TT echnological application ofliving things is as old asbeer and wine. The use ofmicrobial fermentation asan expression system,

however, began in 1973 with thefirst genetic engineering experiment:A gene from the African clawedtoad was inserted into laboratoryEscherichia coli bacteria. Since then,those easily cultivated microbes haveseen numerous research andindustrial uses. Microbialfermentation is the most widelyapplied biotechnology method.

Besides the familiar alcoholicbeverages, products made bymicrobial fermentation (somerecombinant, others not) are far-ranging: from ethanol, dyes, andother chemicals to enzymes, foodsand food additives, vitamins, soyproducts, vaccines for animals andpeople, antibiotics and antifungalagents, steroids, diagnostic agents,and enzyme inhibitors (1). Bacteriaand yeast grow fast in low-costmedia; offer high expression levelsof the proteins they can make,which they sometimes secrete intotheir circulating medium; and bothcan withstand rough treatmentcompared with animal cells (asdescribed in Chapter 2).

THE SCIENCE OF FERMENTATION

“In its simplest form, the bioprocesscan be seen as just mixingmicrooganisms with a nutrientbroth and allowing the componentsto react. . . . More advanced and

sophisticated processes operating atlarge scale need to control the entiresystem so that the bioprocess canproceed efficiently and be easily andexactly repeated with the sameamounts of raw materials andinoculum to produce precisely thesame amount of product.” (1)Fermentation of recombinant celllines begins with geneticengineering of microbes classified“generally recognized as safe”(GRAS), such as E. coli, Bacillussubtilis, Streptomyces species, andSaccharomyces cerevisiae andSchizosaccharomyces pombe yeasts.

Bacterial Fermentation: The mostcommon microbial source forrecombinant protein production is E.coli because the most is known aboutits genetics, and thus it is the specieswith which pioneering molecularbiologists were familiar (2). B. subtilisand its relatives have been used,mainly because of their greater

tendency to secrete proteins intotheir environment. Many of thosenative proteins have industrial uses aswell, but unfortunately some of themare proteases, enzymes that break upother proteins (such as the foreigntherapeutic molecules the bacteriamay be genetically engineered tomake). So the ultimate product yieldscan be lower than companies wouldlike. Various Streptomyces species areunder study in recombinantfermentation, but so far they havedemonstrated low expression levels.Pseudomonas fluorescens may havegreater potential. But E. coli “remainsone of the most attractive because ofits ability to grow rapidly and at highdensity on inexpensive substrates, itswell-characterized genetics, and theavailability of an increasingly largenumber of cloning vectors andmutant host strains” (3).

Bacterial genes are contained in acircular genome and on small circularpieces of extragenomic double-stranded DNA elements calledplasmids in their nucleus-free cells.Self-replicating plasmids containregulatory regions (promoter regionsand origins of replication) that makethem ideal candidates for use ingenetic engineering. They can bemanipulated using restrictionenzymes, cloning vectors (such asbacteriophage viruses), and relativelysimple procedures. Certain genesegments with the ability to promote(promoters), direct, or terminatetranscription of the foreign DNA areoften involved as well.

C H A P T E R ONE

8 BioProcess International JUNE 2004 SUPPLEMENT

NEW BRUNSWICK SCIENTIFIC CO., INC.(WWW.NBSC.COM)

Microbial FermentationThe Oldest Form of Biotechnology

by Cheryl Scott

SUPPLEMENT

Several biopharmaceuticals on the market are producedby recombinant bacteria: somatostatin, insulin, bovinegrowth hormone for veterinary use, �-1 antitrypsin,interleukin-2, tumor necrosis factor, beta and gammainterferon, and some antibody fragments. Expressionlevels range from less than 0.05% to 25% of total cellularprotein. Bacteria can make collagen protein domains at100 mg/L, which suggests a cost-effective productionmethod (4). However, if whole collagen glycoproteins areneeded (in a biomedical device application, for example),this convenient system would not be the best choice.

Very high expression levels of heterologous proteinsexpressed in bacteria may lead to the formation ofinclusion bodies. In such cases, the protein molecules clumptogether (aggregate) in the cytoplasm to create irregularorganelle-like structures (about 1 µm in diameter). Thispresents a good-news–bad-news scenario: Dense inclusionbodies are easily separated from broken cells bycentrifugation, thus facilitating product purification afterthe cells are homogenized. But the aggregated, misfoldedproteins are also insoluble, which can make furtherprocessing difficult. Organic solvents, detergents, orchaotropic substances can be used to denature thoseclumps and solubilize the proteins. But the next problemencountered will be that their three-dimensional structureis almost always wrong by that point.

To get correct, biologically active proteins, arenaturation step must follow: Inclusion bodies aredissolved using chaotropes, then diluted so the proteinscan properly refold. And the larger the protein, themore difficult that usually is. More than 90% of the 200people working at Eli Lilly’s Humulin plant are involvedin recovery processes (1). So along with prokaryotes’general inability to perform complex posttranslationalmodifications (such as glycosylation, discussed in moredetail below and in Chapter 2), inclusion bodies are alimiting factor in the use of bacteria for recombinantbioproduction.

A few bacterial expression options are available thatavoid the inclusion body issue entirely (3). For example,some species secrete products rather than retainingthem within cellular walls. Cultivation of E. coli at lowertemperatures (30 °C rather than 37 °C) sometimesprevents aggregation of heterologous proteins.Coexpression of chaperone proteins (or increasedproduction of innate cofactors) that encourage properfolding of the biotherapeutic molecule may also help.And combining the gene for the protein of interest withone expressing a highly soluble native cytoplasmicprotein (a fusion partner) may offer an answer. But theresulting fusion proteins must be chemically orenzymatically cleaved in downstream processing so thefusion partner can be purified away.

There may even be an answer to the glycosylationproblem. In March 2004, researchers in Austria reportedthe discovery of three strains of bacteria that build sugarcoatings on their cell surfaces in response to highertemperatures. A team at the Center forNanoBiotechnology (part of the University of Natural

CONSTANTSYSTEMS LTD

Cell DisruptionSolutions

constant@score-group.com www.celldisrupter.co.uk

Reproducible High Pressure CellDisruption

Controlled Temperature

High Throughput

40 Kpsi max pressureVolume per shot 0.1ml - 20ml

■ Allows loading of solidsample, cell paste orsuspended sample

■ Protein extraction■ Enzyme release

■ Single cell isolation■ DNA shearing■ Selective breakage■ Liposomes

■ Cell rupture■ Tissue disruption

CONSTANT CELL DISRUPTION SYSTEMS deliveruniquely constant and repeatable pressuredelivering high yielding cell lysis.

With 14 years experience, our products fromresearch through to production can accommodateyour every cell disruption need.

Cooling jacket

Research Scale

40 Kpsi max pressureFlow rate 1.8-188L/hrAutomatic CIP/SIP optionsavailable

■ GMP validation ■ Automatic supply of sample■ Full bio-containment

options availableConstant Systems LimitedLow March, Daventry, Northants. NN11 4SDUNITED KINGDOM

Tel: +44 (0)1327 314146 Fax: +44 (0)1327 314147

Cell Lysis Success

C.albicansafter one pass.

■ Digitally displayedtemperature control

■ Data capture packageavailable

Resources and Applied Life Sciencein Vienna, Austria) has been studyinga mechanism for glycosylation inbacteria (5). Their work could lead toa method of specifically directing it.

One problem that can beresolved only through effectivedownstream processing is thepresence of pyrogenic bacterialendotoxins (proteins that inducefever). To avoid them entirely, youmay want to consider a yeast-basedexpression system.

Fungal Fermentation: For about$50–100/gram of final product, youcan make insulin or a subunit vaccine,among other things, by fermentinggenetically engineered yeast.Although their product yields may berelatively low (around 5%), they canbe as high as 40% of total solubleprotein (6), and yeasts offer certainadvantages over bacteria as arecombinant expression system. Chiefamong these is their ability aseukaryotic organisms to performcertain posttranslational modificationson the proteins they make.

For many of the same reasons

that E. coli is popular amongbacterial production systems, S. cerevisiae (baker’s or brewer’syeast) is the fungal species mostcommonly found in fermentation

processes, recombinant or not. It isthe one with which people have themost experience (thousands of years’worth) and thus the bestunderstanding (the full genome wassequenced in 1996). S. cerevisiae hasbeen genetically engineered toproduce a wide range of proteins:antigens to heptitis B, influenza,and polio; human growth hormone

and insulin; antibodies and antibodyfragments; human growth factors;interferon and interleukin; bloodcomponents such as human serumalbumin; and tissue plasminogenactivator (3). Other species thathave been studied include S. pombeand Pichia pastoris.

In yeast, the expression ofrecombinant proteins to at least 150 mg/L (6) has beendemonstrated in shake flasks. Yeastscan make some collagen proteinscost-effectively, building wholecomplex glycoproteins better thanbacteria can. Expression levels runaround 15 mg/L. However, “giventhe various sets of modifications thata fibrillar collagen molecule mustundergo, the mammalian systemthat is closest to the normal in vivocollagen secretion pathway seemsthe most suitable. Yeasts . . . do notpossess all the posttranslationalequipment required to fulfill thesemodifications” (4). Coexpression ofa necessary enzyme has helped someyeasts produce correctly foldedcollagen molecules. Some

One aspect ofprotein expressionthat has proven tobe very tricky isglycosylation.

����

heterologous proteins can be lethalto the yeast cells that make them —but a methanol-induced promoterhas been developed to meet thatchallenge. Once the yeast culturehas reached a certain cell density,changing its growth medium tomethanol induces expression of thetherapeutic protein in large amounts(effectively ending the batch).Similar inducible promoter geneshave been used with bacteria and S. cerevisiae (7), but P. pastoris onlyworks with methanol, and it is theonly one that does.

Compared to other eukaryoticexpression systems, Pichia [pastoris]offers many advantages because itdoes not have the endotoxinproblem associated with bacterianor the viral contaminationproblem of proteins produced inanimal cell culture. Furthermore, P. pastoris can utilize methanol as acarbon source in the absence ofglucose. . . . Since the proteinsproduced in P. pastoris are typicallyfolded correctly and secreted into

the medium, the fermentation ofgenetically engineered P. pastorisprovides an excellent alternative toE. coli expression systems. Anumber of proteins have beenproduced using this system,including tetantus toxin fragment,Bordatella pertussis pertactin,human serum albumin, andlysozyme. . . . Another advantageof Pichia pastoris is its prolificgrowth rate. (8)

Yeasts that grow fast can presenta challenge in process control (asdiscussed in the “FermentationTechnology” section below).

“Optimum protein production in P. pastoris occurs at 30 °C, and thatall protein expression ceases at 32 °C” (8). That’s a pretty narrowtemperature range to hold an active

CAMBREX BIO SCIENCE INC. (WWW.CAMBREX.COM)

TTaabbllee 11:: Fermentation process control settings (9)

Parameter Bacteria Yeast

Temperature 32–35 °C 30 °CpH 7.0–7.2 5.0Agitation speed 200–300 rpm* 200–300 rpm*Dissolved oxygen 30% 30%Gas sparge rate** 0.5–1.0 VVM 0.5–1.0 VVM

* Maximum value 1000 rpm** Air introduced at a rate measured as volume � vol–1 � min–1 (VVM)

culture to when “it is notuncommon for the temperature [inshake-flask cultures] to increase 25 °C if left uncontrolled” (8).Many people who have made breadfrom scratch have experienced someindication of this while letting thedough rise. But well-managedfermentation processes are expectedin biotechnology, as one reviewerpointed out after reading thischapter: “Temperature is readilymanaged at industrial scale within atight band of at least �0.05 °C.”

One aspect of protein expressionthat has proven to be very tricky isglycosylation. Fungi don’t do it (orother posttranslational processingsuch as phosphorylation,carboxylation, acetylation, acylation,and amidation) quite the same waythat the cells of humans and othermammals do. S. cerevisiae has beenshown in many cases tooverglycosylate; that is, add toomany sugars to the polypeptidechain. Certain other species, such asP. pastoris and S. pombe, have less ofa tendency to do so. A“humanized” yeast (one that canmake fully human glycoproteins) isin development at GlycoFi(Lebanon, NH; www.glycofi.com).Both Saccharomyces and Pichiaspecies tend to hyperglycosylate.Pichia yeasts are known for theirtendency to produce proteaseenzymes as well.

As with bacterial expressionsystems, molecular fusion can solvesome problems. Fusion proteinsresult from engineering the genesfor a therapeutic protein togetherwith another protein or peptide thatmakes it more soluble, more stable,or more biologically available. DeltaBiotechnology Ltd. (Nottingham,UK, www.deltabiotechnology.com),for example, bases its fusion proteintechnology on human serumalbumin made using recombinantyeast. Albumin fusion proteins aremore stable than the active proteinsalone, and they last longer in thebody once given to a patient.

Certain filamentous fungi alsofind use in biotechnology, mostnotably the Aspergillus species niger

and nidulans, which make bothhuman interferons and bovinechymosin. Along with A. oryzae,they have extensive application in themanufacture of industrial enzymes,vitamins, organic acids, antibiotics,and alkaloids. Recombinantexpression levels depend on couplingthe gene of interest to a highlyspecialized fungal promoter gene,and they have typically ranged in thehundreds of milligrams per liter ofsurrounding medium. The biggestproblem so far has been that mostfilamentous fungi naturally secreteproteases in large quantity rightalongside the protein of interest —but genetic engineers are working todevelop strains that do not.

FERMENTATION TECHNOLOGY

Microbiologists define fermentationas “any process for the productionof a product by means of the massculture of a microorganism” (9). AtBioProcess International, we’re mostinterested in the recombinanttechnology that uses it to produceenzymes, amino acids, and humanand animal therapeutics. The actualfermentation process is the same —although details such as scale,temperature, and growth media maybe different — and functions ineither a batch, fed-batch, orcontinuous (perfusion) mode.Despite its limitations, batch modeis still the most common type offermentation found in industry.

A batch fermentation is a closedsystem (9): Microbes are added to asterilized nutrient solution in thefermentor, then allowed to incubate.Nothing more is added exceptoxygen (most microorganisms usedin biotechnology are aerobicspecies), an antifoam agent (toprevent bubble formation), and acidor base to control the solution pH.Consequently, the mixture changesconstantly as a result of cellularmetabolism, with waste productsaccumulating and biomassincreasing over time.

Four growth phases followinoculation: a lag phase(microorganisms physicochemicallyequilibrating with their environment),a log phase (cells have adapted totheir new surroundings and begindoubling their numberlogarithmically), a stationary phase(available food has been used up andgrowth slows or stops; this is whenmetabolites or recombinant proteinsare best harvested), and a death phase(their energy depleted, the cells dieoff). Doubling times during the logphase vary according to the size andcomplexity of the microbe: bacteriadouble in less than an hour, yeasts inone or two hours, filamentous fungiin two to six hours (1).

Fed-batch fermentation is “anenhancement of the closed batchprocess” (9) in which nutrients areadded in increments as growthprogresses. Certain ingredients arepresent at inoculation, and theycontinue to be added in small dosesthroughout production. Continuousfermentation is an open system.Sterile nutrient solution iscontinuously introduced while anequal amount of waste products areremoved. Cell growth may or maynot be adjusted:

In the chemostat in the steadystate, cell growth is controlled byadjusting the concentration of onesubstrate. In the turbidistat, cellgrowth is kept constant by usingturbidity to monitor the biomassconcentration and the rate of feedof nutrient solution isappropriately adjusted. (9)

12 BioProcess International JUNE 2004 SUPPLEMENT

NEW BRUNSWICK SCIENTIFIC CO., INC.(WWW.NBSC.COM)

14 BioProcess International JUNE 2004 SUPPLEMENT

The nutrients involved caninclude carbohydrates, fatty acids,amino acids, and sources of nitrogenand sulfur. Sugars such as glucose,lactose, sucrose, and starch providecarbs and nitrogen. Vitamins,minerals, or growth factors may benecessary for some microbe species.Stirring and mixing adds an airsupply, removes carbon dioxide, anddistributes nutrients, but anantifoam chemical agent is necessaryto keep excess bubbles fromforming. Fermentation is amultiphasic reaction in whichgaseous components (N2, O2, andCO2) must be mixed continuouslywith the liquid medium and solidcells. For optimal yields, the wholeprocess must be carried out at aconstant temperature.

During fermentation, cellulargrowth and metabolism produceheat through the energy released bythe chemical reactions involved (10).Unless the resulting increase inoverall temperature within thefermentor is counteracted, cellscould die or at least fail to produce.Mechanical agitation of the mediumalso adds heat to the system. So acooling system is necessary and mustbe closely monitored and controlled.

Fermentors in the biotechnologyindustry are “complex, aseptic,integrated systems involving varyinglevels of advanced computer inputs”(1). And World Health Organization(WHO) guidelines and FDAregulations describing goodmanufacturing practices (GMPs)require contamination control tomake an end product that is bothsafe and effective. Biotech processesmust be monitored and wellcontrolled (Table 1), whichnecessitates the use of both on-linesensors (measuring pH, CO2, anddissolved O2 levels, for example)that require calibration and specialtreatment and off-line analysis(periodic assays that test for nucleicacids and enzymes, the quality andamount of product present, and soon). On-line sensors are preferred, ofcourse, because of their immediacy.A problem identified quickly can becorrected; some problems identifiedtoo late may be the death of a whole

batch. Many vendor companies areworking to develop more and bettersensors for bioprocess monitoring.

International laws requirecontainment of geneticallyengineered organisms. Recombinantcells, once they have beengenetically engineered andbioanalytically characterized, arevery valuable. A master cell bank iscreated and frozen at –80 °C orfreeze-dried. Frozen storage is themost common method of preservingthese valuable microbes withoutallowing them to divide and grow,but the process of freezing andthawing can kill cells, socryoprotectants such as glycerol maybe used. Freeze-drying(lyophilization) offers a gentler,more stable method of storage.

A working cell bank is createdfrom the master cell bank byreviving the live cells: thawing andthen culturing them on agarmedium. From a frozen bank, itmay take two days to grow enoughnew cells to begin fermentation.From a lyophilized bank, it can takelonger. Refrigerated cultures needonly a day or so.

Bioprocesses typically proceedthrough three stages: laboratoryscale, pilot scale, and productionscale (1). Initially, basic work isperformed using Petri dishes,Erlenmeyer flasks, and other simplelaboratory culture techniques. Next,

pilot-plant studies help processdevelopers determine the optimaloperating conditions at a volume of5–200 L. The final stage transfersthose conditions into 500-L to1000-L (or bigger) fermentors forcommercial manufacturing.

INSULIN PAVED THE WAY

Genentech, Inc. (South SanFrancisco, CA, www.gene.com)developed the original bacterialfermentation process that waslicensed by Eli Lilly and Co. (St.Louis, MO; www.lilly.com) toproduce the first recombinant humaninsulin. Much of Genentech’s currentwork is in mammalian cell culture(see Chapter 2).

Traditionally, insulin was purifiedfrom the pancreases of slaughteredcattle or pigs. As early as the 1920s,diabetics were receiving frequent,painful injections that neverthelesssaved their lives (11). Twocompanies have traditionallycompeted for dominance in theinsulin market: Lilly and Denmark’sNovo Nordisk (www.novonordisk.com). Early side effects were oftencaused by the presence of certainpancreatic peptides (includingsomatostatin, glucagon, andvasoactive intestinal peptide), so bythe 1980s the Danes had come upwith a way to better purify theirinsulin preparation, calling theresults “monocomponent insulin.”

BIOGEN IDEC (WWW.BIOGENIDEC.COM)

16 BioProcess International JUNE 2004 SUPPLEMENT

REGULATIONS AROUND THE WORLD

The World Health Organization(www.who.int) has published a generalguide to Good Manufacturing Practicesfor Pharmaceutical Products (TechnicalReport Series No. 823, Geneva, 1992),which is a good place to start whenlooking for regulatory guidance. Thefollowing documents will be of specificinterest to companies working onrecombinant microbial sourcedproducts.

US FDA DOCUMENTS21 CFR 120: Hazard Analysis andCritical Control Point Systems

21 CFR 210: Current GoodManufacturing Practice (GMP) inManufacturing, Processing, Packing, orHolding of Drugs, General

21 CFR 211: Current GMP for FinishedPharmaceuticals

21 CFR 610: General BiologicalProducts Standards

21 CFR 809: In Vitro DiagnosticProducts for Human Use

Draft Guidance for Industry: DrugSubstance — Chemistry, Manufacturing,and Controls Information (6 January2004)

Draft Guidance for Industry:Comparability Protocols — Protein DrugProducts and Biological Products:Chemistry, Manufacturing, and ControlsInformation (3 September 2003)

Draft Guidance for Industry: SterileDrug Products Produced by AsepticProcessing — Current GoodManufacturing Practice (3 September2003)

Draft Guidance for Industry:Comparability Protocols — Chemistry,Manufacturing, and ControlsInformation (20 February 2003)

Draft Guidance for Industry: DrugProduct — Chemistry, Manufacturing,and Controls Information (28 January2003)

Draft Guidance for Industry: PreventiveMeasures to Reduce the Possible Risk ofTransmission of Creutzfeldt-JakobDisease (CJD) and Variant Creutzfeldt-Jakob Disease (vCJD) by Human Cells,Tissues, and Cellular and Tissue-BasedProducts (HCT/Ps) (14 June 2002)

Draft Guidance for Industry: BiologicalProduct Deviation Reporting forLicensed Manufacturers of BiologicalProducts Other than Blood and BloodComponents (10 August 2001)

Guidance for Industry: Content andFormat of Chemistry, Manufacturing andControls Information and EstablishmentDescription Information for a BiologicalIn Vitro Diagnostic Product (8 March1999)

Guidance for Industry: Content andFormat of Chemistry, Manufacturing andControls Information and EstablishmentDescription Information for a Vaccine orRelated Product (5 January 1999)

FDA Guidance ConcerningDemonstration of Comparability ofHuman Biological Products, IncludingTherapeutic Biotechnology-DerivedProducts (April 1996)

FDA Guidance Document ConcerningUse of Pilot Manufacturing Facilities forthe Development and Manufacturing ofBiological Products

Draft Points to Consider in theCharacterization of Cell Lines Used toProduce Biologicals (12 July 1993)

Supplement to the Points to Consider inthe Production and Testing of NewDrugs and Biologics Produced byRecombinant DNA Technology: NucleicAcid Characterization and GeneticStability (6 April 1992)

Guideline on General Principles ofProcess Validation (May 1987)

Points to Consider in the Productionand Testing of New Drugs andBiologicals Produced by RecombinantDNA Technology (10 April 1985)

EUROPEAN UNION DOCUMENTSRegulation No. 1946/2003 of theEuropean Parliament and of the Councilof 15 July 2003 on TransboundaryMovements of Genetically ModifiedOrganisms

21/04/02 EudraLex Volume 4: MedicinalProducts for Human and Veterinary Use— Good Manufacturing Practices

2000/608/EC: Commission Decision of27 September 2000 Concerning theGuidance Notes for Risk AssessmentOutlined in Annex III of Directive90/219/EEC on the Contained Use ofGenetically Modified Microorganisms

(notified under document numberC(2000) 2736)

Commission Directive 91/356/EEC, of13 June 1991, Laying Down thePrinciples and Guidelines of GoodManufacturing Practice for MedicinalProducts for Human Use.

Council Directive 90/219/EEC of 23April 1990 on the Contained Use ofGenetically Modified Microorganisms

EU/EC Biotechnology Guidelines3AB1A: Production and Quality Controlof Medicinal Products Derived byRecombinant DNA Technology

3AB2A: Quality of BiotechnologicalProducts — Analysis of the ExpressionConstruct in Cells Used for Productionof R-DNA Derived Protein Products

3AB3A: Production and Quality Controlof Cytokine Products Derived byBiotechnological Process

3AB6A: Gene Therapy Product QualityAspects in the Production of Vectorsand Genetically Modified Somatic Cells

3AB10A: Minimizing the Risk ofTransmitting Agents CausingSpongiform Encephalopathy viaMedicinal Products

3AB14A: Harmonization ofRequirements for Influenza Vaccines

ICH DOCUMENTSICH Draft Guidance: Q5EComparability ofBiotechnological/Biological ProductsSubject to Changes in TheirManufacturing Process (29 March2004)

ICH Guidance: Q7A GoodManufacturing Practice Guide forActive Pharmaceutical Ingredients (25September 2001)

ICH Guidance on Quality ofBiotechnological/Biological Products:Derivation and Characterization of CellSubstrates Used for Production ofBiotechnological/Biological Products(21 September 1998)

ICH Final Guideline on Quality ofBiotechnical Products: Analysis of theExpression Construct in Cells Used forthe Production of r-DNA DerivedProtein Products (February 1996)

Lilly’s response was Humulin. NovoNordisk introduced its ownrecombinant version (produced byyeast fermentation) in 1987.Recombinant human insulin is nowproduced industrially on a largescale: 50,000-L fermentors are usedin routine production (2).

Human growth hormone (hGH)was once available only from thepituitary glands of cadavers. It isused to combat dwarfism, butbecause of a lack of availability itdidn’t see a lot of use before themid-1980s, when it was bannedanyway. Several people hadcontracted Creutzfeldt-Jakob disease(a spongiform encephalopathypreviously thought to be onlyinheritable) from contaminatedhGH preparations. Genentech’sProtropin arrived on the market in1985 just in time. It is made byrecombinant E. coli. Because of theresulting increased availability (thereare now eight other brand-nameversions of the same hormone forsale in the United States), clinicalstudies are examining its potentialfor treating burns, peptic ulcers,osteoporosis, and other conditions(2). Bovine growth hormone is alsoproduced in E. coli, and it has founduse in agriculture to increase themilk production of dairy cows.

The only topical biotherapeuticon the market, platelet-derivedgrowth factor (PDGF) produced inrecombinant S. cerevisiae, is made byChiron Corporation (www.chiron.com) for Johnson & Johnson(www.jnj.com). It is formulated as agel for use in wound treatment,specifically diabetic foot ulcers.

Galenus Mannheim, BoehringerMannheim, and Centocor makerecombinant tissue plasminogenactivator (rt-PA) using E. colifermentation (2). The Retiplaseproduct was approved in 1996 foruse as a blood clot destroyer(thrombolytic agent) in patientssuffering heart attacks. The rt-PAforms inclusion bodies in the bacteriacells, which necessitates denaturationand refolding, but thisunglycosylated form of the protein isjust as thrombolytic as the

glycosylated form. Thus, the bacterialprocess is still preferable to a muchmore expensive mammalian cellculture in this case (see Chapter 2).

As a glycoprotein, ß-interferon canbe made correctly only by animalcells. But Betaferon and Betasteron— European and US names for thesame product from Berlex Inc.(Montville, NJ; www.berlex.com) andSchering AG (Berlin, Germany;www.schering.de) — are multiplesclerosis treatments produced in E. coli. This unglycosylated versionhas proved to be medically effective.Actimmune, a �-interferon productfrom InterMune (Brisbane, CA;www.intermune.com) used in treatingchronic granulomatous disease (agenetic immune system disorder), isalso made by recombinant E. coli. Soare Chiron’s Proleukin, arecombinant interleukin-2 (IL-2)used to treat metastatic renal cellcarcinoma, and Neumega fromWyeth (Madison, NJ; www.wyeth.com), an interleukin used to treatchemotherapy-inducedthrombocytopenia. IL-1 made byrecombinant bacteria is used to fighta variety of cancers, as is a form of itsreceptor sourced the same way.Glycosylation — or lack thereof —does not seem to make a difference inthe efficacy of these cytokineproducts. Tumor-necrosis andcolony-stimulating factors producedin E. coli are currently in clinical trials.

The Future Is Now: As shownabove, wherever glycosylation is notnecessary, bacterial (or yeast)fermentation will be the productionmethod of choice. Several companiesare working on ways to improve theprotein refolding steps necessitatedby inclusion body formation.“Successful recovery of insulin,bovine growth hormone, and tissueplasminogen activator from inclusionbodies demonstrates the commercialviability of this technology” (12).

Pfizer Inc. (New York, NY;www.pfizer.com) is focusing onimproving assays for detecting themany undesirable protein andnucleic acid contaminants that resultfrom processing bacterial cells. “Thecontaminants may be present atlevels that span a millionfold range;

they are numerous andheterogeneous; tests must have ahigh level of sensitivity; and thereare numerous materials that mayblock quantitation,” explains oneresearch scientist (12). Regulationsrequire all contaminants to beremoved (or reduced to extremelysmall amounts) and/or inactivated,which can facilitate downstreamprocessing as well.

Contract manufacturerDowpharma (Midland, MI;www.dowpharma.com) hasintroduced a new bacterial expressionsystem using P. fluorescens. Because ofmetabolic differences and a specificpromoter, it is characterized by highexpression levels and simplifieddownstream processing whencompared with E. coli and yeastsystems, which the company alsouses. The modified P. fluorescensbacteria produce normally solublerecombinant proteins without certainmetabolic byproducts that cannegatively affect cell growth in E. coli.

Neose Technologies Inc.(Horsham, PA; www.neose.com) isaddressing the issue of glycosylationwith enzymatic modification ofbacterial-sourced recombinantproteins. A PEGylatederythropoietin product currentlyunder study appears to be its firstsuccess. “The company plans toprocess reagents on a commercialscale for glycoprotein drugs in afuture GMP manufacturing facility” (12).

Meanwhile, Delta Biotechnologyhas become a recognized expert inthe use of S. cerevisiae. Thiscompany has solved problemsincluding expression instability (13)and posttranslational modifications,such as proteolytic degradation,while demonstrating high secretionlevels (13–15). Delta sells the firstbiological excipient produced byrecombinant technology:Recombumin (recombinant humanalbumin), which is used as astabilizer and drug delivery vehiclein drug formulations. It is madeusing a commercially available yeastexpression technology that offersrelatively high expression levels,

18 BioProcess International JUNE 2004 SUPPLEMENT

SUPPLEMENT

stable genetics, and a high-densityfermentation based on growthmedia free of animal- or human-derived components.

New Brunswick Scientific Co.,Inc. (Edison, NJ; www.nbsc.com), amanufacturer of fermentors andbioreactors, works with itscustomers, including many academicresearchers, to optimize processesusing its equipment. For example, acollaboration with University ofCalifornia scientists led to a hugeincrease in production of arecombinant thrombomodulinfragment by P. pastoris: from 1.4 mg/L cultured in shake flasks to200 mg/L in a fed-batch BioFlo3000 fermentor — and up to agram per liter in a continuousculture process (16).

PRODUCTION IS

ONLY THE BEGINNING

The goal of biotechnology is for thecells or organisms to do the hardwork of making product, which isthen processed and formulated byhuman efforts. The right expressionsystem will synthesize proteins inmuch larger amounts than would beavailable any other way. With most ofthe products mentioned in this andthe following chapters, there is noother economical (or even feasible)choice. Chemical synthesis is themost economical method of makingsmall peptides such as oxytocin andvasopressin (both nonamers, peptidesmade of nine amino acids), mainlybecause they are almost alwaysexpressed by cells in extremely smallamounts. It cannot be used to makelarge proteins — even those thatdon’t require glycosylation. And theproblems with indigenous sources arenumerous — beginning with the lackof control over their compositionand availability. For many products,genetically engineered expressionsystems offer the only possiblemethod of production. With theproper knowledge in hand, it ispossible to go from introducing aforeign gene into bacteria to proteinproduction in five days (about twoweeks for yeasts). But a lot of workmust go into developing a

fermentation process, and only someof it has been touched on in thischapter.

The actual act of geneticengineering may be relatively simpleif you’ve got the right equipmentand reagents — I’ve done it myselfin a laboratory classroom setting —but doing it on the industrial levelrequires a deep understanding ofgenetics and molecular biology,methodical planning andprioritization of needs, and detaileddocumentation. Any home-brewercan set up a fermentation batch andmake some beer or wine (of widelyvarying quality). But industrialbioprocesses require cell-linecharacterization, scale-up and scale-down studies for processoptimization, close monitoring andcontrol, and regulatory compliance.The better your choice of expressionsystem, the easier all of those will be.

Bacterial and fungal fermentationmay offer the cheapest option, andsome criteria will make them theobvious answer. But as with chemicalsynthesis (often used to make peptideproducts as well as small-moleculedrugs), there are some things theysimply cannot do. Bacteria (17) andyeast (18) have been shown tosuccessfully express active Fabfragments and single-chainantibodies. But MAbs and other verylarge, complex proteins must bemade by animal cells — andsometimes human cells are the onlychoice. Our next chapter examinesbiotechnological cell culture in detail.

ACKNOWLEDGMENTSCompany-specific information and quotesnot attributed in this article most often camefrom the companies themselves by way oftheir web sites (Thank you, webmasters!) or,in some cases, from the FDA online.

REFERENCES1 Smith, JE. Biotechnology (Third

Edition). Studies in Biology Series. CambridgeUniversity Press: Cambridge, UK, 1996.

2 Walsh, G. Proteins: Biochemistry andBiotechnology. John Wiley and Sons, Inc.:New York, NY, 2002.

3 Baneyx, F. Recombinant ProteinExpression in Escherichia Coli. CurrentOpinion in Biotechnology 1999, 10: 411–421.

4 Ruggiero, F; Chanut, H; Fichard, A.Production of Recombinant Collagen forBiomedical Devices. BioPharm 2000, 13(5):32–37.

5 Messner, P; et al. Glycobiology ofSurface Layer Proteins. Biochimie 2004, 83:591–599.

6 Sleep, D; et al. SaccharomycesCerevisiae Strains that OverexpressHeterologous Proteins. Bio/Technology 1991,9: 183–187.

7 Cousens, DJ; et al. Construction of aRegulated PGK Expression Vector. NucleicAcids Res. 1990, 18(5): 1308.

8 Cino, J. High-Yield Protein Productionfrom Pichia Pastoris Yeast: A Protocol forBenchtop Fermentation. New BrunswickScientific Co., Inc.: Edison, NJ;www.nbsc.com/files/papers/abl_pichia.pdf.

9 Pumphrey, B; Julien, C. FermentationBasics. In An Introduction to Fermentation.New Brunswick Scientific Co. Inc.: Edison,NJ, May 1996; www.nbsc.com/files/papers/fermbas.pdf.

10 Chatrathi, K. Metabolic CoolingCapacity of Fermentors. BioProcessInternational 2(4): 44–47.

11 What’s Hot: Insulin — Its History andFuture. Diabetic-Lifestyle.com: November1998; www.diabetic-lifestyle.com/articles/nov98_whats_1.htm.

12 Morrow, KJ. Cell Lines forRecombinant Protein Production:Optimizing Processes to Reduce Cost andImprove Quality. Genetic Eng. News 2003,23(17), 1 October 2003: 50, 55.

13 Chinery, SA; Hinchliffe, E. A NovelClass of Vector for Yeast Transformation.Curr. Genetics 1989, 16: 21–25.

14 Kerry-Williams, SM; et al. Disruptionof the Saccharomyces Cerevisiae YAP3 GeneReduces the Proteolytic Degradation ofSecreted Recombinant Human Albumin.Yeast 1998, 14: 161–169.

15 Sleep, D; Belfield, GP; and Goodey,AR. The Secretion of Human SerumAlbumin from the Yeast SaccharomycesCerevisiae Using Five Different LeaderSequences. Bio/Technology 1990, 8: 42–46.

16 140-Fold Increase of Protein Achievedin Pichia Fermentation. New BrunswickScientific Co., Inc.: Edison, NJ;www.nbsc.com/files/papers/pichia.pdf.

17 Bowering, LC. Microbial Systems forthe Manufacture of Therapeutic AntibodyFragments. BioProcess International 2004,2(6), in press.

18 Mead, D. Delta Biotechnologies, Ltd.,Nottingham, UK; personal communication, 5 May 2004. ��

20 BioProcess International JUNE 2004