Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Enzymatic harvesting of glycosyl phosphatidylinositol anchored recombinant proteins from mammalian cells Sunderji, Rumina 1994

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

Item Metadata


831-ubc_1994-0149.pdf [ 2.12MB ]
JSON: 831-1.0058606.json
JSON-LD: 831-1.0058606-ld.json
RDF/XML (Pretty): 831-1.0058606-rdf.xml
RDF/JSON: 831-1.0058606-rdf.json
Turtle: 831-1.0058606-turtle.txt
N-Triples: 831-1.0058606-rdf-ntriples.txt
Original Record: 831-1.0058606-source.json
Full Text

Full Text

ENZYMATIC HARVESTING OFGLYCOSYL PHOSPHATIDYLINOSITOL ANCHOREDRECOMBINANT PROTEINS FROM MAMMALIAN CELLSbyRUMINA SUNDERJIB.Sc. (Eng), University ofDar Es Salaam, 1986A THESIS SUBMITTED IN PARTIAL FULFILLMENT OFTHE REQUEST FOR THE DEGREE OFMASTER IN APPLIED SCIENCEinTHE FACULTY OF GRADUATE STUDIESDepartment of Chemical EngineeringWe accept this thesis as conformingto the required standardTHE UNIVERSITY OF BRITISH COLUMBIAFebruary 1994© Rumina Sunderji, 1994In presenting this thesis in partial fulfilment of the requirements for an advanceddegree at the University of British Columbia, I agree that the Library shall make itfreely available for reference and study. I further agree that permission for extensivecopying of this thesis for scholarly purposes may be granted by the head of mydepartment or by his or her representatives. It is understood that copying orpublication of this thesis for financial gain shall not be allowed without my writtenpermission.(Signature)Department of CcLThe University of British ColumbiaVancouver, CanadaDate 9 ClDE-6 (2/88)11AbstractControlled release of recombinant proteins from mammalian cells enables protein productharvesting at increased concentrations and purity by separating protein expression fromthe protein recovery. The chinese hamster ovary (CHO) cell line investigated wasgenetically engineered to express glycosyl phosphatidylinositol (GPI) anchored humanmelanoma antigen p97 on the outer surface of the cell membrane. At intervals the cellswere treated with a phosphatidylinositol phospholipase C (PT-PLC) harvest solution toselectively cleave the GPI anchor and recover the protein at high concentration and purity(cyclic harvesting).The growth of the recombinant CHO cells was investigated in the serum-freemedia CHO-S-SFM I, CHO-S-SFM II, HBCHO, DMEMJF12 and Ham’s F12.CHO-S-SFM II supplemented with DNase and CHO-S-SFM I achieved single suspensioncells at high densities (approximately 6x10 cells/mL). CHO-S-SFM I was selected for allfurther investigations.A repeated harvesting technique which involved re-using the PT-PLC enzymesolution to harvest separate batches of cells was investigated. This approach furtherincreased the concentration of the desired protein product after each harvest. Preliminaryrepeated harvesting experiments recovered 140 jig/mL p97 from 7 harvests of 1 O cellseach. The first harvest recovered approximately 60 jig/mL p97, therefore theoretically 7harvests should have recovered 420 Since this process did not achieve theexpected concentrations, the stability of p97 and PT-PLC were investigated. However,p97 was found to be stable at 37 °C in the harvesting medium (PBS with I mg/mL BSA)for 24 h. Since a suitable assay for P1-PLC was not available, a new PT-PLC assay wasdeveloped. PT-PLC was stable at 37 °C for a period of 14 days. PT-PLC and p97 werealso stable in the pH range of 6.0 - 7.9. Instability of the proteins involved in the111harvesting process was not the cause for the low product concentrations recovered fromrepeated harvesting.Loss of P1-PLC due to adsorption to the cells was studied. An equilibrium wasestablished between P1-PLC in solution and P1-PLC adsorbed to the cell surface within3 minutes. Since PT-PLC was adsorbed to each batch of cells in the repeated harvestingprocess, enzyme was removed with the cells after each harvest. Therefore loss of PT-PLCby adsorption was considered the cause for the reduced protein recovery.The repeated harvesting process was repeated with PT-PLC replenishment aftereach harvest. Addition of 30 or 300 mU/mL PT-PLC recovered respectively 294 or343 ig/mL p97 from 5 consecutive harvests. The estimated purity of p9’7, based on totalprotein, was approximately 30 %.A continuous harvesting process was also investigated. This approach involvedaddition of PT-PLC to the growth medium resulting in the continuous release of p97 intothe medium. The continuous harvesting process was carried out simultaneously with 0, 3and 30 mU/niL PT-PLC. The 0 mU/niL PT-PLC control produced approximately3.8 ig/mL p97 in a batch culture of 11 days. During the same period of time cultureswith 3 and 30 mU/niL P1-PLC yielded 11.6 and 15.3 ig/mL p97 respectively.For the continuous harvesting process at 30 mU/niL PT-PLC the p97 productivitywas 5. 75x 1 0 fig/cell-day or approximately 2-fold higher than that achieved by the cyclicand repeated harvesting processes. However, the repeated harvesting process usedapproximately 10 times less PT-PLC and recovered 20-fold higher p97 concentrations.ivTable of ContentsAbstract iiTable of Contents ivList of FiguresList of Tables ixAcknowlegdementsCHAPTER 1Introduction ICHAPTER 2Literature Review 42.1 Mammalian recombinant protein production 42.1.1 Conventional protein production processes 52.1.2 Regulated secretion 62.1.3 Enzymatic harvesting of membrane bound proteins 72.2 Application of model harvesting system 92.3 Melanotransferrin 92.4 Glycosyl phosphatidylinositol anchors 112.4.1 Structure of GPI anchor 122.4.2 Attachment of protein to GPI anchor 142.4.3 Functions of OPT anchors 142.4.4 Functions of OPT-anchored proteins 162.5 Phosphatidylinositol phospholipase C 192.5.1 Properties of P1-PLC from B. thuringiensis 192.5.2 P1-PLC assays 212.6 Protein adsorption 23CHAPTER 3Materials and Methods 243.1 Cell line 243.2 Tissue culture 243.3 P1-PLC production 253.4 Production of monoclonal antibody against p97 26V3.5 Analytical methods.263.5.1 Cell count 263.5.2 Flow cytometry 273.5.3 Immunofluorescence assay 273.4.4 Glucose analysis 293.4.5 P1-PLC assay 313.4.5.1 P1-PLC assay based on flow cytometry 313.5.4.2 P1-PLC assay based on immunofluorescence 333.5 Experiments 333.5.1 Growth profiles for CHO cells 333.5.2 Repeated harvesting 343.5.3 p97 stability 343.5.4 P1-PLC stability 363.5.5 Adsorption of PT-PLC on cells 363.5.6 Continuous harvesting 37CHAPTER 4Results and Discussions 394.1 Growth media for CHO cells 394.2 P1-PLC assay 464.2.1 Effect of harvesting conditions on p97 removal 464.2.1.1 Incubationtime 464.2.1.2 Enzyme volume 464.2.2 P1-PLC assay based on flow cytometry 474.2.3 P1-PLC assay based on immunofluorescence 494.2.4 Reliability of P1-PLC assay 504.3 Repeated harvesting 514.3.1 p97 stability 534.3.2 P1-PLC Stability 544.3.3 P1-PLC adsorption 564.3.4 Non-specific binding of P1-PLC 614.4 Continuous harvesting 664.4 Comparison of protein production processes 70CHAPTER 5Conclusions 73CHAPTER 6Future Work 75Abbreviations 76References 77viAppendix IGrowth of CHO cells in different growth media 85Appendix 2Variation of initial cell density 89Appendix 3Effect of harvesting conditions on p97 removal 92Appendix 4P1-PLC assay results 93Appendix 5Repeated harvesting process results 94Appendix 6p97 and PT-PLC stability 96Appendix 7Adsorption and desorption of PT-PLC to the cell surface 98Appendix 8Continuous harvesting process results 100Appendix 9Calculations for p97 yield on glucose and P1-PLC 102viiList of Figures2.1 A schematic of addition of hydrophobic sequence to a non-GPI-anchoredprotein 102.2 p97 attached to the outer cell membrane by a GPI anchor 132.3 A schematic model illustrating addition of protein to the GPI anchor in theendoplasmic reticulum 152.4 Hydrolysis of P1 into diglycerides and myo-inositols (Ikezawa, 1986) 173.1 Schematic showing the procedure of sample preparation for measuringthe cell surface p97 283.2 A schematic of the reaction taking place in the wells 303.3 A schematic for the sample preparation procedure for PT-PLC assays 323.4 The schematic of the repeated harvesting process 354.1 Growth profile of CHO cells growing in CHO-S-SFM I. The figureshows the cell concentration (A), glucose concentration (El) andviability of the cells (V) 394.2 Growth profile of CHO cells growing in Ham’s F12 with 5 ig/mL ofinsulin, 5 p.g/mL transferrin, 10 nM sodium selenite, 50 .tg/mL BSA and50 j.ig/mL DNase. The figure shows the cell concentration (A), glucoseconcentration (El) and viability of the cells (V) 404.3 Growth profile of CHO cells growing in Ham’s F12 with 5 j.ig/mL ofinsulin, 5 jig/mL transferrin, 10 nM sodium selenite, 300 .tg/mL BSAand 50 DNase. The figure shows the cell concentration(A), glucose concentration (LI) and viability of the cells (V) 404.4 Growth profile of CHO cells growing in DMEM!F12 with 50 jig/mLDNase. The figure shows the cell concentration (A), glucose concentration(LI) and the viability of the cells (V) 414.5 Growth profile of CHO cells growing in DMEMIF12 with 5 p.g/mL ofinsulin, 5 j.ig/mL transferrin, 10 nM sodium selenite, 50 g/mL BSAand 50 .Lg!mL DNase. The figure shows the cell concentration (A),glucose concentration (LI) and viability of the cells (V) 41viii4.6 Growth profile of CHO cells growing in HBCHO. This figure showsthe cell concentration (A), glucose concentration (LI) and viability of thecells (V) for a batch culture 424.7 Growth profile of CHO cells growing in CHO-S-SFM II. The figureshows the cell concentration (A), glucose concentration (LI) and viabilityof the cells (V) 434.8 Growth profile of CHO cells growing in CHO-S-SFM TI and 100 j.ig/mLDNase. The figure shows the cell concentration (A), glucose concentration(LI) and viability of the cells (V) 444.9 Effect of different incubation times on p97 harvest. Each sample of4x106 cells was treated with 200 jiL of PT-PLC at 37 °C 464.10 Effect of PT-PLC volume on p97 harvest. A sample of 4x106 cells wastreated with 50, 100, 200 and 400 iiL of P1-PLC solutions of5 mUImL. Incubation time of 1 h was used 474.11 Typical standard curve obtained by PT-PLC from Boehringer Mannheimat pH 7.5. 200 iiL of P1-PLC solution and approximately 4x106 cellsper sample were used. A cell incubation time of 1 h at 37 °C was used toobtain this curve 484.12 Typical standard curves obtained from assays based on flow cytometryand immunofluorescence. The open triangles represents p97 concentration(A) and the open squares represents p97 removal (LI). 200 tL of PT-PLCsolution was used to treat 4x106 cells. Incubation time of 1 h at 37 °Cwas used 494.13 Concentration of p97 recovered for 5 repeated harvests. Each harvestwas conducted with 108 cells for 30 minutes. P1-PLC enzyme (0.5 mL)at an initial concentration of 30 mU/mL was used. Open circles (0)represent the measured values and the solid circles (.) represent calculatedvalues which account for sampling dilution 514.14 Percent of p97 removal from the cells after each harvest. Approximately4x106 cell sample was taken from the cell pellet and labeled withthe fluorescent antibody 524.15 p97 stability at 37°C for a period of 48 h. Initial p97 concentration was9.65 in the immunoassay buffer. The error bars represent the standarderror mean 53ix4.16 P1-PLC stability at 37 °C. PT-PLC from the bacterial supernatant at52 mU/niL was used. 200 iiL of two dilutions of each sample were usedto treat 4x106 cells for 1 h at 37 °C 544.17 Effect of pH on P1-PLC activity and p97. 16 mU/mL P1-PLC wasused to harvest 4x106 cells for I h at 37 °C. PBS buffer with I mg/mLof BSA at pH = 6.0, 6.5, 7.0, 7.5 and 7.9 was used for these harvests.The open squares (El) stand for p97 concentration and the open triangles(t?) stand for p97 removal 554.18 Adsorption of P1-PLC (0) to the cell surface over a period of 60minutes. cells/mL of P1-PLC enzyme at 18 mU/mL. A control onadsorption (•) experiment was done with P1-PLC without cellsincubated at 37 °C for 1 h 574.19 Desorption of PT-PLC from the cell surface (0) in I mL of PBS.PT-PLC activity is less than 1 mU/mL. The cell concentration of5x107 cells/mL was used 584.20 Adsorption equilibrium of PT-PLC at 37 °C. Different enzymeconcentrations at 1 cells/mL were used to study adsorption 604.21 Cumulative p97 from repeated harvesting with PT-PLC replenishment.The initial harvesting was started with 0.5 mL of 300 mU/niL ofPT-PLC and then 25 !.IL of P1-PLC enzyme at 3000 mU/niL was addedafter the 5th and 6th harvests. Incubation time of 30 mm at 37 °C wasused for each harvest 624.22 p97 removal from the cell surface during the repeated harvestingprocess. The initial harvesting was started with 0.5 mL of 300 mU/mLof P1-PLC and then 25 .tL of PT-PLC enzyme at 3000 mU/mL wasadded after the 5th and 6th harvests. Incubation of 30 mm at 37 °Cwas used for each harvest 624.23 Percent p97 removal from the cell surface for 5 harvests lx 108 cellswas monitored. Harvesting solution of 300 (0) and 30 (zX) mU/mLwere used. Incubation of I h at 37 °C was used for each harvest 644.24 Open symbols represents cumulative p97 measured after each harvestand the solid symbols represents the calculated values of the p97 harvestwhich includes p97 lost due to sampling. This figure shows p97 recoveredfrom 300 (0) and 30 (A) mU/mL of P1-PLC. Incubation of 1 h at 37 °Cwas used for each harvest 64x4.25 Continuous harvesting of CHO cells in the growth media. The solidsymbols represent cell concentration and the open symbols stand forviabilities. Cultures with P1-PLC at 0 (LI) 3 (A) and 30 (0) mUImLof medium 664.26 Glucose concentrations during the continuous harvesting process with 0 (LI),3 (A) and 30 (•) mU/mL of PT-PLC enzyme 664.27 Fluorescence due to cell surface p97 during the continuous harvesting processwith 0 (LI), 3 (A) and 30 (0) mU/mL of P1-PLC enzyme 524.28 Cumulative p97 in the supernatant from the continuous harvestingprocess with PT-PLC at 0 (LI), 3 (A) and 30 (0) mU/mLof medium 67xiList of Tables2.1 Functions of GPI-anchored proteins found on the cell surface 184.1 Comparison of P1-PLC assay based on flow cytometry andimmunoflourescence assay Each sample was assayed in duplicateat 3 different concentrations 504.2 Stability of p97 in the harvesting solution of PT-PLC in PBS (1 mg/mLBSA). The samples were assayed at 3 different dilutions. The error barswere less than ± 1.5 .ig/mL 534.3 P1-PLC change in the supernatant after 15 mm. 1x108 cells/mL ofP1-PLC solutions at different concentrations were incubated at37°C 594.4 Comparison for transfected and untransfected (CHOWTB) CHOcells. Incubation time of 15 mm and I mL of P1-PLC solution wasused 614.5 Comparison of p97 harvests from different harvesting processes 70xliAcknowledgmentsI would like to thank my supervisors Jamie Piret and Malcolm Kennard for theircontinuous support in completing this thesis. I am grateful for the financial supportprovided by Aga Khan Foundation, Geneva.I would also like to thank all my co-workers at Biotechnology Laboratory,especially Marta for their support and help at all times. Last but not least, I would like tothank Jurgen for his constructive criticisms and for being a good friend.CHAPTER 1IntroductionMammalian cells are used for producing vaccine, therapeutic and diagnostic recombinantproteins. They are able to carry out post-translational modifications such as glycosylation,acetylation and proteolytic processing that is often necessary to produce functionalproteins. However, in many processes, the desired protein is secreted at lowconcentrations (i.e. 0.01 mg/mL) into the cell culture medium containing contaminatingproteins at much higher concentrations (i.e. 0.1 - 1 mg/mL). These contaminants consistof medium proteins which promote cell growth, other secreted proteins and proteinsreleased by cell lysis. Hence, extensive downstream processing is necessary to purify thedesired protein which results in losses of product and high process costs. The use ofserum-free media has reduced the contaminating protein levels, but these media often stillcontain proteins in higher concentrations than the product protein.The initial purity and concentration of the desired proteins can be increased bycontrolled release techniques (Sambanis eta!., 1990a; Kennard eta!., 1993). Thesetechniques separate the cell growth and protein expression from the harvesting of theproduct. A growth medium is used to provide nourishment for cell growth and proteinsynthesis. In the growth medium the protein product remains cell-associated. Theperiodic replacement of growth medium with harvesting medium releases the product intoa low protein containing buffer. This results in a more concentrated and purer product,hence decreasing downstream processing costs.In this thesis the controlled release technique developed by Kennard eta!. (1993)was investigated. The model recombinant protein system studied was the humanmelanoma tumor antigen p97 (melanotransferrin) expressed by chinese hamster ovary2(CHO) cells. The antigen p97 is a glycosyl phosphatidylinositol (GPI) anchored proteinexpressed on the outer cell membrane which can be harvested by a specific bacterialenzyme phosphatidylinositol phospholipase C (PT-PLC). p97 is useftul as a potential anticancer vaccine (Hu eta!., 1988, Estin et a!., 1988; 1989), Production of p97 by thecontrolled release method represents a model system that can be applied to other GPIanchored proteins such as Leishmania gp63 and Malaria gp42 vaccines and alsoartificially made GPI-anchored proteins (Lin eta!., 1990; Moran and Caras, 1991;Scallon et a!., 1992).The aim of this project was to explore different harvesting techniques to improvethe controlled release process developed by Kennard et a!. (1993). Before studying theharvesting processes a number of different types of serum-free media were investigated(Section 4.1) to obtain high density single CHO cell suspensions to facilitate thesubsequent cell analysis.The alternative harvesting processes investigated were (i) repeated and(ii) continuous harvesting. (i) The repeated harvesting technique (Section 4.3) involvedmultiple harvesting of cells using the same enzyme solution for up to 10 separate batchesof cells. This process increased the recovered concentration of the desired proteinproduct. (ii) The continuous harvesting technique (Section 4.4) involved harvesting ofthe product during cell growth and protein expression. PT-PLC was added to the growthmedium and p97 released into the culture continuously. In this harvesting process thecells grew in the presence of PT-PLC, allowing p97 cleavage as soon as it was expressedon the outer cell membrane. This approach increased the cell specific p97 productivity.To help understand the performance of the new harvesting processes the stabilityof p97 (Section 4.3.1) and PT-PLC were investigated. Available PT-PLC assays werelaborious and not sensitive enough for our needs. Hence, an improved PT-PLC assay wasdeveloped (Section 4.2). The stability of P1-PLC (Section 4.3.2) at 37 °C wasdetermined. The effect of pH (Section 4.3.2) on p97 harvest was also investigated.3Phospholipases tend to bind non-specifically to cell surfaces (Deems et al., 1975).The PT-PLC adsorbed to the cell would be lost after each harvest thus depleting theenzyme in the harvesting solution. Hence, P1-PLC adsorption to the cells at 37 °C wasstudied (Sections 4.3.3 and 4.3.4). Finally, the performance of the cyclic, continuous andrepeated harvesting techniques were compared (Section 4.5).4CHAPTER 2Literature Review2.1 Mammalian recombinant protein productionMammalian cells have been used to produce viral vaccines against smallpox, rabies (before1930s), yellow fever (1930) and poliomyelitis (1949). In the 1970’s, recombinant DNAtechnology was developed and enabled the expression of mammalian proteins in bacterialcultures. The simplicity of culturing bacteria provided an attractive alternative tomammalian cell culture. Bacterial media are usually based on single carbon energysources and are less costly than the media needed by mammalian cells, which containexpensive growth factors and serum. A typical bacterial growth doubling time is15 - 30 minutes as compared to 15 - 30 h for mammalian cells. This allows higherproductivities in bacterial cultures (approximately 100 times). However, bacterial cells areunable to carry out many of the post-translational modifications of mammalian cellproteins. These modifications include proteolytic cleavage and addition reactions such asglycosylation or carboxylation that are essential for the correct biological functions ofmany proteins. For example, glycosylation could protect a protein against proteolyticbreakdown, hence maintain its structural stability (Butler, 1987). Many recombinantmammalian cell proteins are secreted into the medium and can be recovered and purifiedfrom the spent medium. The recovery of secreted proteins was more convenient than thecommonly used recovery of protein from lysed bacterial cells. Lysing the cells resulted inextensive contamination of the protein of interest and could also release endotoxins. Forthese reasons, the use of mammalian cells for recombinant protein production has beenexpanding rapidly.5Most commercially produced proteins and glycoproteins are used as diagnostic,therapeutic or veterinary products and hence need extensive purification. Productssecreted from mammalian cells include erythropoietin, tissue plasminogen activator andmonoclonal antibodies. Extensive downstream processing can lead to high productioncosts. Therefore, a careful study and optimization of the production processes isimportant, so as to meet the demand at affordable costs.2.1.1 Conventional protein production processesThe most commonly used mammalian cell protein production processes involve secretionof proteins into the culture medium. The cells are normally grown in stirred tankbioreactors and the proteins recovered from the spent media. The protein productivitiesof mammalian cells are low and the protein is secreted at low concentrations (usuallyI - 100 .tg/mL). Spent media contains many other contaminating proteins often atrelatively higher concentrations, resulting in low purity of the desired protein.Different strategies have been implemented in attempts to increase bioreactorproductivity and initial protein product concentration. For example:(i) Growing adherent cells on microcarriers. Microcarriers are able to achieve higher celldensities, which result in higher productivities.(ii) Gene amplification. The number of gene copies per cell is increased, resulting inincreased RNA and recombinant protein production.However, there has been limited success in increasing the protein product concentrations.Since, the major use of proteins produced by mammalian cells is for medicalpurposes, it is important to achieve high purity. This results in extensive purification andhigh production costs. A typical purification process consists of the following steps(Bailey and Ollis, 1986):6(i) Removal of solid particles such as cells. This may require one of the followingoperations: filtration, centrifugation or sedimentation.(ii) The protein is then isolated and concentrated by one of the following operations:solvent extraction, precipitation or ultrafiltration.(iii) The protein is further purified by either fractional precipitation, chromatography oradsorption.(iv) The final purification step involves recovering the product in the form required by theconsumer. This usually requires crystallizing and drying.To attain higher protein concentrations and optimize the production process,controlled release techniques have been explored. There are two types of controlledrelease techniques that have been studied so far: (i) regulated secretion of intracellularproteins (Sambanis et a!., 1990a) and (ii) enzymatic harvesting of membrane boundproteins (Kennard eta!., 1993).2.1.2 Regulated secretionThe use of regulated secretion for protein production was first reported in 1990 bySambanis et at. Mouse AtT-20 cells can store secretory proteins intracellularly insecretory vesicles and release them when stimulated by an inducer. This cell line wasgenetically engineered to express recombinant human insulin and growth hormone.Secretion of these proteins was induced by 8-bromo cyclic AMP. In a cyclic secretionprotocol the cells were exposed alternatively to growth and secretion medium. Secretionwas induced in low protein medium The secretion rates of human insulin were increasedup to 6-fold during the induction phase. However, this increase was only observed forthree cycles after which the cells started detaching from the culture surface and inducedsecretion decreased (Sambanis et at., 1990a). Induced secretion increased the productionof human growth hormone approximately 4-fold. Up to 60 % of the total cellular7production of human growth hormone was recovered during the induction phase. Theinduced secretion rates decreased after the first five cycles (Sambanis eta!., 1990b).Efforts also were made to apply this technique to a f3TC3 cell line. Signal transductionpathways were manipulated to reduce secretion of insulin during the production processby adding cyclic AMP. However, this decreased the expression of insulin (Gramp eta!.,1992).2.1.3 Enzymatic harvesting of membrane bound proteinsAn alternative production process based on glycosyl phosphatidylinositol (GPI) anchoredmembrane proteins was developed by Kennard eta!. (1993). A CHO cell line geneticallyengineered to express the GPI-anchored protein, p97, on the outer cell membrane wasused as a model system. p97 was harvested by PT-PLC, a specific enzyme that cleaved theGPI anchor releasing p97 into the medium. The harvesting medium consisted ofphosphate buffered saline (PBS) containing 10 mU/mL (0.02 Ig/mL) of P1-PLC enzyme.Over 35 jig/mL of p97 at 30 - 40 % purity was repeatedly recovered from 5x107 cells/mL,harvested in a cyclic fashion over a period of 44 days. The contaminating proteins arebelieved to be mainly other GPI-anchored proteins released from CHO cell surface. Afterharvesting, the cells were returned to fresh medium and they re-expressed the proteinwithin approximately two days. The cells could then be reharvested. The repeatedharvesting of the cells did not affect the growth rate, viability or protein production of thecells. This controlled release technique was developed using a suspension CHO cell line.The repeated centrifugation and washing of the cells required during the harvestingprocess was considered impractical for industrial scale protein production. To facilitatemedium changes harvesting from surface attached CHO cells was investigated (Kennardand Piret, in press).8Initially harvesting p97 from adherent CHO cells cultured in T-flasks was studied.Only 1.5 - 4 .tg/mL of p97 was harvested from these cells (Kennard and Piret, in press).There were two reasons for these low p97 harvests: (i) relatively large volumes ofPT-PLC solution were required to cover the entire surface area containing the cells for theharvesting process. (ii) after achieving 100 % confluency, the cells tended to lift off thesurface decreasing the total number of cells available for harvest. Similar culture stabilityproblems were encountered by Sambanis eta!. (1990a; 1990b). To overcome culturestability problems, the adherent CHO cells were grown on porous microcarriers.Microcarriers have many advantages over conventional culture methods such as: (i) Theycan be used in packed-bed and fluidized-bed bioreactors; (ii) Higher cell densities can beachieved, resulting in higher productivities; (iii) Cells are protected from mechanicalagitation and sparging and shear stresses; (iv) Cells can be maintained at reduced serumlevels. This system resulted in high cell densities and stable cultures. Of particularimportance for the controlled release process, rapid sedimentation of the microcarriersfacilitated easy handling when replacing medium or washing the cells for the controlledrelease process. The concentration of the product recovered increased to approximately100 tg/mL at 25 - 30 % purity. Stable protein production was attained for 15 harvestcycles over a 30 day period. Besides being expensive, the serum necessary for cell growthcontributes to trace contamination of the product. Therefore, efforts were made tomonitor the effect of reduced serum levels in the growth media on harvested p97.Decreasing serum from 10 to 0.5 % did not effect the cell specific p97 production ofporous microcarrier immobilized cells (Kennard and Piret, in press).92.2 Application of model harvesting systemThe technique of controlled release harvesting of GPI-anchored proteins could be used fora wide range of proteins of interest. There are many naturally occurring proteins such asthe potential vaccines Leishmania gp63 and Malaria gp42 that have medical importance.It was determined that the GPI pre-anchor recognition sequence was containedwithin the 37 amino acids at the COOH terminus (Caras et at., 1987a,b). Addition of thissequence to a secretory protein resulted in GPI anchoring and targeting of the fusionprotein to the plasma membrane (Caras et at., 1 987b). Recent work has shown that a pairof amino acids serving as a cleavage/attachment site positioned 10-12 residues from theNH2 side of the pre-anchor sequence were required for GPI anchoring(Moran and Caras, 1991). Figure 2.1 shows the process of a non-GPI-anchored proteinfused with the hydrophobic pre-anchor sequence from a naturally occurring GPI-anchoredprotein. This produces GPI-anchored fusion proteins. Human growth hormone is anexample of the secretory proteins which have been targeted to the plasma membrane asGPI-anchored proteins by using a gene fusion from GPI-anchored decay acceleratingfactor (Caras and Weddell, 1989; Lisanti et al., 1989).2.3 MelanotransferrinMelanotransferrin, the model protein used in this thesis, is also known as the melanomatumor-associated antigen p97. It is a cell surface glycoprotein (Molecular weight 97 kDa)expressed by human melanoma cells (Brown et al., 1981a), but present in only traceamounts in normal adults (Brown et at., 198 ib). The human melanoma, cell line SK-MEL28, expresses approximately 4x105 molecules of p97 per cell (Brown et a!., 1981a), whilethe CHO cell line, genetically engineered to express p97 produces approximately 2 - 5x10610Non GPI anchoredprotein of interestGPI anchored proteine.g. Decay acceleratingfactor (DAF)I DAFI IHydrophobicsignal sequenceI PP - DAF fusion proteinFigure 2.1 A schematic of addition of hydrophobic sequence to a non-GPI-anchoredIIprotein11molecules per cell Although the amino acid sequence of p97 is approximately 40 %identical to human transferrin and lactoferrin (Brown et a!., 1982), unlike these proteins itis membrane bound (GPI-anchored) and differs in amino acid residues which are involvedin iron binding. It has been proposed that p97 has an intact transferrin type iron bindingsite in its N-terminal domain and a possibly defective site in its C-terminal domain(Baker eta!., 1987). Baker eta!. (1992) confirmed that there was only one iron bindingsite present on p97. Rose et a!. (1986) reported that p97 played a role in irontranslocation while Richardson eta!. (1991) refuted this observation.Since most human melanomas express high levels of GPI-anchored p9’7, itseffectiveness as an anti-cancer vaccine has been explored (HellstrOm and HelistrOm, 1969;Hu eta!., 1988). Immunization with a recombinant vaccinia virus, v-p97NY, did inducehumoral and cell-mediated immunity against melanoma-associated antigens in monkeys(Estin eta!., 1988).2.4 Glycosyl phosphatidylinositol anchorsIn eukaryotic cells, many other proteins are anchored to the external surface of the plasmamembrane by covalently attached glycolipids containing inositol. These anchors areknown as glycosyl phosphatidylinositols (GPI).In 1963, Slein et a!., noted that alkaline phosphatase on mammalian cell surfacecould be released by PT-PLC and proposed the presence of GPI anchors. Almost 15 yearslater, it was confirmed that mammalian cells have GPI-anchored proteins attached to theplasma membrane (Ikezawa et a!., 1976; Low & Finean, 1977). Due to low expressionlevels of the GPI-anchored proteins such as alkaline phosphatase and acetylcholinesterase(i.e. approximately 1,000 - 10,000 molecules per cell) it was difficult to study the structureand release mechanisms of the anchor. In the mid 1980’s the Tiypanosonza brucei variantsurface glycoprotein (VSG) was found to be anchored to the cell surface by GPT at12approximately io molecules per cell and thus could be more easily purified(Ferguson eta!., 1985). The structure and release mechanisms of GPI anchors weredetermined by the late 1980’s (Ikezawa, 1986; Low, 1989).2.4.1 Structure of GPI anchorThe GPI anchor consists of a phosphoethanolamine and a variable glycan portion. Theglycan moiety of the anchor varies with cell type. The following is the breakdown of theGPI anchor assembly (Figure 2.2):(i) The protein is covalently bound to the GPI anchor via the x-carboxyl of theC-terminal amino acid. This C-terminal amino acid is not specific for GPIanchoring.(ii) The ct-carboxyl group of the C-terminal amino acid is amide linked to the aminogroup of a phosphoethanolamine moiety.(iii) The phosphoethanolamine is linked to a variable glycan section that consistsmainly of mannose and glucosamine.(iv) The glycan portion is glycosidically linked to an inositol-containing phospholipid,phosphoinositol.(v) Phosphatidylinositol anchors the protein to the membrane.PT-PLC cleaves the GPI anchor at the phosphodiester bond between the phosphoinositolgroup and the lipid portion (Figure 2.2). The protein is released into the medium in asoluble form with the C-terminal linked to ethanolamine, the glycan moiety andphosphoinositol. Diacylglycerol is presumably left in the membrane (Ikezawa, 1986).13NH2p97CH2—CH CH,0 0MEMBRANE(Cr (CHCH CHNHI EthanolamineCH,CH,QE—Q0Inositol0P1-PLCcleavage site0Figure 2.2 p97 attached to the outer cell membrane by a GPI anchor142.4.2 Attachment of protein to GPI anchorThe attachment of GPI anchor to the protein is a rapid post-translational process(Low, 1989) occurring soon after protein synthesis. This process occurs in theendoplasmic reticulum, and not in the Golgi apparatus or plasma membrane (Low, 1987;Doering et at., 1990; Boivin and Delaunay, 1991). A schematic model of protein additionto GPI is shown in Figure 2.3. The protein is synthesized with a pre-anchor sequence,which contains the cleavage/signal site. The GPI anchor is synthesized independently inthe endoplasmic reticulum. Part of the cleavage/signal sequence is cleaved and replacedby the pre-assembled GPI anchor (Ferguson and Williams, 1988). This replacement leavesbehind a short C-terminal hydrophobic domain of the pre-anchor sequence(Lisanti eta!, 1990; Doering et a!., 1990). The rapid addition of the anchor suggests thatthe removal of hydrophobic C-terminal domain and addition of GPI may be catalyzed bythe same transamidase enzyme (Doering et a!, 1990).2.4.3 Functions of GPI anchorsThere are no apparent common functions of GPI-anchored proteins that could help explainthe presence of the anchor compared to the more common transmembrane peptidesequences. The following are a few of the possible functions of GPI anchors which havebeen suggested:(i) Protein motility: GPI anchored proteins have about ten-fold increased lateral mobilityand diflujsion coefficients on the order of 1 - 4x 1 0 cm2lsec have been measured(A. Ishihara et al., 1987; Low, 1987; Boivin and Delaunay, 1991). Transmembraneglycoproteins have diffusion coefficients of 0.5 - 6x 10b0 cm2/sec (Gall and Edelman,1981).1510cc-d0zI100z+r-)I0+-cFigure 2.3 A schematic model illustrating addition of protein to the GPI anchor in theendoplasmic reticulum.16(ii) Protein removal: GPI anchors regulate the removal of the protein based on theirsusceptibility to mammalian phospholipase C enzymes (Boivin and Delaunay, 1991). Byan endocytotic pathway, GPI-anchored proteins are recycled into compartmentscontaining cellular P1-PLC. PT-PLC cleaves the protein, which is then either degraded byproteases or recycled back to the membrane and released into the supernatant(Ferguson and Williams, 1988).(iii) Protein secretion and sorting: GPI anchors can play a specific role in translocation ofproteins across membranes. Studies have shown sorting of GPI-anchored proteins to theapical membrane of polarized epithelial cells (Cross, 1990).(iv) Cell protection: The glycan group lies along the plane of the membrane, hence itcould form a diffi.ision barrier which protects the cells from foreign materials(Boivin and Delaunay, 1991).(v) Cell signaling: GPI anchors were found to be involved in signaling events evoked byproteins such as insulin (Low and Saltiel, 1988). These proteins triggered the productionof GPI specific phospholipase C (mammalian PT-PLC). Action of P1-PLC onphosphoinosides results in diacylglycerols (Figure 2.4), an activator of protein kinase Cand inositol phosphates. Protein kinase C and inositol phosphates evoke specificresponses to insulin (Boivin and Delaunay, 1991).2.4.4 Functions of GPI-anchored proteinsDuring controlled release harvesting of p97, P1-PLC cleaves other GPI-anchored proteinsfrom the cell membrane. Hence, it was important to review the functions of GPI-anchoredproteins, to evaluate the potential impact of harvesting on cell function. These proteinshave been identified and characterized according to their biochemical functions. Thebiological functions of most of the proteins are still unknown.0 Cl) Cl) 0 0 0-C) CD -t 0 CD Cl) 0-myo-inositol-1,2-cyclicphosphateOHPhosphatidylinositolCH,OCOR0HCOCOR’El—p-OCH.OHPT-PLC-P-OHOR9 0 Cl) 0 Cl) CD ICH,OCORHCOCOW±CH,OH0O--OHOHDiglyceridesmyo-inositol-1-phosphate18Table 2.1 Functions of GPI-anchored proteins found on the cell surfaceProtein Source Function ReferenceAlkaline phosphatase Mammalian tissues Hydrolase Low and Finean, 1977;Taguchi and Ikezawa,19785’ - Nucleotidase Mammalian tissues Hydrolase Low and Finean, 1978;Shukia eta!., 1980Acetylcholinesterase Mammalian blood Hydrolase Low and Finean, 1977;cell Low eta!., 1987Alkaline phospho- Rat tissues Hydrolase Nakabayashi and Ikezawa,diesterase I 1984, 1986Variant surface Trypanosorna Protective Ferguson et al., 1985;glycoprotein Brucei coat Low et al., 1987Thy-I Mammalian brain Antigen Low and Kincade, 1985;and T lymphocytes Tse eta!., 1985Trehalase Rabbit tissues Hydrolase Takesue et aL, 1986Decay accelerating Human blood and Complement Davitz eta!., 1986, 1987factor HeLa cells regulatoryproteingp63 Leishniania major Protease Bordier et at., 1986;Etges eta!., 1986RT-6 Rat lymphocytes Antigen Koch eta!., 1986Qa Mouse T Antigen Stiernberg eta!., 1987lymphocytesThB Mouse Antigen Stiernberg et at., 1987lymphocytesT-cell activating Mouse T Antigen Reiser et at., 1986protein lymphocytesN-CAM120 Rat, mouse and Cell-cell He eta!., 1986;chicken brain interactions Hemperly et a!., 1986Heparan sulphate Rat liver Cell-cell and M. Ishihara et aL, 1986proteoglycan Cell-matrixinteractions19According to Turner (1990), over 50 GPI-anchored proteins are present on cell-surfaces.However, Kennard et a!. (1993) found that removal of these GPI-anchored proteins didnot effect the viability and growth of the cells.2.5 Phosphatidylinositol phospholipase CP1-PLC is produced by a number of bacteria including: Staphylococcus aureus(Doery eta!., 1965), Bacillus cereus (Slein and Logan, 1965), Clostridium novyl(Taguchi and Ikezawa, 1978) and Bacillus thuringiensis (Taguchi et a!., 1980).P1-PLC hydrolyzes phosphatidylinositol (P1) and lysophosphatidylinositol, butdoes not recognize other more common membrane phosphatidyls such asphosphatidyicholine, phosphatidylethanol amine and phosphatidylglycerol (Ikezawa andTaguchi, 1981). This enzyme specifically recognizes the inositol-phosphate structurepresent in PT, such that it does not hydrolyze even the more highly phosphorylatedderivatives of PT (for example P1-4-phosphate and P1-4,5 bisphosphate) (Kuppeeta!., 1989). P1-PLC cleaves most GPI anchors, however, some proteins with GPIanchors are partially or completely resistant to PT-PLC (Ferguson and Williams, 1988). Inthis study, Bacillus subtilis transfected with PT-PLC gene from Bacillus thuringiensis wasused to produce P1-PLC.2.5.1 Properties of P1-PLC from B. thuringiensisB. thuringiensis was found to be the highest producing bacterial source of PT-PLC.PT-PLC could be purified to a homogeneous state by polyacrylamide gel electrophoresis(Taguchi eta!., 1980).P1-PLC produced by B. thuringiensis has a molecular weight of approximately23 ± 1 kDa (Taguchi et a!., 1980; Ikezawa and Taguchi, 1981; Ikezawa, 1986). The20optimum pH is between 5 - 8.5 with maximal activity at 7.5 (Ikezawa and Taguchi, 1981).The isoelectric point (p1) is approximately 4.9 - 5.4 (Ikezawa and Taguchi, 1981;Kupke eta!., 1989; GrifflthetaL, 1991).P1-PLC stored in 20 mM Tris HC1 and pH 7.5 is stable after many freeze-thawcycles at 0.1 - 3 mg of proteinlmL and at 37 °C during prolonged incubations(Griffith eta!., 1991). Fresh bovine serum albumin (BSA) protects the enzyme at hightemperatures. P1-PLC diluted in 0.1 % fresh BSA retains greater than 60 % of its activitywhen exposed to temperatures from 70 to 100 °C for 10 minutes. However, when dilutedin freeze-thawed B SA, it became thermolabile and retained only 20 % of its activity in10 minutes at the same temperatures (Kume eta!., 1992).PT-PLC enzyme activity is inhibited by divalent metal ions such as Ca2, Mg2,J44fl2 and Zn2 at concentrations above 10 M (Low and Finean, 1976;Taguchi etal., 1980). This inhibition is pH dependent (Sundler eta!., 1978) and possiblydue to interactions with the substrate rather than the enzyme (Griffith eta!., 1991).Potassium chloride and sodium chloride are inhibitory at concentrations higher than 10*2 M(Sundler et al., 1978; Taguchi et aL, 1980; Ikezawa and Taguchi, 1981). Othercompounds that completely inhibit the enzyme activity are mercuric chloride (HgC12) at0.5 mM and p-chloromercuriphenyl sulfonic acid (PCMBS) at 5 mM. The inhibitoryactivity of HgC12 and PCMB S treated enzyme is completely restored by excess addition ofthe reducing agent dithiothreitol (DTT).Triton X-100 and sodium deoxycholate stimulate the activity of P1-PLC(Taguchi et al., 1980). P1-PLC cleaves GPT anchors more efficiently when treating cellsfrom a culture at high cell density (Berridge, 1987).212.5.2 P1-PLC assaysThe reported PT-PLC assays are based on monitoring the change in substrate or productconcentrations. One unit of P1-PLC is defined as the enzyme activity that hydrolyzes1 iimol phosphatidylinositol per minute at 37 °C and pH 7.5 to phosphoinositol anddiacylglycerol:PT-PLCPT > phosphoinositol + diacyiglycerol (2.1)There are 5 methods reported for determining PT-PLC activity:(1) The determination of water-soluble inositol phosphate from radiolabelledphosphatidylinositol (Griffith et al., 1991). This assay involves radiolabelling of rat livermicrosomes with[3H]inositol. Lipids extracted from these microsomes are used as asubstrate. Radiolabelled lipids are treated with PT-PLC at 37 °C and the reaction stoppedafter 10 minutes. Radioactivity of[3Hjinositol in the supernatant is determined andconverted to PT-PLC activity. This assay is time consuming and expensive.(2) Quantitation of phosphate released from phosphoinositol by the Eibl and Landsmethod (1969). In this assay phosphatidylinositol containing lipids from rat liver aretreated with PT-PLC at 37 °C. The reaction is stopped after 10 minutes and myo-inositol(product from the reaction above, Figure 2.4) is decomposed to phosphorous.Phosphorous is then measured by Eibl and Lands method. This is the most used PT-PLCassay. However, it is lengthy and laborious.(3) Quantitation of soluble GPI-anchored proteins from biological membranes (Ikezawaand Taguchi, 1981). This assay is based on proteins such as GPI-anchored alkalinephosphatase which can be quantified by their enzyme activity. In this assay, tissues from22rat kidneys are homogenized and treated with P1-PLC at 37 °C for 10 - 100 minutes. Thesupernatant is then assayed for alkaline phosphatase.(4) Continuous fluorometric assay using a fluorescent substrate, 2-naphthyl-myo-inositol- 1-phosphate (2NIP) or 4-nitrophenyl myo-inositol- 1-phosphate(Shashidhar eta!., 199 la,b). P1-PLC was added to the fluorescent substrate and thefluorescence monitored at 403 nm wavelength. This assay is simple and gives immediateresults. However, it can only measure activities above 14 mU/mL.(5) Boehringer Mannheim assay # 5646. P1 is treated with PT-PLC and diglycerideproduced from this reaction is reacted with lipase to produce glycerol. A series ofreactions are then triggered with glycerol kinase, pyruvate kinase and lactatedehydrogenase. Decrease in absorbance at 365 or 340 nm due to NADH depletion(Equation 2.6) is measured. Absorbance is then converted to P1-PLC activity. Thelimitation of this assay was its sensitivity, it could only measure activities above150 mU/mL.P1-PLCPT + H20 > Diglyceride + phosphorlinosine (2.2)LipaseDiglyceride + H20 > fatty acids + glycerol (2.3)Glycerol kinaseGlycerol + ATP > glycerol-3-P + ADP (2.4)pyruvate kinasePEP + ADP > pyruvate + ATP (2.5)23Lactate dehydrogenasePyruvate + NADH + W > lactate + NAD (2.6)2.6 Protein adsorptionCells adsorb proteins rapidly and non-specifically in less than a minute (Missirlis eta!.,1990; Shin eta!., 1993). Deems et at., (1975) and Shin et a!. (1993) proposed that thisadsorption of proteins to cell membranes was due to hydrophobic and not electrostaticinteractions. Phospholipase A2 binds to the surface before reacting with the substrate andthis binding is reversible. The enzyme then moves on the surface as an enzyme-surfacecomplex until it collides with a phospholipid (Deems et at., 1975). Deems et a!. alsosuggest that binding studies can be applied to other phospholipases and lipases.Transferrin is also a widely investigated molecule because of its thnction as agrowth stimulator. Studies on transferrin adsorption to a specific receptor resulted in asaturated state at approximately 48,000 molecules/cell (Reed, 1990).24CHAPTER 3Materials and Methods3.1 Cell lineThe wild type chinese hamster ovary (CHO) cell line CHOWTB (obtained from Dr.Maxfield, New York University) was co-transfected with the p97 expression vector,pSV2p97a and the G418 resistance vector, pWJ218 (Kennard et al., 1993). The entirecoding region of p97 cDNA was present in the pSV2p97a vector and was driven by theSV4O promoter (Food et al., 1993).3.2 Tissue cultureThe CHO cell line was maintained in suspension in serum free medium, CHO-S-SFM I(Gibco, Grand Island, N.Y). The cells were cultured in either 75 cm2 T-flasks (Nunc,Gibco) or 150 mL and 250 mL spinner flasks (Belico, Vineland, N.J.). The cultures wereincubated at 37 °C and under a 5 % CO2 humidified atmosphere.The CHO cells also were grown in CHO-S-SFM II (Gibco), HBCHO (IrvineScientific, Santa Ana, CA), Ham’s F12 (Gibco) and DMEMJF12 (Gibco), supplementedwith 50 - 100 Ig/mL of DNase (Boehringer Mannheim, Laval, Quebec), 5 tg/mL ofinsulin (Gibco), 5 jig!mL of transferrin (Gibco), 10 nM of sodium selenite (Gibco) and50 - 300 g/mL of bovine serum albumin (Sigma, St. Louis, MO).The cells stocks were stored in liquid nitrogen at approximately 1xL07 cells/mL in1 mL aliqouts of a solution of 72 % CHO-S-SFM I, 20 % newborn calf serum (Gibco)and 8 % dimethyl sulfoxide (Fisher Scientific, Fair Lawn, NJ.).253.3 P1-PLC productionBacillus subtilis (BG2320) transfected with the PT-PLC gene from Bacillus thuringiensiswas provided by Dr. M. Low of Columbia University, N.Y. The cells were maintained ina medium containing 10 g/L of polypeptone (Becton Dickinson, Cockeysville, M.D.),10 g/L of yeast extract (Difco, Detroit, Michigan), 0.4 g/L of potassium biphosphate(BDH, Toronto, ON), 5 g/L sodium chloride (Fisher Scientific) and 15 .tg/mL ofchioramphenicol (BDH) for selection. The pH adjusted to 7.0 with 1 N NaOH (FisherScientific).100 mL of medium was inoculated with Bacillus subtilis and grown overnight.This culture was then used to inoculate 2 L of medium at 2 - 3 % (v/v). The cultures weremaintained in an Erlenmeyer flask at 37 °C at 150 rpm in a shaker bath (Lab-lineInstruments, Meirose Park, IL.). After 12 h (- late log phase) the culture was centrifugedand the supernatant filtered using a 0.2 jtm membrane (VacuCap, Gelman Sciences). Topartially purify the P1-PLC, 500 mL of filtered supernatant was precipitated using 600 g/Lammonium sulfate (Baker, Phillipsburg, N.J.). The precipitate was separated bycentrifugation (Silencer, Japan) and the filtrate resuspended in 100 mL of 20 mM Tris HC1and 3 mM EDTA buffer. The mixture was then concentrated to 10 mL using a 30,000molecular weight cut off YM 30 membrane in an ultrafiltration cell (model 8400, Amicon).This was repeated two more times and the final 10 mL was diluted by 30 mL of the buffersolution. The enzyme solution was then aliquoted to I mL samples and stored at - 20 °C.For short term storage at 4 °C, the enzyme was diluted to the required concentrations inphosphate buffered saline (PBS) containing 1 mg!mL of BSA. PBS consists of 8.0 g/Lsodium chloride (NaCl), 2.16 g/L sodium hydrogen phosphate (Na2HPO4), 0.2 g/L ofpotassium hydrogen phosphate (KH2PO4)and 0.2 g/L of potassium chloride (KCI) withpH adjusted to 7.4.26The standard 77.8 U/mL P1-PLC solution (Boehringer Mannheim) was stored in50 mM triethanolamine buffer with 10 mM EDTA and 10 mM sodium azide.3.4 Production of monoclonal antibodies against p97The 33B6E4 monoclonal antibody was provided by Dr. M. Kennard. L235 hybridoma cellline (ATCC NB 8446 L235 (M-19) was grown in roller bottles (Nunc, Gibco). Themedium used to culture these cells was RPMI (Gibco) supplemented with non-essentialamino acids, 10 g/L fetal calf serum (Gibco), 1 g/L mercaptoethanol (Sigma), 2 mML-glutamine (Gibco), 2 mM proline (Gibco) and 0.1 mg/mL penicillinlstreptomycin. Thecell culture was maintained in a batch culture until the viability decreased to 50 %. Thecells were removed by centrifugation and the supernatant filtered using 0.2 tm membrane(Gelman Sciences). The monoclonal antibody was then purified using a protein G affinitycolumn (MAbTrap G, Pharmacia LKB Baie d’Urfe, PQ) and concentrated to 1 - 2 mg/mLusing 10,000 MW ultrafilter (Centricon-lO, Amicon Danvers, MA).3.5 Analytical methods3.5.1 Cell countThe haemocytometer (Hausser Scientific, U.S.A.) and trypan blue (8 g/L trypan blue stainand 8.8 g/L NaCI, Gibco) dye exclusion method was used to monitor the cell density andviability of the cultures. Viable cells excluded the dye while the non viable ones took upthe dye and were dyed reddish-blue. The haemocytometer was loaded with the samplediluted with 50 % trypan blue. The cells were then counted under the microscope (Nikon,Missisauga, ON).273.5.2 Flow cytometryImmunofluorescence labeling of p97 and a flow cytometer analyzer (FACScan, BectonDickinson, CA) were used to monitor the cell surface expression of p97.To prepare cells for flow cytometry (FACS) analysis, approximately 4 - 5x106 cellsper sample were spun in a 6 mL (12x75 mm) polystyrene tube (Becton DickinsonLabware, Lincoln Park, N.J.), the spent medium was removed and the cell pellet washed in200 iiL of FACS buffer. The FACS buffer contained 0.5 % (w/v) bovine serum albumin(Sigma), 20 mM Hepes (Sigma) and 20 mM sodium azide (Baker) in Dulbecco’s ModifiedEagle Medium (DMEM, Gibco). The cells were then spun again and the cell pelletresuspended in 100 j.tL of FACS buffer and labeled with fluorescinated (fluorescinisothiocyanate) antibody against p97 (33B6E4) obtained from Dr. M. Kennard (Section3.4). 2 i.iL of fluorescinated antibody (4 mg/mL) was added to the cells and incubated at4 °C for 45 minutes. The cells were then spun and the pellet washed with 1 mL of FACSbuffer followed by 1 mL of PBS. After spinning the cells again, the cells were fixed in1 mL of 1.5 % (v/v) p-formaldehyde (JBS, Pointe Claire-dorval, PQ) (Figure 3.1). 5000events per sample were measured by the flow cytometry and average fluorescence per cellobtained.3.5.3 Immunofluorescence assayA pandex fluorescence concentration analyzer (Idexx, Portland, ME) was used todetermine the concentrations of p97 solubilized by the action of PT-PLC.The coated capture particles and the labeled antibody were provided by Dr. M.Kennard. Carboxyl-polystyrene capture particles (0.87 tm diameter, 5 % w/v, Baxter,Mundelein, IL) were ãoated with anti-p97 IgG (ATCC HB8446 L235 (M-19)). A secondantibody to p97 IgG (33B6E4) was labeled with fluorescin isothiocyanate (FITC, Sigma).FLOWCYTOMETRYCD-CDC200jiLFACSBuffer5E6Ucellsl0ocIspinandremove1000Isupernatantoç1mLFACsBuffer0 CD 0 Cd C CD -t C 0 CD -t CD C Cd, CD-e 1 CD-e-t C C -t CD Cl) -t C CD 0 CDwash,spinandremovesupernatantspinandremovesupernatant100p.LFACSbuffer&2p.Lfluorescinatedantibodyap97(4mg/mL)incubateat4°Cfor45mm.1mLPBSwash,spinandremovesupernatantI0.5mLPBS+0.5mLp-formaldehyde(3%v/v)FACSCANANALYSISleaveovernightat4°Cwash,spinandremovesupernatant/00Fixthecells29The assay was performed in 96 well plates (Idexx). These plates have a membraneat the bottom which enables removal of liquids from the wells using vacuum filtration.The p97 samples were diluted in immunofluorescence buffer (DMEM with 1 g/L sodiumazide and 10 g/L bovine serum albumin) to within 0.15 - 1.5 tg/mL (linear range ofimmunofluorescence assay), 20 iiL was added to each well and incubated with 20 iiL ofthe coated capture particles for 20 minutes at room temperature (-- 20 °C). Then a further20 tL of the second fluorescinated antibody (50 .tg/mL) was added to each well andincubated for another 20 minutes at room temperature (Figure 3.2). The plate was thendrained and washed 3X with PBS within the pandex analyser and the fluorescence in thewells read using a 485/535 nm filter at lox gain. Standards with known p97concentrations ranging from 0.15 - 1.5 Ig/mL were used to produce a calibration curvefor each plate.3.4.4 Glucose analysisGlucose analyser 2 (Beckman Instruments, Fullerton CA) was used to measure glucoseconcentrations of the cultures. It can measure the glucose concentrations fromapproximately 0.1 - 4.5 g/L. Glucose measurement is based on monitoring the decrease inoxygen concentration. It uses an enzymatic glucose electrode to carry out the followingreaction:Glucose oxidaseJ3-D-glucose + 02 > Gluconic acid + H20 (3.1)All measurements were done in triplicates and the average of three taken as therepresentative value.Cl) C) CD C) 0 CD CD C) 0 C) CDIMMUNOFLUORESCENCEASSAY20iLsampleforp97analysisp9720tLfluorescinatedantibodyCLp97(50g/mL) 120mmatroomtemperatureCD CD20.tLcaptureparticles(0.25%wlv)coatedwithIgGCLp97\97p979JMeasurefluorescencePANDEXANALYSIS20mm.ICDrain313.4.5 P1-PLC assayP1-PLC assays developed were based on cleaving p97 from the cell surface andmonitoring the cell surface p97 using the flow cytometry or measuring the solubilized p97in the supernatant using the immunofluorescence assay. P1-PLC assay based on flow cytometryApproximately 4 - 5x106 cells were washed with FACS buffer and incubated with200 j.i.L of known and unknown samples of P1-PLC for 1 h at 37 °C. The cells were thenwashed and labeled with fluorescinated antibody according to the previously describedmethod for preparing cells for analysis by flow cytometry. The fluorescence of the cellsdue to remaining cell surface p97 was measured using the flow cytometry (Figure 3.3) andthe percentage removal of p97 calculated relative to positive (untreated cells labeled withfluorescinated antibody) and negative (untreated cells with no label) controls. Thenegative control gave the cell autofluorescence and the positive control gave thefluorescence due to autofluorescence plus cell surface p97. Percentage removal of p97was calculated as follows:% Removal ofp97 4 (positive control) - (unknown) 1 100 (3.2)[(positive control) - (negative control) JThe percentage removal of p97 was then converted to PT-PLC concentration usinga standard curve obtained from cells similarly treated with known enzyme concentrations.—.fIQPT-PLCASSAY000+spent°°mediumwash,spinandremovesupernatant200ilLFACsBuffer(knownorunknownsample)200tLofP1-PLCsolution5E6cells0 CD C-) CDspinandremovesupernatantincubateat37°Cfor1h. ISP:cells/supemant‘MMUNOASSJ333.5.4.2 P1-PLC assay based on immunofluorescenceApproximately 4 - 5x106 cells were washed with FACS buffer and treated with 200 iiL ofknown and unknown P1-PLC samples (same as flow cytometry assay), for 1 h at 37 °C.After the incubation, the cells were removed by centrifuging and the cell free supernatantassayed for p97 concentration (Figure 3.3) as described in immunofluorescence assaysection earlier. A standard curve of p97 in the supernatant against P1-PLC was obtainedfrom samples treated with known enzyme solutions. Unknown P1-PLC activities werethen obtained from the standard curve.3.5 Experiments3.5.1 Growth profiles for CHO cells7 different types of serum-free media were investigated and compared to the growthprofile of CHO-S-SFM I. Cell density, viability and glucose concentrations weremonitored over a period of 8 - 12 days. Media investigated are (i) Ham’s F12 withinsulin, transferrin, sodium selenite, 50 .tg/mL of BSA and 50 tg/mL DNase (ii) Ham’sF12 with insulin, transferrin, sodium selenite, 300 ig/mL of BSA and 50 .tg/mL DNase(iii) DMEM!F12 with insulin, transferrin, sodium selenite, 50 ig/mL of BSA and50 jtg/mL DNase (iv) DMEMJF12 and 50 .tg!mL DNase (v) HECHO(vi) CHO-S-SFM II (vii) CHO-S-SFM II with 100 tg/mL DNase.CHO cells were grown in CHO-S-SFM I in 250 mL spinner flask to approximately2x106 cells/mL. 25 mL of this culture was spun and resuspended in a mixture of twomedia (25 mL CHO-S-SFM I and 25 mL HBCHO) and maintained in a 75 cm2 T-flask.The reason for using the mixture of 2 media was to adapt the cells to the new medium.Slowly, the medium was replaced by 100 % HBCHO. 150 mL spinner flask containing34HBCHO was then inoculated by cells from this T-flask and its growth profile monitoredover a batch culture. Similar process was carried out to adapt the cells to Hams F12,DMEMJFI2 and CHO-S-SFM II.3.5.2 Repeated harvestingCHO cells were grown in CHO-S-SFM I in 250 mL spinner flasks to 4x106 cells/mL.25 mL of the culture (108 cells) in 50 mL centrifuge tube (Sarstedt, St. Laurent, PQ) wasspun and the cell pellet transferred to a 15 mL centrifuge tube (Sarstedt). The cells werewashed with 5 mL PBS and then resuspended in 0.5 mL of PT-PLC solution. In thisprocess, the same enzyme solution was used to harvest 5 - 10 separate samples of cells.Hence after 30 - 60 minutes incubation at 37 °C, the cells were spun and the supernatantused to treat a fresh sample of 10 cells. A 20 - 50 iL sample was drawn after eachharvest. These samples were spun and the cell free supernatant stored at - 20 °C and lateranalysed for p97 using the immunofluorescence assay. For preliminary experiments, thevolume of PT-PLC solution removed from sampling was replaced by PBS and for laterexperiments, it was replaced by concentrated P1-PLC (3000 - 300 mU/mL). Thisexperiment was carried out under sterile conditions.3.5.3 p97 stabilityp97 harvested from CHO cells was diluted in immunofluorescence buffer to approximately9.7 jig/mL. p97 solution was then sterile filtered by using 0.2 p.m membrane (VacuCap,Gelman Sciences). 10 eppendorfs (Sarstedt) with 1 mL of p97 solution were set up at37 °C and one eppendorf removed at time 0, 1, 2, 3, 24 and 51 h and stored at 4 °C.These samples were then analysed for p97 concentration using the immunofluorescenceassay.REPEATEDHARVESTINGH108cells108cells_____1!111111111CD0.5mLToP1-PLCmmiiiSupernatantHarvest#3aEnzymeL‘IcbIiHarvest#2aHarvest#1CellsWashBIOREACTORvi363.5.4 P1-PLC stabilitya) At 37°CPT-PLC from the bacterial supernatant at 55 mU/mL was set-up in 1 mL eppendorf tubesand incubated at 37 °C. Microcentrilhige tubes were removed at different time intervals(0 - 14 days) and stored at - 20 °C. The samples were then assayed for P1-PLC activity.b) pH200 ILL of 16 mU/mL PT-PLC in PBS (1 mg/mL) buffer with pH of 6, 6.5, 7, 7.5 and 7.9was used to treat 4x106 cells each. The cell surface fluorescence of the treated cells wasmeasured using the flow cytometry and solubilized p97 concentrations in supernatantmeasured by immunofluorescence assay.c) Freeze-thawSince PT-PLC samples were stored at - 20 °C and then thawed prior to use stability wasdetermined at 2 different dilutions (33 and 3.3 mU/mL). The enzyme was rapidly thawedin hot water (water bath at 37 °C) and frozen in the freezer (- 20 °C). The PT-PLC wasstable. This implied that PT-PLC solution can be stored at - 20 °C and thawed up to 10times without degradation In a separate experiment when PT-PLC samples were thawedin air slowly at room temperature, approximately 50 % of the activity was lost after eachfreeze-thaw cycle. Hence, it was important to thaw the enzyme very rapidly in hot water.3.5.5 Adsorption of P1-PLC on cellsCHO-S-SFM I was used to grow CHO cells in a 250 mL spinner flask to a density of4x106 cells/mL. 25 mL of this culture was used for each sample. The cell pellet of37approximately 108 cells was washed in 10 mL of PBS and then resuspended in I mL ofvarying concentrations (10 - 1500 mU/mL) of P1-PLC in PBS (1 mg/mL BSA) at 37 °C.For preliminary experiments, P1-PLC in the supernatant was monitored over aperiod of 60 minutes. Samples of 250 tL were drawn at different time intervals(30 sec - 60 mm). The first sample was taken as soon as the enzyme and the cells hadbeen mixed. The mixture of enzyme and cells was shaken before drawing a sample, tokeep the concentration of the cells constant throughout the experiment. Recoveredsamples were spun at 2000 rpm for a minute and the supernatant stored at - 20 °C. Thewhole process of removing the first sample, centrifuging and storing took less than3 minutes. After 60 minutes, the cells were spun and washed twice in 1 mL of PBS(1 mg/niL BSA). The cells were then resuspended in I mL of PBS and enzyme activity inthe supernatant monitored over a period of 60 minutes. The samples were then assayedfor PT-PLC activities.For later experiments, 108 cells were washed and incubated in 1 mL ofPT-PLC solution at 37 °C. After 15 minutes the cells were spun and the supernatantassayed for P1-PLC activity. PT-PLC solution was incubated without cells at 37 °C and itsactivity monitored over a period of 60 minutes as a control. Each sample was analysed induplicate at 3 different dilutions and the average taken as the representative value forPT-PLC activity.3.5.6 Continuous harvesting150 niL spinner flasks with 80 mL of CHO-S-SFM I medium were inoculated withapproximately 1x105 cells/mL from a common source (250 niL spinner flask). Onespinner flask was maintained without any P1-PLC as a control. Two spinner flaskscontained 3 and 30 mU/rnL PT-PLC each. All 3 cultures were maintained for a period of11 days (batch culture). The cultures were sampled everyday to monitor the glucose38concentration, cell concentration and viability, cell surface p97 and soluble p97concentration.39CHAPTER 4Results and Discussions41 Growth media for CHO cellsCHO cells growing in suspension tend to aggregate at high cell densities. For this study itwas important to maintain single cells so that the cell concentrations could be accuratelydetermined. Flow cytometry used to monitor cell surface protein also required singlecells, since large aggregates of cells block the flow tubes and make it impossible to analysecells individually.Originally the transfected CHO cells were grown attached to the surfaces usingHam’s F12 supplemented with newborn calf serum (NCS). The presence of NCS causesthe suspension cells to aggregate and adhere to the surface of tissue culture flasks. Thesuspension CHO cell line used in this work was selected (Kennard eta!., 1993) usingCHO-S-SFM I, serum-free media.A culture of cells grown in CHO-S-SFM I was inoculated at lx cells/mL andgrew to approximately 6x10 cells/mL of medium. The cells did not aggregate, even atthe highest cell density (Figure 4.1). CHO-S-SFM I is a proprietary medium whoseformulation is secret. A number of other serum-free media were screened in an attempt tofind a less expensive alternative of known composition.Cells cultured in DMEMIF12 or Ham’s F12 media without serum formedaggregates, such that the cells could not be enumerated. The cell aggregates consisted ofmore than 20 - 50 cells each. Renner etal. (1993) found that aggregating of CHOsuspension cells at high densities was caused by DNA released from lysed cells and byadding 50 - 100 jigi’mL of DNase the cell aggregation could be eliminated. Hence an40attempt was made to avoid the aggregating of the CHO cells by adding DNase to themedia.Ham’s F12 with DNase and 50 or 300 jig/mL of BSA were used to grow the CHOcells. In the medium with 300 .tg/mL of BSA the cells grew to 1 .2x106cells/mL(Figure 4.2), while the cells in Ham’s F12 with 50 .tg/mL of BSA only grew to8x1 cells/mL (Figure 4.3). The cells started aggregating at approximately6x105 cells/mL. These aggregates consisted of 2 - 5 cells each.DMEMJF 12 containing only DNase or DMEMIF 12 with insulin, transferrin,sodium selenite, BSA and DNase resulted in cell densities of approximately2x106 cells/mL of medium (Figures 4.4 and 4.5). The viability of above 90 % wasmaintained for 8 days in both media. Cell aggregates of approximately0C•)CD100800)60 -40%.200Time (day)4.1 Growth profile of CHO cells growing in CHO-S-SFM I. The figure showsconcentration (A), glucose concentration (LI) and viability of the cells (V).Figurethe cell411.00x106 I I 2.51002.0U) 7.50x10 8001.5C C 60-o 0CU 500x10 ° =U)CDCCD 40o—C I-o.2 2.50x10.5 20C)0.00 I I I I 0.0 .00 2 4 6 8 10 12Time (day)Figure 4.2 Growth profile of CHO cells growing in Ham’s F 12 with 5 jig/mL of insulin,5 ig/mL transferrin, 10 nM sodium selenite, 50 jig/mL BSA and 50 jtg!mL DNase. TheFigure shows the cell concentration (A), glucose concentration (LI) and viability of thecells (V).1.5x106 I 2.0100-JE—5 1.5• 80o 1.OxlO Qc C 0)o 0 601.00CUI.- CD_C —.C) CD 40o -5C50x10 00•0 0.5- 20C)__________________________________________•0.0 Q0.0 I I0 2 4 6 8 10 12Time (day)Figure 4.3 Growth profile of CHO cells growing in Ham’s F12 with 5 jtg/mL of insulin,5 p.g/mL transferrin, 10 nM sodium selenite, 300 p.g!mL BSA and 50 jig/mL DNase. Thefigure shows the cell concentration (A), glucose concentration (El) and viability of the cellscv).422.5x10 I I I1002.102 3—80Cl)C-)—‘i.5x106 QC Co60-CD1.0x106 40C)C) I5.cc10 20C-)II I I I I 0 00 2 4 6 8 10 12Time (day)Figure 4.4 Growth profile of CHO cells growing in DMEMIF12 with 50 j.ig/mL DNase.The figure shows the cell concentration (A), glucose concentration (Li) and the viabilityof the cells (V).100.1 063 80QC.) 6‘—‘1 5x10 CC)25x1062g.106CDC.a)’ c. .4OC)C0C.)5 OxlO 2000.0 I I I I I - 00 2 4 6 8 10 12Time (day)Figure 4.5 Growth profile of CHO cells growing in DMEMJF12 with 5 tg/mL of insulin,5 jtg/mL transferrin, 10 nM sodium selenite, 50 jg/mL BSA and 50 ig/mL DNase. Thefigure shows the cell concentration (A), glucose concentration (LI) and viability of the cells(V).43-JEC,)CDC-)C0(U-I-,Ci)C-)C0C-)CD05 - 10 cells per aggregate were observed at higher densities (approximately1x106 cells/mL). The increased cell growth in DMEM/F12 (Figure 4.5) compared toHam’s F12 (Figure 4.2) was probably due to the higher concentrations of nutrients inDMEMJF 12.Another medium that was investigated was HBCHO. In this experiment a celldensity of about 1.5x106 cells/mL was achieved (Figure 4.6). Besides attaining such a lowdensity, the cells started clumping at approximately lxi 06 cells/mL. At 1. 5x 106 cells/mL,the viability of the cells had already dropped to 78 %. Hence, this medium was notsuitable for our work.1.501.0 CC)0C,)CDCDI10080.6O20.0 0Time (day)Figure 4.6 Growth profile of CHO cells growing in HBCHO. This figure shows the cellconcentration (A), glucose concentration (Li) and viability of the cells (V) for a batchculture.Another medium that was investigated was CHO-S-SFM II. This medium haslower protein content and is also cheaper than CHO-S-SFM I. CHO-S-SFM II wasinoculated with approximately 6x1 cells/mL and grew up to 3. 5x 106 cells/mL of medium44(Figure 4.7). In this medium, most cells existed as aggregates of 2 - 10 cells at high celldensities (— 2x106 cells/mL). Hence, an attempt was made to select for the single cells.The aggregates of cells were allowed to settle down in the culture growing in the spinnerflask and 10 mL of culture with single cells was decanted from the suspension andtransferred to a T-flask. Single cells that were transferred to a T-flask were grown toapproximately 2x106 cells/mL. At this point the cells had already started aggregatingagain. Single cells were again transferred to a new T-flask and grown to high cell densityand the process repeated. Although this process was repeated over five times, we did notobtain a population of cells which would not aggregate. To eliminate the problem of cellaggregating, 100 .tg/mL of DNase was added to CHO-S-SFM II. This culture wasinoculated at 2x105 cells/mL. A cell density of 7x106 cells/mL of medium was achievedand this culture maintained single cells throughout the batch run. (Figure 4.8).10080Q6040200Time (day)Figure 4.7 Growth profile of CHO cells growing in CHO-S-SFM II. The figure showsthe cell concentration (z\), glucose concentration (Li) and viability of thecells (V).4510080Cl)ci6040C)20C-)0Time (day)Figure 4.8 Growth profile of CHO cells growing in CHO-S-SFM II and 100 .tg/mLDNase. The figure shows the cell concentration (A), glucose concentration (Li) andviability of the cells (V).Of the seven different serum-free media investigated, CHO-S-SFM II with100 jig!mL of DNase and CHO-S-SFM I were able to achieve the highest cell densitywhile maintaining single cells. Since CHO-S-SFM Ills also a proprietary medium,CHO-S-SFM I was used to grow the cells for all the investigations that followed.The experiments performed above show that DNA released into the media is notthe only cause of aggregation, because addition of DNase to Ham!s F12 and DMEM]F12did not completely eliminate the aggregating problem. However, aggregating of cells inCHO-S-SFM II was successfully eliminated by the addition of DNase.The effect of varying the initial cell density was also investigated (Appendix 2).C)CC.)00CD464.2 P1-PLC assaySince, available P1-PLC assays were too insensitive, too labour intensive and/or tooexpensive to perform routinely, it was necessary to develop a new assay. The assaydeveloped was based on monitoring, in the presence of P1-PLC, either cell surface p97 orsolubilized p97 in the supernatant.4.2.1 Effect of harvesting conditions on p97 removalp97 harvest is dependent on the conditions prevailing in the harvesting process.Harvesting conditions were varied to find the suitable parameters for the P1-PLC assay.The parameters investigated were incubation time and the volume of PT-PLC solution persample. Incubation timeKennard et at. (1993) reported that in 10 mU/mL P1-PLC, the viability of i08 CHOcells/mL did not decline up to 1 h. However, after 1 h the viability of the cells starteddropping. Hence one hour incubation time was used as an upper limit. Figure 4.9 showsthat p97 harvest increases with time, and that 60 minutes standard curve had the greatestrange of p97 removal. Enzyme volumeThe effect of enzyme volume used per sample was investigated. Samples of 4x106cellswere treated with 50, 100, 200 and 400 p.L of 5 mU/mL of P1-PLC enzyme solution for474o02I.—0)0.P1-PLC (mU/mL)Figure 4.9 Effect of different incubation times on p97 harvest. Each sample of4x106 cells was treated with 200 iiL of PT-PLC at 37 °C.1 h at 37 °C. Figure 4.10 shows that p97 harvest increases with P1-PLC volume (that is,increases with the ratio of PT-PLC to cells). At 50 !.iL of PT-PLC cleaved approximately16 % of the cell surface p97, while 400 1iL of PT-PLC solution cleaved 28 % of the totalcell surface p97. Hence, efforts were made to keep a constant enzyme volume duringeach P1-PLC assay to ensure that the ratio of cells to enzyme is maintained.Therefore for all PT-PLC assays, 200 p.L of PT-PLC enzyme solution, 4x106cellsand 1 h incubation time were used, unless otherwise specified.4.2.2 P1-PLC assay based on flow cytometryThe first assay developed was based on monitoring the cell surface p97, using flowcytometry (Section Enzyme solution (200 jiL) with a range of PT-PLCconcentrations was used to treat 4x106 cells at 37 °C for 1 h and cell surface p97604840- • • •30-o 10-0— • I • I • I0 100 200 300 400Pt-PLC volume (p. L)Figure 4.10 Effect of PT-PLC volume on p97 harvest. A sample of 4x106 cells wastreated with 50, 100, 200 and 400 p.L of PT-PLC solutions at 5 mU/mL. Incubation timeof 1 hwas used.measured, before and after the harvest. A standard curve of p97 removal from cells vsP1-PLC concentration was obtained with the range between 0.8 mU/mL and 80 mU/mL(Figure 4.11). Unknown P1-PLC samples were diluted to fall within this range and usedto harvest cells in parallel with known standards. p97 removal of unknown samples wasconverted to PT-PLC activity using the standard curve obtained with the known standards.PT-PLC at 8 and 16 mU/mL were used to treat cells in triplicates. The standard error wasless than 2 % showing that these results are reliable and reproducible (Figure 4.11).This assay was time consuming because of the need for repeated washing andcentrifhging of the cells (Figure 3.1). Hence, an attempt was made to develop anotherPT-PLC assay.49iOo4020Ui0.1 1 10 100P1-PLC (mU/mL)Figure 4.11 Typical standard curve obtained for P1-PLC at pH = 7.5 using 200 p.L ofPT-PLC solution and approximately 4x106 cells per sample. A cell incubation time of 1 hat 37 °C was used to obtain this curve.4.2.3 P1-PLC assay based on immunofluorescenceThis assay made use of immunofluorescent analysis (Section of soluble p97concentration in the supernatant recovered from PT-PLC incubated with cells. p97 in thesupernatant of the samples treated with known quantities of P1-PLC was used to obtain astandard curve of p97 recovered (jig/mL) against P1-PLC activity (mU/mL). PT-PLCactivity of an unknown sample was determined from the concentration of p97 in thesupernatant and the standard curve (Figure 4.12).Standard curves obtained from the assays based on flow cytometry andimmunofluorescence were identical to each other (Figure 4.12). The immunofluorescenceassay was less time consuming and involved less work.80605012• 10010’•808_J CDE•60 -,CDzi. °•402 -20•1 . -00.1 1 10 100P1-PLC (mU/mL)Figure 4.12 Typical standard curves obtained from assays based on flow cytometry andimmunofluorescence. The open triangles represents p97 concentration (A) and the opensquares represents p97 removal (LI). 200 iiL of P1-PLC solution was used to treat4x106 cells. Incubation time of I h at 37 °C was used.4.2.4 Reliability of P1-PLC assayAfter developing two different assays, their reliability was investigated. Two unknownPT-PLC samples at three different dilutions were tested and their average taken as therepresentative enzyme activity of that solution (Table 4.1).The actual PT-PLC activities of the samples according to Boehringer Mannheimwere 6.2 mU/mL and 31 mU/mL respectively. The PT-PLC activities determined for thesamples by flow cytometry assay were 6.23 and 34.2 mU/mL respectively. Theimmunofluorescence assay determined the samples to be 5.7 and 33.5 mU/mLrespectively. PT-PLC activities obtained by both the assays were within ± 10 % of theactual values.51Flow CytometryMeasured CalculatedP1-PLC PT-PLC[mU!mLl [mU/mLl1.8 7.23.0 65.5 5.50.9 361.63 32.5- 34ImmunofluorescenceMeasured CalculatedPT-PLC P1-PLC[mU/mLl [mU/mLl1.30 5.203.20 6.405.40 5.401.20 30.001.75 35.003.55 35.504.3 Repeated harvestingThe first production process to be investigated was the repeated harvesting technique.The repeated harvesting process involves re-using the PT-PLC enzyme solution forharvesting multiple samples of cells in an attempt to further increase the productconcentration.In the first experiment, 7 samples of 1 cells were harvested consecutively with0.5 mL of 30 mU/mL of P1-PLC in PBS with 1 mg/mL BSA (Figure 4.13). After eachsampling, 20 jiL of PBS (1 mg/mL BSA) was added to the harvesting solution to replacethe sample volume removed to maintain a constant volume. The cumulative p97recovered increased for the first two harvests and then remained approximately constant.Figure 4.13 also shows the corrected values, which take into account the p97 dilution dueto sampling. There also may have been dilution effects due to the liquid associated withthe cell pellet. The p97 recovered from the first harvest was approximately 59.7 tg!mL.Hence, after 7 harvests theoretically 7-fold more or 417.8 tg/mL p97 could have beenTable 4.1 Comparison of P1-PLC assay based on flow cytometry andimmunofluorescence assay. Each sample was assayed induplicate at 3 different concentrations.Sample Dilutionin PBS1 4212 40201052recovered. However, there was only a 2-fold increase in the p97 concentration in theproduct. These results indicated that there was loss of either P1-PLC or p97 in thesuccessive harvests.-JE3 4 5Harvest (#)Figure 4.13 Concentration of p97 recovered for 5 repeated harvests. Each harvest wasconducted with i08 cells for 30 minutes. P1-PLC enzyme (0.5 mL) at an initialconcentration of 30 mU/mL was used. Open circles (0) represent the measured valuesand the solid circles (.) represent calculated values which account for sampling dilution.After each harvest, a sample of approximately 4x106 cells was washed in PBS andthe cell surface p97 measured by flow cytometry. p97 removal from the cell surfacedecreased from 60 % to 20 % over 7 harvests (Figure 4.14).Although the recovered p97 concentration was not as high as expected, the 2-foldincrease in p97 showed that the technique had potential. Further investigations werecarried out to determine the reasons for not achieving the predicted amount of p97.53100-80-60-40-20-0-.I0 1 2 3 4 5 6 7Harvest #Figure 4.14 Percent of p97 removal from the cells after each harvest. Approximately4x106 cell sample was taken from the cell pellet and labeled with the fluorescent antibody.4.3.1 p97 stabilityThe p97 stability was investigated to determine if degradation of the protein could accountfor part of the decreased harvest yield. Since, the harvesting experiments were done at37 °C, we investigated p97 stability at that temperature. Firstly, purified p97(Kennard et at, 1993) stability was studied in immunoassay buffer (storage buffer). Tenp97 harvests of 1 h each take approximately 16 h, however p97 was found to be stable inthis buffer over a period of 48 h (Figure 4.15).The next experiment was done under harvesting conditions, that is in the presenceof PT-PLC and possible proteases released from cell lysis. Two samples of p97 harvestedfrom cells were incubated at 37 °C for 4 and 24 h. There was slight decline of p97 over aperiod of 24 h (Table 4.2). However, the decline in p97 was not enough to account forthe insufficient harvesting. Hence, p97 degradation was not the cause for the low levelsobtained from repeated harvesting.II•I•I•II’ 15420 .i •181614121o. Ii80)Q- 64.20• •..i I1 10Time (h)Figure 4.15 p97 stability at 37°C for a period of 48 h. Initial p97 concentration was9.65 ig/mL in the immunoassay buffer. The error bars represent the standard error mean.Table 4.2 Stability of p97 in the harvesting solution of PT-PLC in PBS (1 mg/mLBSA). The samples were assayed at 3 different dilutions. The errorbars were less than ± 1 5 ig/mL.Time Initial FinalConcentration Concentration[h] [jig/mL] [ig/mL]4 26.4 23.324 23.2 P1-PLC StabilityThe PT-PLC enzyme stability was studied to determine if the loss of activity could explainthe repeated harvesting results. P1-PLC at 52 mU/mL was incubated at 37 °C for a periodof 2 weeks and samples drawn at different time intervals. These samples were stored at4 °C and then analyzed for PT-PLC activity. PT-PLC proved to be stable over 14 day55period (Figure 4.16). Griffith eta!. (1991) also found that PT-PLC was stable at 37 °C forprolonged periods (they did not report the exact time).(U.60-0 0 0 0EP.o 30--J0..L 20-0.10-O• ....... •uI11•uI1 111111111 111111111 I0.01 0.1 1 10Time (day)Figure 4.16 P1-PLC stability at 37 °C. PT-PLC from the bacterial supernatant at52 mU/mL was used. 200 p.L of two dilutions of each sample were used to treat4x106 cells for 1 h at 37 °C.The pH of the harvesting solution after 10 repeated harvests of adherent CHO cellson porous microcarriers falls to approximately 6.5 (Kennard, personal communication).Hence, effect of pH on PT-PLC activity was investigated. The volumes of PT-PLC usedfor the harvests in this work were too small (0.5 mL) to monitor pH routinely during theactual harvesting process. Therefore, PT-PLC enzyme diluted in PBS (1 mg/mL BSA) atpH = 6.0, 6.5, 7.0, 7.5 and 7.9 was used to treat 4x106cells (Figure 4.17). Thepercentage removal of p97 based on flow cytometry analysis of cell surface p97 increasedwith decreasing pH. Therefore, changes in pH do effect PT-PLC activity, however it doesnot inactivate it. According to Ikezawa and Taguchi (1981), the optimum pH is between5 - 8.5, with maximal activity at 7.5. However that trend did not agree with this work, in56contrast PT-PLC activity increased monotonically from 7.9 - 6 pH. Insufficient harvestingcan not be attributed to decreases in pH.Supernatant from these harvests were then analysed for p97. Figure 4.17 showsthat p97 recovered at 16 mU/mL of PT-PLC solution also decreased with increasing pHfrom 6.0 - 8.0. This trend agrees with that of p97 removal from the cell surface (Figure4.17). Therefore, the change in pH of the harvesting solution was not the cause for therepeated harvesting results.100 1680—126O>4030) r20-0 I I I I6.0 6.5 7.0 7.5 8.0pHFigure 4.17 Effect of pH on P1-PLC activity and p97. P1-PLC at 16 mU/mL was used toharvest 4x106 cells for 1 h at 37 °C. PBS buffer with 1 mg/mL of BSA at pH = 6.0, 6.5,7.0, 7.5 and 7.9 was used for these harvests. The open squares (E) stand for p97concentration and the open triangles (ba) stand for p97 removal.4.3.3 P1-PLC adsorptionAn effort was made to investigate the loss of P1-PLC enzyme due to adsorption to the cellsurface. Phospholipases are a special type of membrane interacting proteins. AccordingI I I57to Hendrickson and Dennis (1984), phospholipase A2 (PLA2) adsorbs to the cell surfacebefore reacting with the substrate. In the repeated harvesting process the cells wereremoved from the harvesting solution and replaced with a fresh set of cells after eachharvest, any PT-PLC adsorbed to the cell surface would be lost at each harvest. Since, itwas not possible to measure the P1-PLC adsorbed to the cell surface, PT-PLC changes inthe supernatant were monitored over time and PT-PLC adsorbed to the cell surfacecalculated as follows:[PIPLCJa = [PTPLC]0 - [PI”PLC]s (4,1)where [PI-PLC]a = P1-PLC adsorbed on the cell surface at time t[P1-PLC]0 = Initial P1-PLC concentration in the solution[P1-PLC]5 = P1-PLC concentration in the supernatant after time tThe first experiment was performed to monitor the change in PT-PLCconcentration over a period of 1 h in the mixture of 18 mUImL enzyme and 108 cells/mL.Since preliminary repeated harvesting process used 108 cells per harvest, the same numberof cells were used for the adsorption experiments. The cells and enzyme were mixedthoroughly before removing a sample, so as to maintain a constant cell concentration inthe mixture. P1-PLC activity decreased rapidly to 4.2 mUImL within 3 minutes and thenremained constant over a period of 60 minutes (Figure 4.18). The decrease of PT-PLCcould be attributed to the rapid binding of the protein to the cell surface. Althoughsamples were drawn at 30 s, the cells were centriftiged and separated from the supernatantwithin 3 minutes. Thus the first data point is reported at 3 minutes in Figure 4.18. As acontrol, P1-PLC solution without any cells was incubated at 37 °C for a period of I h. Nochange in enzyme activity was observed over the period of 1 h (Figure 4.18).58The cells were then washed twice with 2 mL of PBS and resuspended in 1 mL ofPBS and incubated at 37 °C for an hour to investigate the desorption of P1-PLC from thecell surface. PT-PLC activities in the wash solutions I and II were 2.4 and 0.8 mU/mLrespectively.-JDC)-J040Time (mm)Figure 4.18 Adsorption of P1-PLC (0) to the cell surface over a period of 60 minutes.108 cells/mL of PT-PLC enzyme at 18 mUImL. A control on adsorption (•) experimentwas done with P1-PLC without cells incubated at 37 °C for 1 h.The ratio of volume of cells to liquid for desorption was lower than that foradsorption experiment because of the loss of cells due to sampling. Desorptionexperiment was done over a period of 2 h to investigate the effect of increased incubationtime on the equilibrium between PT-PLC in solution and P1-PLC on the cell surface. Inthis case PT-PLC activity in the supernatant increased rapidly (within 30 sec) and thenremained constant (Figure 4.19). Figure 4.19 shows p97 removal on the vertical axisinstead of the P1-PLC activity, because PT-PLC activity in the desorption solution wasbelow the linear range (less than 1 mU/mL P1-PLC) on the standard curve. Hence it wasnot possible to convert percentage removal of p97 to PT-PLC activities. Figures 4.18 and594.19 show that adsorption and desorption of P1-PLC was at equilibrium within less than3 minutes.60 80Time (mm)Figure 4.19 Desorption of P1-PLC from the cell surface (0) in 1 mL of PBS. PT-PLCactivity is less than 1 mU/mL. The cell concentration of 5x107 cells/mL was used.Study of PT-PLC desorption was also important for the controlled release cyclicharvesting technique developed by Kennard et a!. (1993). The cyclic harvesting processinvolves recycling the harvested cells into the growth medium for re-expression of p97(Figure 3.4). Presence of residual PT-PLC in the growth medium would cleave p97 duringthe growth cycle resulting in loss of protein to the growth medium. Since, PT-PLC willrapidly reach an equilibrium between the cells and the supernatant, two or three quickwashes will remove most of the PT-PLC from the cells surface. Leaving the cells in washsolution for 5 - 10 minutes will not help the desorption process, but using large volume ofwash solutions will remove more p97 from the cell surface.[00) 560Enzyme activities of the samples from adsorption experiment were analyzed at 3different dilutions and their average taken as the representative value for the activity ofeach sample.The adsorption experiment was repeated with different enzyme concentrations toinvestigate whether adsorption resulted in saturation or equilibrium. In case of saturationa fixed amount of P1-PLC would be adsorbed to the cell surface, while for equilibrium aproportional amount to initial PT-PLC concentration will be adsorbed. Using the datafrom Table 4.3 and equation 4.1, an equilibrium curve was established (Figure 4.20).Table 4.3 P1-PLC change in the supernatant after 15 mm. 1 ü cells/mLof PT-PLC solutions at different concentrations were incubatedat 37 °C.P1-PLC at time = 0 P1-PLC at time = 15 mm{mU/mL] [mU/mU2.50 0.959.40 5.5011.1 9.5038.0 15.580.0 30.0128 37,5560 156P1-PLC added to the cells resulted into an equilibrium between PT-PLC on the cellsurface and PT-PLC in solution. Figure 4.20 shows that PT-PLC lost due to celladsorption was directly proportional to the enzyme concentration added to the cell pelletup to approximately 150 mU/mL. The horizontal axis on Figure 4.20 shows PT-PLC insolution after adsorption.615x1064x1OD3x106- 2x106Cu0-J0... 1x1060050 75 100 125 175P1-PLC in solution (mU/mL)Figure 4.20 Adsorption equilibrium of PT-PLC at 37 °C. Different enzymesconcentrations at cells/rnL were used to study adsorption.4.3.4 Non-specific binding of PT-PLCSince PT-PLC selectively cleaves the GPI anchor it may be that the PT-PLC specificallybinds to the anchor of recombinant proteins. Therefore, an attempt was made toinvestigate whether P1-PLC adsorption was due to specific or non-specific binding. Thiswas done by studying P1-PLC adsorption on untransfected CHO cells (CHOWTB -Food et a!, 1994) in comparison to the genetically engineered cells used in this study(Table 4.4).Table 4.4 shows that both the cell lines adsorbed 1 .8x107 mU P1-PLC per cell,hence protein binding was not increased by the recombinant p97 on the surface of thetransfected cells. In addition literature reviewed on specific binding of manganese (Suárezand Eriksson, 1993) and lactoferrin (Maneva etal., 1993) show that binding takes placeslowly over a period of 30 - 60 minutes. Therefore, rapid binding of PT-PLC to cellsurface was most probably non-specific.62Table 4.4 Comparison for transfected and untransfected (CHOWTB) CHOcells. Incubation time of 15 mm and I mL of PT-PLC solution wasused.CHO Cell P1-PLC at PT-PLC at P1-PLCCell line concentration 0 mm 15 mm adsorbed[cells/mL] [mU/mU [mUImL] [mU/cell]Untransfected 5.5x107 32 22 1.82x107Transfected 1x108 32 14 1.8x107To improve the repeated harvesting process, one option was to add excessPT-PLC so that even after the losses due to adsorption, there is still enough P1-PLC tocleave approximately 80 - 90 % of the p97 from the cell surface. However, this will resultin additional cost for P1-PLC and also increase contamination of the product by theenzyme. Hence as an alternative, the repeated harvesting process was started with lowenzyme concentration and PT-PLC replenished. A repeated harvesting process was thencarried out for 10 consecutive harvests with enzyme replenished after the 5th and 6thharvest to confirm the loss of P1-PLC due to adsorption. The first harvest was startedwith 0.5 mL of 300 mU/mL of P1-PLC and 25 iL of 3000 mU/mL PT-PLC was added tothe harvesting solution after the 5th and 6th harvest (Figure 4.21) to replenish the lostPT-PLC. The reason for using a small PT-PLC volume at high concentration forreplenishment was to minimize the dilution of the protein product.Figure 4.21 shows that p97 in the product increased 4-fold from approximately72 .ig/mL in the first harvest to 290 tg/mU. The decrease in p97 concentrations at the3rd and 9th harvest are believed to be due to the dilution effect due to the fluid associated-JE8F.—0)0Harvest NumberFigure 4.21 Cumulative p97 from repeated harvesting with P1-PLC replenishment. Theinitial harvesting was started with 0.5 mL of 300 mU/mL of PT-PLC and then25 jiL of P1-PLC enzyme at 3000 mU/mL was added after the 5th and 6th harvests.Incubation time of 30 minutes at 37 °C was used for each harvest.Cu20)Figure 4.22 p97 removal from the cell surface during the repeated harvesting processThe initial harvesting was started with 0.5 mL of 300 mU/mL of P1-PLC and then25 iL of P1-PLC enzyme at 3000 mU/mL was added after the 5th and 6th harvests.Incubation of 30 minutes at 37 °C was used for each harvest.630 2 4 6 8 101008060402O0I I2 4 6 8 10Harvest Number64to the cell pellet. Figure 4.22 shows that adding PT-PLC enzyme after the 5th and 6thharvests restored the percentage removal of p97 from the cell surface. Hence Figures 4.21and 4.22 confirm that insufficient harvesting of p97 during the repeated harvesting processwas caused by the loss of P1-PLC due to adsorption to the cell surface when the harvestedcells were removed.Having determined that PT-PLC adsorption to the cell surface had caused reducedharvesting of p97, the next repeated harvesting experiment was carried out using 2different initial concentrations (300 and 30 mU/mL) and enzyme supplemented at eachharvest. 50 tL of the harvesting solution was drawn after each harvest for p97 analysisand replaced by 50 j.L of concentrated enzyme (Figures 4.23 and 4.24). The repeatedharvesting experiment that was started with 300 mU/mL PT-PLC was supplemented with50 p.L (approximately 300 mUImL total volume) of 3000 mU/mL PT-PLC after eachharvest. Similarly the one that was started with 30 mUImL P1-PLC was supplementedwith 50 p.L (approximately 30 mU/mL) of 300 mU/mL PT-PLC.Figure 4.23 shows that repeated harvesting process was successful and p97concentration was increased approximately 4-fold with five harvests. In the case ofindustrial production, reduced volumes of concentrated PT-PLC can be added to minimizethe dilution of the protein in the product. Also there is no need to remove a sample fromthe harvesting solution and this will prevent loss of p97 due to sampling.The calculated values on Figure 4.24, only account for p97 lost by sampling, itdoes not account for the dilution effects caused by the liquid associated with the cells.Cumulative p97 recovered from 300 and 30 mU/mL does not differ considerably. Hence,it is advisable to use lower concentrations of PT-PLC. Using less PT-PLC results in lessproduction costs and also purer product.65120’100’80’&601- 40C)a20I I I I I0 1 2 3 4 5Harvest NumberFigure 4.23 Percent p97 removal from the cell surface for 5 harvests of 108 cells wasmonitored. Harvesting solution of 300 (0) and 30 (z) mU/mL were used. Incubation of1 h at 37 °C was used for each harvest.I I I I350--J— 300-‘- 2500-20O1508 100C) 50a0 I I I I I0 1 2 3 4 5Harvest NumberFigure 4.24 Open symbols represents cumulative p97 measured after each harvest andthe solid symbols represents the calculated values of the p97 harvest which includes p97lost due to sampling. This figure shows p97 recovered from 300 (0) and30 (A) mU/mL of PT-PLC. Incubation of 1 h at 37 °C was used for each harvest.664.4 Continuous harvestingThe second harvesting process investigated was a continuous harvesting technique. Thistechnique involved harvesting of p97 into the growth media and a batch run over a periodof 11 days. This production process is similar to the typical protein production process bymammalian cells where the desired protein is secreted into the growth media. Anadvantage of this harvesting process was the reduction in p97 losses due to membraneturnover. Approximately 1 - 3 % of the cell membrane is internalized per minute(Steinman eta!., 1976; Burgess and Kelly, 1987). This phenomena resulted ininternalization and possibly degradation of the desired membrane protein, resulting indecreased product recovery. Theoretically, PT-PLC in the growth media will cleave p97as soon as it appears on the cell surface preventing this loss. Therefore media containingthree different PT-PLC concentrations of 0, 3 and 30 mU/mL PT-PLC were inoculated withapproximately 2x105 cells/mL. Cell viability (Figure 4.25), cell surface p97 (Figure 4.27),glucose concentration (Figure 4.26) and p97 in the supernatant (Figure 4.28) weremonitored over an 11 day period.Figure 4.25 shows that PT-PLC in the growth media does not have an effect on cellconcentration or the viability of the cells in the culture. The glucose consumption wasalso approximately the same for all three cultures (Figure 4.26). However, cell surfacefluorescence of cells growing in the cultures with 3 and 30 mU/mL decreases rapidly,while the cells growing in a culture with no (0 mU/mL) PT-PLC decreases steadily (Figure4.27) as Kennard eta!. (1993) reported over a batch culture.674.0x10’ I I • I -100-J-80= 6.-60o 1)2.0x106c8 1.0x106-20a)C-)0.0- I I • I • I -00 4 6 8 10 12Time (day)Figure 4.25 Continuous harvesting of CHO cells in the growth media. The solid symbolsrepresent cell concentration and the open symbols stand for viabilities. Cultures withPT-PLC at 0 (El) 3 (A) and 30 (0) mU/mL of medium.I • I • I • I4.-JE3I:• I • I • I • I • I0 2 4 6 8 10 12Time (day)Figure 4.26 Glucose concentrations during the continuous harvesting process with 0 (LI),3 (A) and 30 (•) mUIrnL of PT-PLC enzyme.68I • I • I • I •300— 2258CG)1500DLL.0 • I • I • I •0 2 4 8 10 12Time (day)Figure 4.27 Fluorescence due to cell surface p97 during the continuous harvestingprocess with 0 (El), 3 (A) and 30 (0) mU/mL of P1-PLC enzyme.I I • I • I •16-J140)o 10C86C84N0•)00 • I I • Io 2 4 6 8 10 12Time (day)Figure 4.28 Cumulative p97 in the supernatant from the continuous harvesting processwith PT-PLC at 0 (LI), 3 (A) and 30 (0) mU/mL of medium.69Figure 4.28 shows that approximately 15 tg/mL of p97 was recovered in theculture with 30 mU/mL P1-PLC. A control culture with no P1-PLC enzyme producedapproximately 4 ig/mL of p97. Membrane bound p97 may have been released in thesupernatant due to cell lysis. Kennard et a?. (1993) and Food et a?. (1994) reported thattransformed CHO cell line used in this study also produces a secreted form of p97. Hencep97 recovered in the spent media was the result of the secreted and membrane bound p97.According to Figure 4.27 and the calculations, based on theoretical values, we haveharvested all the p97 present on the cell surface by using 30 mU/mL P1-PLC.The calculation for approximate p97 production in to the media was done by thefollowing equation:d[p97] = (4.2)dtwhere r = average rate of p97 production (jig/cell-h)x number of cellst = time (h)Integrating equation 4.3, we get:[p97]=rt (4,3)where x = average number of cellsAccording to Kennard eta?. (1993), approximately 6x107 jig of p97 is expressed per cellwithin 48 h after P1-PLC treatment.Therefore,r= [p97 per cell] = [6x1o Lg/cell]= L25x108 tg!cell-h (4.4)For a batch culture of approximately 11 days and an average of 3x106 cells/mL,substituting equation 4.4, we get:70[p97] = (1.25x108p.g/cellIh) x (3x106 cells/mL) x (11 day) x (24 h/day). (4.5)= 9.9 ig/mLTotal p97 attained experimentally is 15 .tg/mL, of which 4 jig/mL (0 mU/mL in Figure4.28) is the secreted form of p97. Therefore, p97 produced due to PT-PLC action is:15 -4 = 11 jig/mLTaking into account the errors due to p97 analysis, p97 released from the cells due to celllysis and the loses recovered due membrane turnover rate, we have harvested the majorityof p97 that was expressed on the cell surface.4.4 Comparison of protein production processesThree harvesting processes, cyclic (Kennard eta!., 1993), repeated and continuousharvesting techniques were compared based on their p97 productivity, yield on glucoseconsumed, PT-PLC used and p97 purity (Table 4.5). The data for cyclic harvestingprocess has been estimated from the work of Kennard eta!. (1993), based on harvesting108 cells in 0.5 mL P1-PLC solution. The comparison was based on p97 harvests over 8day period in a 250 mL bioreactor.The continuous harvesting process attained a yield of 3.4 jig p97 per mg ofglucose consumed. It also attained the highest p97 cell specific productivity(5.75x10 jig/cell-day) (Table 4.5). However, it achieved the lowest p97 concentration ofthe 3 production processes.The repeated harvesting process carried out by using 30 mU/mL of PI-PLproduced 295 jig/mL p97, while the one carried out with 300 mU/mL of PT-PLCproduced 343 p97. Hence, it is not advantageous to increase the enzymeconcentration lox to increase the protein in the product by approximately 8 %. As weknow from the adsorption experiments that P1-PLC lost to the cell surface will beproportional to the amount added. Therefore, harvesting with the lower PT-PLC71concentrations (30 mU/mL or less) is recommended. Cyclic and repeated harvesting arevery similar except with repeated harvesting the PT-PLC volume is reduced and p97concentration is increased.Table 4.5 Comparison of p97 harvests from different harvesting processesHarvesting p97 Average p97 p97 Volume Estimatedprocess p97/cell yield on yield on of the Purityper day PT-PLC glucose harvestingsolution[p.g/mL] [tg/cell-day] [jj.g/mU] [ig/mgJ [mU [%]Cyclic5 harvests 62 1.93x107 2.07 0.39 25 3030 mU/mLRepeated5 harvests 343 2.14x107 0.23 0.43 5 30300 mU/mLRepeated5 harvests 295 1.84x107 1.97 0.37 5 3030 mU/mLContinuouslbatchculture 11.5 5.75x10 0.38 3.40 250 3.630 mU/mLFor calculations for Table 4.5, see appendix 9In addition to the dilute and impure product obtained from continuous harvestingprocess, the quantity of P1-PLC required was high. To harvest 2875 tg of p97, 7500 mUof PT-PLC were required (ratio of 2.6:1). The repeated harvesting process recovered721474 p.g of p97 using 750 mU of PT-PLC (ratio of 0.5:1). Therefore, the yield of p97 onPT-PLC was higher for repeated harvesting process than for the continuous harvestingprocess.Hence, the question is to choose a harvesting process with high yield or highprotein concentration. This will depend on the relative magnitudes of the production(including enzyme cost) and the downstream processing costs. However, the majormotivation for this work was to achieve high initial purity and concentration to decreaseextreme downstream processing. Therefore the recommended production process is therepeated harvesting technique.73CHAPTER 5ConclusionsThe growth of CHO cells was investigated in CHO-S-SFM I, CHO-S-SFM II, HBCHO,DMEMJF12 and Ham’s F12. CHO cells have a tendency to aggregate at high celldensities, making it difficult to determine cell surface p97 expression. DNA released fromthe lysed cells triggers cell aggregation at high densities. DNase was added to these mediato reduce aggregation. Addition of DNase to HBCHO, DMEMIF12 and Ham’s F12 didnot eliminate cell aggregation. However, addition of DNase toCHO-S-SFM II eliminated the aggregation problem altogether.The repeated harvesting process involves re-using the PT-PLC enzyme solution toconsecutively harvest multiple samples of cells. Initially only 139 jig/mL p97 wasrecovered from 7 harvests of 108 cells. The first harvest recovered approximately60 .tg/mL, therefore theoretically 7 harvests should have recovered 420 ig/mL.In an attempt to understand the low product recovery, the stability of p97 andP1-PLC were studied. Whereas the repeated harvesting process was complete within 16 h,p97 was found to be stable at 37 °C for a period of over 24 h. Since a suitable assay forquantifying P1-PLC was not available, PT-PLC assays based on flow cytometry andimmunofluorescence analysis of p97 were developed. PT-PLC activity was found to bestable at 37 °C for a period of 14 days. Thus the low product harvests were not due top97 or PT-PLC degradation.Loss of PT-PLC due to adsorption to the cell surface was investigated. Cellsurface adsorption and desorption of PT-PLC was found to take place in less than3 minutes. An equilibrium was established between the P1-PLC in the solution andPT-PLC adsorbed to the cell surface. Since the repeated harvesting process re-uses the74PT-PLC solution, adsorption to cell surfaces resulted in a loss of enzyme upon cell removalafter each harvest. This depleted the P1-PLC enzyme in the solution and decreasedsubsequent p97 harvests.Repeated harvesting experiments were then done with P1-PLC replenishment aftereach harvest to maintain the PT-PLC concentration. From 5 consecutive harvests of1O cells a 4-fold increase in p97 concentration in the product was achieved.A continuous harvesting process also was investigated. This process involvedharvesting of p97 in the cell growth medium at the same time as protein expression. Froma batch culture 15.3 .tg/mL of p97 was recovered.Three production processes: cyclic, repeated and continuous harvesting werecompared. The greatest yield on glucose was obtained by the continuous harvestingprocess (3.4 p.g p97/mg glucose), while the greatest yield on PT-PLC was obtained by thecyclic harvesting process (2 .tg p97/mg glucose). The continuous harvesting process alsoachieved the highest p97 productivity per cell (5.75x10 tgIcell-day), approximately 2-fold higher than the cyclic and repeated harvesting processes. Increased productivity ofthe CHO cells in the continuous harvesting process was probably due to reduced lossesdue to the membrane turnover. The repeated harvesting process achieved the highest p97concentration and its purity was comparable to that of cyclic harvesting process.75CHAPTER 6Future Work1, In this work P1-PLC adsorption to the cell surface was studied up to approximately150 mU/mL to cover the range of enzyme concentration used to harvest the CHO cells inthese production processes. It would be interesting to develop a complete adsorptionisotherm to investigate the saturation of P1-PLC on cell surface.2. Different GPI-anchored proteins have been reported to respond differently to cleavageby PT-PLC. Hence, it is important to apply this technique to other GPI-anchored proteinsand determine if the p97 model system is representative of other proteins.3. It would be worthwhile to do an economic analysis to compare the relative benefits ofthe 3 controlled release processes.4. The controlled release harvesting process is laborious and involves multiple washingand centrifuging. These problems could be reduced by designing a specialized bioreactorto accommodate all steps of the harvesting process.76AbbreviationsADP Adenosine DiphosphateATP Adenosine TriphosphateCHO Chinese Hamster OvaryGPI Glycosyl PhosphatidylinositolNAD Nicotinamide Adenine Dinucleotide (oxidized)NADH Nicotinamide Adenine Dinucleotide (reduced)NCS Newborn Calf Serump97 MelanotransferrinPBS Phosphate Buffered SalinePEP Phosphoethanol PyruvateP1 PhosphatidylinositolP1-PLC Phosphatidylinositol Phospholipase C77ReferencesBailey J.E., Ollis, D.F., Biochemical engineeringfundamentals, 2nd ed., McGraw HillPublishing Co., 1986.Baker, E.N., Baker, H.M., Smith, C.A., Stebbins, M.R., Kahn, M., HelistrOm, K.E.,Hellstrom, I., “Human melanotransferrin (p9’7) has only one functional iron-binding site”,FEBsLett., 298(2,3): 215-218 (1992).Baker, E.N., Rumball, S. V., Anderson, B .F., “Transferrins: Insight into structure andfunctions from studies on lactoferrin” Trends Biochem. Sci., 12: 3 50-353 (1987).Berridge, M.J., “Inositol lipids and cell proliferation”, Biochim. Biophys. Acta, 907: 33-45(1987).Boehringer Mannheim Biochemica, Bulletin No. 0189, B 635.5, 1144 286.Boivin, P., Delaunay, I., “La membrane du globule rouge”, Xle congrès de la sociétéFrancaise d’Hématologie, 3(1), 125-128 (1991).Bordier, C., Etges, R.J., Ward, J., Turner, M.J., Cardoso de Almeida, M.L., “Leishmaniaand Trypanosoma surface glycoproteins have a common phospholipid membrane anchor”,Proc. Nat!. Acad. Sd., 83(15): 5988-5991 (1986).Brown, J.P., Hewick, R.M., HellstrOm, I., HelistrOm, K.E., Doolittle, R.F., Dreyer, W.J.,“Human-melanoma associated antigen p97 is structurally and functionally related totransferrin”, Nature, 296: 171-173 (1982).Brown, J.P., Niyashama, K., HellstrOm, I., HellstrOm, K.E., “Structural characterization ofhuman-melanoma-associated antigen p97 with monoclonal antibodies”, I ImmunoL, 127:539-546 (1981a).Brown, J.P., Woodbury, R.G., Hart, C.E., HellstrOm, I., HellstrOm, K.E., “Quantitativeanalysis of melanoma associated antigen p97 in normal and neoplastic tissues”, Proc Nat!.Acad. Sci. USA. 78: 539-543 (1981b).Burgess, T.L., Kelly, R.B., “Constitutive and regulated secretion of proteins”, Ann. ReiCe!! Biol., 3: 243-293 (1987).Butler, M., Animal cell technology: Principles and Products, Open University Press,1987, U.S.A.78Caras, I.W., Davitz, M.A., Rhee, L., Weddell, G., Martin, D.W., Nusseinzweig, V.,“Cloning of decay accelerating suggests novel use of splicing to generate two proteins”,Nature, 325: 545-548 (1987a).Caras, I.W., Weddell, G., Davitz, MA,, Nusseinzweig, V., Martin, D.W., “Signal forattachment of a phospholipid membrane anchor in decay accelerating factor”, Science,238: 1280-1283 (1987b).Caras, I.W., Weddell, G., “Signal peptide for proteins secretion directingglycophospholipid membrane anchor attachment”, Science, 243: 1196-1198 (1989).Cross G.A.M., “Glycolipid anchoring of plasma membrane proteins”, Ann. Rev. Cell Biol.,6: 1-39 (1990).Davitz, M.A., Low, M.G., Nusseinzweig, V., “Release of decay-accelerating factor (DAF)from the cell membrane by phosphatidylinositol-specific phospholipase C(PT-PLC). Selective modification of a complement regulatory protein”, I Exp. Med., 163:1150-1161 (1986).Davitz, M.A., Gurnett, A.M., Low, M.G., Turner, M.J., Nussein.zweig, V., “Decay-accelerating factor (DAF) shares a common carbohydrate determinant with the surfaceglycoprotein (VSG) of the African Trypanosoma brucei”, I. Immunol., 138: 520-523(1987).Deems, R.A., Eaton, B.R., Dennis, E.A., “Kinetic analysis of phospholipase A2 activitytoward mixed micelles and its implications for the study of lipolytic enzymes”, J Biol.Chern., 250(23): 9013- 9020 (1975).Doering, T.L., Masterson, W.J., Hart, G.W., Englund, P.J., “Biosynthesis of glycosylphosphatidylinositol membrane anchors”, J. Biol. Cheni., 265: 61 1-614 (1990).Doery, H.M., Magnusson, B.J., Gulasekharam, J., Pearson, J.E., “The properties ofphospholipase enzymes in Staphylococcal toxins”, J. Gen Microbiol., 40: 283-296 (1965).Eibl, H., Lands, W.E.M., “A new, sensitive determination of phosphate”, Anal. Biochern.,30: 5 1-57 (1969).Estin, C.D., Stevenson, U.S., Plowman, G.D., Hu, S.L., Sridhar, P., HelistrOm, I., Brown,J.P., HellstrOm, K.E., “Recombinant vaccinia virus against human melanoma antigen p97for use in immunotherapy”, Biochern., 85: 1052-1056 (1988).Estin, C.D., Stevenson, U.S., Kahn, M., HelistrOm, I., HelistrOm, K.E., “Transfectedmouse melanoma lines that express various levels of human melanoma-associated antigenp9’7”,J. Natl. Cancer Inst., 81: 445-448 (1989).79Etges, R., Bourier, J., Bordier, C., “The major surface protein of Leishmaniapromastigotes is a protease”, J. Blot. Chern., 261(20): 9098-9101 (1986).Ferguson, A.J., Williams, A.F., “Cell-surface anchoring of proteins via glycosylphosphatidylinositol structures”, Ann. Rev. Biochern., 57: 285-320 (1988a).Ferguson, M.A.J., Homans, S.W., Dwek, R.A., Rademacher, T.W., “Glycosylphosphatidylinositol moeity that anchors Trypanosoma brucei variant surface glycoproteinto the membrane”, Science, 239: 753-759 (1988b).Ferguson, A.J., Low, M.G., Cross, G.A.M., “Glycosyl-sn-1,2-dimyristyl-phosphatidylinositol is covalently linked to Trypanosoma brucei variant glycoprotein”, J.Biol. Chern., 260: 14547-14555 (1985).Food, M.R., Rothenberger, S., Gobathuler, R., Haidi, I.D., Reid, G.S., Jefferies, W.A.,“Transport and expression in human melanomas of a transferrin-like glycosylphosphatidylinositol-anchored protein”, I Biot. Chern., 269: 1-7 (1994).Gall, W.E., Edelman, G.M., “Lateral diffusion of surface molecules in animal cells andtissues”, Science, 213: 902-905 (1981).Gramp, G.E., Sambanis, A., Stephanopoulos, G.N, “Use of regulated secretion in proteinproduction from animal cells: an overview”, Adv. Biochern. Eng., Biotechnol., 46: 35-62(1992).Griffith, O.H., Voiwerk, J.J., Kuppe, A., “Phosphatidylinositol specific phospholipase Cfrom Bacillus cereus and Bacillus thuringiensis”, Meth. Enzyrnol., 197, 493-502 (1991).He, H.T., Barbet, J., Chaix, J.C., Goridis, C., “Phosphatidylinositol is involved in themembrane attachment of N-CAM120 the smallest component of the neural cell adhesionmolecule”, E1’vIBO J, 5: 2489-2494 (1986).HellstrOm, K.E., Hellstrôm, I., “Cellular immunity against tumor specific antigens”, Adv.Cancer Res., 12: 167-223 (1969).Hemperly, J.J., Edelman, G.M., Cunningham, B.A., “cDNA clones of neural cell adhesionmolecule (N-CAM) lacking a membrane-spanning resin consistent with evidence formembrane attachment via a phosphatidylinositol intermediate”, Proc. Nail. A cad. Sci., 83:9822-9826 (1986).Hendrickson, H.S., Dennis, E.A., “Kinetic analysis of the dual phospholipid model forphospholipase A2 action”, JBiot. Cheni., 259(9): 5734-5739 (1984).80Hu, S.L., Plowman, G.D., Sridhar, P., Stevenson, U.S., Brown, J.P., Estin, CD.,“Characterization of recombinant vaccinia virus expressing human melanoma-associatedantigen p97”, J. Viral., 62(1): 176-180 (1988).Ikezawa, H., “Phosphatidylinositol-specific phospholipase C”, J. Toxicol. Toxin Rev., 5: 1-24 (1986).Ikezawa, H., Taguchi, R., “Phosphatidylinositol-specific phospholipase C from Bacilluscereus and Bacillus thuringiensis”, Meth. Enzymol., 71, 73 1-741 (1981).Ikezawa, H., Yamanegi, M., Taguchi, R., Miyashita, T., Ohyabu, T., “Studies ofphosphatidylinositol phosphodiesterase (phospholipase C type) of Bacillus cereus I.Purification, properties and phosphatase releasing activity”, Biochin. Biophys. Acta, 450:154-164 (1976).Ishihara, A., Hou, Y, Jacobson, K., “The Thy-i antigen exhibits rapid lateral diffusion inthe plasma membrane of rodent lymphoid cells and fibroblasts” Proc. Nail. Acad Sd., 84:1290-1293 (1987).Ishihara, M., Fedarko, N.S., Conrad, H.E., “Involvement of phosphatidylinositol andinsulin in the coordinate regulation of proteoheparan sulfate metabolism and hepatocytegrowth”, J. Biol. Chern., 262: 4708-47 16 (1987).Kennard, ML., Food MR., Jefferies, W.A., Piret, J.M., “Controlled release process torecover heterologous glycosyl phosphatidylinositol membrane anchored proteins fromCHO cells”, Biotechnol. Bioeng., 42: 480-486 (1993).Kennard, M.L., Piret, J.M., “Glycolipid membrane anchored recombinant proteinproduction from CHO cells on porous microcarriers”, BiotechnoL Bioeng., in press(1994).Koch, F., Thiele, H.G., Low, M.G. “Release of the rat T cell alloantigen RT - 6.2 fromcell membranes by phosphatidylinositol-specific phospholipase C”, I. Exp. Med., 164:1338-1343 (1986).Kume, T., Taguchi, R., Tomita, M., Tokuyama, S., Morizawa, K., Nakachi, 0., Hirano,J., Ikezawa, H., “The study of phosphatidylinositol-specific phospholipase C from Bacillusthuringiensis: Synthesis of homogeneous substrates, substrate specificity and otherproperties”, Chem. Pharm. Bull., 40(8): 2133-2137 (1992).Kupke, T., Lechner, M., Kaim, G., GOtz, F., “Improved purification and biochemicalproperties of phosphatidylinositol - specific phospholipase C of Bacillus thuringiensis”,Eur. J. Biochem., 185: 15 1-155 (1989).81Kuppe, A., Evans, L.M., McMillen, D.A., Griffith, O.H., “Phosphatidylinositol - specificphospholipase C of Bacillus cereus: Cloning, sequencing and relationship with otherphospholipases”, J. Bacteriol., 171(11): 6077-6083 (1989).Leigh, A.J., Volwerk, J.J., Griffith, OH., Keana, J.F.W., “Substrate stereospecificity ofphosphatidylinositol-specific phospholipase C from Bacillus cereus examined using theresolved enantiomers synthetic myo-inositol 1-C4-nitriphenyl phosphate”, Biochern., 31:8978-8983 (1992).Lisanti, M.P., Caras, I.W., Davitz, M.A., Rodriguez-Boulan, E., “A glycophospholipidmembrane anchor acts as an apical targeting signal in polarized epithelial cells”, J. Cell.Biol., 109: 2145-56 (1989).Lisanti, M.P., Rodriguez-Boulan, E., Saltiel, A.R., “Emerging functional roles of theglycosyl-phosphatidylinositol membrane protein anchor”, I Menibr. Blot., 117: 1-10(1990).Low, M.G., “Biochemistry of the glycosyl-phosphatidylinositol membrane proteinanchors”, Biochern. J., 244: 1-13 (1987).Low, M.G., “The glycosyl-phosphatidylinositol anchor of membrane proteins”, Biochirn.Biophys. Acta, 988: 427-454 (1989).Low, MG., Finean, J.B., “The anchor of phosphatidylinositol specific phospholipases onmembranes”, Biochern. J., 154: 203-208 (1976).Low, M.G., Finean, J.B., “Release of alkaline phosphatase from membranes byphosphatidylinositol - specific phospholipase C”, Biochern. 1, 167: 281-284 (1977).Low, M.G., Finean, J.B., “Specific release of plasma membrane enzymes by aphosphatidylinositol-specific phospholipase C”, Biochirn. Biophys. Acta., 508: 565-570(1978).Low, M.G., Futerman, A.H., Ackermann, K.E., Sherman, W.R., Silman, I., “Removal ofcovalently bound inositol from Torpedo acetyicholinesterase and mammalian alkalinephosphatases by deamination with nitrous acid”, Biochern. 1, 241: 615-619 (1987).Low, M.G., Kincade, P.W., “Phosphatidylinositol is the membrane-anchoring domain ofthe Thy-1-glycoprotein”, Nature, 318: 62-64 (1985).Low, M.G., Saltiel, A.R., “Structural and functional roles of glycosyl phosphatidylinositolin membranes”, Science, 239: 268-275 (1988b).Missirlis, Y.F., Lemm, W., “Modern aspects of protein adsorption on biomaterials”,Kinetics ofprotein adsorption, Kiuwer Academic Publishers, London, 1991.82Moran, P., Caras, I.W., “Fusion sequence from non-anchored proteins to generate a frillyfunctional signal for glycophosphatidylinositol membrane anchor attachment”, J. CellBiol., 115(6): 1595-1600(1991).Nakabayashi, T, Ikezawa, H., “Release of alkaline phosphodiesterase I from rat kidneyplasma membrane produced by the phosphatidylinositol-specific phospholipase C ofBacillus thuringiensis”, Cell Struct. Fund., 9: 247-263 (1984).Nakabayashi, T, Ikezawa, H., “Alkaline phosphodiesterase I release from eukaryoticplasma membranes by phosphatidylinositol-specific phospholipase C. I. The release fromrat organs”, J.Biochern., (Tokyo), 99(3): 703-712 (1986).Reed, R., “Spurious cell surface receptors: inadequate correction for saturable, nonspecific binding mimics receptor binding”, Anal. Biochem., 185: 160-163 (1990).Reisser, H., Oettgen, H., Yeh, E.T.H., Terhorst, C., Low, M.G., Benacerraf, B.,Rock, K.L., “Structural characterization of the TAP molecule: a phosphatidylinositollinked glycoprotein distinct from the T cell receptor/T3 complex and Thy-i”, Cell, 47:365-3 70 (1988).Renner, W.A., Jordan, M., Eppenberger, H.M., Leist, C., “Cell - cell adhesion andaggregation: influence of growth behaviour of CHO cells”, Biotechnol. Bioeng., 41: 188-193 (1993).Richardson, D., Baker, E., “The uptake of inorganic iron complexes by human melanomacells”, Biochirn. Biophys. Ada., 1093: 20-28 (1991).Rose, T.M., Plowmann, G.D., Teplow, D.B., Dreyer, W.J., HellstrOm, K.E., Brown, J.P.,“Primary structure of the human melanoma associated antigen p97 (melanotransferrin)deduced from the mRNA sequence”, Proc. Natl. Acad. Sci., 83: 1261-1265 (1986).Sambanis, A., Stephanopoulos, G., Sinskey, A.J., Lodish, H.F., “Use of regulatedsecretion in protein production from animal cells: an evaluation with the AtT-20 modelcell line”, Biotechnol. Bioeng., 35: 771-780 (1990a).Sambanis, A., Stephanopoulos, G., Lodish, H.F., “Multiple episodes of induced secretionof human growth hormone from recombinant AtT-20 cells”, Cytotechnol., 4: 111-119(1990b).Scallon, B.J., Kado-Fong, H., Nettleton, M.Y., Kochan, J.P., “A novel strategy forsecreting proteins: Use of phosphatidylinositol-glycan specific phospholipase D to releasechimeric phosphatidylinositol-glycan anchored proteins”, Bio/Technol., 10: 550-556(1992).83Shashidhar, M.S., Volwerk, J.J. Keana, J.F.W., Griffith, O.H., “A fluorescent substrate forcontinuous assay of phosphatidylinositol-specific phospholipase C: synthesis andapplication of 2-naphthyl myo-inositol-1-phosphate”, Anal. Biochem., 198: 10-14 (1991a).Shashidhar, M.S., Voiwerk, J.J., Griffith, O.H., Keana, J.F.W., “A chromogenic substratefor phosphatidylinositol-specific phospholipase C: 4-nitrophenyl myo-inositol-1-phosphate”, Chem. Phys. Lipids, 60: 101-110(1991b).Shin, Y., Levinthal, C., Levinthal, F., Hubell, W.L., “El binding to membranes: Timeresolved studies of spin labeled mutants”, Science, 259: 960-963 (1993).Shukia, S.D., Coleman, R., Finean, J.B., Michell, R.H., “Selective release of plasma-membrane enzymes from rat hepatocytes by a phosphatidylinositol-specific phospholipaseC”, Biochem. J., 187: 277-280 (1980).51cm, M.W., Logan, G.F. Jr., “Partial purification and properties of two phospholipases ofBacillus cereus”, J. Bacteriol. 85: 369-381 (1963).Slein, M.W., Logan, G.F., “Characterization of the phospholipases of Bacillus cereus andtheir effects on erythrocytes, bone and kidney cells”, J. Bacteriol. 90: 69-8 1 (1965).Steirnberg, J.. Low, M.G., Flaherty, L., Kincade, P.W., “Removal of lymphocyte surfacemolecules with phosphatidylinositol-specific phospholipase C: Effects on mitogenresponses and evidence that ThB and certain Qa antigens are membrane-anchored viaphosphatidylinositol”, J. Immunol., 38: 3877-84 (1987).Steinman, R.M., Brodie, S.E., Cohn, Z.A., “Membrane flow during pinocytosis, asteriologic analysis”, I Cell Blot., 68: 665-687 (1976).Sundler, R., Alberts, A.W., Vagelos, P.R., “Enzymatic properties of phosphatidylinositolinositol phosphohydrolase from Bacillus cereus”, J. Biot. Cheni., 253, 4175-4179 (1978).Taguchi, R., Ikezawa, H., “Phosphatidylinositol-specific phospholipase C fromClostridium novyi Type A”, Arch. Biocheni. Biophys., 186, 196-201 (1978).Taguchi, R., Asahi, Y., Ikezawa, H., “Purification and properties of phosphatidylinositolspecific phospholipase C from Bacillus thuringiensis”, Biochini. Biophys. Acta, 619, 48-57(1980).Takasue, Y., Yokota, K., Nishi, Y., Taguchi, R., Ikezawa, H., “Solubilization of trehalasefrom rabbit renal and intestinal brush-border membranes by a phosphatidylinositol-specificphospholipase C”, FEBsLeIt., 210: 5-8 (1986).84Tse, A.G.D., Barclay, A.N., Watts, A., Williams, A.F., “A glycophospholipid tail at thecarboxyl terminus of the Thy-i glycoprotein of neurons and thymocytes”, Science, 230:1003-1008 (1985).Turner, A.J., “Molecular and cell biology of membrane proteins - Glycolipid anchors ofcell surface proteins”, Series of Molecular Biology, Ellis Horwood Publishers, Toronto,ON (1990).Voiwerk, J.J., Wetherwax, P.B., Evans, L.M., Kuppe, A., Griffith, O.H.,“Phosphatidylinositol-specific phospholipase C from Bacillus cereus: Improvedpurification, amino acid composition and amino-terminal sequence”, J. Cell. Biochern., 39:3 15-325 (1989).85Appendix 1Growth of CHO cells in different growth mediaTables labelled according to the corresponding figuresFigure 4.1 Growth of CHO cells in CHO-S-SFM ITime Cell Concentration Glucose Viability[day] [106 Cells/mL] [gfL] [%]0 0.10 3.80 100.02 3.60 3.50 99.73 1.24 3.20 97.64 3.00 2.40 98.45 4.90 1.40 94.26 5.20 0.75 92.97 5.80 0.58 90.68 5.90 0.30 89.79 5.50 0.18 88.010 4.70 0.15 82.511 3.50 0.14 71.412 3.10 0.12 76.5Figure 4.6 Growth of CHO cells in HBCHOTime Cell Concentration Glucose Viability[day] [106 Cells/mL] [g/L] [%]0 0.07 1.31 68.01 0.06 1.20 74.02 0.42 1.00 94.43 0.58 0.58 84.15 1.30 0.12 92.26 1.50 0.11 78.17 0.76 0.10 74.086Figure 4.2 Growth of CHO cells in Ham’s F12 with 5 ig/mL insulin, 5 tg/mL transferrin,10 nM sodium selenite, 50 BSA and 50 .tg/mL DNase.Time Cell Concentration Glucose Viability[day] [106 Cells/mU [g/L] [%]0 0.20 1.70 87.01 0.25 1.61 96.22 0.27 1.45 96.43 0.37 1.40 90.24 0.48 1.28 92.35 0.61 1.14 95.36 0.65 1.10 81.37 0.82 0.98 79.68 0.83 0.85 80.69 0.58 0.67 65.910 0.50 0.60 50.011 0.30 0.45 33.012 0.20 0.25 23.5Figure 4.3 Growth of CHO cells in Ham’s F12 with 5 tg/mL insulin, 5 tg/mL transferrin,10 nM sodium selenite, 300 BSA and 50 j.ig/mL DNase.Time Cell Concentration Glucose Viability[day] [106 Cells/m.LJ [g/L] [%]0 0.25 1.55 75.81 0.34 1.44 97.12 0.71 1.27 91.03 0.89 1.30 88.14 0.83 1.15 87.45 0.98 1.00 83.16 1.25 1.01 82.87 1.22 0.70 81.98 1.23 0.71 75.09 0.73 0.54 65.810 0.70 0.32 56.011 0.56 0.20 51.812 0.50 0.15 51.5Figure 4.4 Growth of CHO cells in DMEM!F12 with 50 jig/mL DNase.Time Cell Concentration Glucose Viability[day] [106 Cells/mL] [g/L] [%]0 0.20 2.92 87.01 0.21 2.82 95.52 0.36 2.57 97.33 0.50 2.50 93.64 0.88 2.30 96.35 1.29 2.30 98.16 1.56 2.10 96.37 1.93 1.80 95.88 2.10 1.70 96.89 1.86 1.50 90.710 1.64 1.43 87.711 1.48 1.25 92.512 1.38 0.95 87.9Figure 4.5 Growth of CHO cells in DMEMIFI2 with 5 pg/mL insulin, 5 tg/mL transferrin,10 nM sodium selenite, 50!mL BSA and 50 p.g/mL DNase.Time Cell Concentration Glucose Viability[day] [106 Cells/mL] [g/L] [%]0 0.20 2.92 87.01 0.28 2.78 96.62 0.35 2.60 97.23 0.56 2.50 94.94 0.83 2.45 96.55 1.18 2.35 96.76 1.47 2.25 97.07 1.85 1.97 97.18 1.70 1.86 94.79 1.87 1.64 9L210 1.58 1.62 85.411 1.50 1.45 86.612 1.26 1.33 81.08788Figure 4.7 Growth of CHO cells in CHO-S-SFM IITime Cell Concentration Glucose Viability[day] [106 Cells/mLj [g/L] [%J0 0.06 3.74 81.81 0.09 3.68 89.72 0.31 3.40 93.93 0.90 2.87 96.85 3.46 1.68 96.76 2.82 1.18 92.87 2.12 0.81 74.18 2.02 0.48 64.79 0.83 0.32 33.210 0.00 0.18 0.0Figure 4.8 Growth of CHO cells in CHO-S-SFM II and 100 tg/mL DNaseTime Cell Concentration Glucose Viability[day] [106 Cells/mU [g/L] [%J0 0.20 3.80 100.01 0.45 3.50 97.82 1.10 3.35 97.43 1.54 2.95 96.94 3.00 2.50 98.45 5.00 2.20 98.06 7.00 1.60 98.27 5.20 0.98 89.48 4.60 0.58 78.089Appendix 2Variation of initial cell densityThe effect of varying the initial cell density on the growth profiles was investigated. The aimwas to increase the final cell concentration by varying inoculation concentration. Figure 7.1shows that cultures started with higher inoculum grew to higher cell densities. Cell culturestarted with 3.5x104 cells/mL of medium reached a maximum of 4x106 cells/mL and theculture that was started with greater lx 106 cells/mL achieved a maximum of about7x106 cells/mL. There appears a limit to the high density inoculum that can be used to boostthe growth of cells. A cell culture that was inoculated with 2.2x106 cells/mL, achieved7x106 cells/mL only.2 7.0x106Ej 6.Ox1OC-)5.1Oj 4.0x1063.0x1061)0C)a)C-)Time (day)Figure 7J Growth profile of cells growing in CHO-S-SFM I. Cultures started at differentinoculum densities.90-Ja)U)00CDTime (day)Figure 7.2 Cells growing in CHO-S-SFM I. Glucose concentrations for cultures that werestarted at different initial cell density.Inoculating the cells at high cell concentration is a poor scale-up technique as theinoculum culture will have to be maintained for longer periods to achieve high cellconcentrations. The four initial cell concentrations that were investigated gave the followingfinal to initial cell concentration ratios: (i) 7/1.5 (ii) 3/0.03 (iii) 7/1 (iv) 7/2.2. Theculture inoculated with 3x104 cells/mL gave a 100-fold increase. Starting at very low celldensities (approximately 1x104 cells/mL) will also increase the total time to achieve themaximum cell density because of the initial lag time, therefore we will lose up to two days.91Tables labelled according to the corresponding figuresFigures 7.1 and 7.2 Effect of innoculum concentration on final cell densityTime Cell Concentration Glucose[day] [106 Cells/mLj [mg/mL]A B C D A B C D0 1.48 0.35 1.06 2.16 3.69 3.67 3.81 3.881 2.76 0.44 2.38 4.35 2.22 3.25 2.39 1.852 4.20 1.60 4.86 6.70 0.84 2.33 1.20 0.753 6.48 3.90 7.00 7.00 0.32 1.05 0.57 0.304 6.88 3.00 7.04 3.00 0.14 0.19 0.145 6.50 4.60 0.12 0.1392Appendix 3Effect of harvesting conditions on p97 removalTables labelled according to the corresponding figuresi) Figure 4.9 Effect of incubation time on p97 harvestPT-PLC p97 removal [%J[mU/mL] 60 mm 30mm 15 mm10 52.4 38.9 23.95 318 23.2 13.61 9.5 6.1 2.70.5 6.2 3.6 0.90 6.2 4.5 3.6ii) Figure 4.10 Effect of P1-PLC volume on p97 harvestVolume of p97PT-PLC removal[jiLl [%]50 16.7100 19.7200 22.5400 28.093Appendix 4P1-PLC assay resultsTables labelled according to the corresponding figuresi) Figure 4.11 P1-PLC assay based on flow cytometryPT-PLC p97 removal Error[mU/mU [%J ±156.0 97.5 -78.0 94.1 -15.6 67.7 1.97.8 52.5 0.51.6 15.1 -0.8 7.9 -0.2 2.7 -ii) Figure 4.12 Standard curve based on flow cytometry and immunofluorescencePT-PLC p97 p97 removal[mU/mU [ig/mL] [%]156.0 10.60 98.978,0 9.80 97.915.6 9.50 87.17.8 8.80 74.71.6 4.50 36.50.8 2.98 25.20.2 1.20 10.394Appendix 5Repeated harvesting process resultsTables labelled according to the corresponding figuresFigures 4.13 and 4.14Harvest p97 removal Cumulative Calculatedp97 p97[#] [%j [p.g/mLJ [p.g/mLJ0 0.0 0.00 0.001 58.1 59.7 59.72 34.3 109.0 111.43 36.6 110.5 117.24 28.0 105.5 116.65 28.0 103.6 118.96 20.4 105.8 121.17 20.4 115.3 139.0Figures 4.21 and 4.22Harvest p97 removal Cumulative p97 Calculated p97{#] {%] [jig/mL] [.tg/mL]0 0 0.0 0.01 75.7 72.6 72.62 58.0 148.6 150.43 48.4 118.1 123.64 31.3 148.5 157.05 26.1 148.9 161.16 91.3 196.0 211.97 92.0 250.1 270.98 79.4 298.5 325.69 55.1 247.2 281.710 52.1 275.7 316.495Figures 4.23 and 4.24300 mU/mL P1-PLC 30 mU/mL P1-PLCHarvest p97 Cumulative Calculated p97 Cumulative Calculatedremoval p97 p97 removal p97 p97[#J [%] [ig/mL] [jig/mU [%] [jig/mL] [jig/mU1 97.1 85.2 85.2 84.9 62.1 62.12 97.5 147.6 162.4 91.9 129.7 135.93 96.5 202.7 226.0 93.3 160.3 179.54 98.2 249.5 293.7 95.2 193.4 228.75 98.1 274.3 343.3 93.1 240.3 294.896Appendix 6p97 and P1-PLC stabilityTables labelled according to the corresponding figuresi) Figure 4.15 p97 stability at 37°CTime p97 Error[hi [jig/mU ±0 9.65 0.01 10.51 1.412 9.69 0.613 10.25 0.9424 9.92 0.4151 8.60 0.30ii) Figure 4.16 P1-PLC stability at 37 °CTime PT-PLC[day] [mU/mL]0.00 60.780.004 54.480.007 54.480.01 50.160.02 55.730.04 55.110.08 55.110.21 49.931.00 56.365.00 46.5314.0 50.16iii) Figure 4.17 Effect of pH on PT-PLC and p97. P1-PLC solution of 16 mU/mL wasused.pH p97 removal p97 Error[%J {.tg/mL] ±6.00 93.5 12.97 0.886.50 84.7 10.23 0.857.00 75.0 9.08 0.357.50 64.2 10.86 0.337.90 51.2 6.85 1.679798Appendix of P1-PLC on CHO cells over 120 mm.p97 removal[%]012.012.810.412.0Adsorption and desorption of P1-PLC to the cell surfaceTables labelled according to the corresponding figuresi) Figure 4.18 Adsorption of PT-PLC on CHO cells over a period of 60 mm. A control ofPT-PLC incubation without cells was also done for 60 mm.Time P1-PLC P1-PLCAdsorption Control[mini [mU/mL] [mU/mL]0 18 183 2.0 - Figure 4.19Time[mini0315.060.0120.099iii) Figure 4.20 Data for the P1-PLC equilibrium curvePT-PLC in solution PT-PLC adsorbed[mU/mU [ (Y7 mU/cellj0.0 0.01.0 0.165.5 0.399.5 1.5615.5 2.2530.0 5.037.5 9.05155.0 4.05100Appendix 8Continuous harvesting process resultsTables labelled according to the corresponding figuresFigure 4.25Time Cell Concentration Viability[day] [106 Cells/mU [%]0 mU/mL 3 mU/mL 30 mU/mL 0 mU/mL 3 mU/mL 30 mU/mL1 0.12 0.14 0.15 80.0 93.3 93.82 0.34 0.34 0.40 91.9 94.4 95.23 0.53 0.56 0.76 91.4 90.3 97.44 1.60 1.32 1.52 95.8 95.7 94,75 1.90 2.20 2.80 93.1 91.7 94.06 1.80 1.50 1.90 90.5 87.2 83.07 1.20 1.10 1.75 72.3 68.8 70.38 1.60 0.69 1.65 67.8 48.3 73.39 0.77 0.68 1.40 43.5 57.6 64.510 0.29 0.67 1.30 16.2 44.7 41.911 0.04 0.24 0.62 2.2 19.4 19.3101Figure 4.26Time Cell surface fluorescence Glucose Concentration[day] [fluorescence/cell] [mglmL]0 mU/mL 3 mU/mL 30 mU/niL 0 mU/mL 3 mU/mL 30 mU/mLPT-PLC P1-PLC P1-PLC P1-PLC P1-PLC PT-PLC0 320.0 320.0 320.0 3.90 3.90 3.901 314.4 168.1 129.1 3.70 3.40 3.722 257.6 150.7 180.8 3.33 3.37 3.353 241.2 98.4 128.5 2.88 2.87 2.624 190.2 47.4 40.5 1.88 2.07 1.915 161.3 35.6 15.3 1.55 1.50 1.346 135,7 23.7 8.50 1.32 1.33 1.237 89.8 26.5 10.9 0.95 0.98 0.838 55.5 20.0 7.5 0.92 1.04 0.509 32.1 22.6 9.2 0.63 0.93 0.3810 27.7 16.2 8.2 0.66 0.87 0.1711 25.8 13.6 10.6 0.68 0.77 0.13Figure 4.27Time Cumulative p97[day] [ig/mL]0 mU/mL 3 mU/mL 30 mU/mLPT-PLC PT-PLC PT-PLC0 0.3 0.4 0.41 0.3 0.4 0.42 0.4 0.8 0.83 0.7 1.5 2.34 1.4 3.4 4.55 2.3 5.8 7.96 3.0 8.5 8.87 3.1 8.6 9.78 3.1 9.2 11.59 3.2 9.3 11.510 3.5 10.5 13.411 3.8 11.6 15.3102Appendix 9Calculations for p97 yield on glucose and P1-PLCi) Cyclic harvestingReactor Volume: 250 mL at 4x106 cells/mLTotal time: 8 daysNumber of harvests: 5Cells per harvest: 1x109 cellsPT-PLC concentration: 30 mUImLPT-PLC volume per harvest: 5 mLp97 recovered from single harvest (estimated from the work of Kennard et al., 1993;based on harvesting cells in 0.5 mL P1-PLC): 62 .tg/mLCHO-S-SFM I medium per harvest: 250 mLInitial glucose concentration: 4 mg/mLFinal glucose concentration: 0.8 mg/mLTotal glucose consumed = (250 mL/harvest * 5 harvests) *(4 - 0.8) mg/mL glucose= 4000 mgTotal PT-PLC used = (30 mU/mL) * (5 mL per harvest) *(5 harvests)750 mU P1-PLCTotal p97 harvested (62 .ig/mL/harvest) * (5 mL/harvest) * (5 harvests)= 1550 igTherefore:Yield on glucose = (1550 ig p97)/Ql000 mg glucose)= 0.39 i.g p97/mg glucoseYield on PT-PLC = (1550 jig p97)/(750 mU P1-PLC)= 2.07 jig p97/mU PT-PLCAverage p97 /cell-day = 1550 jig p97/(1x10 cells * 8 days)103ii) Repeated harvesting processa) Initial P1-PLC concentration 300 mU/mLReactor volume: 250 mL at 4x106 cells/mLTotal time: 8 daysNumber of harvests: 5Cells per harvest: 1x109 cellsInitial P1-PLC concentration: 300 mU/mLPT-PLC replenishment after each harvest: 1500 mUTotal PT-PLC volume: 5 mLTotal p97 harvested: 1715 jtgCHO-S-SFM I medium per harvest: 250 mLInitial glucose concentration: 4 mg/mLFinal glucose concentration: 0.8 mg/mLTotal glucose consumed = (1250 mL medium) * (4 - 0.8) mg/mL glucose= 4000 mgTotal P1-PLC used (300 mU/niL) * (5 ml) + (1500 mU/harvest * 4 harvests)7500 mU P1-PLCTherefore:Yield on glucose = (1715 .tg p97)1(4000 mg glucose)= 043 .tg p97/mg glucoseYield on PT-PLC = (1715 .tg p97)1(7500 mU PT-PLC)= 0.23 ig p97/mU P1-PLCAverage p97/cell-day = 1715 ig p97/(1x10 cells * 8 days)2.14x10 jig p97/cell-day104b) Initial P1-PLC concentration 30 mU/mLReactor volume: 250 niL at 4x106 cells/mLTotal time: 8 daysNumber of harvests: 5Cells per harvest: 1x109 cellsInitial P1-PLC concentration: 30 mU/mLPT-PLC replenishment after each harvest: 150 mUTotal P1-PLC volume: 5 mLTotal p97 harvested: 1474 tgCHO-S-SFM I medium per harvest: 250 mLInitial glucose concentration: 4 mg/niLFinal glucose concentration: 0.8 mg/mLTotal glucose consumed = (1250 mL medium) * (4 - 0.8) mg/mL glucose= 4000 mgTotal PT-PLC used = (30 mU/mL) * (5 niL) + (150 mU/harvest * 4 harvests)750 mU P1-PLCTherefore:Yield on glucose = (1474 jig p97)1(4000 mg glucose)= 0.37 i’g p97/mg glucoseYield on PT-PLC = (1474 jig p97)1(750 mU P1-PLC)= 1.97 jig p97/mU PT-PLCAverage p97/cell-day = 1474 jig p97/(1x10 cells * 8 days)= 1.84x107pg p97/cell-day105iii) Continuous harvesting processReactor volume: 250 mL (batch culture)PT-PLC concentration: 30 mU/mLTotal time: 8 daysTotal p97 harvested was: 11.5 .tg/mL * 250 mL = 2875 igAverage initial glucose concentration: 3.9 mg/mLAverage final glucose concentration: 0.5 g/LTotal glucose consumed = 250 mL * 3.4 mg/mL= 850 mg glucoseTotal PT-PLC used = 250 mL * 30 mU/mL7500 mU P1-PLCTherefore:Yield on glucose = (2875 pg p97)1(850 mg glucose)= 3.4 jig p97/mg glucoseYield on PT-PLC = (2875 jig p97)1(7500 mU PT-PLC)= 0.38 jig p97/mU P1-PLCTotal cells = 250 mL * 2.5x106 cells/mL = 6.25x108cellsAverage p97/cell-day = 2875 jig p97/(6.25x 108 cells * 8 days)= 5.75x1W jig p97/cell-day


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