phage-displayed peptides as biosensor reagents

Phage-displayed peptides as biosensor reagents

Ellen R. Goldman,1 Mehran P. Pazirandeh,2 J. Matthew Mauro,2 Keeley D. King,3

Julie C. Frey3 and George P. Anderson2*1Georgetown University Medical Center, Department of Biochemistry and Molecular Biology, 3900 Reservoir Road NW, Washington, DC,20007, USA2Center for Bio/Molecular Science and Engineering, Naval Research Laboratory, Washington, DC 20375, USA3Geo-Centers Inc., Rockville, MD 20852, USA

This study investigated the potential to utilize phage-displayed peptides as reagents in sensor applications. Alibrary of random 12-mers displayed on phage was panned against staphylococcal enterotoxin B (SEB), acausative agent of food poisoning. Nine SEB binding phage clones were isolated, all of which share theconsensus sequence Trp His Lys at their amino terminus. Binding of several of these phage was shown to beinhibited when they were assayed in a competitive enzyme-linked immunosorbent assay (ELISA) formatwith synthesized peptide corresponding to the peptide-encoding region of one of the clones. Whole phagewere labeled with the dye Cy5, and incorporated into fluoroimmunoassays. Labeled phage were able todetect SEB down to a concentration of 1.4 ng/well in a fluorescence-based immunoassay. When incorporatedinto an automated fluorescence-based sensing assay, Cy5-labeled phage bound to probes coated with SEBgenerated a robust signal of about 10,000 pA, vs a signal of 1000 pA using a control fiber coated withstreptavidin. These results demonstrate the potential for development of phage-based sensor reagents.Copyright # 2000 John Wiley & Sons, Ltd.

Keywords:phage display; biosensors; staphylococcal enterotoxin B

Received 22 March 2000; revised 26 July 2000; accepted 27 July 2000

INTRODUCTION

A vast number of bio-assays and biosensors depend onantibodies as recognition reagents. While antibodiesfrequently have the desired sensitivity and selectivity, therecan be problems with antibody reagents. In some cases,antibodies may be unobtainable due to the non-antigenicnature of the analyte, or the target of interest may need to beanalyzed in a sample matrix not compatible with antibodyfunction. This latter limitation can be especially importantin environmental testing applications, where compoundsmust be extracted from soil or groundwater with organicsolvents. An additional drawback to conventional poly-clonal antibody technology is the time- and labor-intensiveprocess of obtaining antisera from animals, which can yielda variable product. While monoclonal antibodies provide amore consistent product, their development and productionis even more difficult. Several lines of research can addressthese limitations, including the use of nucleic acid aptamers(Kleinjung et al., 1998; Bruno and Kiel, 1999) or phage-

displayed peptides (Smith, 1985; Scott and Smith, 1990) asreagents in sensors.

We have explored the feasibility of using short peptidesdisplayed on the minor coat protein (pIII) of filamentousbacteriophage (Smith, 1985) as reagents in biosensors.Peptides are expressed fused to pIII on the surface of theviral particle, while the DNA encoding the fusion is presentinside the phage. This linkage between the peptidephysically attached to a coat protein of the phage and theDNA encoding the peptide facilitates development ofpeptide ligands for a variety of targets (Scott & Smith,1990). In order to isolate selective peptides, librariescontaining random peptide sequences displayed on phageare incubated with a solid substrate coated with target. Non-binding phage are washed away and selectively boundphage are eluted and subsequently amplified in bacterialcells. Selected phage can then be taken through additionalbinding/amplification cycles to enrich for desired bindingsequences. In this way, phage with specificity for a desiredtarget can be isolated from a random combinatorial libraryin a matter of days. In addition to reducing production timecompared with development of antibody reagents, thismethod can allow selection of peptides that bind tosubstances which might be difficult to obtain antibodiesagainst.

In the present work, we isolated bacteriophage displayingpeptides that bind to staphylococcal enterotoxin B (SEB), acausative agent of food poisoning. We have shown thatwhole phage, selected as described above, can be fluores-cently labeled for use in immunoassays. Labeled phage

JOURNAL OF MOLECULAR RECOGNITIONJ. Mol. Recognit.2000;13:382–387

Copyright# 2000 John Wiley & Sons, Ltd.

* Correspondence to: G. Anderson, Center for Bio/Molecular Science andEngineering, Naval Research Laboratory Code 6900, 4555 Overlook Ave SW,Washington, DC 20375, USA.E-mail: [email protected]/grant sponsor:Naval Research Laboratory.Contract/grant sponsor:Office of Naval Research.

Abbreviations used: SEB, staphylococcal enterotoxin B SEB; SEA,staphylococcal enterotoxin A; SEC 1, staphylococcal enterotoxin C1; SED,staphylococcal enterotoxin D.

were studied by enzyme-linked immunosorbent assay(ELISA), fluorescence microplate assays,and testedin afiber optic biosensor.To the bestof our knowledgethis isthe first time whole phagehave been labeled for use influoroimmunoassays;the results show promise for usingphage-displayedpeptidesasreagents in biosensors.

EXPERIMENTAL

Reagents

A library of random peptides12aminoacidslongdisplayedon the minor coat protein, geneIII, of the bacteriophageM13 waspurchasedfrom New England Biolabs(Ph.D.-12Phagedisplay library kit). The library contains1.9� 109

electroporated sequences. SEB, SEA, SED, SEC1 andaffinity-purified rabbit anti-SEBIgG were purchasedfromToxin Technology Inc. (Sarasota,FL). Peptides weresynthesizedand purified to 98% purity by PeptoGeneticResearch andCo.,Livermore,CA.

Selection of SEB binding phage in a random peptidelibrary

Biopanning experiments were carried out essentially asdescribed in the New England Biolabs Ph.D.-12 Phagedisplay library kit. Briefly, a well of a 24-well plate(Linbrotissueculture plates, ICN Biomedicals, Inc., Aurora, OH)wascoated overnight at 4°C with 100�g/ml SEB in 0.1MNaHCO3 (pH 8.6). Af ter completeremovalof the coatingsolution, eachcoatedwell was filled with blocking buffer[0.1 M NaHCO3 (pH8.6), 5 mg/ml BSA, 0.02%NaN3] andincubated at 4°C for 2 h. Af ter removing the blockingsolution and washing each well six times with TBST(50mM Tris–HCl pH 7.5,150mM NaCl, 0.1%Tween-20),approximately 4� 1010 phage from the random displaylibrary diluted in TBST were incubatedin the SEB-coatedwell for 1 h at room temperature. Unbound phage wereremovedby washing10 times with TBST. Bound phagewereelutedwith 0.2M glycine (pH 3) containing 1 mg/mlBSA and neutralized with 1 M Tris–HCl pH 9.1. Elutedphagewere titered and amplified in E. coli ER2537. Thebinding andelution procedureswererepeatedtwice exceptthatthepercentage.Tween-20wasincreasedto 0.5%for thesecond and third rounds. The phage input varied from5� 109 phagein thesecond roundto 1� 1011 phagein thethird round depending on the amplification level of theeluted phage. Phage were amplified and purified asdescribed in the NEB manual and in Rozinov and Nolan(1998).

PhageELISA

Ninety-six-well plates were coated with 100�l 0.1MNaHCO3 (pH 8.6) or TBS containing 5�g/ml SEB or100�l of buffer solution containingno SEB,andincubatedat4°C overnight.After discarding theSEBor blank solutionfrom the wells, plateswere blocked at 4°C for 2 h withblocking buffer. Plateswere then washedfive times in an

automatic plate washer (MAX line microplate washer,Moleculardevices,Menlo Park,CA) with washing/reactionbuffer (TBST with 0.5% Tween-20) and 5� 1010 phageparticlesper100�l wash/reaction buffer wereincubated ineachwell for 1.5–2h, shaking gently (60 rpm), at roomtemperature.Wellswerewashedfivetimes,then100�l of a1:5000 dilution of HRP-conjugated anti-M13 antibody(Pharmacia) in wash/reaction buffer wasincubatedin eachwell for 1 h at room temperature.Plateswerethenwashed10timesand100�l SigmafastOPD(Sigma)mix wasaddedto each well and incubated for about 10min at roomtemperature. Reactions were stoppedby the addition of100�l 4 N H2SO4. Absorbanceat490nmwasmeasuredin amicroplatereader (nmax Kinetic Microplatereader,Molecu-lar devices,Menlo Park,CA). In competitive ELISAs, thesynthetic peptides, at the indicated concentrations, wereaddedto the phagebeforethe phagebinding step.

Single-strand DNA preparation

Af ter plaque amplification, 0.5ml of phage-containingsupernatant was set aside for the preparation of single-stranded DNA template. Phagewere precipitated at roomtemperatureby theaddition of 200�l PEG/NaCl (20%w/vpolyethyleneglycol 8000,2.5 M NaCl). After 10min, thephagewerepelletedby centrifuging10minutesat full speedin a microfuge, and the supernatantwas discarded. Tubeswerecentrifugedfor 10s,andresidual supernatantremoved.The pellets were resuspended in 100�l iodide buffer[10 mM Tris–HCl (pH 8.0), 1 mM EDTA, 4 M NaI] andthe single-strandedDNA precipitated by the addition of250�l ethanol. After 10min, the mixture was centrifugedfor 10min. Thesupernatantswerediscardedandthepelletswashed with 70% ethanoland dried briefly in a vacuumcentrifuge. Thefinal pelletswere suspendedin 30�l water.Single-strandedDNA was sent to the Molecular GeneticsInstrumentation Facility at the University of Georgia forsequencing.

Fluorescentlabeling of phage

An aliquotof 300�l phagesuspendedin PBS(from a stockcontainingat least1� 1013 phage/ml) wasaddedto 300�llabeling buffer [50 mM sodium tetraborate (pH 9.0),40mMNaCl]. The phagewere addedto one vial of the Cy5 dye(FluroLink2 bisfunctional CY5Dye2, Amersham LifeScience),a NHS-esterwhich will covalentlyattachto freeamino groups on the phagesurface. Af ter incubation atroomtemperaturein thedark for 30min, with brief mixingevery 5 min, labeled phage were purified using sizeexclusionchromatographyon a BioGel P-10column(Bio-Rad, Hercules, CA). The column was equilibrated withPBS-0.5%Triton X-100,andthephage/Cy5 mix wasloadeddropwise. Fractions were collected asmonitored by a UVdetector; the first fraction contained labeled phagewhilesubsequentpeakscontainedunboundfluorophore.Labeledphagewere storedat 4°C. The number of Cy5 per phagewas calculated from the absorbance of labeled phageat650nm (molar extinction coefficient of Cy5, 250,000)assuming no lossof phageduring the labelingstep.

PHAGE-DISPLAYEDPEPTIDES 383

Copyright# 2000JohnWiley & Sons,Ltd. J. Mol. Recognit.2000;13:382–387

Fluorescence microplate assay

SEB wasadsorbedto 96-well membraneplates, with 3.0�biodyneA membrane-bottomed wells (Pall, Portsmouth,England), in amountsfrom 1000 to 1.4ng/well in 50mMTris–HCl (pH 8.0), 100mM NaCl, overnightat 4°C. Af terdiscarding the SEB solution, plates were washedthreetime with PBS–0.1% Tween 20, and then blocked with asolution of 1 mg/ml eachof caseinand BSA with 0.1%Triton X-100. Plates were then incubated with 100�l ofCy5-labeledrabbit anti-SEBantibody(20�g/ml; 133 nM)or Cy5-labeledPhage14 (�3.3� 1010/ml; �56 pM), bothdiluted into thepreviously describedblockingbuffer for 1 hat 4°C. Wells were washed three times with PBS–0.1%Tween 20. The amount of bound Cy5-antibody or Cy5-Phage14 wasdeterminedusing a SpectrofluorPlus(Tecan,Salzburg, Austria) with a 620nm (10nm band pass)excitation filter and a 670 (10nm band pass) emissionfilter. Triplicatesweremeasuredfor eachSEBconcentrationand the average background (no SEB adsorbed)wassubtracted.

Fiber optic immunoassaymethods

Optical probes for use in the fiber optic biosensorwerecoated with SEB or streptavidin, at 10�g/ml in 0.1MNa2CO3 (pH 9.6) overnight at 4°C. The optical probeswerethenrinsedwith distilled waterfor usein theRAPTORautomated fiber optic biosensor (Anderson et al., 2000;ResearchInternational Inc.,Woodinville, WA). Fouropticalprobes wereassembledinto a disposablechamber, referredto as a coupon,which when insertedinto the RAPTORprovidesall therequired fluidics connectionsandalignmentof the optical probesin the instrument.

To perform an assaya coupon was inserted into theRAPTOR device, filled with PB-Triton buffer (8.3mMphosphate pH 7.4,0.5%Triton 100)anda baselinereading

taken beforesample introduction. Buffer wasremoved, and1 ml labeled phage(1010–1012 phage/ml) was loaded ontothecoupon.The laser wasturnedon to takean initial Cy-5reading and then the optical probeswere incubated for 1hwith the sample. At the end of the hour a second Cy-5reading was taken before washing the coupon with PB-Triton. After thewashstep,afinal readingwastakenandthebound signal calculated by subtracting the initial bufferbaseline reading from the signal observed after washing.The signalsfrom theSEB-coatedprobes werecomparedtocontrol probescoatedwith streptavidin.

RESULTS AND DISCUSSION

Selection of phagedisplayedpeptidesthat bind SEB

The toxin SEB was the targetchosen to test the ability ofphagedisplayedpeptides to beincorporatedinto bioassays.Sincephage-displayedrandompeptideshavebeenselectedfor binding to a variety of protein targets (Cesareni et al.,1999; Rodi and Makowski, 1999) including toxic shocksyndrometoxin-1 (Satoet al., 1996),which is structurallysimilar to SEB(Papageorgiouet al., 1990),it wasreasonedthatspecificpeptide-binding sequencescouldalso befoundfor SEB.Additionally, high quality antibodiesagainstSEBareavailablefor usein sensorapplications (Tempelmanetal., 1996;Wadkinsetal., 1998);thusSEBofferedatargettowhich peptide binding effectiveness could be easilyevaluatedcompared with the antibody.

Ninerepresentative phageclonesfrom thethird roundofbiopanningwereamplifiedseparately, purified andanalyzedfor SEB binding by ELISA. All of these phage weredetermined to bind to immobilized SEB, but not touncoated, blocked wells (Fig. 1). Control wild-type phage(M13mp19), when similarly tested, did not bind to SEB.

Figure 1. Nine clones shown by ELISA to bind SEB. Absorbance at 490 isshown for each clone when examined against an SEB coated well (solidbars) and a blank coated well (striped bars). One hundred microliters of5� 1011 phage/ml were used in each sample. Measurements were donein triplicate; error bars represent the standard deviation.

384 E. R. GOLDMAN ET AL.


Sequenceanalysis of isolated phage

Thededucedaminoacidsequencesof therandom regionofthenineclones(corresponding to theELISA datain Fig. 1)arereported in Table1. Sevenout of the nine phagesharethe consensussequenceTrp His Lys at the aminoterminalendof thefused12-mer.Theother two phageisolatesshowconservative substitutions of this motif with a Trp to Phesubstitution in phage15 anda Lys to Arg changein phage16. The sequencesof thesephageareall proline-rich withtwo to four prolines per 12-mer, which may conferimportant secondary structure. Interestingly, only one ofthesesequencescontains a negatively chargedaminoacid,but all contain at leasttwo positively chargedresidues.

Synthetic peptide competitio n studies

A peptidecorresponding to thesequencefrom phage14wassynthesized(designatedpeptide1: NH2-Trp His LysAla ProArg Ala ProAla ProLeu Leu-COOH)to testfor inhibitionof binding of purified phage 14 to SEB. Inhibition ofbinding of phage14 was tested by co-incubation in theELISA formatwith threeconcentrationsof peptide1 (7�M,60�M, 1 mM). No inhibition of bindingto immobilizedSEBwasfoundatpeptideconcentrationsof 7 or 60�M; howeverwhen the peptide concentration was increasedto 1 mM,competition was observed(Fig. 2). No competition wasobserved with a control peptidepresent (peptide 2: NH2-AsnLysAsnSer PheAspAla Trp LeuGln SerPhe-COOH,isolatedin a panning againstimmobilized streptavidin) atidentical concentrations. Competition for SEBbindingwasalso seen when phage 18 was incubated with 1 mM ofpeptide 1 (Fig. 2), which indicatesthat the Trp His Lysfamily of peptides isolated recognizes a common oroverlapping binding site on SEB. That peptide1 wasableto competewith thephagefor bindingto SEBshowsthatthe12-meralone,independent of phageprotein III, bindsto thetarget. Thusit maybefeasibleto movethepeptideto otherplatforms for future uses.

Crossreactivity with other toxins

To investigate possible cross reactivity of the selectedpeptide ligands,threerepresentativeclones(phages14, 18and 20) were examined by ELISA for binding to otherstaphylococcal enterotoxins (SEA,SEC,SED)(Fig. 3). Thethree-dimensional structures of these toxins are similar(Papageorgiou et al., 1998). Minimal crossreactivity wasfoundwith SEAandSED.Thesetwo toxinsarehomologousbut shareonly about28% aminoacid sequenceidentity toSEB. Significant cross-reactivity, however, was observedwith SEC1. This is notsurprising consideringthatSEBandthe SECsshareabout60–70% sequence identity. As bothSEB and SEC are often present in staphylococcol foodpoisoning, it could be advantageousin anticipatedsensorapplications for the phage-displayed peptide reagent tocross-react.Interestingly, several of theantibodiescurrentlyusedto detectSEBhavea specificityprofile similar to thatof the phagewith no cross reactivity against SEA or SED(Tempelmanet al., 1996) and about20% cross-reactivityagainst the SECs.

Reactivity of Cy5 labeled phage

In preparation for usein fiber optic assays,we wereabletolabel phagewith 300–2000molecules of the dye Cy5 perviron (seeexperimental section). The reactiveform of thedye Cy5 usedfor labeling reacts mainly at lysine residuesandis probably reacting with accessiblelysinesonthemajorcoatprotein (proteinVI II) of thebacteriophage.To testthelabeled phage for retention of reactivity, we performedELISA assayswith labeled andunlabeledphageandfoundthat the Cy5 labeled material retainedits ability to bind toimmobilized SEB (datanot shown). This suggeststhat the

Table 1. SEB binding sequences isolated from a phagedisplay library

Clone Sequence

12 W H K P P P S A L G P K13 W H K T PK A T T Q P L14 W H K A P R A P A P L L15 F H K E W R P R P Y A F16 W H R P T PK P T L T I17 W H K P P V R P P S T Q18 W H K I P Q K A P L N P19 W H K Q K P M T A P Y P20 W H K F P P R P P S L G

Conserved motif is in bold type. Conservativesubstitutions in WHK motif are in bold and underlined.Prolinesareunderlined.

Figure 2. Competitive inhibition of selected phage binding to SEBby a synthetic peptide (peptide 1). Peptide 1 (1 mM, light graybars) was used to compete with phage 14 and 18 for binding tothe immobilized SEB in ELISA format. Relative absorbancevalues at 490 nm are shown, 100% corresponds to absorbanceat 490 nm in the absence of peptide. A control peptide (peptide 2,1 mM, striped bars) showed 100% signal for both phage. The datarepresent the mean of triplicate wells.



Lys of the Trp His Lys motif was not modified in asignificant fraction of the labeledwhole phage.

SEB binding phage in fluoroimmunoassayapplications

TheSEBbinding phageweretestedasfluorescentreagentsfor assaysconductedusing a fluorescencemicrotiter platereader and the RAPTOR fiber optic biosensor. SEBimmunoassaysusingtheCy5-labeledphageweredevelopedfor a fluorescence microtiter platereader. The Cy5-labeledphage-14 and a Cy5-labeled anti-SEB antibody bothproduced signal over background down to the lowestconcentration of SEB tested (1.4ng/well), Fig. 4. Thisresult shows the potential utility for selected phage asfluorescent reagents. In parallel assays similar molarconcentrationsof Cy5 were usedfor both the phageandthe antibody; however, as the phage are more highly

labeled, the molar ratio of antibody to phageused wasapproximately 2000:1. The concentration of phage wasselectedfrom a preliminaryexperiment thatdeterminedtheamount of phage which produced the best signal overbackground upon binding to SEB at 1000ng/well. Evenoptimized, the backgroundusing Cy5-labeledphagewastwice that observed for the antibody. High backgroundsignals generatedby Cy5-phagein these assaysresultedinincreaseddatauncertainty.

We are currently developingassays for field portableinstruments such as the RAPTOR fiber optic biosensor.Using thefiber optic biosensor,very largesignals(10,000–12,000 pA) were observedafter Cy-5-labeled phage(at aconcentrationof about1012 phage/ml) boundto SEB-coatedoptical probes. This comparesto a backgroundof 1000–2000 pAmp observedwhen the same phageweretestedonprobes coated with streptavidin. Cy-5 labeled anti-SEBantibody, howeverhasvery little background(0–20pA) andasignalof 80pA overbackgroundis consideredpositiveforantibody assays on this sensor (Anderson et al., 2000).Antibody binding to SEB coatedfibersgenerated signal inexcess of 22,000pA. Although phagegive a lower signalthan that generatedby the labeledantibody, the phagedogenerate a robust signal, but since the large numbers ofphages (1012) required resulted in high backgrounds,detection of small amountsof SEB wasproblematic. Thiswasindeedourobservation.Whensandwich assaysusingananti-SEBantibodyasthecapture, andCy-5 labeled phagetogenerate the signal, differentiation of signal over back-groundwasnot possible dueto therelatively smallamountof SEBthat wasboundin theseassays.Assaysfor SEBonthe microtiter plate were successful, howeverperhaps theimprovedsensitivity of thephagefor SEBon themembranein comparisonto the optical probesis an additive effect ofpeptide–SEB binding andnonspecificbinding forces.

Having shown that the isolatedphagedisplay peptideisselective for SEB and SEC and under certain assayconditions can be usedto sensitively detect SEB down to

Figure 3. Cross-reactivity of three clones with enterotoxins SEA, SEC and SEDshown by ELISA. One hundred microliters of 5� 1011 phage/ml were used ineach sample. Results are reported as the percentage of the phage-SEB signalrecorded on the same ELISA plate. The data represent the mean of triplicatewells.

Figure 4. Comparison of Cy5-labeledphage 14 (3.3� 1010/ml,squares) and Cy5-labeled anti-SEB antibody (20 �g/ml, triangles)binding to SEB adsorbed to biodyne A membrane. The averagesignal minus background for triplicates is shown �SE.

386 E. R. GOLDMAN ET AL.


a concentrationof 1.4ng/well in a fluoroimmunoassay,weplan to examine different constructs that may furtherimprove the utili ty of this SEB binding peptide. In thecurrentphagedisplaysystem,thepeptideis displayedonallfive copies of pIII , so avidity effects likely make asignificant contribution to phage performance. A moreuseful reagentmight employ the peptide fusedwith geneVIII, the major coatprotein of M13 (Ilyichev et al., 1989;Markland et al., 1991), in order to generatephage-basedreagentswith evenhigher avidity, asgeneVIII systems candisplay hundredsof copiesof apeptideperphage.A secondapproach would be to optimize the affinity of the peptideligands by designing and panning secondary libraries(Rozinov et al., 1998; McConnell et al., 1998) in whichthe initial ly selected motifs are conserved and flankingresidues re-randomized. In some systems useof a disulfide-constrained library has yielded peptides which bind withorders of magnitude higher affinity than their non-constrained linear counterparts (Giebel et al., 1995). Aconstrained peptide library could be screenedfor SEBbinding; howevera library of cyclic hexapeptidesfailed to

yield any phagethat boundto TSST-1 (Sato et al., 1996).Higher affinity peptidescould be usedasgeneIII fusions,transferredto geneVIII, or usedon a monovalent(Zwick etal., 1998) or multivalent (Terskikh et al., 1997) proteinscaffold for use as sensing reagents. An advantage ofexpressing the peptide as a fusion with a smaller proteinwould be that the peptidecould be usedat much higherconcentrationsthanpossible on the phage.

The presentresearchdemonstratesthat phagedisplayedpeptides show promise as reagents for biosensorapplica-tions. In optimized formsthese phage-basedreagentscouldpotentially perform at levels comparable to their antibodycounterparts.

Acknowledgements

This work wassupportedby theNavalResearchLaboratory andtheOfficeof Naval Research.The views,opinions,and/orfindingsdescribedin thisreport are thoseof the authorsand should not be construedas officialDepartmentof theNavy positions,policy or decision.

REFERENCES

Anderson GP, King KD, Gaffney KL, Johnson LH. 2000. Multi-analyte integration using the ®ber optic biosensor. Biosens.Bioelectron. 14: 771±778.

Bruno JG, Kiel JL. 1999. In vitro selection of DNA aptamers toanthrax spores with electrochemiluminescence detection.Biosens. Bioelectron. 14: 457±464.

Cesareni G, Castagnoli L, Cestra G. 1999. Phage displayedpeptide libraries. Comb. Chem. High Throughput Screen. 2:1±17.

Giebel LB, Cass RT, Milligan DL, Young DC, Arze R, Johnson CR.1995. Screening of cyclic peptide phage libraries identi®esligands that bind streptavidin with high af®nities. Biochem-istry 34: 15430±15435.

Ilyichev AA, Mienkova OO, Tatkov SI, Karpyshev NN, EroshkinAM, Petrenko VA, Sandakhchiev LS. 1989. Production of aviable variant of phage M13 with incorporated foreignpeptide in the major envelope protein. Dokl. Biochem. (Proc.Acad. Sci. USSR)- Eng. Transl. 307: 196±198.

Kleinjung F, Klussmann S, Erdmann VA, Scheller FW, Furste JP,Bier FF. 1998. High-af®nity RNA as a recognition element in abiosensor. Anal. Chem. 70: 328±331.

Markland W, Roberts BL, Saxena MJ, Guterman SK, Ladner RC.1991. Design, construction and function of a multicopydisplay vector using fusions to the major coat protein ofbacteriophage M13. Gene 109: 13±19.

McConnell SJ, Dinh T, Le MH, Brown SJ, Becherer K, Blumeyer K,Kautzer C, Axelrod F, Spinella DG. 1998. Isolation oferythropoietin receptor agonist peptides using evolvedphage libraries. Biol. Chem. 379: 1279±1286.

Papageorgiou AC, Tranter HS, Acharya KR. 1998. Crystalstructure of microbial superantigen staphylococcal entero-toxin B at 1.4 A resolution: implications for superantigen

recognition by MHC class II molecules and T-cell receptors. J.Mol. Biol. 227: 61±79.

Rodi DJ, Makowski L. 1999. Phage-display technologyÐ®nding aneedle in a vast molecular haystack. Curr. Opin. Biotechnol.10: 87±93.

Rozinov MN, Nolan GP. 1998. Evolution of peptides that modulatethe spectral qualities of bound, small-molecule ¯uorophores.Chem. Biol. 5: 713±728.

Sato A, Ida N, Fukuyama M, Miwa K, Kazami J, Nakamura H.1996. Identi®cation from a phage display library of peptidesthat bind to toxic shock syndrome toxin-1 and that inhibit itsbinding to major histocompatibility complex (MHC) class IImolecules. Biochemistry 35: 10441±10447.

Scott JK, Smith GP. 1990. Searching for peptide ligands with anepitope library. Science 249: 386±390.

Smith GP. 1985. Filamentous fusion phage: novel expressionvectors that display cloned antigens on the virion surface.Science 228: 1315±1317.

Tempelman LA, King KD, Anderson GP, Ligler FS. 1996.Quantitating staphylococcal enterotoxin B in diverse mediausing a portable ®ber-optic biosensor. Anal. Biochem. 233:50±57.

Terskikh AV, LeDoussal J, Crameri R, Fisch I, Mach J, Kajave AV.1997. ªPeptabodyº: a new type of high avidity bindingprotein. Proc. Natl Acad. Sci. USA 94: 1663±1668.

Wadkins RM, Golden JP, Pritsiolas LM, Ligler FS. 1998. Detectionof multiple toxic agents using a planar array immunosensor.Biosens. Bioelectron. 13: 407±415.

Zwick MB, Bonnycastle LL, Noren KA, Venturini S, Leong E,Barbas CF III, Noren CJ, Scott JK. 1998. The maltose-bindingprotein as a scaffold for monovalent display of peptidesderived from phage libraries. Anal. Biochem. 264:87±97.



phage-displayed peptides as biosensor reagents

Documents