower volumes to beprocessed by salt precipitation, which makes this technique appear quite feasible.Chapter 8. Purification of IgY^ 1028.3.3 UltrafiltrationHPLC analysis of ultrafiltered eluate showed no significant improvement in purity wasobtained under the conditions used. A slight reduction in intensity of the two small peaks(see Figure 7.24) representing smaller molecular weight proteins occurred, but the ratioof IgY to a-livetin remained essentially unchanged. Thus under the conditions tested, a-livetin displays no tendency to pass through the membrane of the filter.Further work would be necessary to determine whether or not ultrafiltration of theelutate from cation exchange chromatography could result in increased IgY purity. Itmay be that the lipid concentration in the eluate is still sufficiently high enough to causefouling of the membrane. Also, some type of complex formation between a-livetin andsome other protein, or perhaps even aggregation of a-livetin itself, may be preventing itfrom passing through the membrane.8.4 ConclusionsUltrafiltration and anion exchange chromatography can be used to improve the purity ofthe eluate obtained after the second step of the process only marginally. Salt precipita-tion using sodium sulphate, however, can improve IgY purity to around 95% after onlyone step, and is therefore a more appropriate technique to use as the third step in thepurification protocol.Chapter 9Breakthrough Curve Analysis9.1 IntroductionMuch information useful for predicting large scale operation of a chromatographic columncan be obtained by running small scale experiments (Gosling, 1987). Various mathemat-ical models have been dereloped to take advantage of small-scale experimental data forpredicting large-scale operation, but these models are often verified with ideal protein orsugar solutions involving molecular species with well defined iso-electric points and molec-ular weights. A somewhat less elegant, but nevertheless practical approach is to carryout small-scale experiments under conditions similar to those expected for the large-scaleoperation, and to vary key parameters in order to study their influence on the separa-tion. In chromatographic processes in general, a doubling of column length results in adoubling of the pressure drop across the column but an improvement in the resolutionbetween peaks by a factor of only .4. Increasing column diameter, on the other hand,generally results in a broadening of the peaks'. Thus the scale up of even a simple binaryseparation on a chromatography column can be quite complicated to predict, and as aresult, in practical applications it is often easier to maintain a set bed height, increasecolumn diameter to allow for increased capacity, and then connect several other "scaledup" columns in series in order to counteract peak broadening effects.'This can be a complex function of a number of parameters involving flowrate, particle geometry, thepresence of "short circuiting" channels in the column and other factors.103Chapter 9. Breakthrough Curve Analysis^ 104Adsorption chromatography involves a very different type of separation process. Inadsorption chromatography, the desired molecule is usually preferentially bound to thecolumn, while undesireable contaminants have little or no affinity for the column andpass through with the mobile phase. Whereas in standard chromatography only a smallsection of the upper portion of the column is used to bind the sample, in adsorptionchromatography the majority of the column capacity can be used. Once the column issaturated, a step change in ionic strength or pH is used to elute the adsorbed species.Although there are many published reports involving mathematical modelling of theadsorption of one species of protein molecule to different adsorbents, few have reportedthe more realistic situation of multicomponent adsorption involving more than one pro-tein (Skidmore and Chase, 1990). In the case of the binding of IgY from the WSF toa cation exchanger (HC-2), the heterogeneity in the pI of the antibody, as well as theoverlapping pI range of a-livetin results in a much more complex binding pattern thanwould be predicted by even the more complicated models involving adsorption from bi-nary protein mixtures. Figure 9.26 illustrates graphically the expected binding patternsfor IgY and a-livetin compared to standard chromatography with homogeneous molecu-lar species. It is proposed that the strongly binding subclasses of each protein bind nearthe top of the column and gradually displace the weakly binding subclasses towards thebottom. Since the pI's of the two proteins overlap, it is likely that their subclasses areintermingled on the column during adsorption.In adsorption chromatography, the simplest approach to modelling for purposes ofscale-up is the analysis of breakthrough curves for the species to be purified. The shapeof the breakthrough curve (Vermeulen and Hiester, 1959) provides information aboutthe strength with which a protein binds to the exchanger (ie.: equilibrium and rate ofbinding), as well as the capacity the exchanger has to bind the protein of interest (ie.:EMIAdsorptionChromatographyof 2 componentswith mit'subclassesA+B^B3B2A3A2BIAl# 111111111111Chapter 9. Breakthrough Curve Analysis^ 105Figure 9.26: Schematic Representation of Protein Binding to Adsorption ColumnTime - 0^Time > 0^Breakthrough Curves4StondcrdChromatographyA+B+CCB111111111111111A4 Tine4IdealAdsorptionChromatographyIIIEN A4 Time4^TimeChapter 9. Breakthrough Curve Analysis^ 106the stoichiornetry). The break through curve for a pilot of large-scale column is neededto determine the point at which it is no longer economical to continue the binding.9.2 Materials and Methods9.2.1 Cation Exchange ChromatographyIn order to obtain breakthrough data for binding at different flowrates, 5 cc syringes(1.2 cm i.d.) were used as miniature chromatography columns. Figure 9.27 shows aphotograph of one of these small columns packed with 3 nil of HC-2 exchanger. Glasswool was packed into the bottom of the syringe, followed by the equilibrated exchangerand excess equilibration buffer. The plunger with its rubber seal was used as the top capfor the column and a syringe needle was forced through the rubber seal to allow entry ofbuffer into the top of the column Silicone tubing was used to carry buffer to the column,and was inserted into the top of the plastic funnel portion of the syringe needle. Solutionexited the column under gravity and was collected by a fraction collector (not shown).Sample was applied at a controlled flowrate by an Ismatec peristaltic pump.The same buffers as described previously were used for equilibration, washing, elutionand saltwash. Sample was yolk supernatant at pH 5.5 prepared according to the protocoldeveloped in Chapter 4.Experiments were also carried out using a 1 cm i.d. column packed to a heightof 15 cm (about 12 ml bed volume of exchanger). Figure 9.28 shows the experimentalapparatus including the chromatography column, fraction collector, peristaltic pump andsample flask. Again, buffers were as described for the 3 ml column experiments.All experiments were carried out at room temperature using WSF prepared accordingto the procedure described at the end of Chapter 4.Chapter 9. Breakthrough Curve Analysis107Figure 9.27: A Small Column Made from a Disposable 5 cc SyringeChapter 9. Breakthrough Curve Analysis108Figure 9.28: Apparatus used for Breakthrough ExperimentsChapter 9. Breakthrough Curve Analysis^ 1099.2.2 Methods of AnalysisRID as described in Chapter 4 was used to determine the IgY concentration of selectedfractions.9.3 Results and DiscussionAt first the absorbance at 280 nm was monitored at the column exit in order to determinethe breakthrough of IgY. However, this proved impractical because IgY is only one ofmany proteins in the applied sample, and no distinct increase in absorbance occurred.It was determined that only an assay specific for IgY, such as RID, would allow themonitoring of IgY breakthrough.Figure 9.29 shows a typical breakthrough curve for IgY using one of the 3 ml columnsof HC-2. The data points shown represent individual fractions collected from 2 identicalexperiments in which the flowrate was 1.0 ml/min, and the corresponding superficialvelocity 53.1 cm/h.Unlike the sigmoidal breakthrough curve that would be expected for an adsorbingspecies with a distinct isoelectric point, the curves obtained here show an early break-through of IgY. An attempt was made to fit the model of Arnold et al. (1985), which isa simple but practical model developed for affinity chromatography based on principlesof specific adsorption, but this was quite unsuccessful. Another simple yet very practicalmodel for adsorption chromatography developed by Skidmore and Chase (1988) was alsoconsidered, but rejected, since it also assumes that the desired species has a well-definedisoelectric point. An attempt to develop a more complex mathematical model capable ofdealing with potential interactions between IgY and other proteins present, and takinginto account the heterogeneity in the isoelectric point (pI) of the molecule would haveChapter 9. Breakthrough Curve Analysis^ 110been beyond the resources of the author. Therefore, an empirical model that wouldclosely approximate the breakthrough curve, and allow comparison of results from var-ious experiments seemed preferable. Since the general shape of the breakthrough curveappeared similar to that of the Langmuir adsorption isotherm (Langmuir, 1918), thismodel was tested and found to provide an adequate fit. Thus the fitted curve is of theform:y aXb+Xwhere Y is the dimensionless IgY concentration, X is the number of equivalent bedvolumes of WSF applied to the column, and a and b are empirical parameters chosen tobest fit the curve to the data. Since the shape of the breakthrough curve is also a functionof the flowrate, the parameters cannot be equated directly to those of an adsorptionisotherm, which represents equilibrium binding conditions. However, the fitted curveis useful in comparing the results from experiments involving different flowrates (andtherefore different superficial velocities).Figure 9.30 shows the fitted curves for experiments carried out at 2.0 ml/min (CurveA) and 0.175 ml/min (Curve B) using the 3 ml column It is clear that at the lowerflowrate, more of the IgY is binding2.Experiments at 5 different flowrates (0.175, 0.225, 0.345, 1.0 and 2.0 ml/min) werecarried out using the 3 ml HC-2 columns, and the resulting data points are shown inFigure 9.31, essentially bounded by the two curves of Figure 9.30 which represent thetwo extremes in flowrate used.Since the binding of IgY is greatly influenced by both the amount of sample appliedand the flowrate of sample application, a plot of the amount of the protein bound as2The area above the curves is directly proportional to the quantity of IgY bound.1 .010^20^30^40^500 .0o 60Chapter 9. Breakthrough Curve Analysis^ 111Figure 9.29: Break Through Analysis using 3 ml Column at 1.0 ml/minNumber of Bed Volumes AppliedChapter 9. Breakthrough Curve Analysis^ 112Figure 9.30: Comparison of Breakthrough Curves at Two Different FlowratesNumber of Bed Volumes Applied0 10^20^30 40 50Legend0.175 ml/min0.5 ml/min0.221 ml/min2.0 ml/min1.0 ml/minChapter 9. Breakthrough Curve Analysis^ 113Figure 9.31: Plotted Data from Breakthrough Experiments at Five Different Flowrates(the solid lines represent curves fitted to data for the highest and lowest flowrates).1 .000..4.,0^0.8s.a)000C-)^0 . 6›-.izzmrna)-al0^0.40'FA0cvE.-.0 . 20.0Number of Bed Volumes AppliedChapter 9. Breakthrough Curve Analysis^ 114a function of bed volumes applied at various flowrates would provide a practical toolfor estimating recovery. If curves are fitted at all flowrates, and the areas above eachone integrated in order to estimate the amount of IgY bound, then the recovery as afunction of the number of bed volumes applied can be estimated by dividing the totalamount bound by the total amount applied. The resulting curves (fitted using the modeldescribed above) are shown for 3 flowrates in Figure 9.32. In all three cases, the recoverydrops off rapidly with increased loading. However, it is clear that as the loading rate indecreased, a significant improvement in recovery occurs. This work was all done usingthe small columns, so the next step was to determine how to use this information topredict large-scale results. This is the central topic of the following chapter.9.4 ConclusionsDue to the heterogeneity of the IgY molecule, it is not possible to obtain completebinding, even after the application of only one bed volume of sample. For this reason,it has been suggested that ion-exchange chromatography is unlikely to prove feasible forIgY separation (Higgins, 1976). However, the breakthrough analysis studied here showsthat under the right conditions of pH and flowrate, it is possible to bind a significantquantity of the molecule. Since not all subclasses may be desired, the fact that a sub-fraction of the IgY does not bind in cation exchange chromatography may actually bedesireable. The simple empirical model used to fit the experimental curves is useful inestimating the recovery at various sample loadings and flowrates.I^I^1^I^i5 10^15^20^2500 30Chapter 9. Breakthrough Curve Analysis^ 115Figure 9.32: Recovery as a Function of Loading for 3 Different FlowratesNumber of Bed Volumes AppliedChapter 10Scale-up of the Separation Process10.1 IntroductionThis chapter deals with efforts to use the results from the small-scale experiments carriedout in Chapter 9 to predict the performance of a pilot-scale column The bed volumeof the column was increased by a factor of 500 (from the 3 ml columns to a pilot-scale1.5 litre column, 100 cm2 cross-sectional area and 15 cm bed height). The pilot-scalework was carried out in both batch and column mode in order to compare these twotechniques.As well, a test of the pressure drop across the pilot-scale column as a function ofthe superficial velocity of the mobile phase was done to determine whether or not therewould be any danger of bed compaction at high flowrates.10.2 Materials and Methods10.2.1 Flow Testing a Column of HC -2A pressure gauge was connected just before the inlet to the column in order to measurethe pressure drop at various flowrates. The column was equilibrated with 0.2 M PBpH 5.0, and the same buffer pumped through the column at various flowrates using aperistaltic pump. The flowrate was varied stepwise from 30 ml/min to 500 ml/min. Oncethe pressure drop across the column had stabilized, a reading was taken and the flowrate116Chapter 10. Scale-up of the Separation Process^ 117then increased to the next level.10.2.2 Batch Separation Using HC -2Approximately 1.0 kg of HC-2 cation exchanger, previously used in the 1.5 litre columnexperiments, was poured into a 4 litre flask and allowed to equilibrate with 0.2 M phos-phate buffer (pH 5.0). In order to ensure proper mixing and contact of the buffer with theexchanger particles, an 8 cm long magnetic stirring bar was added, and the flask placedon a Fisher Model 11-500-7311 stirring/hot plate, with the stir control set on maximum.It was necessary to stir manually at first in order to fluidize the exchanger particles. Afterequilibration, the buffer was decanted, and the remaining buffer removed by filtration.To accomplish this, the flask contents were poured into an empty 1.5 litre columnand air was pumped through with a pressure drop across the column of approximately17 kPa gauge until the excess aqueous buffer was removed. This was considered to haveoccurred when the flowrate leaving the column had dropped to less than 1 ml/minWashing buffer, 0.01 M phosphate buffer (pH 5.4), was then used to wash the ex-changer back into the 4 litre flask. Mixing, decanting, and replacement of spent bufferwith fresh continued until the ionic strength of the aqueous phase, after significant mix-ing, approached that of fresh buffer. The procedure outlined above for removal of bufferthen repeated, and the exchanger was then ready for sample loading.The equilibrated and washed exchanger was first mixed with 1.5 litres of dilutedyolk supernatant resulting in a total volume of about 2200 ml. Following 50 minutesof agitation with the stir bar, the exchanger was allowed to settle for 10 minutes. Asample of the clear supernatant above the exchanger was taken, and the contents of theflask then poured into the 1.5 litre column for aqueous phase removal. Since a significantamount of the exchanger adhered to the sides of the flask, 200 ml of washing buffer wasChapter 10. Scale-up of the Separation Process^ 118used to help wash it into the column. A total of 1.6 litres of liquid was recovered fromthe exchanger.The exchanger was then washed back into the 4 litre flask with 1.5 litres of 0.2 Mphosphate beffer (pH 5.0, hereafter called elution buffer). After another 50 minutesof mixing and 10 minutes of settling, a sample of the supernatant was again taken,representing the eluted fraction. Buffer removal resulted in the collection of a total of1.8 litres of eluted fraction. Following buffer removal, a saltwash (0.01 M phosphatebuffer/0.5 M NaCl/pH 4.8) was used to remove any remaining protein and prepare theexchanger for equilibration. This completed one cycle of the batch process.10.2.3 Column Separation Using HC -2The automated chromatography system developed in Chapter 3 was used to control theseparation of IgY on a 1.5 litre column (i.d. approximately 11.3 cm, bed height 15 cm)packed with HC-2 cation exchanger. A photograph of the system including the pilot scalecolumn is shown in Figure 10.33. The breakthrough experiment was carried out underthe same set of conditions used in Chapter 9. Samples were automatically collected atthe column exit after every bed volume applied, and the IgY concentration for selectedfractions determined by RID.10.3 Results and Discussion10.3.1 Flow Test ResultsMany cellulose-based ion-exchangers are suitable only for laboratory work involving rel-atively low flowratesl. In order to determine the limiting flowrate for sample application1-For example, attempts to vary the flowrate in initial experiments carried out using DEAE-Sephacelresulted in some cases in column adapters being blown to their extreme positions by the build-up inChapter 10. Scale-up of the Separation Process^ 119Figure 10.33: Automated Chromatography System Including Pilot Scale ColumnChapter 10. Scale-up of the Separation Process^ 120and elution with HC-2, an experiment was carried out with a pilot-scale column and thepressure drop versus the superficial velocity plotted. Figure 10.34 shows the results ofthe flow test using a 1.5 litre column packed with HC-2. In this experiment the columnwas subjected to superficial velocities far beyond those generally recommended (ie: under50 cm/h). At a pH of 5.0, the exchanger was able to withstand a superficial velocity ofover 300 cm/h, at a pressure drop across the column of about 20 kPa, without irreversiblecompaction. This demonstrates that the exchanger has excellent flow properties.10.3.2 Batch SeparationA batch experiment indicated that recovery and purity, 53 and 57% respectively, werelower than for the column experiments. To verify these values, the experiment wasrepeated and similar results were obtained. The lower recovery may be due to loss of IgYwhen removing the mobile phase, since a portion may be loosely bound to the exteriorof the cellulose exchanger particles. In any case, the performance of the batch separationdid not appear interesting enough to warrant further study.10.3.3 Column SeparationThe 1.5 litre column was equilibrated and washed with the usual buffers and then 31equivalent bed volumes of yolk supernatant (approximately 46.7 litres) was applied ata flowrate of 30 ml/min. The dimensionless IgY concentration was calculated by divid-ing the IgY concentration determined for each sample by that of the yolk supernatant,resulting in values between 0 and 1. The resulting breakthrough curve is illustrated inFigure 10.35.pressure caused by compaction of the bed.Chapter 10. Scale-up of the Separation Process^ 121Figure 10.34: Pressure Drop vs. Superficial Velocity for a Pilot Scale HC-2 Column 25500^60^120^180^240^300^360Superficial Velocity of Mobile Phase (cm/h)1.00.00^5^10^15^20^25^30^35Chapter 10. Scale-up of the Separation Process^ 122Figure 10.35: Breakthrough Curve for the Pilot Scale ColumnNumber of Bed Volumes AppliedChapter 10. Scale-up of the Separation Process^ 123The values of the two fitted parameters and their standard errors where determinedto be:Parameter Value Standard Errora^0.696^0.017b^1.603^0.267Since the shape of the breakthrough curve is a function of the flowrate, the param-eters cannot be equated directly to those of an adsorption isotherm which representsequilibrium binding conditions. However the parameter 'a' does indicate that based ondata from the first 31 bed volumes applied approximately 30% of the IgY present in thesupernatant is binding very strongly to the exchanger. An additional 10% or so bindsstrongly enough to remain on the column under the conditions and flowrate used in thisexperiment. Thus 20 to 30 bed volumes can be applied with a recovery of around 40%and purity of 60-63% according to HPLC analysis.Figure 10.36 shows the fitted curve for the pilot scale run plotted together with thefitted breakthrough curve for the 12 ml column. Although both curves are rather steepduring binding of the initial 5 bed volumes, once 25 to 30 bed volumes have been applieda greater percentage of the IgY is binding to the pilot scale column than to the smallercolumn, despite the fact that the diameter of the pilot scale column is 15 times that ofthe 12 ml column, and one would expect the velocity profile across the cross-section ofthe smaller column to be more even than in the large column. The only other differencebetween the two cases is the superficial velocity, which is 18 cm/h for the pilot scalecolumn and 25 cm/h for the small column. Thus it would appear that the superficialvelocity has such a great influence on the binding that it outweighs the effect of increasingChapter 10. Scale-up of the Separation Process^ 124column diameter by a factor of 15.As was shown in Figure 9.32, the recovery for the 3 ml columns is strongly dependenton the flowrate used to apply sample. If the recovery after the application of a specifiednumber of bed volumes is plotted against the superficial velocity (rather than flowrate,so that columns of various dimensions can be compared), then the result is a plot suchas shown in Figure 10.37a, which shows the points obtained by plotting recovery after 30bed volumes of sample application. An exponential curve has been fitted to the data byfirst taking the logarithm of the superficial velocity and carrying out a linear regressionas shown in Figure 10.37b.This strong correlation between superficial velocity and extent of binding suggeststhat diffusion of IgY into the pore spaces of the adsorbent particles may be the limitingfactor. A similar attempt to correlate recovery to the residence time of the sample in thecolumn failed to yield a significant regression. For scale-up purposes then, it is possibleto obtain a reasonable estimate of the recovery for a large column by using a smalllaboratory-scale column to prepare a plot of recovery as a function of superficial velocityfor a given number of bed volumnesImprovements to Binding of IgY Once it became evident that recoveries of over60% like those obtained for applications of 3 to 5 bed volumes could not be reproducedwhen binding closer to the capacity of the exchanger, a small scale experiment using the15 cm column was repeated using supernatant adjusted with 0.1 N HC1 to pH 5 insteadof the usual pH 5.5. HPLC analysis indicates that a significant increase in binding of IgYoccurs, but at the expense of increased impurity. One of the problems with using a lowerpH is that a- and /3-livetin bind more strongly. As well, ovalbumin has an isoelectric pointof about 4.7 and may also be bound more strongly. Using HPLC analysis, the purity12m1 column25 cm/h1500m1 column18 cm/h0 . 80 . 60 . 40 . 20.0Chapter 10. Scale-up of the Separation Process^ 125Figure 10.36: Breakthrough Comparison for Pilot Scale and Small Scale Columns1.0 ^I^i^I^I^I^I0^5^10^15^20^25^30^35Number of Bed Volumes Applied-i^1^1^I^I20^40^60^80^1000 1 20Chapter 10. Scale-up of the Separation Process^ 126Figure 10.37: Recovery as a Function of Superficial Velocity for Several Experiments6050..--.R......a)o0a)401:4 302 0Superficial Velocity (cm/h)2 3^4Ln(Superficial Velocity)5Chapter 10. Scale-up of the Separation Process^ 127Table 10.13: Example Protocol for IgY Purification. Extraction was using distilled wa-ter and pH adjustment; separation using CEC; and purification using sodium sulphateprecipitation.Column Separation Batch SeparationStep IgY Cumulative IgY Purity IgY Cumulative IgY PurityRecovery (%) (%) Recovery (%) (%)1 (Extraction) 80* 18 80 182 (Separation) 51 61 42 573 (Purification 50 95 42 95*Assuming two-step dilution/extraction is used, or recycling of the precipitated LSF inorder to improve recovery.dropped to around 42%, and the recovery increase to over 60%. Thus it is necessaryas usual to trade purity in order to obtain better recovery, and the more importantparameter is determined by the end use of the product.Summary of the Process Although many combinations of the isolation methodsdescussed in this dissertation are possible, to complete the discussion of scale-up anexample protocol is summarized in Table 10.13. This procedure would be relativelysimple to scale up, including a small number of steps (3), and yet yield a relatively highpurity product. A brief discussion of economic considerations is given in Appendix C.10.4 ConclusionsThe degree of binding of IgY to the cation exchanger is primarily affected by pH and thesuperficial velocity of the mobile phase. Even when scaling-up the process by a factor of500 with respect to exchanger bed volume and volume of sample applied, it is possibleto obtain a reasonable estimate of the recovery and breakthrough curve based on small-scale experiments. Column operation results in better binding and therefore recoveryChapter 10. Scale-up of the Separation Process^ 128of IgY than continuously stirred batch operation under the same conditions. Lower pHallows higher recovery of the molecule (probably due to stronger binding of additionalsubclasses), but at the expense of lower purity of the eluted fraction.Chapter 11ConclusionsAs IgY is a heterogeneous polyclonal antibody with multiple isoeletric points, it is notpossible, using ion-exchange chromatography, to completely isolate it from the otherwater soluble proteins in egg yolk. It is necessary to trade off purity for recovery, or viceversa, if the intact molecule is desired. Cation exchange chromatography is a potentiallyuseful technique for the separation of IgY from the water soluble fraction of egg yolk, andhas been shown to allow better recovery and purity than other ion-exchange methods.A simple yet efficient 3 stage process has been developed which is easily scaled-up forindustrial application.Since subclasses of IgY exist and it appears that approximately 1 of the IgY presentin the WSF does not bind to the column under the conditions used here, it is possi-ble that the technique could be used to efficiently fractionate some of these subclasses.Only further study using egg yolk from immunized birds that can be analyzed usingimmunoassays such as ELISA's would allow clarification of this hypothesis.The major contaminant remaining after cation exchange chromatography of the WSFis a-livetin. This protein may be effectively separated from IgY by one-step salt precipi-tation with Na2SO4, yielding a final product with a purity of at least 95%.HPLC gel filtration can be used effectively to determine IgY concentration and purityin relatively pure (greater than 50%) samples, and can also provide a rapid, rough esti-mate of molecular weight. The automated system assembled for this study is a practical129Chapter 11. Conclusions^ 130means of speeding up the testing of exchangers under various experimental conditions.The key parameters in determining the extent of binding of IgY in cation exchangechromatography are the pH, ionic strength, and superficial velocity of the mobile phase.Breakthrough curve analysis of laboratory-scale columns has been shown to provide apractical method of estimating the performance of a much larger column.Chapter 12RecommendationsThe LSF precipitated in the extraction stage of the process is substantial in volumeand further work on its utilization would be desireable before the process is used at anindustrial scale. Waste disposal is becoming a major concern and it would be prefereablenot to waste the valuable proteins still present in the LSF.An interesting extension of this work would be to study the effect of dilution inthe extraction stage using previously extracted WSF so as to reduce water useage andincrease IgY concentration in the supernatant to be applied to the chromatography col-umn. Another potential variation would be to use the sample application fraction fromchromatography to dilute yolk in the extraction stage.Further study of ultrafiltration for purification of IgY from the eluate of the chro-matography step would be valuable. It is possible that under certain conditions of saltconcentration and pH the a-livetin might pass through a 100 kDa membrane more read-ily. Also, a larger membrane (perhaps 150 kDa NMWC) would likely allow a-livetin topass while still retaining the majority of the IgY. Although ultrafiltration is generallymore energy intensive than either sedimentation or centrifugation, it would allow contin-uous processing in the third stage of the process and is less labour intensive than batchcentrifugation.Detailed study of the binding of IgY subclasses using yolks of hens immunized againstspecific antigens would provide better insight into the potential of cation exchange asa technique for separation of IgY from the WSF. As well, further work is needed on131Chapter 12. Recommendations^ ) 3 2the aggregation of IgY in aqueous solution, such as the eluate from the cation exchangecolumn, in order to determine whether or not such aggregation is reversible. Optimizationof the cation exchange portion of the process would also be beneficial before attemptingto use it at a large scale.Bibliography[1] Abelson, P.II., 1989, Synthetic membranes, Science, 244(2411), 1421.[2] Akita, E. and S. Nakai, 1992, Isolation and purification of immunoglobulins fromegg yolk, J. Food Sci., 57, 629-634.[3] Al Moudallal, Z., D. Altschuh, J.P. Briand and M.H.V. Van Regenmortel, 1984,Comparative sensitivity of different ELISA procedures for detecting monoclonalantibodies, J. Immunol Methods, 68, 35-43.[4] Allen, J.C., and C.J. Smith, 1987, Enzyme-linked immunoassay kits for routinefood analysis, Tibtech, 5, 193-199.[5] Altschuh, D., G. Hennache and M.H.V. Van Regenmortel, 1984, Determination ofIgG and IgM levels in serum by rocket immunoelectrophoresis using yolk antibodiesfrom immunized chickens, J. Immunol. Methods., 69, 1-7.[6] Ansari, A.A., and T.S. Chang, 1983, Immunochemical studies to purify rabbit andchicken immunoglobulin G antibody by protein A-Sepharose chromatography, Am.J. Vet. Res., 44(5), 901-906.[7] Arnold, F.H., J.J. Chalmers, M.S. Saunders, M.S. Croughan, H.W. Blanch andC.R. Wilke, 1985, A rational approach to the scale-up of affinity chromatography,In: Purification of Fermentation Products, LeRoith, D., J. Shiloach and T.J Leahy,eds., Amer. Chem. Soc. Symp. Ser., 271, 113-122.[8] Awdeh, A.L., A.P. Williamson and B.A. Askonas, 1968, Isoelectric focusing inpolyacrylamide gel and its application to immunoglobulins, Nature, 219, 66.[9] Bade, H., and H. Stegemann, 1984, Rapid method of extraction of antibodies fromhen egg yolk, J. Imrnunol. Methods, 72, 421-426.[10] Ballabriga, A., 1982, Immunity of the infantile gastrointestinal tract and implica-tions in modern infant feeding, Acta Paediatr. Jpn., 24(2), 235-239.[11] Bar-Joseph, M., and M. Malkinson, 1980, Hen egg yolk as a source of antiviralantibodies in the enzyme-linked immunosorbent assay (ELISA): A comparison oftwo plant viruses, J. Virol. Methods, 1, 179-183.[12] Barker, P.E., and G. Ganetsos, 1988, Chemical and biochemical separations usingpreparative and large scale batch and continuous chromatography, Separation andPurification Methods, 17(1), 1-65.[13] Bauwens, R.M., J.A. Kint, M.P. Devos, K.A. Van Brussel and A.P. De Leen-heer, 1987, Production, purification and characterization of antibodies to 1,25-dihydroxyvitamin D raised in chicken egg yolk, Clinica Chimica Acta, 170, 37-44.133[14] Bauwens, R.M., M.P. Devos, J.A. Kint and A.P. De Leenheer, 1988, Chicken eggyolk and rabbit serum compared as sources of antibody for radioimmunoassay of1,25-dihydroxyvitamin D in serum or plasma, Clinical Chemistry, 34(10), 2153-2154.[15] Benedict, A.A., 1967, Production and purification of chicken immunoglobulins, In:Methods in Immunology and Immunochemistry, Vol.', Williams, C.A. and M.W.Chase, eds., Academic Press, NY, 229-237.[16] Bornstein, S., A. Rautenstein-Arazi and Y. Samberg, 1952, Some aspects of con-genital passive immunity to Newcastle disease in chicks. I. The transfer of hemag-glutinin inhibitors from maternal yolk to chick, Am. J. Vet. Res., 13, 373-378.[17] Brambell, F.W.R., 1970, The Transmission of Passive Immunity from Mother toYoung, Elsevier/North Holland Publishing Co., Amsterdam.[18] Brandly, C.A., H.E. Moses and E.L. Jungherr, 1946, Transmission of antiviralactivity via the egg and the role of congenital passive immunity to newcastle diseasein chickens, Am. J. Vet. Res., 7(7), 333-342.[19] Briles, W.E., 1948, Induced hemolytic disease in chicks., Genetics, 33, 96-97 (ab-stract).[20] Burger, D., M.-A. Ramus and M. Schapira, 1985, Antibodies to human plasmakallikrein from egg yolks of an immunized hen: Preparation and characterization,Thrombosis Research, 40, 283-288.[21] Burley, R.W. and J.F. Back, 1987, The proteins in avian eggs: A test of proteinseparation methods on HPLC, In: Proteins: Structure and Function. L'Italien,J.L., ed., Plenum Press, NY, 771-777.[22] Burley, R.W., and D.V. Vadehra, 1979, Chromatographic separation of the solubleproteins of hen's egg yolk: an analytical and preparative study, Anal. Biochem.,94, 53-59.[23] Burley, R.W., and D.V. Vadehra, 1989, The Avian Egg - Chemistry and Biology,John Wiley and Sons, New York.[24] Buxton, A., 1951, Transfer of bacterial antibodies from the hen to the chick, Nature,168, 657-658.[25] Buxton, A., 1952, On the transference of bacterial antibodies from the hen to thechick, J. Gen. Microbiol., 7, 268-286.[26] Carroll, S.B., and B.D. Stollar, 1983, Antibodies to calf thymus RNA polymeraseII from egg yolks of immunized hens, J. Biol. Chem., 258(1), 24-26.[27] Cser, L., I.A. Gladkih, D. Hadge and H Ambrosius, 1982, X-ray small-angle scat-tering study of general structure of chicken immunoglobulin Y, Immunology Letters,4, 15-19.134[28] Davis, D.R., 1987, The theory and practice of cross-flow filtration, Brewers Digest,August Issue, 34-36,43.[29] de Maat-Bleeker, F., A.G. van Dijk and L. Berrens, 1985, Allergy to egg yolkpossibly induced by sensitization to bird serum antigens, Annals of Allergy, 54,245-248.[30] DePalma, A., 1993, Sepracor Inc. Develops novel Process Separations Technique,Genetic Engineering News, 13(1), 1.[31] Devergne, J.C., L. Cardin, J. Burckard and M.H.V. Van Regenmortel, 1981, Com-parison of direct and indirect ELISA for detecting antigenically related cucu-moviruses, J. Virol. Methods, 3, 193-200.[32] Dreesman, G. and A.A. Benedict, 1965, Properties of papain-digested chicken 7 S7-globulin, J. Immunol., 95(5), 855-866.[33] Dzierzgowski, S.K., 1901, EM Beitrag zur Frage der Vererbung der kunstlichenDiphtherie-immunitat, Ztschr. fur Bact. I. Abteil., 30, 2263-2265.[34] Dziezak, J.D., 1990, Membrane separation technology offers processors unlimitedpotential, Food Technology, 44(9), 108-113.[35] Ehrlich, P., 1892, Ueber immunitat durch verebung und saugung, Z. Hyg. Infeckt-Krankh, 12, 183-203.[36] Eichler, D., and K. Rubach, 1986, Gewinnung von Antikorpern gegen Sojaproteinaus Eiern immunisierter Legehennen, Z. Lebensm. Unters. Forsch, 183, 8-11.[37] Erhard, M.H., I. Von Quistorp, I. Schranner, A. Jungling, B. Kaspers, P. Schmidtand R. Kuhlmann, 1992, Development of specific enzyme-linked immunosorbentantibody assay systems for the detection of chicken Igs G, M, and A using mono-clonal antibodies, Poultry Science, 71(2), 302-310.[38] Ewing, 1903, Med. News, 83, 871.[39] Fertel, R., J.Z. Yetiv, M.A. Coleman, R.D. Schwarz, J.E. Greenwald and J.R.Bianchine, 1981, Formation of antibodies to prostaglandins in the yolk of chickeneggs, Biochem. and Biophys. Res. Comm., 102(3), 1028-1033.[40] Fichtali, J., E.A. Charter, K.V. Lo and S. Nakai, 1992, Separation of egg yolkimmunoglobulins using an automated liquid chromatography system, Biotech. andBioeng., 40, 1388-1394.[41] Flodin, P., and J. Killander, 1962, Biochimica et Biophysica Acta, 63, 403.[42] Fraser, D.T., T.H. Jukes, H.D. Branion and K.C. Halpern, 1934, The inheritanceof Diphtheria immunity in ducks, J. Immunol., 26, 437-446.[43] Friend, B.A., K.M. Shahani and B.N. Mathur, 1983, Newer advances in humanmilk substitutes for infant feeding, Journal of Applied Nutrition, 35(2), 88-115.135[44] Gardner, P.S., and S. Kaye, 1982, Egg globulins in rapid virus diagnosis, J. Virol.Methods, 4, 257-262.[45] Gassmann, M., P. Thommes, T. Weiser and TI. Hubscher, 1990, Efficient produc-tion of chicken egg yolk antibodies against a conserved mammalian protein, FasebJournal, 4(8), 2528-2532.[46] Gassmann, M., T. Weiser, P. Thommes and U. Hubscher, 1990, The chicken eggas a source of polyclonal antibodies, Schweizer Archiv Fur Tierheilkunde, 132(6),289-294.[47] Gast, R.K., and C.W. Beard, 1991, Research note: Detection of Salmonellaserogroup D-specific antibodies in the yolk of eggs laid by hens infected withSalmonella enteritidis Poultry Science, 70, 1273-1276.[48] Gosling, I.S., 1987, The design of adsorptions in downstream processing: The im-portance of data from small-scale experiments, In: Biochemical Engineering - AChallenge for Interdisciplinary Cooperation, Chmiel, H., et al., eds., Gustav FischerVerlag, Stuttgart, 396[49] Gross, 1899, Dissertation, Strassburg.[50] Hassl, A. and H. Aspock, 1988, Purification of egg yolk immunoglobulins - A twostep procedure using hydrophobic interaction chromatography and gel filtration, J.Immunol. Methods, 110, 225-228.[51] Hassl, A., H. Aspock and H. Flamm, 1987, Comparative studies on the purity andspecificity of yolk IgY, Zbl. Bakt. Hyg. A, 267, 247-253.[52] Hatta, H., J.S. Sim and S. Nakai, 1988, Separation of phospholipids from egg yolkand recovery of water-soluble proteins, J. Food Sci., 53(2), 425-431.[53] Hatta, H., M. Kim and T. Yamamoto, 1990, A novel isolation method for hen eggyolk antibody, "IgY", Agric. Biol. Chem., 54(10), 2531-2535.[54] Hau, J., J.G. Westerfaard, P. Svendsen, A. Bach and B. Teisner, 1980, Compar-ison of the pregnancy-associated murin protein-2 and human pregnancy-specific/31-glycoprotein, J. Reprod. Fert., 60, 115-119.[55] Hektoen, L., 1918, The production of precipitins by the fowl, Journal of InfectiousDiseases, 22, 561-566.[56] Heller, E.D., 1975, The immune response of hens to multiple Escherichia coli injec-tions and transfer of immunoglobulins to the egg and hatched chick, Research inVeterinary Science, 18, 117-120.[57] Hersh, R.T. and A.A. Benedict, 1966, Aggregation of chicken -y-G immunoglobulinin 1.5 M sodium chloride solution, Biochimica et Biophysica Acta, 115: 242-244.[58] Higgins, D.A., 1976, Fractionation of fowl immunoglobulins, Research in VeterinaryScience, 21, 94-99.136[59] Howell, H.M., H.E. Conrad and E.W. Voss, 1973, Hexose and hexosamine content ofpurified chicken anti-2, 4-dinitrophenyl antibody by radiochromatographic analysis,Immunochemistry, 10, 761-766.[60] Ingham, K.C., T.F. Busby, Y. Sahlestrom and F. Castino, 1980, Separation ofmacromolecules by ultrafiltration- Influence of protein adsorption' protein-proteininteractions, and concentration polarization. In: Ultrafiltration Membranes andApplications, Cooper, A.R., ed., Plenum Press, New York, 141-158.[61] Jensenius, J.C., I. Andersen, J. Hau, M. Crone and C. Koch, 1981, Eggs: Con-veniently packaged antibodies. Methods for purification of yolk IgG., J. Immunol.Methods, 46, 63-68.[62] Johansson, H., M Ostling, G. Sofer, H. Wahlstrom and D. Low, 1988, Chromato-graphic equipment for large-scale protein and peptide purification, In: Advances inBiotechnological Processes, Mizrahi, A., ed., 8, 127-157.[63] Jukes, T.H. and H.D. Kay, 1932, The immunological behaviour of the second pro-tein (livetin) of hen's egg yolk, J. Exp. Med., 56, 469-482.[64] Jukes, T.H., D.T. Fraser and M.D. Orr, 1934, J. Immunol., 26, 353.[65] Kang, C.B.H., and K.K. Ho, 1991, Characterization of a soluble inorganic pyrophos-phatase from Microcystis aeruginosa and preparation of its antibody, Archives ofBiochemistry and Biophysics, 289(2), 281-288.[66] Kay, H.D. and P.C. Marshall, 1928, The second protein (livetin) of egg-yolk, Bio-chemical Journal, 22, 1264-1269.[67] Kekwick, R.A., 1940, Biochem. J., 34, 1248.[68] Klemperer, F., 1893, Uber naturliche Immunitat und ihre Verwerthung fur dieImmunizierungstherapie, Arch. f. Exp. Path. u. Pharmak., 31, 356-382.[69] Knight, P.F. and A.M. Schechtman, 1954, The passage of heterologous serum pro-teins from the circulation into the ovum of the fowl, J. Exp. Zool., 127, 271-303.[70] Kroner, K.H., H. Schutte, H.Hustedt and M.R. Kula, 1984, Cross-flow filtration inthe downstream processing of enzymes, Process Biochemistry, April, 67-74.[71] Kuhlmann, R., V. Wiedemann, P. Schmidt, R. Wanke, E. Linckh and U. Losch,1988, Chicken egg antibodies for prophylaxis and therapy of infectious intestinaldiseases. I. Immunization and antibody determination, J. Vet. Med., B 35, 610-616.[72] Kunz, U.S., A. Pospischil and M.F. Paccaud, 1991, Imrnunohistochemical detec-tion of chlamydiae in formalin-fixed tissue sections: comparison of a monoclonalantibody with yolk derived antibodies (IgY), Zentralblatt Fur Veterinarmedizin -Reihe B., 38(4), 292-298.[73] Kwan, L., E. Li-Chan, N. Helbig and S. Nakai, 1991, Fractionation of water-solubleand -insoluble components from egg yolk with minimum use of organic solvents, J.Food Sci., 56(6), 1537-1541.137[74] Langmuir, I., 1918, The constitution and fundamental properties of solids andliquids. Part 1. Solids., J. Amer. Chem. Soc., 38, 2221-2295.[75] Larsson, A., and J. Sjonquist, 1988, Chicken antibodies: a tool to avoid falsepositive results by rheumatoid factor in latex fixation tests, J. Immunol. Methods,108, 205-208.[76] Larsson, A., and J. Sjoquist, 1990, Chicken IgY: utilizing the evolutionary dif-ference, Comparative Immunology, Microbiology and Infectious Diseases, 13(4),199-201.[77] Lee, Y.Z., J.S. Sim, S. Al-Mashikhi and S. Nakai, 1988, Separation of immunoglob-ulins from bovine blood by polyphosphate precipitation and chromatography, J.Agric. Food Chem, 36, 922.[78] Leslie, G.A., 1975, Ontogeny of the chicken humoral immune mechanism, Am. J.Vet. Res., 36(4), 482-485.[79] Leslie, G.A., and L.W. Clem, 1969, Phylogeny of immunoglobulin structure andfunction, J. Exp. Med., 130(6), 1337-1352.[80] Leslie, G.A., and L.W. Clem, 1970, Chicken imm-unoglobulins: Biological half-livesand normal adult serum concentrations of IgM and IgY, Proc. Soc. exp. Biol. Med.,134, 195-198.[81] Loeken, M.R., and T.F. Roth, 1983, Analysis of maternal IgG subpopulations whichare transported into the chicken oocyte, Immunology, 49, 21-28.[82] Losch, U., I. Schranner, R. Wanke and L. Jurgens, 1986, The chicken egg, anantibody source, J. Vet. Med., B 33, 609-619.[83] Mandalla, M.M., A.L de Weck and C.A. Dahinden, 1988, Bird-egg syndrome -Cross reactivity between bird antigens and egg-yolk livetins in IgE-mediated hy-persensitivity, Int. Arch. Allergy Appl. Immunol., 87, 143-150.[84] Mandeles, S., 1960, Use of DEAE-cellulose in the separation of proteins from eggwhite and other biological materials, J. Chromatogr., 3, 256-264.[85] Martin, W.G. and W.H. Cook, 1958, Preparation and molecular-weight of -y-livetinfrom egg yolk, Canadian J. Biochem. Physiol., 36, 153-160.[86] Martin, W.G., J.E. Vandegaer and W.H. Cook, 1957, Fractionation of livetin andthe molecular weights of the a- and fl- components, Canadian J. Biochem. Physiol.,35(4), 241-250.[87] McBee, L.E., and O.J. Cotterill, 1979, Ion-exchange chromatography and elec-trophoresis of egg yolk proteins, J. Food Sci., 44, 656-660, 667.[88] McCannel, A.A. and S. Nakai, 1989, Isolation of egg yolk immunoglobulin-richfractions using copper-loaded metal chelate interaction chromatography, Can. Inst.Food Sci. Technol. J., 22(5), 487-490.138[89] McCannel, A.A., and S. Nakai, 1990, Separation of egg yolk immunoglobulins intosubpopulations using DEAE-ion exchange chromatography, Can. Inst. Food Sci.Technol. J., 23(1), 42-46.[90] McIndoe, W.M. and J. Culbert, 1979, The plasma albumins and other livetin pro-teins in egg yolk of the domestic fowl (gallus domesticus), International Journal ofBiochemistry, 10: 659-663.[91] Meisel, H., 1990, Enzyme-immunoassay (ELISA) using IgY antibodies againstlact oferrin., Milchwissenschaft , 45(8), 510-512.[92] Meyer, R.E., 1977, Reports full of gobbledygook, Journal of Irreproducible Results,22(4), 12.[93] Nace, G.W., 1953, Serological studies of the bolld of the developing chick embryo,J. Exp. Zool., 122, 423-448.[94] Nakai, S., E. Li-Chan and V. Lo, 1992, Separation of immunoglobulin from eggyolk, Presented at the International Symposium on Non-conventional Egg Usesand Newly Emerging Processing Technologies, Banff Springs Hotel, Banff[95] Orlans, E., 1967, Fowl antibody VII. A comparison of natural, primary and sec-ondary antibodies to erythrocytes in hen sera: Their transmission to yolk and chick,Immunology, 12, 27-37.[96] Orlans, E. and M.E. Rose, 1972, An IgA-like immunoglobulin in the fowl, Immuno-chemistry, 9, 833-838.[97] Osborne, T., and G.F. Campbell, 1900, The proteins of the egg yolk, J. Amer.Chem. Soc., 22, 413-422.[98] Otake, S., Y. Nishihara, M. Makimura, H. Hatta, M. Kim, T. Yamamoto and M.Hirasawa, 1991, Protection of rats against dental caries by passive immunizationwith hen-egg-yolk antibody (IgY), Journal of Dental Research, 70(3), 162-166.[99] Otani, H., K. Matsumoto, A. Saeki and A. Hosono, 1991, Comparative studies onproperties of hen egg yolk IgY and rabbit serum IgG antibodies, Lebensm.-Wiss.u.-Technol , 24, 152-158.[100] Parkinson, T.L., 1972, Separation of the proteins of egg white and egg yolk and astudy of their interactions in whole egg I. Quantitative fractionation of protein J.Sci. Fd. Agric., 23, 649-658.[101] Parkinson, T.L., 1967, Effect of pasteurisation on the chemical composition of liquidwhole egg. I. Development of a scheme for the fractionation of the proteins of J.Sci. Fd. Agric., 18, 208-213.[102] Patterson, R., J.S. Youngner, W.O. Weigle and F.J. Dixon, 1962a, Antibody pro-duction and transfer to egg yolk in chickens, J. Immunol., 89, 272-278.139[103] Patterson, 11., J.S. Youngner, W.O. Weigle and F.J. Dixon, 1962b, The metalbolismof serum proteins in the hen and chick and secretion of serum proteins by the ovaryof the hen, Journal of General Physiology, 45: 501-513.[104] Peila, T.H., C.M. Gulka, V.J. Yates and P.W. Chang, 1984, Use of egg yolk inseological tests (ELISA and HI) to detect antibody to Newcastle Disease, InfectiousBronchitis, and Mycoplasma gallisepticum, Avian Diseases, 28(4), 877-883.[105] Plimmer, R.H.A., 1908, The proteins of egg-yolk, Chemical Society Transactions,93(Part II), 1500-1506.[106] Poison, A., 1990, Isolation of IgY from the yolks of eggs by a chloroform polyethy-lene glycol procedure, Immunological Investigations, 19(3), 253-258.[107] Poison, A., 1991, Zone electrophoresis of anti-human sperm antibodies in concen-tration gradients of sucrose, Immunological Investigations, 20(5-6), 451-459.[108] Poison, A., M.B. von Wechmar and G. Fazakerley, 1980, Antibodies to proteinsfrom yolk of immunized hens, Immunological Communications, 9(5), 495-514.[109] Poison, A., M.B. von Wechmar and M.H.V. van Regenmortel, 1980, Isolation of vi-ral IgY antibodies from yolks of immunized hens, Immunological Communications,9(5), 475-493.[110] Poison, A., T. Coetzer and J. Kruger, 1985, Improvements in the isolation of IgYfrom the yolks of eggs laid by immunized hens, Immunological Investigations, 14(4),323-327.[111] Powrie, W., and S. Nakai, 1985, Characteristics of edible fluids of animal origin:Eggs. In: Food Chemistry, 2nd ed., Fennema, O.R., ed., Marcel Dekker, Inc., NewYork, pp. 829-855.[112] Ramon, G, 1928, De l'immunite natureile et acquise vis-a-vis de l'intoxicationtetanique chez la poule, C.R. Soc. Biol., 99, 1473-1475.[113] Ramon, G., 1928, Sur le passage de la toxine et de l'antitoxirie tetanique de la poulea l'oeuf et au poussin, C.R. Soc. Biol., 99, 1476-1478.[114] Rassi, Z.E., and C. Horvath, 1986, Metal chelate-interaction chromatography ofproteins with iminodiacetic acid-bonded stationary phases on silica support, J.Chromatogr., 359, 241-253.[115] Regnier, F.E., and I. Mazaroff, 1987, A theoretical examination of adsorption pro-cesses in preparative liquid chromatography of proteins, Biotechnology Progress,3(1), 22-26.[116] Ricke, S.C., D.M. Schaefer, M.E. Cook and K.H. Kang, 1988, Differentiation ofruminal bacterial species by enzyme-linked immunosorbent assay using egg yolkantibodies from immunized chicken hens, Applied and Environmental Microbiology,54(2): 596-599.140[117] Rokhlin, 0.V., R.S. Nezlin and H.K.G. Machulla, 1975, Allotype ratio in isoelectricfractions of anti-azobenzoate antibodies from heterozygous rats. A distinct fractionwith absolut allotypic p Immunochemistry, 12, 311.[118] Rose, M.E., E. Orlans and N. Buttress, 1974, Immunoglobulin classes in the hen'segg: their segregation in yolk and white, Eur. J. Immunol., 4, 521-523.[119] Schmidt, P., A. Hafner, G.H. Reubel, R. Wanke, V. Franke, U. Losch and E.Dahme, 1989, Production of antibodies to canine distemper virus in chicken eggsfor immunhistochemistry, Zentralblatt Fur Veterinarmedizin - Reihe B., 36(9), 661-668.[120] Schmidt, P., R. Kuhlmann and U. Losch, 1988, A competitive inhibition enzymeimmunoassay for detection and quantification of organophosphorus compounds,Zeitschrift Fur Naturforschung - Section C - Biosciences, 43(3/4), 167-172.[121] Schmittle, S.C., 1950, Studies on Newcastle Disease V. A comparison of NewcastleDisease antibody titers in blood and egg yolk as measured by the hemagglutination-Am. J. Vet. Res., 11(4), 226-230.[122] Schram, A.C., S.-J. Hwong and C.P. Christenson, 1971, Purification of chickenanti-bovine serum albumin antibody, Immunology, 21, 637-648.[123] Seideman, W.E., O.J. Cotterill and C.W. Gehrke, 1969, Ion-exchange chromatog-raphy of egg yolk - 1. Separation methods, Poultry Science, 48(3), 884-894.[124] Seideman, W.G., and O.J. Cotterill, 1965, Poultry Science, 44, 1414.[125] Sell, S., 1987, Immunology, Immunopathology and Immunity. 4th ed., ElsevierScience Publishing Co., Inc., New York.[126] Shepard, C.C., and G.A. Hottle, 1949, Studies of the composition of the livetinfraction of the yolk of hen's eggs with the use of electrophoretic analysis, J. Biol.Chem., 179, 349.[127] Shimizu, M., R.C. Fitzsimmons and S. Nakai, 1988, Anti-E. Coli immunoglobulinY isolated from egg yolk of immunized chickens as a potential food ingredient, J.Food Sci., 53(5), 1360-1366.[128] Shirman, G.A., N.A. Graevskaya, I.K. Lavrova, V.A. Ginevskaya, L.E. Baeva, A.P.Savinov, V.A. Kazantseva and S.G. Drozdov, 1988, Separation of yolk immunoglob-ulins from eggs of chickens immunized with enteric viruses, Voprosy Virusologii, 6,725-729.[129] Skidmore, G.L., and H.A. Chase, 1988, A study of ion exchangers for proteinpurification, In: Ion Exchange for Industry, Streat, M., ed., Ellis Horwood Ltd.,Chichester, UK, 520-532.[130] Skidmore, G.L., and H.A. Chase, 1990, Two-component protein adsorption to thecation exchanger S Sepharose FF, J. Chromatogr., 505, 329-347.141[131] Smith, L.W., N.M. Lenoc and M.J. Boland, 1986, Development and applicationsof cellulose ion exchangers in the food and related industries. A review., WaitakiInternational Biosciences, Christchurch, New Zealand.[132] Song, C.-S., J.-H. Yu, D.H. Bai, P.Y Hester and K.-H. Kim, 1985, Antibodies to thea-subunit of insulin receptor from eggs of immunized hens, J. Immunol., 135(5),3354-3359.[133] Steward, M.W., 1984, Antibodies: Their structure and function, Chapman andHall, London.[134] Streat, M., ed., 1988, Ion Exchange for Industry, Ellis Horwood Ltd., Chichester,520-532.[135] Sturmer, A.M., D.P. Driscoll and D.E. Jackson-Matthews, 1992, A quantitativeimmunoassay using chicken antibodies for detection of native and recombinant a-amidating enzyme, J. Irnmunol. Methods, 146(1), 105-110.[136] Sutherland, 1910, Scientific Memoirs of Officers of the Medical and Sanitary De-partments of the Government of India, No. 39.[137] Tenenhouse, H.S. and H.F. Deutsch, 1966, Some physical-chemical properties ofchicken 7—globulins and their pepsin and papain digestion products, Immunochem-istry, 3, 11-20.[138] Ternes, W., 1989, Characterization of water soluble egg yolk proteins with, J. FoodSci., 54(3), 764-765.[139] Vaag, P. and L. Munck, 1987, Immunochemical methods in cereal research andtechnology, Cereal Chem., 64(2), 59-72.[140] Van Orden, D.E. and H.P. Treffers, 1968b, A demonstration of the coprecipitationof 0-lipoproteins with specific prcipitates of chicken antibodies and human serumalbumin, J. Immunol , 100(3), 664-674.[141] Van Orden, D.E. and H.P. Treffers, 1968a, The effect of salt-induced aggregationon the gel filtration of chicken 7 S antibodies, J. Immunol., 100(3), 659-663.[142] Vermeulen, T., and N.K. Hester, 1959, Kinetic relationships for ion exchange pro-cesses, Chemical Engineering Progress Symposium Series, 55(24), 61-69.[143] Vieira, J.G., et al., 1984, Egg yolk as a source of antibodies for human parathyroidhormone (hPTH) radioimmunoassay, J. Immunoassay, 5(1/2), 121-129.[144] Wallmann, J., C. Staak and E. Luge, 1990, A simple method for the isolation ofimmunoglobulin (Y) from the eggs of immunized hens, Zentralblatt Fur Veterin-armedizin - Reihe B., 37(4), 317-320.[145] Wang, K.-Y., C.A. Hoppe, P.K. Datta, A.F. Fogelstrom and Y.C. Lee, 1986, Iden-tification of the major mannose-binding proteins from chicken egg yolk and chickenserum as immunoglobulins, Proc. Natl. Acad. Sci. USA, 83, 9670-9674.142[146] Weir, D.M., 1983, Immunology, 5th ed., Churchill Livingstone, Edinburgh, UK.[147] Wiedemann, V., E. Linckh, R. Kuhlmann, P. Schmidt and U. Losch, 1991, Chickenegg antibodies for prophylaxis and therapy of infectious intestinal diseases. V. Invivo studies on protective effects against Escherichia coil diarrhea in pigs, Zentral-blatt Fur Veterinarmedizin - Reihe B., 34(4), 283-291.[148] Wilkinson, P.C., and V.I. French, 1969, Antigenic heterogeneity of chicken 7S im-munoglobulin, Immunochemistry, 6, 498-501.[149] Williams, J., 1962, Serum proteins and the livetins of hen's egg yolk, BiochemicalJournal, 83, 346-355.[150] Woodward, S.A., and O.J. Cotterill, 1983, Electrophoresis and chromatography ofheat-treated plain, sugared and salted whole egg, J. Food Sci., 48, 501-506.[151] Yamamoto, H., H. Watanabe, G. Sato and T. Mikami, 1975, Identification of im-munoglobulins in chicken eggs and their antibody activity, Jap. J. Vet. Res., 23,131-140.[152] Yazawa, S., 0. Hosomi and A. Takeya, 1991, Isolation and characterization of anti-H antibody from egg yolk of immunized hens, Immunological Investigations, 20(7),569-581.[153] Yolken, R.H., F. Leister, S.-B. Wee, R. Miskuff and S. Vonderfecht, 1988, Antibod-ies to rotaviruses in chicken's eggs: A potential source of antiviral immunoglobulinssuitable for human consumption, Pediatrics, 81(2), 291-295.143Appendix ASchematic Diagram of ChemResearch/Solenoid Valve Interface'Schematic Diagram showing electrical connectionsbetween ChemResearch Interface and Solenoid Valve'First described in: March, A.C., Computer automation of a novel ion-exchange process for thesimultaneous recovery of lysozyme and avidin from chicken egg albumen, Masters Thesis, 1988, UBC,Vancouver, Canada144Appendix BSample Chromatograms of Molecular Weight MarkersmAU IGY -0229.D: ADC CHANNEL Aocoo162- cvOVOMUCO/D160;158-156-154-152-150148-1.•c,Ablio,-2-1Time ->^5.00 10.00^15.00 20.00^25.00145Appendix B. Sample Chromatograms of Molecular Weight Markers^146LC 1^ Thu Apr 29 17:32:26 1993^ Page -1-mAU IGY-0230.D: ADC CHANNEL AC:,—cv210 cvCARBONIC ANHYDRASE200190180170160N—150Time -> 5.00^10.00^15.00 20.00 25.00LC 1^ Thu Apr 29 17:34:19 1993^ Page -1-270-260-250:.240,230-,220-2101,IGY-0221.D: ADC CHANNEL A(3- LAcroaLoBuLIN200190 -1802170-16011501Time 5.100^10.00 15.00^20.00 25.00^1Appendix CEconomics of IgY PurificationFrom an economic standpoint, separations of biological molecules are significant in thatthey account for 30 to 50% of bio-processing costs - possibly the largest discrete segment inthe manufacture of biomolecules (DePalma, A., 1993). The following is a brief discussionof some of the costs that would influence the feasibility of large-scale production.The economic analysis of the purification of IgY is dependent on a number of factorsincluding raw material, equipment, operating and labour costs. As well, the market (orlack thereof) for the purified product cannot be neglected, since the most technologicallysound and efficient process is of no use if the end product is not in demand. However,the market demand for IgY of various purities is uncertain at present, and until it hasbeen approved by organizations such a the U.S. Food and Drug Administartion for usein food or medicines, it is unlikely that there will be a large enough demand to makelarge-scale processing feasible in North America.Based on a market fax' the cost for a tanker load of liquid unpasteurized yolk wasrunning at about 41.5 cents US per pound in mid 1992. Assuming a density of separatedyolk of 1 0 mg/ml, a US/Canadian exchange rate of 1.282 and an IgY concentration of8 mg/ml, the raw material cost would be approximately $0.15 Canadian per gram of IgY.With an overall recovery for the purification process of 50%, the material cost of finishedproduct would be $0.30 per gram.The estimation of equipment and operating costs requires a detailed calculation thatlUrner Barry Publications, dated June 23, 1992.2The approximate rate as of the end of 1992.147Appendix C. Economics of IgY Purification^ 148is beyond the scope of this work. However, the equipment required would be similar tothat found in most food processing plants (stainless steel tanks of various dimensions,centrifugal and/or positive displacement pumps, filters, etc.) and if IgY were to be anadditional product for a company that was already producing several other food products(the most likely case), then this equipment might be shared. This would help to offsetthe cost somewhat. A significant operating cost could be attrition of the exchange media.Company literature from Gibco CEL, New Zealand, suggests that their exchangers havea low attrition rate, and in the case of whey protein recovery, an exchanger based onthe same solid matrix as HC-2 (S-2) has been used repeatedly more than 18,000 times(Smith et al, 1986). However such claims would have to be verified for use with HC-2under the conditions of separation of IgY. Utilities costs should be reasonably low forthe process developed in this work since energy intensive operations have been avoided.Labour is likely to represent another major cost in the separation and purification ofIgY, and based on the experience of a local biotechnology firm, would probably accountfor a larger percentage of the overall costs involved than equipment and operating costs.If IgY were to be prepared on a large scale (ie. metric tonne quantities per year)for use in a food product such as infant formula, then a purity of 40 to 50% mightbe quite acceptable'. This would significantly reduce the cost of separation as the saltprecipitation step could be avoided resulting in substantial savings in chemical costs andlabour. For the highly purified product, the costs would be somewhat higher, but couldeasily be offset by the much higher price generally paid for products of >99% purity.3The major contaminant, a-livetin, would have to be studied with regards to infant formula to verifythat it too would not cause an allergic response in infants.


Citation Scheme:


Citations by CSL (citeproc-js)

Usage Statistics



Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            async >
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:


Related Items