Cost
analysis
based
on
bioreactor
cultivation
conditions:
Production
of
a
soluble
recombinant
protein
using
Escherichia
coli
BL21(DE3)
Valdemir
M.
Cardoso
a,
Gilson
Campani
a,b,
Maurício
P.
Santos
a,
Gabriel
G.
Silva
a,
Manuella
C.
Pires
c,
Viviane
M.
Gonçalves
c,
Roberto
de
C.
Giordano
a,
Cíntia
R.
Sargo
a,d,
Antônio
C.
L.
Horta
a,
Teresa
C.
Zangirolami
a,*
a
GraduateProgramofChemicalEngineering(PPGEQ),FederalUniversityofSãoCarlos(UFSCar),RodoviaWashingtonLuís,km235,13565-905,SãoCarlos, SP,Brazil
bDepartmentofEngineering,FederalUniversityofLavras,37200-000,Lavras,MG,Brazil
cLaboratoryofVaccineDevelopment,ButantanInstitute,Av.VitalBrasil1500,05508-900,SãoPaulo,SP,Brazil d
BrazilianBiorenewablesNationalLaboratory(LNBR),BrazilianCenterforResearchinEnergyandMaterials(CNPEM),13083-970,Campinas,SP,Brazil
ARTICLE INFO Articlehistory:
Received31October2019
Receivedinrevisedform6February2020 Accepted21February2020
Keywords: Batchculture Energyconsumption Processeconomicanalysis Stirred-Tankreactor Bioprocessengineering
ABSTRACT
Theimpactofcultivationstrategyonthecostofrecombinantproteinproductioniscrucialfordefining
cost-effective bioreactoroperationconditions. Thispaperpresents amethodology toestimateand
comparecostimpactsrelatedtoutilitiesaswellasmediumcomposition,usingsimpledesignequations
andaccessibledata.Datafrombatchbioreactorcultureswereusedascasestudyinvolvingtheproduction
of pneumococcal surface protein A, a soluble recombinant protein, employing E. coli BL21(DE3).
Cultivationstrategiesandcorrespondingprocesscostscoveredawiderangeofoperationalconditions,
includingdifferentmedia,inducers,andtemperatures.Thecoreexpenseswererelatedtothemedium
andcooling.WhenthepriceofpeptonewasabovethethresholdvalueofUS$30/kg,definedmedium
becamethebestchoice.IPTGandtemperaturesaround32CledtoshorterculturesandlowerPspA4Pro
productioncosts.Theprocedureoffersasimple,accessibletheoreticaltooltoidentifycost-effective
productionstrategiesusingbioreactors.
©2020TheAuthors.PublishedbyElsevierB.V.ThisisanopenaccessarticleundertheCCBYlicense
(http://creativecommons.org/licenses/by/4.0/).
1.Introduction
Therapeutic recombinant proteins are used for treating a varietyofdiseases [1,2]. Recombinantproteinsarealsousedas antigensofsubunitvaccines,amongwhichrecombinantantigens ofNeisseriameningitidisandcholeratoxinBsubunitareproduced inE.coli[3].Thereisalsoanenormousnumberofsubunitvaccine candidates for different pathogens, such as dengue virus [4], Leptospira [5], Leishmania [6], Streptococcus pneumoniae [7], Mycobacteriumtuberculosis[8],andHelicobacterpylori[9],among others,severalofthemproducedasrecombinantproteinsinE.coli. Almostathirdoftherapeuticrecombinantproductshavebeen producedusingE.coli[2,10,11],demonstratingtheimportanceof thismicroorganismasanexpressionhost.E.coli notonlyhasa well-characterizedgenome,butisalsoeasilygenetically manipu-lated[1],which hasenabledtheexpressionofseveral different
recombinant proteins [12,13] and metabolites [14,15]. Further-more,E.colipresentsrapidgrowthandhighbiomassformation yield,aswellaslowproductioncosts[16,17],favoringitsindustrial scaleuse,comparedtootherpossiblehosts[18].
TheproductionofrecombinantproteinsusingE.colihasbeen widelyreported.Ahugenumberofstudieshavefocusedonlyon cloning and heterologousproteinproduction atthe shakeflask scale[19–21].Atthebioreactorscale,mostworkshavefocusedon cultivation strategies to achieve high cell and protein concen-trations[22,23].Therehavealsobeenafewstudiesaddressingthe impactofdifferentcultivation techniquesonprocesseconomics [24–31].Definedorcomplexmediacanbeusedduringcultivations [23].Ononehand,adefinedmediumhasbeensuggestedasthe best choicefor production of therapeuticrecombinant proteins [32],since itsknowncomposition enablestheconcentrationof individual components to be followed, resulting in better bioreactorcontrol.Ontheotherhand,complexmediacanboost product formation [22] and provide higher growth rates [33], comparedtodefinedmedia.Forthesereasons,complexmediaare widely used in industry [18], including for the production of therapeutics.However,inthiscase,complexmediumformulations requireprion-freecertifiedanimal-basednitrogensources(suchas
* Corresponding author at: Department of Chemical Engineering, Federal UniversityofSãoCarlos(UFSCar),RodoviaWashingtonLuís,km235,CEP 13565-905,SãoCarlos,SP,Brazil.
E-mailaddress:teresacz@ufscar.br(T.C. Zangirolami).
https://doi.org/10.1016/j.btre.2020.e00441
2215-017X/©2020TheAuthors.PublishedbyElsevierB.V.ThisisanopenaccessarticleundertheCCBYlicense(http://creativecommons.org/licenses/by/4.0/).
ContentslistsavailableatScienceDirect
Biotechnology
Reports
tryptones),orcertifiednitrogensourcesderivedfromplantsand microbes(suchasyeastextractandsoybeanpeptones).Hence,the choicebetweencomplexanddefinedmediaiscontroversialandits economic impact goes beyond the corresponding costs of the mediumcomponents.Infact,togetherwithtemperature,culture medium and inducerwill influencethe duration ofcultivation, proteinsynthesis,andbiomassformation.So,sheddinglightonthe impactofmedium formulationonoverallbioreactoreconomics would help in decidingwhich medium touse for recombinant proteinproductioninE.coli.
Sincetheaimofindustrialfermentationistoobtainhighyields usinginexpensiverawmaterials,withlowcapitalcosts[34],cost evaluation may provide a convenient guide for selection of a cultivationstrategy[32].Differentapproachescanbeusedforthis purpose. It has been reported that the replacement of batch processes by continuous ones may be a natural option for decreasingcostsandraisingproductivity[30].Otherstudieshave focusedonthesimulationof newconfigurationsof biopharma-ceuticalplantsandestimatesofpayofftimes[25,31].Comparison hasbeenmadeofalternativemediumformulationsforgrowthofE. coli as a hostorganism[26,28]. Although Fong and Wood [28] consideredtheuseofisopropyl-β-D-thiogalactopyranoside(IPTG)
andcheapernitrogenandcarbonsources,thecostimpactsofthe inducerand alternativenitrogenand carbon sources wereonly evaluatedqualitatively.Ferreiraetal.[27]estimatedthe quantita-tiveeconomicimpactsofthesematerialsbysimulation,although their influences on process dynamics were not demonstrated experimentally.ThestudiesundertakenbyCampanietal.[24]and Knollet al.[29] focusedonbioreactorutilities,withtheoretical analysesofcostsrelatedtomixing,compression,andthesupplyof pureoxygen. Until now,for recombinantproteinproduction in bioreactors, thecombined cost implications of differentmedia, inducers,nitrogensources,andcultivationtemperatures,together withtheassociatedutilitiesexpenses,havenotbeenconsideredor publishedinasinglepaper,inordertofacilitatetheincorporation ofsuchamethodologyintheselectionofbioreactorcultivation conditionsatlaboratoryorindustrialscales.
Althoughdifferenttypesofrecombinantproteinsmaypresent particularprocesssetsthatenableachievementofbetter produc-tion results [35,36], efforts are needed to understand how bioreactor operational conditions impact the overall economic performance of heterologousproteinproduction processes. The untaggedfragmentofpneumococcalsurfaceproteinA(PspA4Pro) isapotentialcandidateforaserotype-independentvaccineagainst Streptococcuspneumoniae[37]anditsproductionhasbeenstudied usingrecombinantE.colicultivationscarriedoutunderdifferent strategies[24,38,39].PspA isa singlechainα-helix-richprotein thatbindstohumanlactoferrin[40]andtotheC3moleculeofthe serumcomplement[41],contributingtoS.pneumoniaeevadingthe host immune system. Recombinant fragments of PspA, which containthe N-terminalα-helix domain,theproline-richregion, andnodisulfidebonds,arehighlysoluble[42–44]andpresenta Nomenclature
ANS Animalnitrogensource
CkComp Concentrationofcomponentkinthemedium
CD Circulardichroism CM Complexmedium
Cx Biomassconcentration,gDCW/L
Di Impellerdiameter,0.078m
DCW Drycellweight DM Definedmedium
DR Reactordiameter,0.160m
DOT Dissolvedoxygentension
EC Totalenergyconsumptionofcompressor,kWh
ES Totalenergyconsumptionofstirring,kWh
EM Totalmetabolicenergygeneration,kWh
Exp# Experimentnumber fC Correctionfactor
g Gravitationalacceleration,m/s2 Hi Impellerheight,0.267m
HR Reactorheight,m
IPTG Isopropyl-β-D-thiogalactopyranoside K Constantofproportionality,50kJ/mmolO2
LAC Lactose MC Mediumcost
N Stirringspeed,rpmors1
Np Powernumber,5.2(Rushtonimpellerinturbulent
regime) Pn
AIR Totalairamountduringtheexperiment,mol
nCmax Maximumairamountintocompressor,mol
nCmin Minimumairamountintocompressor,mol
OD Opticaldensity
OUR Oxygenuptakerate,mmolO2/h
p Pressureofenteringgas,atm P Powerofungassedstirring,kW
p1 Inletabsolutepressureofcompressor,Pa
p2 Outletabsolutepressureofcompressor,Pa
PB Powerconsumptionofthermostaticbath,0.710kW
PC Powerconsumptionofcompressor,kW
PP PspA4Proproduction,g/L
PS Powerconsumptionofstirring,kW
PspA4Pro Untaggedpneumococcalsurface ProteinA from clade4
Q1 Totalenteringvolumetricgasflow,L/min
QAIR Airvolumetricflowrate
QG Totalgasvolumetricflowrate,m3/s
QHeat Metabolicheatreleaserate,kJ/h
QO2 Oxygenvolumetricflowrate
R Gasconstant,0.082atm.L/mol.K
$i Costof i,where iis gases,cooling, stirring, or
medium,US$/(gDCWorgPspA)
$kComp Costofcomponentkinthemedium
STR Stirredtankreactor T Culturetemperature
T1 Temperatureofenteringgasstreams,C
tf Cultivationtime
VNS Vegetalnitrogensource
VO2 Totaloxygenconsumptionvolume,m3
VR Bioreactorculturevolume,L
Greekletters
r
b Brothdensity,kg/m3h
B Efficiencynumberofthermostaticbath,0.7h
C Efficiencynumberofcompressor,0.70h
S Energyefficiency,0.65g
Isoentropicexponent,1.4foroxygenSuperscripts
max Relatedtoamaximumvalue min Relatedtoaminimumvalue
‘^’ IndicatorthatgeometriesarefromtheSTRusedinthe experiments
Subscript
coiled-coilstructure[45,46].Thisstructureconfersstabilityand thermoresistantcharacteristicstothemolecule[37].
Thus, this work presents a procedure for implementing an overallcostanalysis,consideringtheintegratedeffectsofdifferent media,complexnitrogensources,temperatures,andinducerson theassociated costs (for cooling,stirring, air compression, and supplyofpureoxygen),basedonbioreactorcultivationconditions andtheoreticaldesign equations.Asacase study,theproposed calculationprocedurewasappliedtocarryoutanoveralleconomic analysisofprocessperformanceforbiomassandPspA4Proprotein production during batch stirred-tank bioreactor cultures. The procedurecouldbeeasilyextendedtoeconomicassessmentsfor otherbioproductsandbioreactorscales.
2.Materialsandmethods 2.1.Experimentaldata
All experiments were carried out with E. coli BL21(DE3) (Invitrogen, Carlsbad, CA, USA) harboring the plasmid pET37b (+)/pspA4Pro, which carries the gene encoding an untagged fragment of the pneumococcal surface protein A (PspA) gene fromfamily2,clade4,namedPspA4Pro[44].Theexpressionofthe PspA4ProgeneiscontrolledbythelacUV5andT7lacpromoters, withthesolublerecombinantproteinbeingaccumulated intracel-lularlyaftertheinducer(IPTG or lactose)is added [37,38]. The recombinantE.colistrainwaskindlyprovidedbyDr.ElianeMiyaji from the Laboratory of Molecular Biology, Butantan Institute (SãoPaulo,Brazil).
Although there are other factors that can play a role in recombinantproteinproductionusingE.colibioreactorcultures, such as duration of induction, timing of induction, inducer concentration, and dissolved oxygen tension, among others [36,47],inthiswork,temperature,typeofinducer,andmedium composition were chosen to evaluate the application of the proposedmethodologytotheeconomicsofPspA4Proproduction andtocomparecultivationstrategies.Experimentaldataforthe costanalysiswereobtainedfromninebioreactorculturescarried out under different cultivation and induction conditions, using complexanddefinedmedia,differentcomplexnitrogensources, lactoseandIPTG(indefinedmedium)asinducers,andcultivation temperaturesfrom27to37C(Table1).
Thefirststageofthestudyusedafactorialdesignapproachto evaluatethecosteffectsoftemperatureandinducersinPspA4Pro production using batch bioreactor cultivations employing
moderate biomass concentrations. Experiments #1 to #5were batchesperformedwithHDFmedium[48],at27,32,and37C, usingglycerolassolecarbonsourceandisopropyl-β-D
-thiogalac-topyranoside or lactose as inducer. After identifying the most promising temperature, additional experiments using complex mediumatthistemperaturewerecarriedoutinordertoenable comparisonoftheeffectsofdefinedorcomplexcultivationmedia onthecostofPspA4Proproductioninthebioreactor.Inthesecond stage,investigationwasmadeoftheperformanceachievedusing differentcomplexnitrogensourcesobtainedfromthehydrolysisof animalandvegetalproteins.Experiments#6to#8comprisedaset ofbioreactorbatchescarriedoutwithmodifiedZYM-5052 auto-inductioncomplexmedium[49].Tryptone(enzymatichydrolysate ofcasein)wastheanimalnitrogensourceusedinExperiment#6, whilesoypeptone(enzymaticpapaindigestofsoybeanprotein) andahomemadesoybeanproteinsupplementextractwereusedin Experiments #7(a andb, asduplicates) and#8,as thevegetal nitrogensources.Thelatterexperimentwasperformedusingthe extract of an inexpensive soybean protein source, in order to determine the influence of feedstock quality on recombinant protein production and its related costs. Lactose was used as inducerinallexperimentscarriedoutwithcomplexmediaunder theauto-inductionstrategy.
Detailsabouttheprocedureandtheresultsofeachexperiment can be found in the references provided in Table 1, or in the Supplementary Material. The compositions of all the media employedintheexperimentsarepresentedinTable2.
2.1.1.Bioreactorcultivation
The following experimental procedure was employed in all experiments.Apre-inoculumwaspreparedfromasinglebacterial colony transferred from an LB-Miller-agar-kanamycin plate to 50mLofthedesiredliquidmediumpluskanamycin(50mg/L),ina 0.5LErlenmeyerflask.Theflaskwaskeptovernightat37C,with agitationat250rpm,inacontrolledtemperatureincubator-shaker (New Brunswick),untilreaching optical density(OD) of2.0 (at 600nm).Underthesameconditions,aninoculumwasprepared fromavolumeofthepre-inoculumnecessarytoreachODof0.1in 300mL of the fresh medium under investigation, distributed equally in three 0.5L Erlenmeyer flasks. When OD of 2.0 was reached,theinoculumwastransferredtothebioreactorcontaining approximately4Lofthemediumunderinvestigation.A5L stirred-tankreactor(STR),describedindetailelsewhere[50],wasusedfor allthecultureslistedinTable1.Thebioreactorwascontrolledand monitored usingSuperSys_HCDC softwareand theon-line data wereloggedusingananalog-to-digitalconverter(ModelcFP2020, NationalInstruments).ThepHwasautomaticallycontrolledat6.7 (growthphase)or6.9(inductionphase),usingsolutionsofNH4OH
(15%v/v)andHCl(9%v/v).Dissolvedoxygentension(DOT)was measuredwithanamperometricprobe(ModelInPro6830,Mettler Toledo)andwasmaintainedat30%ofsaturationbyadjustingthe stirrer speed (from 200 to 900rpm) and the two mass flow controllers(GFC,Aalborg)governingtheairandoxygenflowrates. Airandoxygenweresuppliedtothebioreactorfromacompressor (Model MAW-60/425, GMEG) and a high-pressuregas cylinder (WhiteMartins),respectively,reachingamaximumtotalgasflow rate of 4–6 std L/min (at 21.1C and 1atm). The culture temperaturewasmeasuredwithathermocouple(Pt-100,Exacta) andmaintainedatthedesiredset-pointbycontinualindirectheat transfer using water from a thermostatic bath (0.7kW, Solab) passedthroughthebioreactorjacket.
2.1.2.Analysesofculturesamples
ThebioreactorculturebiomassandPspA4Proconcentrations usedinthecostestimationarealsoshowninTable1.Thebiomass concentration (CX) was monitored by optical density (OD)
Table1
Main E. coli BL21(DE3) pET37b(+)/pspA4Pro cultivation conditions for batch bioreactor Experiments #1 to #8, cultivation time (tf) to reach the highest
PspA4Proproduction(PP),andthecorrespondingbiomassconcentration(CX).Data
forExperiments#2,#5,#7(IandII),and#8are availableasSupplementary Material.
Exp# Medium/T/Inducer tf CX PP Reference
h gDCW/L gPspA4Pro/L
1 DM/27C/IPTG 25.5 32.81.8 3.10.5 [38] 2 DM/27C/LAC 27.5 34.50.6 2.60.3 Thiswork
3 DM/32C/IPTG 16.3 31.31.6 3.80.4 [38]
4 DM/37C/IPTG 14.6 27.01.0 3.40.3 [38]
5 DM/37C/LAC 19.5 20.40.3 1.30.0 Thiswork
6 CM-ANS/31C/LAC 16.4 38.00.7 5.40.3 Thiswork
7.I CM-VNS/31C/LAC 16.6 38.30.7 5.60.3 [37] 7.II CM-VNS/31C/LACa
16.0 39.20.7 6.00.3 Thiswork 8 CM-VNS/31C/LACb 15.0 31.82.3 4.70.7 Thiswork
CM-ANS:Complexmediumwithanimalnitrogensource;DCW:Drycellweight; DM: Defined medium; Exp #: Experiment number; LAC: lactose; PspA4Pro: PneumococcalsurfaceProteinAfromclade4;T:Culturetemperature;CM-VNS: Complexmediumwithvegetalnitrogensource.a:Replicationofexperiment#7.I;b: Alternativevegetalnitrogensource.
measurements at 600nm and by the dry cell weight method [51,52].Plasmidlossduringthecultivationswasevaluatedusing diluted samples (1:106 or 1:107) of culture broth spread onto
LB-Miller-agarplates[24].
PspA4Proproduction(PP)wasassessedbybanddensitometry
aftercelldisruption(bysonication)andclarification[24].Thetotal solubleproteinconcentrationfortheknowncellconcentrationof thedisrupted samplewas determined bythe Bradfordmethod [53].Theproteinspresentintheclarifiedsamplewereidentifiedby 12%SDS-PAGE[54].Thesolublerecombinantproteinfractionin the clarified sample was estimated using ImageJ software to analyzeimages of thegel stainedwith CoomassieBlue R[55], togetherwithdeterminationofthetotalsolubleprotein concen-trationasdescribedbyCampanietal.[24].Inordertoconfirmthe absenceofinsolublePspA4Proasinclusionbodies,pelletsobtained aftersonicationwereresuspendedin2mLofbuffer(20mMTRIS; 250mM NaCl; pH 8.0), mixed with Laemmli buffer(1:2) [54], boiledat100Cfor15min,andappliedto12%SDS-PAGEplates.
2.1.3.PspA4Prosecondarystructureandlactoferrinbindingassay Additionalmethodstocharacterizethequalityoftheproteins obtained by different cultivation strategies were applied to samplesfromExperiments #2,#3,#4,#7,and #8.Tothis end, cellpelletsharvestedattheendofbioreactorcultures#2,#3,#4, #7, and #8 were frozen (at 80C), followed by processing
accordingtothepurificationproceduredescribedpreviously[37]. ThepurerecombinantPspA4Proobtainedfromthesolublefraction of each biomass was analyzed by circular dichroism (CD) to confirm the PspA4Pro secondary structure, according to the methoddescribedpreviously[37].
ThePspA4Probiologicalactivitywasevaluatedbyalactoferrin bindingassayusing96-wellflat-bottomplates(MaxiSorp,Nunc). The plates were coated with 2
m
g/well of human lactoferrin (L1294,Sigma),incubatedovernightat4C,washed3timeswithphosphatebufferedsaline+0.05%Tween20PBS-T,andblocked
with5%skimmedmilk.TheplateswerethenwashedwithPBS-T andincubatedat37Cfor2h,withserialdilutions,in5%skimmed milk, of PspA4Pro purified from each experiment. Blank wells receivedonly5%skimmedmilk,whilethepositivecontrolwasa PspA4Pro standard. After washing with PBS-T, anti-PspA4Pro rabbitserum1:5000in5%skimmedmilkwasaddedandtheplates wereincubatedat37Cfor1h,washedwithPBS-T,andincubated with anti-rabbit IgG conjugated to peroxidase A0545, Sigma, 1:5000in5%skimmedmilk.Thecolorwasdevelopedfor15min,at roomtemperature,in the presenceof o-phenylenediamineand hydrogenperoxide,andthereactionwasstoppedwith4MH2SO4.
The absorbance measured at 492nm was plotted against the PspA4Proconcentration,withtheslopebeingusedasaparameter for the lactoferrin binding property of PspA4Pro. The anti-PspA4Pro antibody was obtained in rabbit immunized with 3 doses, at two week intervals, of 200
m
gof previously purified PspA4Pro [37] absorbed in 5mg of aluminum hydroxide. The immunizationandbleedingprotocolsfollowedtherulesissuedby the National Council for Control of Animal Experimentation CONCEAandwereapprovedbytheEthicsCommitteeonAnimal UseoftheButantanInstituteCEUAIBno7,755,300,718.2.2.Directcostestimation
In order to assess the influence of different cultivation conditionsontheprocesseconomics,thedirectcostsrelatedto cellandrecombinantproteinproductionwereevaluatedbytaking intoaccountthecostsofrawmaterials(pureoxygen,inducers,and mediumcomponents)andutilities(basedonenergyconsumption for cooling, stirring, and air compression). Effective theoretical methodologiestoestimatethesecostswereobtainedfromseveral literaturesourcesandwerecompiledinaspreadsheettoenable simple and accessible implementation of the evaluations. The proposedprocedurecouldbeappliedtoestimatecostsnotonlyfor recombinantproteins,butalsoforanybioproduct,providedthata
Table2
Concentrationsandcostsofeachmediumcomponentusedinthebioreactorcultures.
Component Purchase
Cost(US$/kg)
Concentrationsforeachexperiment
Units 1,3,and4 2and5 6,7,and8
Glucosea 0.7 g/L 10.0 Glycerola 1.0 g/L 60.0 60.0 60.0 Yeastextracta 1.7 g/L 5.0 MgSO4.7H2Oa 0.3 g/L 0.4 0.4 0.5 Na2HPO4a 2.0 g/L 9.0 KH2PO4b 1.2 g/L 17.7 17.7 3.4 NH4Cla 0.2 g/L 2.7 (NH4)2HPO4a 0.4 g/L 5.3 5.3 Citricacida 1.2 g/L 2.3 2.3 Na2SO4a 0.1 g/L 0.7 Kanamycinb 1.0 mg/L 100.0 100.0 100.0 Ferriccitrateb 6.0 mg/L 133.3 133.3 100.8 CoCl2.6H2Ob 12.7 mg/L 3.3 3.3 2.5 MnCl2.4H2Ob 1.7 mg/L 20.0 20.0 15.0 CuCl2.2H2Oa 4.2 mg/L 2.0 2.0 1.5 H3BO3a 0.6 mg/L 4.0 4.0 3.0 NaMoO4.2H2Oa 9.9 mg/L 2.8 2.8 2.1 Zn(CH3CHOOH).H2Ob 1.7 mg/L 33.8 33.8 33.8 EDTAa 1.9 mg/L 18.8 18.8 14.1 Thiaminea 36.1 mg/L 45.0 45.0 45.0 Polypropyleneglycola 3.6 g/L 0.3 0.3 0.3 Peptonea 6.1 g/L 10.0 Soysupplementc 12.2 g/L 10.0 Lactosea 2.0 g/L 20.0 20.0 IPTGb 601 mmol/L 1.0
Totalmediumexpenses US$/L 0.24 0.15 0.21d/0.27e
cultivationmonitoringsystemisavailablefordataacquisition,as describedinthebioreactorcultivationsection.Furthermore,this costanalysismethodologycouldbeappliedtocomparebioreactor operationalconditions,regardlessofthebioreactorscale.
Aflowchartsummarizingthemain stepsofthecostanalysis implementationisprovidedintheSupplementaryMaterial(Figure S1).Inaddition,asimpleversionofthespreadsheetisavailableasa workstationforfastcostevaluationsofbioreactorculturesandfor teachingpurposes.Thecomponentvaluesofthisworkstationmay beupdatedbyusers,inordertomaketheresultsofeachanalysis more realistic, depending on the criteria adopted. Despite contributing to the total cost of a fermentation process, costs relatedto sterilization,equipment depreciation,and employees werenotconsideredinthiswork,becausetheyremainedalmost thesame,regardlessofthecultivationconditionsevaluatedinthe experiments.
Asdescribednext,thedirectcostwasestimatedusingEq.s1to 11anddatafrombench-scalebioreactorcultures(Section2.1.1). Foreachproductionstrategy,thedirectcostwasevaluatedatthe cultivation time corresponding to the maximum recombinant proteinconcentration(Table1).Directcostvalueswerepresented asratios relativetoExperiment#3,which correspondedtothe lowestdirectcostforPspA4Proproductionusingdefinedmedium, inthepresentexperiments.
2.2.1.Mediumcosts
ThemediumcostswereestimatedusingEq.1,consideringthe initial amounts of the components added to formulate the medium, together withtheirmarketprices (Table 2).Although all the experiments were performed with laboratory-grade productspurchasedinsmallquantities,thecostsofmostofthe mediumcomponentswereestimatedusinginformationobtained from business intelligence and e-commerce platforms [56,57], whichprovidedabetterrepresentationofmarketprices.Together withtheconceptof“directcostratio”,thestandardizationofthe price survey was important to ensure the data consistency requiredtocomparethecultivation conditionsinvestigated. An exceptionwasthesoyproteinsupplementusedinExperiment#8, forwhichthesalepricewasuseddirectly.Thepricesofcaseinand soybean-basedpeptone,whichareindicatedinTable2as animal-based(ANS)andvegetal(VNS)nitrogensources,respectively,were also retrieved from the platforms, where they were identified genericallyas“peptone”.Themarketpriceofpeptonevariedfrom 6to300US$/kg,dependingonitsnutritionalvalueasamicrobial culturesupplement,solubility,andqualitycertification,aswellas theamountpurchasedandthesupplier.OxoidLP0042Tryptone was used in Experiment #6, BD BBL Phytone was used in Experiments#7.aand#7.b,andGrowthSupplementssoyprotein supplement was used in Experiment #8. This nitrogen source requiredadditionalpreparationstepsforitssolubilization,before being added to the medium (details are provided in the SupplementaryMaterial). $MEDIUM¼ 1 VRðCX or PPÞ Xn 1ðC Comp K $ Comp K Þ ð1Þ
$MEDIUM:Costofmedium,US$/(gDCWorgPspA4Pro);
CX:Biomassconcentration,gDCW/L;
CkComp:Concentrationofcomponentkinthemedium(Table2);
PP:PspA4Proproduction,g/L;
$kComp:Costofcomponentkinthemedium(Table2)converted
accordingtoCkComp.
2.2.2.Costofairandoxygensupplies(gases)
Inorder tomaintainDOTat30%of saturation,a sequential controlstrategywasappliedusingtheSupersys_HCDCsoftware,
comprisingstepwiseincreasesinstirringspeedandairvolumetric flowrate(QAIR),uptotheirupperlimits,followedbythegradual
enrichment of air with pure oxygen. This enrichment was accomplishedbymanipulationof theoxygen(QO2)and air
volumetricflowrates,sothatthetotalinletflowrateofthegas supplied tothebioreactorwas keptat4–6 stdL/min [58]. The QO2 and QAIR values were recorded by Supersys_HCDC,
enabling determination of thetotal volumeof oxygen supplied (VO2 = ƩQO2.
D
t) and the total energy consumption of thecompressor(EC=ƩPC.
D
t), withthesetwo values beingusedtodetermine thegasescosts.Theoverallcostofsupplyingair and oxygenisgivenbyEq.2,wherethemarketpricesforoxygenand electricitywereUS$0.52/stdm3andUS$0.126/kWh,respectively
[56]. The compressor power consumption (PC) was estimated
theoretically using Eq. 3. This approach was analogous to the procedure of Knollet al.[29], consideringan idealsingle-stage compressor, with electrical and mechanical compression losses takenintoaccountusingtheefficiencynumber(
h
C=0.7)[24].$GASES¼
1
VRðCX or PPÞð0:52V02þ
0:126ECÞ ð2Þ
$GASES:Costofgases,US$/(gDCWorgPspA4Pro);VR:Bioreactorculture
volume,L;VO2:Totaloxygenconsumptionvolume,m3;EC:Total
energyconsumptionofcompressor,kWh.
PC¼ p1QAIR
h
Cg
ðg
1Þ p2 p1 ðg1Þ g ½ 1 ( ) ð3Þ PC:Powerconsumptionofcompressor,kW;QAIR:Airvolumetricflowrate,L/min;p1:Inletabsolutepressureofcompressor,Pa;p2:
Outletabsolutepressureofcompressor,Pa;
h
C:Efficiencynumberofcompressor[24];
g
:Isoentropicexponent,1.4foroxygen[24]. 2.2.3.CostofstirringThecostofstirringwasdeterminedusingEq.4.Thetotalenergy consumedforstirring(ES)wascalculatedbyintegratingovertimethe
theoreticalvaluesofgassedstirringpower consumption,PS(ES=ƩPS.
D
t).Thesevalueswerecorrelatedtothepowerforungassedstirring(P), usingEqs.5or6[59].Theenergyefficiency(
h
S)wasassumedtobe0.65 [24] and the impellers were considered to be operated at maximum velocity(900rpm),sincethisconditionwoulddemandthe highest energy consumption, in order to better represent an overdesignedengine.Consideringthreeimpellersimmersedinthe liquid,thePvaluesweredeterminedusingEqs.7and8,accordingto theRushtonmethodcombinedwiththeMichelandMillercorrelation [60] and the correction factor (fC) used by Campani et al. [24], assuming
standard STR geometries of diameter (DR/Di) and height (HR/Di)
equalto3.ForEqs.5and6,thetotalvolumetricgasflowratevalues were retrieved from the dataset automatically acquired by the SuperSys_HCDCsoftware.
$S¼ 0:126
1 VRðCX or PPÞ
ES ð4Þ
$S:Costofstirring,US$/(gDCWorgPspA4Pro);
ES:Energyconsumptionofstirring,kWh.
PS¼ P
h
S 112:2QG ND3i ! ; valid for QG ND3i <0:037 ð5Þ PS¼ Ph
S 0:621:85QG ND3i ! ; valid for QG ND3i >0:037 ð6Þ P: Power consumption of ungassed stirring, kW;QG: Total gasDi: Impeller diameter, 0.078m;
h
S: Efficiency number ofstirring,0.65[24].
P¼ 3fCNP
r
bN 3D5ig ð7Þ
P: Power consumption of ungassed stirring, kW; Np: Power
number,5.2(Rushtonimpellerinturbulentregime,accordingto theReynoldsnumber);
r
b:Brothdensity,kg/m3;g:Gravitationalacceleration,m/s2.fC¼ ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi DR=Di ð Þ^ðHL=DiÞ^ DR=Di ð ÞðHL=DiÞ s ð8Þ DR:Bioreactordiameter,0.160m;HL:Heightoftheliquid,0.267m;
‘^’indicatesthegeometriccharacteristicsoftheSTRusedinthe experiments.
2.2.4.Costofcooling
In ordertomaintainaconstant bioreactortemperature,it is necessarytoremovetheheatreleasedbycellmetabolism(QHEAT)
andthe heattransferred fromtheimpellers (ES)totheculture
broth.Atheoreticalapproachtoestimatethecostofcoolingwas used, considering that all the heat generated by the cells and impellerswastotallytransferredtothecoolingwater,assuminga negligiblelagphaseduration.Hence,coolingofthebrothbygas stripping and losses to the environment were neglected. The metabolic heat released during the bioreactor cultures was correlatedtotheoxygenuptakerate(OUR)usingEq.9,proposed for aerobic microorganisms by Shuler and Kargi [61], as a generalizationoftheexpressionsreportedbyCooneyetal.[62] andAbbottandClamen[63].
The OUR values wereautomatically calculated using Eq. 10, whichwasobtainedfromthenitrogenandoxygenmolarbalances in thegas phase, based onpressure (p), temperature(T), total volumetricgasflow (Q),and molarfractionsofoxygen(O) and
carbondioxide(C)inthegasstreamsentering(1)andleaving(2) the bioreactor. The total metabolic energy generation (EM)
was calculated by integration of QHEAT over time, up to the
experimentalpointofmaximumPP,andwasaddedtothetotalES
value obtained as described in the previous section. Next, the theoreticalcostofcoolingwasestimatedusingEq.11,assumingan efficiencynumber(
h
B)of0.7[29].QHEAT¼ K:OUR:VR ð9Þ
K:Constantofproportionality,0.50kJ/mmolO2;
OUR:Oxygenuptakerate,mmolO2/(L.h);
QHeat:Metabolicheatreleaserate,kJ/h.
OUR¼ pQ1 RVR yO1yO2 1yO1 ð Þ 1yO2yC2 ð Þ ð10Þ p=Pressureofenteringgas,atm;
Q1=Enteringgasvolumetricflowrate,L/h;
T1=Temperatureofenteringgas,K;
R:Idealgasconstant,8.2105atm.L/(mmol.K). $COOLING¼ 0:126 EMþES ð Þ
h
B 1 VRðCX or PPÞ ð11Þ$COOLING: Cost of cooling, US$/(gDCW or gPspA4Pro); EM: Total
metabolicenergygeneration,kWh;
h
B:Efficiencynumberofthermostaticbath,0.7[29].2.3.Statisticalanalysis
Alltheresultspresentedaretheaveragevalues fordifferent experimentalanalysescarriedoutintriplicateandthepropagation oftheirstandarddeviations.Genuinereplicateswereprovidedby Experiments#7.aand#7.b.TheTukeytestwasusedtoevaluate statisticaldifferences[64],wherep-values<0.05wereconsidered significant.
Fig.1.Directcostratios,usingExperiment#3asareference(ref.),intermsof(a)PspA4Pro(ref.US$96.5/kgPspA4Pro)and(b)biomass(ref.US$11.9/kgDCW).*TheTukeytestwas
3.Resultsanddiscussion
3.1.PspA4Proandbiomassproductioncosts:generalaspects
The medium and cooling, in that order, were the main components of the direct cost, regardless of the cultivation strategyused,representingmorethan80%ofthedirectcostfor any experiment evaluated (Fig. 1). High contributions of the mediumtothecostshavebeenevidencedinotherstudieswith recombinant proteins [25,31]. Initially, the complex media (Experiments#6,#7,and#8)providedthelowestdirectPspA4Pro costs,inagreementwiththefindingsofZhangandGreasham[32] forlow-costproducts,butthiswouldstronglydependonnitrogen source prices (as will be demonstrated in the last section). Experiments #6, #7, and #8, performed with complex media, presentedapproximately70–80%ofthecostsofExperiment#3, forwhich thePspA4Procost wasthebestvalueobtainedusing defined medium (Fig.1.a). Higher soluble recombinant protein productionand ashorter cultivationtime usingcomplexmedia (Table1)seemedtobethemainreasonsfortheseresultsobtained during bioreactor batches, since prolonged protein expression couldincreasecosts[28].
Inthecaseofbiomassproduction(Fig.1.b),Experiments#2,#6, and#7resultedinthelowestoverallcostratios,becausethehigher biomass concentrations contributed to mitigating the costs (Table 1).For the other cultivation strategies investigated, the differencesintheoverallcostratiovalueswerenotstatistically significant.
3.2.PspA4Proandbiomassproductioncosts:specificaspects 3.2.1.Effectoftemperatureondirectcost
Operationatmoderatetemperatureappearedtoreducethecost ofPspA4Proproduction.ComparisonoftheresultsforExperiments #1 (27C), #3 (32C), and #4 (37C) (Fig.1), which were all performedwithdefinedmediumandinductionbyIPTG(Table1), showedthatthelowestcostintermsofproteinwasobtainedfor Experiment #3. For this condition, a better balance between metabolicburdenandbiomassproductionwaslikelytoreducethe cost [38,65]. This possibility was reinforced by the results of Experiments#2and#5(Table1),sincelessproteinwasproduced at 37C than at 27C, possibly as a consequence of a higher metabolic burden on the host cells in the former experiment [66,67],whichcouldbepartiallyevidencedbyhighplasmidlosses duringExperiment#5at37C[47].
3.2.2.Effectofinducersonenergycosts
ThetypeofinducercaninfluencethePspA4Procostintermsof theenergyrequiredduringbioreactorbatches.Theproteincosts
relatedtoenergyconsumption(forcooling,stirring,and compres-soroperations)werelowerwhenIPTGwasusedastheinducer, ratherthanlactose,atthesametemperatures(Fig.2).Comparedto lactose,theuseofIPTGasinducerleadstofasterproteinsynthesis [28,67],withthehigherPspA4Proproductionresultinginshorter inductionphases,hencereducingenergy costs.Generally, faster recombinant protein synthesis has been associated with an increase ininclusionbodies orlowqualityrecombinantprotein (incorrectly folded, or with impaired function and properties) [68,69]. In fact, due to their coiled-coil α-helix structure [46], N-terminalfragmentsofPspAareverysolubleandthermostable. WhenPspA4Prowasheatedfrom15to95C,followedbycooling
to15C,itrecoveredtheoriginalstructure,asmeasuredbyCD,and a relativelythermostable regionwas identifiedbetween40 and 50C [37]. In addition, the presence of PspA4Pro in inclusion bodieswasneverfoundintheculturesamplesanalyzed(Figure S2). For these reasons,fast induction strategies can beapplied during E. coli cultures to produce PspA, without affecting its secondarystructure(FigureS3).Thus,themostsuitableinduction strategy depends onthe protein characteristics and should be investigatedonacase-by-casebasis.
3.2.3.Influenceofcomponentpriceonmediumcost
3.2.3.1.Complexmedium(auto-induction). Forthecostestimation basedontheinformationobtainedfrombusinessintelligenceand e-commerceplatforms[56,57],thenitrogenand carbonsources weremajorcontributorstothecostsinExperiments#6,#7,and #8, performed according to the auto-induction strategy, with lactoseasinducer.Thesecomponentsrepresentedaround70%of themediumcost(Fig.3.aand3.c),whiletheinducer(lactose)and other medium components played less important roles. Considering all the nitrogen sources, the protein hydrolysates (ANSorVNS)hadthegreatestimpactonthecost(Fig.3.b),because theirpriceswereapproximately3timeshigherthanthemarket valueofyeastextract,whichwasalsopresentinthemediumwith halfthehydrolysateconcentration(Table2).
Theimpactofthenitrogensourcepricesoncostscouldbeeven greater,giventhatANSandVNSpricescanvarywidelyaccordingto thesupplier.Forexample,whenpeptonewasreplacedbysoybean proteinsupplementextractinExperiment#8,adoptingitscurrent saleprice,therewasanincreaseof15%inthecontributionofthe nitrogensourcetothecostofthemedium(Fig.3.d).Accordingto sensitivityanalysisofpeptoneprices(Fig.4.a),ifthecommercial certified peptonewiththebest price(US$ 45/kg)foundin our surveywasconsideredinthecalculations(Supplier1),it would leadtoanincreaseof125%inthePspA4Procost,relativetothe valueestimatedusingthecommercialplatformprices.Similarly,if the second best supplier was considered, it would lead to an
Fig.2.Effectofinducersonenergycosts(cooling,stirring,andcompressoroperations).Costratios*intermsof(a)PspA4Pro(ref.US$21.9/kgPspA4Pro)and(b)biomass(ref. US$2.7/kgDCW).Allexperimentsusingdefinedmedium.*Reference:Experiment#3(Table1).*TheTukeytestwasperformedtoevaluatestatisticaldifferences,considering standarddeviationsandapvalue<0.05assignificant.ns:notsignificant.
increaseofaround400%.AsshowninFig.4.a,thecertifiedpeptone pricesofdifferentmarketsuppliersstronglyimpactedthedirect cost.Inthesameway,theglycerolpricehadsomeimpactonthe PspA4Prodirectcost(Fig.4.b).However,thisrawmaterialmaynot offeragreatopportunitytodecreasetheproductioncost,sinceasa commodity[70,71],itiscommonlyofferedbyseveralsuppliersat lowerpricesthanthoseofnitrogensources.Hence,thesefindings indicate that nitrogen sources are more relevant than carbon sources,intermsofthecostsofcultivationusingcomplexmedia,
andthereforerepresentapotentialtargetforloweringthecostsof recombinantproteinandbiomassproduction.
ComparisonofExperiments#6,#7,and#8(Fig.1)showedthat theuseofnitrogenfromanimalorvegetalsourcesledtolowcosts forproducingPspA4Pro.However,forbiopharmaceutical produc-tion,peptonesfromvegetalsourcesseemtobemoresuitable,since theycontainprion-freeproteinandareincompliancewithgood manufacturingpractices(GMP).Tryptoneswithprion-freecerti fi-cationarealsoavailableandcanbeusedforbiopharmaceuticals
Fig.3. Costofcomplexmediumwithpeptones:(a)corecosts;(b)nitrogenandcarbonsourcesindetail.Costofcomplexmediumwithsoysupplement:(c)corecosts;(d) nitrogenandcarbonsourcesindetail.MC:mediumcost;Suppl.Ext:supplementextract.
Fig.4.Sensitivityanalysesofpeptone,glycerol,IPTG,andlactosepricesonPspA4Prodirectcost.Surveyinvolvingcertifiedpeptonesfromthreelargemarketsuppliers. Reference:Experiment#3(US$96.5/kgPspA).
production,buttheyaremoreexpensivethanordinaryenzymatic caseinhydrolysates.Asdiscussedinthenextsection,thepricesof peptonesmaybecrucialcriteriaintermsofmediumselection.
3.2.3.2.Definedmedium. Concerningthecultivationscarriedout withdefinedmedium(Experiments#1-#5,Table1),thechoiceof inducerplaysanimportantroleindeterminingcosts[32].When IPTGwasusedastheinducerinthecultivations,thiscompound presented the highest percentage of the medium cost, with a contribution2-foldhigherthanthatofglycerol,thesolecarbon source used (Fig. 5.a). Although the amount of IPTGadded to inducePspA4Proexpressionwasverysmall(1mM),itwasbyfar themostexpensivecomponent(Table2).Sincelactoseischeaper thanIPTG,whenitreplacedIPTGastheinducerinthecultivations, glycerolbecamethemostrelevantmediumcomponentinterms ofcost(46%),followedbylactose(30%)(Fig.5.b).Thus,theprices of IPTG, glycerol, and lactose may have substantial effects on the direct costs of PspA4Pro production using defined medium(Figs.4.b,4c,and4d).Itisimportanttoemphasizethat definedmediumformulationsalsocomplywithGMPguidelines [32], even using IPTG as inducer, since the presence of this substance as a contaminant in the final product is very unlikely,due tothe extensivepurification stepsused toobtain biopharmaceuticals[37].
3.3.InfluenceofcultivationstrategyonPspA4Proquality
The qualityof the proteinproduced is a crucial issueto be considered when choosing the cultivation strategy, since the biologicalfunctionoftherecombinantproteincouldbeaffectedby cultivation conditions [68]. The analysis of inclusion bodies formation showed that, irrespective of the cultivation strategy adopted, no PspA4Pro was aggregated and accumulated as insolubleinclusionbodies (FigureS2),suggestingthatPspA4Pro wascorrectlyfoldedinallcases.However,PspAhasnoenzymatic activity, soit is not possible toevaluate its biological function (lactoferrinbindingcapability)withoutpreviousproteinpuri fica-tion.Similarly,theanalysisofsecondarystructurebyCDcanonly beperformedwithpureproteinsamples.Therefore,fourdifferent cultivation conditions wereselected for protein purification, in ordertoevaluatetheinfluenceoftheproteinproductionstrategy onPspA4Proquality:Experiment#2,usingdefinedmediumand mildinductionconditions;Experiment#3,which wasthemost cost-effectivecultivationstrategy;Experiment#4,usingdefined mediumandwithfastbiomassgrowth;andExperiment#8,which wascarriedoutwiththealternativenitrogensource(Table1).
ThepurificationofPspA4Profromthebiomassharvestedinthe selected experiments was performed following the procedures describedpreviously[37],with96–98%finalpuritiesobtainedfor allthepurificationprocessesperformed.TheresultsofCDanalyses andlactoferrinbindingassaysforExperiments#2,#3,#4,and#8
are provided in the Supplementary Material. The results for Experiment #7werereportedby Figueiredoet al.[37]. TheCD spectrademonstratedthatPspA4Prohadtheexpectedsecondary structure in allcases (usingdifferentmedia,temperatures, and proteinsyntheses),beingcomparabletothePspA4Prostandardas well as to literaturedata [45,46,72,73], presenting two typical valleyswithminimaat208nmand222nm.
ThePspA4Propurifiedfromallfourexperimentsshowedequal ability to bind lactoferrin, with the same slope for binding accordingtoPspA4Proconcentration,indicatingthattheproduct ofeachexperimenthadthesamequality(FigureS3).Inaddition, the lactoferrin binding assay also indicated that rabbit anti-PspA4Pro antibodies, which bind to the PspA4Pro attached to lactoferrinboundtotheplate,recognizedPspA4Profromallthe experiments.
3.4.PspA4Proproteinandbiomassproductioncosts:pointsto highlight
Afterperformingtheeconomicanalysisoftheproduction of PspA4ProusingE.coliBL21(DE3)inthebioreactor,accordingtothe processfactorsassessedinthiswork,themostimportantpointsto considerin theselectionof a suitablecultivation strategywere identified,asdiscussedbelow.
There is a limit valuefor theprice of nitrogensources that determines whetherdefined or auto-induction complex media shouldbeselectedfortheproductionoftheproteinatlowcost.In ordertofindthisthresholdvalue,thepeptonepriceinExperiment #7(Table1)wasvariedsothatthePspA4Prodirectcostratiowas equivalent to the best performance achieved with a defined medium (Experiment #3), considering all other raw material pricesandprocessparameterstoremainconstant.Followingthe criteria adoptedduringthe presentcost analysis, thethreshold priceobtainedwasaboutUS$30/kgforANS/VNS,whichcouldbe usedtodecidewhentousecomplexordefinedmediumduringa bioreactor batch. However, commercial peptone products with prion-freecertificationsarecommonlysoldforpricesabovethis thresholdvalue(Fig.4),differentfromthee-commerceplatform peptonepricesusedinthepresentcostestimates.Furthermore, thesurveyofdifferentANSand VNSsuppliersshowed thatthe pricescanvarysignificantlyaccordingtothemanufacturer,quality, andquantitypurchased,withtheminimumpriceofsoypeptone being US$ 45/kg (1.5-fold higher than the threshold price). Therefore, the results demonstrated that in this case, useof a definedmediumwasthemostcost-effectiveoption.
Efforts to provide low-cost raw materials with acceptable quality arerequired in order tobroaden the usageof complex mediainrE.colicultures.The“homemade”extractofsoyprotein supplementforhumanconsumptionusedinExperiment#8isan exampleofacheapnutrientthatcanleadtoacompetitivedirect cost.Itisalsoimportanttorememberthatforbio-basedproducts
suchasindustrialenzymesandchemicals,ordinaryANS/VNScan beemployedinrE.colicultures[28],whereascertifiednitrogen sources are a specific requirement for biopharmaceutical production,whichcanmakeproteinproductionmoreexpensive. Moreover,thequalityofthenutrientsusedformedium formula-tion may have different effects on synthesis of a specific heterologousprotein,affectingitsmolecularintegrityandleading toinclusionbodyformation[74].Therefore,qualityassessmentof thefinalproductshouldbeprovidedwhennon-conventionalraw materialsourcesorundefinedmaterials(suchaspeptonesoryeast extract,amongothers)fromdifferentsuppliersareused,inorder toguaranteeproteinfunctionand properties.Forthis reason,it is always important to perform specific studies to evaluate commercialpeptonesandyeastextractfromdifferentsuppliers, in order to assess the impacts of these nutrients on process economicsandproductquality.
Careisneededinselectionofthetemperatureusedtocultivate E.colicellsandproducesolublerecombinantproteins.Although PspA4Propresented thebest production resultsat 32C, other heterologous proteins would be unlikely to show the same behavior[35]. Metabolicburden, formationof inclusion bodies, and protein quality impairment are some important factors to evaluateforeachdifferentrecombinantproduct[68].Studieshave indicatedthatlowtemperatureseemstohelpin avoidingthese issues,or atleast in mitigatingthem during E.colicultivations [75,76].However,theseresultsdidnotconsidercostimpactson processes. Studies with bioreactor cultivations may be time consumingandexpensiveintermsoftheproduction,purification, andanalysisofrecombinantproducts.Theymayalsoleadtoa trial-and-error approach, becauseseveral factorsmay influence this type of production [47]. Nonetheless, efforts aiming at cost-effectiverecombinantproteinproductionshouldbeencouraged,in ordertoimprovedecision-makingbefore implementingprocess strategies.
According tothe findings of thepresent economic analysis, inductionwithIPTGseemedtobethebeststrategyforproduction of PspA4Pro. As discussed previously, induction with IPTG appearedtointensifyandacceleraterecombinantprotein produc-tion,resultinginlowerprocesscosts.Itisclearthatthechoiceof inducer(aswellastheinductionstrategy)mustbeevaluatedona case-by-casebasis, accordingtothecharacteristicsofthetarget recombinant protein. Nonetheless, this contradicts the general opinionthatinductionwithIPTGisnotindustriallyviable,dueto itshighprice[27,65,77,78].Infact,consideringonlytheimpactof the medium on the cost of recombinant protein or biomass production, IPTG would be the most expensive raw material. However,takingthecomprehensiveapproachofthepresentcost analysisforPspA4Proproduction,IPTGwouldbethebestchoice.In addition, the IPTG concentration could be reduced [27,79,80], furtherdecreasingtheproductioncostofthisrecombinantprotein. Besides the application of the proposed methodology to identify cultivation conditions for PspA4Pro production, as discussed above, there are some considerations concerning extending the economic analysis approach developed here to othersituations.Althoughthereareseveraldifferentculturemedia andbioreactoroperationalmodesthatcanbeemployedtoproduce recombinantproteinsinE.coli,theresultsobservedforPspA4Pro may be valid when the recombinant protein solubility is not (orminimally)affectedbythecultureconditions.Infact,thereare many recombinant antigens reported as fully soluble when produced in recombinant E. coli using Lac operon expression systems:pneumococcalproteinssuchasZmpB[81], neuramini-daseA[82],andPotD[83];autolysinfromListeriamonocytogenes andPseudomonasaeruginosa[84,85];transferrinbindingproteins A from N. meningitidis [86]; hyaluronate lyase and PspC from Streptococcussuis[87,88];andthecholesterol-dependentcytolysin
family,includingstreptolysinO,pneumolysin,suilysin[89],and arcanolysin[90].IncommonwithPspA,allthesemoleculesare single-chainpolypeptideswithoutdisulfidebonds.
Eventhoughtheinfluenceoftheoperationalconditionsonthe processeconomicswasbasedonbench-scaleSTRbioreactordata, theproposedapproachmaybeappliedforevaluatingdifferentSTR bioreactorscales,includingthelargeronespresentintheindustry, since the theoretical equations used are general. Besides, the methodology may be easily extended to air-lift and similar bioreactorssimplybysetting thestirringspeed tozeroinsuch cases.Furthermore,theconceptof“directcost ratio” shouldbe mentioned. It was introduced to make the economic analysis generalandless dependentofthebioreactorscaleaswellasof unavoidablefluctuationsinthepricesofmediumcomponentsand energy.Ontheotherhand,ifdesired,thecostanalysiscanbealso basedontheactualpricesofraw-materialsandutilitiesknownfor aspecificcultivationcaseandleadtotheidentificationofthemain targetsforalowercostoperation.
In addition to recombinant proteins, E. coli canbeusedto synthesize value-addedbiomoleculessuchasaminoacids,organicacids,and advancedbiofuels,correspondingtoaglobaltradeofUS$22billion [91–93]. Usingsystems biologytools,this bacteriumhasstrong potentialtobecomethefuturemicrobial factoryforsustainable productionofbio-basedchemicals,especially consideringitsability to grow in inexpensive, abundant, and renewable feedstocks, including lignocellulosic biomass hydrolysates [94] and crude glycerolgeneratedfrombiodieselproduction[95].Thecalculation procedure developed here could also be used to assess the economicsoftheproductionofthesebio-basedchemicals,aswell astoidentifythemostfavorablebioreactorcultivationconditions. Itisalsoimportanttopointoutthattheoutlinedmethodology is suitable for estimation of the direct costs of thecultivation process.Therealcostofarecombinantprotein(oranybioproduct) depends ontheoverallproduction cost,which encompassesall stepsinupstreamanddownstreamprocessingandaccountsfor equipmentdepreciation,capitalcost,andlaborcost,inadditionto thedirectproductioncosts.
4.Conclusions
Based on usual cost estimation equations for cultivation suppliesand energyconsumption,combinedwithdesign corre-lations, the approach developed here was applied in the assessment of process economics using data from 8 different bioreactorprocessstrategies.Inthecasestudydiscussedhere,the beststrategywascharacterizedbyashortinductionphase,carried outatmoderatetemperature(32C)andwithIPTGasinducer,in ordertomaintainhighPspA4Proproductionrates,togetherwith lowercellmetabolicstressandreducedenergyconsumption.
Thepricesofnitrogensourcesusedincomplexmediaseemto bedeterminantforselectionofthetypeofmedia,especiallywhen it is necessary to ensure adherence to GMP standards for biopharmaceutical production, as in the case of PspA4Pro. On theotherhand,ordinaryANS/VNS,suchassoysupplement,may facesomeconstraintsduetothisregulatoryissue,althoughsuch materialsmaybesuitablefortheproductionofbio-basedenzymes andchemicals.Ingeneral,forcertifiedcommercialpeptoneswith pricesaboveUS$30/kg,definedmediumisthemostcost-effective choice. However, the use of complex media is a promising approachforreducingtheproductioncostofthetargetprotein, giventhatresearcheffortsaremadetoaddresscriticalissues,such asthecharacterizationofpeptonesfromdifferentsourcesandwith differentqualitiesand prices, interms of nutrientcontent,and theirimpactsonproductandbiomassyields.Furthermore,itisalso necessary to ensure the commercial availability of certified peptoneswithmoreaccessibleprices.
Selectionofthebestcultivationstrategyclearlydependsonthe targetproduct,hostcell,andproductionscale,requiringanalyses onacase-by-casebasis.However,providedthatbasicinformation about process conditions, biomass concentration, and product formationisavailable,thetheoreticalapproachpresentedhereto analyze the effects of different media, inducers, and complex nitrogensourcesonPspA4Proproductioncanbeeasilyextendedto evaluation of the economics of any bioreactor cultivation, contributingtoidentificationofthemostcost-effectivestrategy. Authorcontributions
(1)theconceptionand designof thestudy,oracquisitionof data,oranalysisandinterpretationofdata,(2)draftingthearticle orrevisingitcriticallyforimportantintellectualcontent,(3)final approvaloftheversiontobesubmitted.
CRediTauthorshipcontributionstatement
ValdemirM.Cardoso:Conceptualization,Methodology, Inves-tigation,Validation,Datacuration,Visualization,Writing-original draft,Formalanalysis.GilsonCampani:Investigation.MaurícioP. Santos:Investigation.GabrielG.Silva:Investigation.ManuellaC. Pires:Investigation.VivianeM.Gonçalves:Supervision,Writing -review&editing,Fundingacquisition.RobertodeC.Giordano: Supervision,Funding acquisition.Cíntia R.Sargo: Investigation. AntônioC.L.Horta:Investigation,Software.TeresaC.Zangirolami: Conceptualization, Methodology,Validation,Writing -review & editing,Supervision,Fundingacquisition.
DeclarationofCompetingInterest
The authors declare that they have no known competing financial interests or personal relationships that could have appearedtoinfluencetheworkreportedinthispaper.
Acknowledgements
This work was in part financed by the Coordenação de Aperfeiçoamento de Pessoalde Nível Superior, Brazil(CAPES – Finance code 001). The authors acknowledge São Paulo State Research Foundation, Brazil (FAPESP, grant #2015/10291-8) for funding the experimental work and São Paulo State Research Foundation, Brazil(FAPESP, grant #2018/13469-0)for MC Pires MSc scholarship. The authors thank Renan Minin de Mori for providing rabbit anti-PspA4Pro antibodies and Cost-Drivers platformforprovidingcommercialdata.
AppendixA.Supplementarydata
Supplementarymaterialrelatedtothisarticlecanbefound,inthe onlineversion,atdoi:https://doi.org/10.1016/j.btre.2020.e00441.
References
[1]M.Baeshen,A.M.Al-Hejin,R.S.Bora,M.M.M.Ahmed,H.A.I.Ramadan,K.S. Saini,N.A.Baeshen,E.M.Redwan,ProductionofbiopharmaceuticalsinE.coli: currentscenarioandfutureperspectives,J.Microbiol.Biotechnol.25(2015) 953–962.
[2]K.M.Wilding,J.P.Hunt,J.W.Wilkerson,P.J.Funk,R.L.Swensen,W.C.Carver,M. L.Christian,B.C.Bundy,Endotoxin-freeE.Coli-basedcell-freeprotein synthesis:pre-expressionendotoxinremovalapproachesforon-demand CancerTherapeuticproduction,Biotechnol.J.14(2019)1–6.
[3]G.Walsh,Biopharmaceuticalbenchmarks,Nat.Biotechnol.32(2014)(2014) 992–1000.
[4]N.K.Tripathi,A.Shrivastava,Recentdevelopmentsinrecombinantprotein– baseddenguevaccines,Front.Immunol.9(2018)1–15.
[5]G. Bashiru, A. Bahaman, Advances & challenges in leptospiral vaccine development,IndianJ.Med.Res.147(2018)15–22.
[6]C.B.Palatnik-de-sousa,NucleosidehydrolaseNH36:avitalenzymeforthe leishmaniagenusinthedevelopmentofT-Cellepitopecross-protective vaccines,Front.Immunol.10(2019)1–12.
[7]T. Lagousi, P. Basdeki, J. Routsias, V. Spoulou, Novel protein-based pneumococcalvaccines:assessingtheuseofdistinctproteinfragments insteadoffull-lengthproteinsasvaccineantigens,Vaccines.7(2019). [8]P.Méndez-Samperio,Developmentoftuberculosisvaccinesinclinicaltrials:
currentstatus,Scand.J.Immunol.88(2018)1–6.
[9]M.Keikha,M.Eslami,B.Yousefi,A.Ghasemian,M.Karbalaei,Potentialantigen candidatesforsubunitvaccinedevelopmentagainstHelicobacterpylori infection,J.Cell.Physiol.234(2019)21460–21470.
[10]A.Berlec,B.Strukelj,Currentstateandrecentadvancesinbiopharmaceutical productioninEscherichiacoli,yeastsandmammaliancells,J.Ind.Microbiol. Biotechnol.40(2013)257–274.
[11]L.Sanchez-Garcia,L.Martín,R.Mangues,N.Ferrer-Miralles,E.Vázquez,A. Villaverde,Recombinantpharmaceuticalsfrommicrobialcells:A2015update, Microb.CellFact.33(2016)1–7.
[12]S.Fayaz,P.Fard-esfahani,M.Golkar,M.Allahyari,S.Sadeghi,Expression, purificationandbiologicalactivityassessmentofromiplostimbiosimilar peptibody,J.Pharm.Sci.24(2016)1–5.
[13]J.S.Tan,R.N.Ramanan,T.C.Ling,M.Shuhaimi,A.B.Ariff,Enhancedproductionof periplasmicinterferonalpha-2bbyEscherichiacoliusingion-exchangeresinfor insituremovalofacetateintheculture,Biochem.Eng.J.58–59(2011)124–132. [14]M.Imielinski,C.Belta, Á.Halász, H.Rubin,Systemsbiologyinvestigating
metaboliteessentialitythroughgenome-scaleanalysisofEscherichiacoli productioncapabilities,Bioinform.21(2005)2008–2016.
[15]S.Pontrelli,T.Chiu,E.I.Lan,F.Y.Chen,P.Chang,J.C.Liao,Escherichiacoliasahost formetabolicengineering,Metab.Eng.50(2018)16–46.
[16]W. Choe, R. Nian, W. Lai, Recent advances in biomolecular process intensification,Chem.Eng.Sci.61(2006)886–906.
[17]B. Jia, C.O. Jeon, High-throughput recombinant protein expression in Escherichiacoli:currentstatusandfutureperspectives,OpenBiol.6(2016) 1–17.
[18]C.Huang,H.Lin,X.Yang,Industrialproductionofrecombinanttherapeuticsin Escherichiacolianditsrecentadvancements,J.Ind.Microbiol.Biotechnol.39 (2012)383–399.
[19]U.Beshay,H.El-enshasy,I.M.K.Ismail,H.Moawad,E.Wojciechowska,S. Abd-El-Ghany,B-Glucanaseproductionfromgeneticallymodifiedrecombinant Escherichiacoli:effectofgrowthsubstratesanddevelopmentofaculture mediuminshakeflasksandstirredtankbioreactor,ProcessBiochem.39 (2003)307–313.
[20]A.Khushoo,Y.Pal,K.J.Mukherjee,Optimizationofextracellularproductionof recombinantasparaginaseinEscherichiacoliinshake-flaskandbioreactor, Appl.Microbiol.Biotechnol.68(2005)189–197.
[21]M.Kunze, R. Huber, C. Gutjahr, S.Mullner, J. Büchs, Predictivetool for recombinantproteinproductioninEscherichiacolishake-flaskculturesusing anon-linemonitoringsystem,Biotechnol.Prog.28(2012)103–113. [22]S.Y.Lee,Highcell-densitycultureofEscherichiacoli,TrendsBiotechnol.14
(1996)98–105.
[23]J. Shiloach, R. Fass, Growing E. Colito high cell density— a historical perspectiveonmethoddevelopment,Biotechnol.Adv.23(2005)345–357. [24]G. Campani, M.P. Santos, G. Gonçalves, A.C.L. Horta, A.C. Badino, R.D.C.
Giordano,V.Maimoni,T.C.Zangirolami,Recombinantproteinproductionby engineeredEscherichiacoliinapressurizedairliftbioreactor:a techno-economicanalysis,Chem.Eng.Process.ProcessIntensif.103(2016)63–69. [25]J.Choi, S.Y.Lee, Processanalysisand economic evaluationfor Poly
(3-hydroxybutyrate)productionbyfermentation,BioprocessEng.17(1997) 335–342.
[26]M.K.Danquah,G.M.Forde,Growthmediumselectionanditseconomicimpact onplasmidDNAproduction,J.Biosci.Bioeng.104(2007)490–497. [27]G.Ferreira,A.R.Azzoni,S.Freitas,BiotechnologyforBiofuelsTechno-economic
analysisoftheindustrialproductionofalow-costenzymeusingE.coli:the caseofrecombinantβ-glucosidase,Biotechnol.Biofuels11(2018)1–13. [28]B.A.Fong,D.W. Wood,Expressionand purificationof ELP-intein-tagged
targetproteinsinhighcelldensityE.Colifermentation,Microb.CellFact.77 (2010)1–12.
[29]A.Knoll,B.Maier,H.Tscherrig,J.Büchs,Theoxygenmasstransfer,carbon dioxideinhibition,heatremoval,andtheenergyandcostefficienciesofhigh pressurefermentation,AdvBiochemEngin/Biotechnol.92(2005)77–99. [30]J.Walther,R.Godawat,C.Hwang,Y.Abe,A.Sinclair,K.Konstantinov,The
businessimpactofanintegratedcontinuousbiomanufacturingplatformfor recombinantproteinproduction,J.Biotechnol.213(2015)3–12.
[31]R.J.V.A.N. Wegen, Y. Ling, A.P.J.M. Member, Polyhydroxyalkanoates using escherichiacoli:aneconomicanalysis,Chem.Eng.Res.Des.76(1998)417–426. [32]J. Zhang, R. Greasham, Chemically defined media for commercial
fermentations,Appl.Microbiol.Biotechnol.51(1999)407–421.
[33]J. Kim, K.H.Kim, Effects of minimal mediavs. Complexmedia on the metaboliteprofilesofEscherichiacoliandSaccharomycescerevisiae,Process Biochem.57(2017)64–71.
[34]J.P.M.Sanders,J.H.Clark,G.J.Harmsen,H.J.Heeres,J.J.Heijnen,S.R.A.Kersten, W.P.M.VanSwaaij,J.A.Moulijn,ProcessIntensificationprocessintensification inthefutureproductionofbasechemicalsfrombiomass,Chem.Eng.Process. ProcessIntensif.51(2012)117–136.
[35]G.Gellissen,etal.,Keyandcriteriatotheselectionofanexpressionplatform, in:P.D.G.Gellissen(Ed.),Prod.Recomb.PROTEINS-Nov.Microb.Eukaryot. Expr.Syst.,WILEY-VCHVerlag-GmbH&CoKGaA,Weinheim,2005,pp.1–6.
[36]J.Kaur,A.Kumar,J.Kaur,Strategiesforoptimizationofheterologousprotein expressioninE.coli:roadblocksandreinforcements,Int.J.Biol.Macromol.106 (2018)803–822.
[37]D.B.Figueiredo,E.Carvalho,M.P.Santos,S.Kraschowetz,R.T.Zanardo,G. Campani,G.G.Silva,C.R.Sargo,A.C.L.Horta,RdeC.Giordano,E.N.Miyaji,T.C. Zangirolami,J.Cabrera-Crespo,V.M.Gonçalves,Productionandpurificationof anuntaggedrecombinantpneumococcalsurfaceproteinA(PspA4Pro)with high-purityandlowendotoxincontent,Appl.Microbiol.Biotechnol. 101(2017) 2305–2317.
[38]G. Campani, G.G. Silva, T.C. Zangirolami, M.P.A. Ribeiro, Recombinant Escherichiacolicultivationinapressurizedairliftbioreactor:assessmentof theinfluenceoftemperatureonoxygentransferanduptakerates,Bioprocess Biosyst.Eng.40(2017)1621–1633.
[39]G.Campani,M.P.A.Ribeiro,T.C.Zangirolami,F.V.Lima,Ahierarchicalstate estimationandcontrolframeworkformonitoringanddissolvedoxygen regulationinbioprocesses,BioprocessBiosyst.Eng.(2019)1–15.
[40]A.Håkansson,H.Roche,S.Mirza,L.McDaniel,A.Brooks-Walter,D.Briles, Characterizationofbindingofhumanlactoferrintopneumococcalsurface proteinA,Infect.Immun.69(2011)3372–3381.
[41]B.Ren,A.Szalai,O.Thomas,S.Hollingshead,D.Briles,Bothfamily1andfamily 2PspAproteinscaninhibitcomplementdepositionandconfervirulencetoa capsularserotype3strainofStreptococcuspneumoniae,Infect.Immun.71 (2003)75–85.
[42]G.C.Barazzone,R.Carvalho,S.Kraschowetz,A.L.Horta,C.R.Sargo,A.J.Silva,T.C. Zangirolami,C.Goulart,L.C.C.Leite,M.M.Tanizaki,V.M.Gonçalves,J. Cabrera-Crespo,Productionandpurificationofrecombinantfragmentof
pneumococcalsurfaceproteinA(PspA)inEscherichiacoli,ProcediaVaccinol.4 (2011)27–35.
[43]R.J.Carvalho,J.Cabrera-crespo,M.M.Tanizaki,V.M.Gonçalves,Development ofproductionandpurificationprocessesofrecombinantfragmentof pneumococcalsurfaceproteinAinEscherichiacoliusingdifferentcarbon sourcesandchromatographysequences,Appl.Microbiol.Biotechnol.94 (2012)683–694.
[44]A.T.Moreno,M.L.S.Oliveira,D.M.Ferreira,P.L.Ho,M.Darrieux,L.C.C.Leite,J.M. C.Ferreira,F.C.Pimenta,A.L.S.S.Andrade,E.N.Miyaji,Immunizationofmice withsinglePspAfragmentsinducesantibodiescapableofmediating complementdepositionondifferentpneumococcalstrainsand cross-protection,Clin.VaccineImmunol.17(2010)439–446.
[45]M.J.Jedrzejas,S.K.Hollingshead,J.Lebowitz,L.Chantalat,D.E.Briles,E.Lamani, Productionandcharacterizationofthefunctionalfragmentofpneumococcal surfaceproteinA,Arch.Biochem.Biophys.373(2000)116–125.
[46]M.J.Jedrzejas,E.Lamani,R.S.Becker,Characterizationofselectedstrainsof pneumococcalsurfaceproteina,J.Biol.Chem.276(2001)33121–33128. [47]C.P. Papaneophytou, G. Kontopidis, Statistical approaches to maximize
recombinantproteinexpressioninEscherichiacoli:ageneralreview, ProteinExpr.Purif.94(2014)22–32.
[48]A.C.L.Horta,C.R.Sargo,A.M.Velez,M.P.Santos,R.C.Giordano,T.C.Zangirolami, On-linemonitoringofbiomassconcentrationbasedonacapacitancesensor: assessingthemethodologyfordifferentbacteriaandyeasthighcelldensity fed-batchcultures,BrazilianJ.Chem.Eng.32(2015)821–829.
[49]F.W.Studier,Proteinproductionbyauto-inductioninhigh-densityshaking cultures,ProteinExpr.Purif.41(2005)207–234.
[50]A.C.L.Horta,A.J.Silva,C.R.Sargo,A.M.Velez,M.C.Gonzaga,R.C.Giordano,T.C. Zangirolami,Asupervisionandcontroltoolbasedonartificialintelligencefor highcelldensitycultivations,BrazilianJ.Chem.Eng.31(2014)457–468. [51]L.Olsson,J.Nielsen,On-lineandinsitumonitoringofbiomassinsubmerged
cultivations,TrendsBiotechnol.15(1997)517–522.
[52]B.Sonnleitner,G.Locher,A.Fiechter,Biomassdetermination,J.Biotechnol.25 (1992)5–22.
[53]M.M.Bradford,Arapidandsensitivemethodforthequantitationmicrogram quantitiesofproteinutilizingtheprincipleofprotein-dyebinding,Anal. Biochem.72(1976)248–254.
[54]U.K.Laemmli,Cleavageofstructuralproteinsduringtheassemblyofthe bacteriophageT4,Nature.227(1970)680–685.
[55]M.D. Abràmoff, P.J. Magalhães, S.J. Ram, Image processing with ImageJ, BiophotonicsInt.11(2004)36–42.
[56]COST-DRIVERS,PriceTrendsandForecasts,(2016).SãoPaulohttps://www. costdrivers.com/.
[57]MOL-BASE,ChemicalSearchEngine,(2016).(accessedJune20,2009)http:// molbase.com.
[58]A.M. Vélez, A.J. da Silva,A.C. Luperni Horta, C.R. Sargo,G. Campani, G. GonçalvesSilva,R.deLimaCamargoGiordano,T.C.Zangirolami, High-throughputstrategiesforpenicillinGacylaseproductioninrE.Colifed-batch cultivations,BMCBiotechnol.14(2014).
[59]R.E.TREYBAL,Mass-TransferOperations,3.ed.,McGraw-Hill,NewYork,1980. [60]B.J.Michel,S.A.Miller,Powerrequirementsofgas-liquidagitatedsystems,
AIChEJ.8(1962)262–266.
[61]M.L.SHULER,F.KARGI,BioprocessBioengineering:Basic Concepts,2.ed., PrenticeHall,NewJersey,2002.
[62]C.L.Cooney,D.I.C.Wang,R.I.Mateles,Measurementofheatevolutionand correlationwithoxygenconsumptionduringmicrobialgrowth*,Biotechnol. Bioeng.11(1968)269–281.
[63]B.J. Abbott, A. Clamen, The relationship of substrate, growth rate, and maintenancecoefficienttosinglecellproteinproduction,Biotechnol.Bioeng. 15(1973)117–127.
[64]J.W. Tukey, Comparing individual means in the analysis of variance, Biometrics.5(1949)99–114.
[65]R.S.Donovan,C.W.Robinson,B.R.Glick,Review:Optimizing inducerand cultureconditionsforexpressionofforeignproteinsunderthecontrolofthe lacpromoter,J.Ind.Microbiol.16(1996)145–154.
[66]S. Carneiro,E.C. Ferreira, I. Rocha, Metabolic responses to recombinant bioprocessesinEscherichiacoli,J.Biotechnol.164(2013)396–408. [67]Z. Li, X.Zhang, T. Tan,Lactose-induced productionof human soluble B
lymphocytestimulator(hsBLyS)inE.Coliwithdifferentculturestrategies, Biotechnol.Lett.28(2006)477–483.
[68]G.L.Rosano,E.A.Ceccarelli,RecombinantproteinexpressioninEscherichiacoli: advancesandchallenges,Front.Microbiol.5(2014)1–17.
[69]S.Ventura,A.Villaverde,Proteinqualityinbacterialinclusionbodies,Trends Biotechnol.24(2006)179–185.
[70]M.Anitha,S.K.Kamarudin,N.T.Kofli,Thepotentialofglycerolasavalue-added commodity,Chem.Eng.J.295(2016)119–130.
[71]M. Ayoub, A.Z. Abdullah, Critical review on the current scenario and significanceofcrudeglycerolresultingfrombiodieselindustrytowards moresustainablerenewableenergyindustry,RenewableSustainableEnergy Rev.16(2012)2671–2686.
[72]S.L.Haughney,L.K.Petersen,A.D.Schoofs,A.E.Ramer-Tait,J.D.King,D.E.Briles, M.J.Wannemuehler,B.Narasimhan,Retentionofstructure,antigenicity,and biologicalfunctionofpneumococcalsurfaceproteinA(PspA)releasedfrom polyanhydridenanoparticles,ActaBiomater.9(2013)8262–8271.
[73]E. Lamani,D.T. McPherson, S.K. Hollingshead, M.J. Jedrzejas, Production, characterization,andcrystallizationoftruncatedformsofpneumococcal surfaceproteinAfromEscherichiacoli,ProteinExpr.Purif.20(2000)379–388. [74]F.Silva,J.A.Queiroz,F.C.Domingues,Evaluatingmetabolicstressandplasmid stabilityinplasmidDNAproductionbyEscherichiacoli,Biotechnol.Adv.30 (2012)691–708.
[75]J.M.Song,Y.J.An,M.H.Kang,Y.H.Lee,S.S.Cha,Cultivationat6-10Cisan
effectivestrategytoovercometheinsolubilityofrecombinantproteinsin Escherichiacoli,ProteinExpr.Purif.82(2012)297–301.
[76]J.A.Vasina,F.Baneyx,Expressionofaggregation-pronerecombinantproteins atlowtemperatures:acomparativestudyofthe,ProteinExpr.Purif.9(1997) 211–218.
[77]A.K.Gombert,B.V.Kilikian,RecombinantgeneexpressioninEscherichiacoli cultivationusinglactoseasinducer,J.Biotechnol.60(1998)47–54. [78]H.G.Menzella,E.A.Ceccarelli,H.C.Gramajo,NovelEscherichiacolistrainallows
efficientrecombinantproteinproductionusinglactoseasinducer,Biotechnol. Bioeng.82(2003)809–817.
[79]K.Einsfeldt,J.Baptista,S.Júnior,A.Paula,C.Argondizzo,M.Alberto,T.Lívio,M. Alves,R.Volcan,A.Leites,CloningandexpressionofproteaseClpPfrom StreptococcuspneumoniaeinEscherichiacoli:studyoftheinfluenceof kanamycinandIPTGconcentrationoncellgrowth,recombinantprotein productionandplasmidstability,Vaccine.29(2011)7136–7143.
[80]M.Lecina,E.Sarró,A.Casablancas,F.Gòdia,J.J.Cairó,RegulararticleIPTG limitationavoidsmetabolicburdenandaceticacidaccumulationininduced fed-batchculturesofEscherichiacoliM15underglucoselimitingconditions, Biochem.Eng.J.70(2013)78–83.
[81]Y.Gong,W.Xu,Y.Cui,X.Zhang,R.Yao,D.Li,H.Wang,Y.He,J.Cao,Y.Yin, ImmunizationwithaZmpB-basedproteinvaccinecouldprotectagainst pneumococcaldiseasesinmice,Infect.Immun.79(2011)867–878. [82]J. Long, H. Tong, T. DeMaria, Immunization with native orrecombinant
Streptococcuspneumoniaeneuraminidaseaffordsprotectioninthechinchilla otitismediamodel,Infect.Immun.72(2004)4309–4313.
[83]T.R.Converso,C.Goulart,D.Rodriguez,M.Darrieux,L.C.C.Leite,Systemic immunizationwithrPotDreducesStreptococcuspneumoniaenasopharyngeal colonizationinmice,Vaccine.35(2017)149–155.
[84]E.M.Scheurwater,J.M.Pfeffer,A.J.Clarke,Productionandpurificationofthe bacterialautolysinN-acetylmuramoyl-L-alanineamidaseBfrom
Pseudomonasaeruginosa,ProteinExpr.Purif.56(2007)128–137.
[85]L.Wang,L.Walrond,T.D.Cyr,M.Lin,AnovelsurfaceautolysinofListeria monocytogenesserotype4b,IspC,containsa23-residueN-terminalsignal peptidebeingprocessedinE.Coli,Biochem.Biophys.Res.Commun.354 (2007)403–408.
[86]J.S.Oakhill,C.L.Joannou,S.K.Buchanan,A.R.Gorringe,R.W.Evans,Expression andpurificationoffunctionalrecombinantmeningococcaltransferrin-binding proteinA,Biochem.J.364(2002)613–616.
[87]A.G.Allen,H.Lindsay,D.Seilly,S.Bolitho,S.E.Peters,D.J.Maskell,Identification andcharacterisationofhyaluronatelyasefromStreptococcussuis,Microb. Pathog.36(2004)327–335.
[88]K.Vaillancourt,L.Bonifait,L.Grignon,M.Frenette,M.Gottschalk,D.Grenier,D. Grenier,Identificationandcharacterizationofanew cellsurfaceprotein possessing factorH-bindingactivityintheswinepathogenandzoonoticagentStreptococcus suisPrintedinGreatBritain,J.Med.Microbiol.62(2013)1073–1080. [89]K.Nakayama,T.Ezoe,HeatIncubationInactivatesStreptococcalExotoxinsand
RecombinantCholesterol-DependentCytolysins:Suilysin,Pneumolysinand StreptolysinO,Curr.Microbiol.69(2014)690–698.
[90]B.Jost,E.Lucas,S.Billington,A.Ratner,D.McGee,Arcanolysinisa cholesterol-dependentcytolysinofthehumanpathogenArcanobacteriumhaemolyticum, BMCMicrobiol.239(2011)1471–2180.
[91]Y.Tsuge,H.Kawaguchi,K.Sasaki,A.Kondo,Engineeringcellfactoriesfor producingbuildingblockchemicalsforbio-polymersynthesis,Microb.Cell Fact.15(2016)1–12.
[92]T.Werpy,G.Petersen,TopValueAddedChemicalsFromBiomassVolumeI— ResultsofScreeningforPotentialCandidatesFromSugarsandSynthesisGas, Washington,2004.https://www.nrel.gov/docs/fy04osti/35523.pdf. [93]X.Chen,L.Zhou,K.Tian,A.Kumar,S.Singh,B.A.Prior,Z.Wang,Metabolic
engineeringofEscherichiacoli:asustainableindustrialplatformforbio-based chemicalproduction,Biotechnol.Adv.31(2013)1200–1223.
[94]Z. Li,P. Hong,Y. Da,L. Li, G. Stephanopoulos,Metabolicengineeringof EscherichiacolifortheproductionofL-malatefromxylose,Metab.Eng.48 (2018)25–32.
[95]S.S. Yazdani, R. Gonzalez, Engineering Escherichia coli for the efficient conversionofglyceroltoethanolandco-products,Metab.Eng.10(2008) 340–351.
