Computers & Electronics

A small bispecific protein selected for orthogonal affinity purification

A small bispecific protein selected for orthogonal affinity purification
of 13
All materials on our website are shared by users. If you have any questions about copyright issues, please report us to resolve them. We are always happy to assist you.
Related Documents
  © 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim605Biotechnol. J. 2010, 5, 605–617DOI 10.1002/ 1Introduction Due to the high cost in producing and purifyingproteins,the production steps as well as the purifi-cation of the target protein need to be optimized.For high-throughput production and purificationof many proteins in parallel,general methods with-out the need for individual optimization are neces-sary.Recombinant techniques allow efficient pro-duction of target proteins and affinity chromatog-raphy offers a selective purification method.Toachieve effective purification for a wide range of different proteins,a selective purification tag canbe used.Different affinity interactions have beenexploited for this purpose,  e.g .,enzyme-substrate,protein-metal,protein-carbohydrate,or protein-protein [1].Depending on the choice of interactionpartners the cost and complexity of the purificationconditions will vary.The choice of affinity tag canalso influence the amount of produced protein,andits solubility,stability and function.A small tag isless likely to interfere with the target protein andits further applications.Another important aspectis that the tag should be robust and fold easily re-gardless of target protein.To achieve a target pro-tein with high purity usually more than one purifi-cation step is necessary.To increase the selectivity of the purification method,and thus enhance thepurity of the sample,an orthogonal purificationstrategy would be beneficial.In classical orthogo- Research Article A small bispecific protein selected for orthogonal affinitypurification Tove Alm, Louise Yderland, Johan Nilvebrant, Anneli Halldin and Sophia Hober  School of Biotechnology, Department of Proteomics, Royal Institute of Technology, AlbaNova University Center, Stockholm,SwedenA novel protein domain with dual affinity has been created by randomization and selection. Thesmall alkali-stabilized albumin-binding domain (ABD*), used as scaffold to construct the library,has affinity to human serum albumin (HSA) and is constituted of 46 amino acids of which 11 wererandomized. To achieve a dual binder, the binding site of the inherent HSA affinity was untouchedand the randomization was made on the opposite side of the molecule. Despite its small size andrandomization of almost a quarter of its amino acids, a bifunctional molecule, ABDz1, with abili-ty to bind to both HSA and the Z 2 domain/protein A was successfully selected using phage dis-play. Moreover, the newly selected variant showed improved affinity for HSA compared to theparental molecule. This novel protein domain has been characterized regarding secondary struc-ture and affinity to the two different ligands. The possibility for affinity purification on two differ-ent matrices has been investigated using the two ligands, the HSA matrix and the protein A-based,MabSelect SuRe matrix, and the new protein domain was purified to homogeneity. Furthermore,gene fusions between the new domain and three different target proteins with different character-istics were made. To take advantage of both affinities, a purification strategy referred to as or-thogonal affinity purification using two different matrices was created. Successful purification of all three versions was efficiently carried out using this strategy. Keywords: Albumin binding domain · Bispecific binder · Orthogonal purification · Phage display · Z domain Correspondence: Professor Sophia Hober, AlbaNova University Center,KTH, School of Biotechnology, Proteomics, 106 91 Stockholm, Sweden E-mail: Fax: +46-5537-8481 Abbreviations:ABD ,albumin-binding domain; ABD *,stabilized ABD; CD ,circular dichroism; CV ,column volume; Z 2 domain ,dimeric Z domainReceived19February 2010Revised12April 2010Accepted14April 2010  BiotechnologyJournal Biotechnol. J. 2010, 5, 605–617606© 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim nal purification design,the succeeding separationcharacteristics are distinct from each other,utiliz-ing different properties of the proteins.Here we in-troduce the term orthogonal affinity purification.Endowing one purification tag with more than oneuseable affinity would increase the efficiency of thepurification.The two binding surfaces in the smallbispecific tag could be used for subsequent purifi-cation without any target-specific optimization.A small protein domain,historically used as anaffinity tag for protein purification,is the albumin-binding domain (ABD) [2].ABD is a three-helicalbundle [3] derived from the streptococcal protein G[4].To render this domain even more usable inaffinity purification it has been engineered into a variant denoted ABD* with increased resistance toalkaline conditions [5].Hence,it is possible to cleanequipment,including the affinity matrix with ABD*coupled as ligand, in situ .To gain more informationabout the protein domain,directed mutagenesis was used to locate the albumin-binding site.It wasshown to be positioned mainly in the second helix[6].This finding was further supported by structur-al investigations using NMR and X-ray crystallog-raphy [7,8].Triple alanine mutants were also madein helices one and three at positions pointing away from the suggested binding site;these showed re-tained or increased binding to human serum albu-min (HSA) [6].This indicated that it would be pos-sible to randomize positions pointing away fromthe HSA-binding surface to create a library that would make selection of new binders with retainedaffinity to HSA possible.Due to the improved char-acteristics of ABD* and its ability to bind HSA,it was chosen as a scaffold when creating the phagedisplay library.Another protein also frequently used for pro-tein purification,both as an affinity ligand and apurification tag,is the staphylococcal protein A (SpA) [9].The high selectivity of protein A for IgGhas made protein A matrices commonly used forantibody purification [10] and a number of matri-ces with protein A as a ligand are available.One of the domains responsible for IgG-binding was cho-sen as target in the phage display selection.Here,for the first time,we present a strategy that makes it possible to select small proteinaceousbinders capable of selectively binding two differentproteins.Moreover,the possibility of purifying a se-lected and characterized variant from the ABD* li-brary using two different affinity matrices was in- vestigated.Finally,the selected domain was intro-duced on the Nterminus of three different targetproteins and utilized as an affinity tag in two suc-ceeding purification steps. 2Materials and methods 2.1Bacterial strains and oligonucleotides For recombinant and phage work the  Escherichiacoli strain RRI ∆ M15 was used [11].The recombi-nant DNA techniques used were carried out essen-tially as described by Sambrook et al .[12].Oligonu-cleotides for the library construction were synthe-sized at Scandinavian Gene Synthesis (SGS;Köping,Sweden).All other primers were synthe-sized at MWG-BIOTECH AG (MWG;Ebersberg,Germany).DNA restriction enzymes were boughtfrom MBI Fermentas (Vilnius,Lithuania).Proteinproduction was carried out in  E. coli strain Rosetta(DE3) (Novagen,Merck KGaA,Darmstadt,Ger-many) and all protein purifications were carriedout using an ÄKTAExplorer (GE Healthcare,Upp-sala,Sweden). 2.2Library construction The phagemid vector pML was constructed frompKN [13] by introduction of an  Eco RI site and adummy fragment using restriction sites  Mlu I and  Xho I.pML was restricted using  Eco RI and  Xho I andthe restricted vector was separated on a 1% agarosegel and extracted from the gel using QIAquick GelExtraction Kit (QIAGEN,Solna,Sweden).The li-brary was assembled using oligonucleotides ABD-libCod3 and ABDlibRev3 having complementary regions in the region encoding helix two.The DNA fragments were hybridized and extended using sixtemperature cycles with AmpliTaq Gold (AppliedBiosystems,Foster City,CA,USA),purified usingQIAquick PCR Purification Kit (QIAGEN) and re-stricted using  Eco RI and  Xho I.After separation andextraction from a 3% low melt agarose gel,the pu-rified vector and fragment were ligated using T4DNA ligase (New England BioLabs,Boston,MA,USA).To remove contaminating proteins the liga-tion mixture was phenol-chloroform extracted andthereafter ethanol precipitated.The plasmids weretransformed into electrocompetent RRI ∆ M15 usingBio-Rad Gene Pulser (Bio-Rad Laboratories,Her-cules,CA,USA) to yield a final library size of ~10 7 members. 2.3Analysis of the ABD* library The library was analyzed by PCR amplification andDNA sequencing was performed on an ABI 3700DNA Sequencer (Applied Biosystems).To assessthe retainment of HSA-binding in the library,tworounds of phage selection against HSA were per-formed.HSA (Sigma-Aldrich,Steinheim,Ger-  © 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim607 many) was biotinylated using EZ-Link Sulfo-NHS-LC-Biotin (Pierce,IL,USA) according to the man-ufacturer’s recommendation.The selections werecarried out as described in Sect.2.4 with a targetconcentration of 1µM with the exception that cycle1 was performed with the target protein capturedon paramagnetic beads during the binding step.A Western blot was done on phage stocks elut-ed from cycles 1 and 2 to analyze the enrichment of HSA binders from the initial library.As a positivecontrol cell lysate with the parental ABD* was in-cluded.The samples were reduced and analyzed by SDS-PAGE on a 4–12% gradient NuPAGE NovexBis-Tris gel (Invitrogen,Paisley,Scotland).Afterseparation the proteins were transferred to a nitro-cellulose membrane (Bio-Rad Laboratories) usingan XCell II™ Blot Module (Invitrogen).The mem-brane was blocked with PBS supplemented withTween (50mM phosphate,100mM NaCl,pH7.2,0.1% Tween 20;PBST) also containing 1% gelatin.The membranes were incubated with biotinylatedHSA (120nM) for 1h,followed by washing (PBST),incubation with the secondary peroxidase-conju-gated streptavidin (DakoCytomation,Glostrup,Denmark;0.6g/L diluted 1:5000),and a final wash-ing step.Detection was carried out using a CCDcamera (Bio-Rad) with SuperSignal West Dura Ex-tended Duration Substrate (Pierce) according tothe manufacturer’s protocol.To investigate the degree of truncated,com-pared to full-length,pIII on the phage surface,an-other Western blot analysis was done as describedabove but detection was carried out using a pri-mary mouse antibody recognizing pIII (NordicBiosite,Täby,Sweden;0.5g/L diluted 1:1000) and asecondary peroxidase-conjugated rabbit antibody (Sigma-Aldrich;diluted 1:5000).The relative inten-sities of the bands were compared using Quantity-One software (Bio-Rad). 2.4Selection Four rounds of selection in solution against adimeric Z domain (Z 2 domain) [14] were per-formed.All tubes used in the selection were pro-tein LoBind tubes (Eppendorf,VWR International,Stockholm,Sweden) blocked with PBST supple-mented with 5% BSA (Sigma-Aldrich;PBSTB 5%).Streptavidin-coated beads or REGEN beads coat-ed with neutravidin according to manufacturer’srecommendations (Dynal Biotech,Oslo,Norway) were used for preselection and selection.Beads were washed twice with 500µL PBST and blocked with 500µL PBSTB 5%.Four series (A–D) with dif-ferent target concentration were performed and,toincrease the stringency of each cycle,the numberof washes and amount of Tween 20 in the washingbuffer (PBST supplemented with 3% BSA) was in-creased for each cycle (Table1).PBS was used inthe last wash of each cycle.For each round of se-lection a new phage stock was prepared by infec-tion with helper phage M13KO7 (New EnglandBioLabs),followed by polyethylene glycol andsodium chloride (PEG/NaCl) precipitation,result-ing in phage titers of 10 11 –10 12 cfu/mL.Each roundof selection started with a negative selection usingthe corresponding blocked paramagnetic beads.The supernatant from this step was moved to anew tube and mixed with the biotinylated Z 2 do-main at varying concentrations (Table1),and in-cubated for 2h at room temperature.The bindingphages were captured on 0.5mg paramagneticbeads and washed as described above.For elution,incubation with 500µL 0.05M glycine-HCl pH2.2for 10min at room temperature was performed.After elution,the pH was adjusted with 450µL PBSand 50µL 1M Tris-HCl pH8.0.The eluate wasmixed with log phase RRI ∆ M15 and incubated at37°C for 30min.Thereafter,the cells were spreadon tryptone yeast extract (TYE) agar plates (15g/Lagar,3g/L NaCl,10g/L tryptone and 5g/L yeastextract),supplemented with 2% glucose and100mg/L ampicillin (Amp;Sigma-Aldrich) and in-cubated at 37°C overnight.The cells were resus-pended in tryptic soy broth (TSB,30 g/L;Merck,Darmstadt,Germany).A fraction of the cell sus-pension was used to create a new phage stock.Elu-tion titers and number of phages introduced ineach cycle were determined by infection of logphase RRI ∆ M15. Biotechnol. J. 2010, 5, 605– Table1 . Biopanning by phage display a) A (nM)B (nM)C (nM)D (nM)Number of washesTween (%) Cycle 1100100100100 (N)30.1Cycle 28020808050.2Cycle 380205050 (N)70.3Cycle 480202020100.4 a)Four series (A, B, C, and D) of biopanning with different target concentration were performed in four cycles (1–4). To increase the stringency in every selectionround, the number of washes and the amount of Tween was increased in each cycle. Streptavidin-coated beads have been used except where (N) indicates thatneutravidin was used.  BiotechnologyJournal Biotechnol. J. 2010, 5, 605–617608© 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim 2.5Cloning 2.5.1Cloning of ABD* variants Colonies from cycles3 and 4 were randomly picked from agar plates supplemented with100mg/L Amp for PCR screening and sequencing.Three constructs were chosen for further investi-gation;ABDz1,ABDz83,ABDz86 (ABDzx,x = 1,83,86).The ABD* library fragments were amplified by PCR,using Phusion High Fidelity DNA poly-merase (New England BioLabs).The cysteine inABDz1 was replaced by a serine using primerABDz1C25S and PCR-mutagenesis to createABDz1C25S.The fragments were purified beforerestriction as described in Sect.2.2.The expression vector pHis (containing a T7-promoter,a His-tag,and a kanamycin resistance gene) and the frag-ments ABDzx and ABDz1C25S were cleaved withthe restriction enzymes  Eco RI and  Xho I.Thepurified fragments were ligated with pHis us-ing T4 DNA ligase (New England BioLabs),result-ing in pHisABDzx and pHisABDz1C25S.A dimeric version of ABDz1C25S was created using primersABDdimerCod and ABDdimerRev with pHisABD-z1C25S as template,resulting in a fragment withflanking  Xho I restriction sites.The purified frag-ment and pHisABDz1C25S were restricted with  Xho I and purified.The linear vector was dephos-phorylated using Antarctic Phosphatase (New England BioLabs).Subsequent ligation resulted inpHisABDz1C25SDim. 2.5.2Cloning of target proteins in fusion with ABDz1 DNA fragments containing only ABDz1 were madeby PCR,using primers ABDz1NotI and ABDz1NcoI,incorporating restriction sites  NcoI  and  NotI  .Therestriction enzymes were used to cleave the puri-fied PCR fragment and vectors containing differenttarget proteins;pZbABP141377,pZbABPHT875,and pZbABPHT2375 [15].Ligation of vector andABDz1-fragment was performed at room tempera-ture using T4 DNA ligase (New England BioLabs)giving pABDz1-141377,pABDz1-HT875 andpABDz1-HT2375.The plasmids were sequence verified. 2.6Protein production Protein production was essentially performed ac-cording to [15].After protein production for 18h,the cell suspensions were gently harvested(2400  ×  g ,8min,4°C) and pellets were frozen forlater use. 2.7Protein purification of His 6 -tagged ABD*variants The frozen pellets were resuspended in 25mL run-ning buffer (50mM sodium phosphate,6M urea,300mM NaCl,pH8.0) and cells were disrupted by sonication (Vibra cell;Sonics and materials,Inc.,Danbury,CT,USA) at 60% amplitude and 1.0/1.0pulses for 3min.Before loading on a 1-mL Talonresin column (Clontech Laboratories,Inc.,Moun-tain View,CA,USA),the samples were centrifuged(35000  ×  g ,20min,4°C) and filtered (0.45µm).Thecolumn was equilibrated with 5 column volumes(CV) of running buffer and 20mL sample wasloaded.After washing with 20CV running bufferbound proteins were eluted with elution buffer(50mM NaAc,6M urea,100mM NaCl,30mM HAc,pH4.5). 2.8Protein analysis The eluted proteins were analyzed by SDS-PAGEusing a NuPAGE Novex Bis-Tris 4–12% gradient gelstained with GelCode Blue Stain Reagent (Pierce).NAP10 columns (GE Healthcare) were used forbuffer exchange to HEPES-buffered saline (10mMHEPES,150mM NaCl,3.4mM EDTA and 0.05%surfactant P20,pH7.4) (HBS-EP).Protein concen-tration was determined by absorbance measure-ments at 280nm and amino acid analysis. 2.9Biosensor analysis Real-time biospecific interaction analysis betweenthe selected variants and the target protein,the Z 2 domain,and ABD’s srcinal binding partner HSA  was performed on a Biacore 2000 instrument (GEHealthcare).The Z 2 domain was immobilized(~130RU) by thiol coupling through a C-terminalCys and HSA was immobilized (~2500RU) by amine coupling onto the dextran layer of a CM5chip according to the manufacturer’s instructions.All analytes were diluted in HBS-EP,and filteredthrough a 0.45-µm filter.ABD* and IgG were in-cluded in all analyses as controls.After every injec-tion the surfaces were regenerated using 15mMHCl.In the initial study each variant (ABDzx,ABDz1C25S and ABDz1C25SDim) was diluted to~500nM and injected over all surfaces at50µL/min.For kinetic studies a new chip was made with an immobilization level of ~200RU and~2000RU on the Z 2 -surface and HSA-surface re-spectively.ABD* and ABDz1C25S were analyzedon the HSA-surface and ABDz1 on the Z 2 -surfacein concentrations ranging from 10 to 1000nM at aflow rate of 50µL/min.The association rate con-  © 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim609 stant (k a ),the dissociation rate constant (k d ),andthe dissociation equilibrium constant (K D ) were es-timated using BIAevaluation 3.2 software (Bia-core). 2.10Circular dichroism Analysis of secondary structure was performed us-ing a J-810 spectropolarimeter (JASCO,Tokyo, Japan).All samples were diluted to approximately 0.22mg/mL except ABD* that was diluted to0.05mg/mL in PBS.The samples were scanned ina 1-mm cell from 250 to 195nm at a speed of 100nm/min.An average from five scans was calcu-lated.Using the same sample the thermal stability  was also investigated by heating the sample from20°to 90°C using a temperature slope of 5°C/min.The circular dichroism (CD) signal at 221nm wasdetected.Thereafter,a new CD spectrum was per-formed as previously described to investigate if theprotein regained its secondary structure. 2.11Protein purification using HSA Sepharose The frozen pellets were resuspended in 25mL run-ning buffer (25mM Tris-HCl,200mM NaCl,1mMEDTA,0.05% Tween20,pH8.0) and supernatantcontaining protein lysate was prepared as de-scribed in Sect.2.7.The protein lysates were loadedon a column with 1mL HSA Sepharose.The col-umn was equilibrated with 10CV of running bufferand 25mL of the sample was loaded.After washing with 5CV running buffer and 5CV washing buffer(5mM NH 4 Ac,pH5.5) the bound proteins wereeluted with 0.5M HAc,pH2.8.The protein concen-tration in the eluted fractions was determined by absorbance measurements at 280nm.The elutedpeaks were analyzed using SDS-PAGE as de-scribed in Sect.2.8.To compare the binding capac-ity of ABD* and ABDz1 to the HSA Sepharose,equal amounts of previously purified protein wereloaded onto the column and purified as describedabove.In the orthogonal purification setup,the twoprotein fractions with the highest absorbance fromthe SuRe purification step were pooled and pH wasadjusted to 7.5 before loading on the HSA Sepharose and purified as described above. 2.12Protein purification using MabSelect SuRe The frozen pellets were thawed and resuspendedin 25mL running buffer (20mM phosphate,150mM NaCl,pH7.2) and the supernatants con-taining protein lysates were prepared as describedin Sect.2.7.The protein solution was loaded on a 1-mL HiTrap MabSelect SuRe column (GE Health-care).The column was equilibrated with 5CV of running buffer and 20mL of sample was loaded.After washing with 5CV of running buffer,the pro-teins were eluted with 0.2M HAc,pH3.3.In the or-thogonal purification setup,the two protein frac-tions with the highest absorbance from the HSA purification step were pooled and pH was adjustedto 7.5 before loading on the SuRe matrix and puri-fied as described above.To investigate if it was pos-sible to achieve effective purification also under re-ducing conditions,one sample was treated with50mM DL -dithiothreitol (DTT;Sigma-Aldrich) at37°C for 30min before filtration and loading ontothe column.Thereafter,the described purificationscheme was used.In an attempt to find milder elu-tion conditions,the samples were eluted using0.2M HAc at pH3.3,pH4.0 and pH4.5.The frac-tions of the eluted peaks were pooled and analyzedas described in Sect.2.8.When purifying fusionproteins using the SuRe matrix,0.2M HAc atpH3.3 was used as elution buffer. 3Results 3.1Rationale for the design of the library Here we present a novel phage display library of ~10 7  variants constructed to create a possibility of obtaining small protein molecules that are able tobind two different target molecules with high se-lectivity.For construction of the library,a small andstable molecule with inherent and selective bind-ing to a protein target,was desired.Due to im-proved characteristics [5] and the inherent bindingproperty,ABD* was chosen as scaffold for the li-brary.In an earlier mutational study [6],it was alsoshown that despite mutations of nine amino acidsin helices one and three,at positions distant fromthe suggested binding site,the HSA-binding wasretained or even improved.Therefore,these nineamino acids in helices one and three,and two addi-tional amino acids,at the end of helices one andthree,respectively,were chosen for randomization when building the library (Fig.1a).The amino acids were randomized using NNK degeneracy codingfor all 20 amino acids and the amber stop (TAG) us-ing 32 different codons.A visualization of the pro-tein domain is shown in Fig.1b,where the three he-lices are shown as ribbons;the molecular surface ispresented in Fig.1c. 3.2Analyses of the unbiased ABD* library To verify that the library was randomized in thecorrect positions and without bias for certain Biotechnol. J. 2010, 5, 605–
Related Search
We Need Your Support
Thank you for visiting our website and your interest in our free products and services. We are nonprofit website to share and download documents. To the running of this website, we need your help to support us.

Thanks to everyone for your continued support.

No, Thanks