Journal of Immunological Methods 321 (2007) 60–69www.elsevier.com/locate/jim

Research paper

Selection and characterization of scFv antibodies against the SinNombre hantavirus nucleocapsid protein

Nileena Velappan a, Jennifer S. Martinez a, Rosa Valero a, Leslie Chasteen a,Liana Ponce b, Virginie Bondu-Hawkins b, Craig Kelly b, Peter Pavlik a,

Brian Hjelle b, Andrew R.M. Bradbury a,⁎

a Los Alamos National Laboratory, TA-43, HRL-1, MS M888, Los Alamos NM 87545, United Statesb University of New Mexico, Department of Pathology, School of Medicine, Albuquerque, NM 87131, United States

Received 16 August 2006; received in revised form 2 December 2006; accepted 7 January 2007Available online 8 February 2007

Abstract

Rodent-borne hantaviruses cause hemorrhagic fever with renal syndrome (HFRS) in the old world and hantavirus cardio-pulmonary syndrome (HCPS) in the new. Most cases of HCPS in North America are caused by Sin Nombre Virus (SNV). Currentviral detection technologies depend upon the identification of anti-viral antibodies in patient serum. Detection of viral antigen mayfacilitate earlier detection of the pathogen. We describe here the characterization of two single-chain Fv antibodies (scFvs), selectedfrom a large naïve phage antibody library, which are capable of identifying the Sin Nombre Virus nucleocapsid protein (SNV-N),with no cross reactivity with the nucleocapsid protein from other hantaviruses. The utility of such selected scFvs was increased bythe creation of an scFv-alkaline phosphatase fusion protein which was able to directly detect virally produced material without theneed for additional reagents.© 2007 Elsevier B.V. All rights reserved.

Keywords: Hantavirus; Sin Nombre Virus; Phage display; scFv; Phage antibody; Nucleocapsid protein

1. Introduction

Hantaviruses are a group of enveloped negative-strand RNA viruses carried by numerous rodent speciesthroughout the world. The hantavirus genome containsthree segments, termed large, medium and small,encoding the polymerase (L segment), two surfaceglycoproteins, G1 and G2, (M segment) and thenucleocapsid protein (S segment). The nucleocapsid

⁎ Corresponding author. Tel.: +1 505 665 0281; fax: +1 505 6653024.

E-mail address: amb@lanl.gov (A.R.M. Bradbury).

0022-1759/$ - see front matter © 2007 Elsevier B.V. All rights reserved.doi:10.1016/j.jim.2007.01.011

protein is responsible for encapsidating genomic RNAsegments into ribonucleoprotein complexes, whichparticipate in genome transcription and replication aswell as in virus assembly.

These viruses cause hemorrhagic fever with renalsyndrome in the Eastern hemisphere and an acuterespiratory disease termed hantavirus cardio-pulmonarysyndrome (HCPS)(Nichol et al., 1993) in the USA, adisease which arose for the first time in 1993 in the FourCorners region (Nichol et al., 1993). Hantaviruscardiopulmonary syndrome is a life threatening zoonoticdisease with a mortality of 50%, as a result of which it isconsidered an NIAID category A bio-threat agent. Since

61N. Velappan et al. / Journal of Immunological Methods 321 (2007) 60–69

the identification of hantaviruses as the causative agentof HCPS, many other novel hantaviruses have beendiscovered. Their ubiquity and the gravity of thediseases they cause make these viruses a serious globalpublic health issue (Schmaljohn and Hjelle, 1997). Mostcases of HCPS in the USA are caused by Sin NombreVirus (SNV), carried asymptomatically by deer mice(Peromyscus maniculatus).

Diagnosis of infection is based on detection of eitherviral components or antibodies produced in the host inresponse to those components. At the time of presen-tation, patients often have strong serum antibodyresponses to the nucleocapsid protein and the G1surface glycoprotein (Jenison et al., 1994; Vapalahtiet al., 1996), as a result of which, the detection ofantibodies recognizing the SNV nucleocapsid protein(SNV-N) has been shown to be both sensitive andspecific for SNV infection (Schmidt et al., 2005), withclinical outcome often related to the levels of theseantibodies (low antibody responses are associated with apoor outcome (Bharadwaj et al., 2000)). Reliablediagnostic assays for anti-SNV antibodies in the serum(Schmidt et al., 2005), with very high sensitivity andspecificity for acute SNV infection (Green et al., 1998;Hjelle, 2002) are available, as is a robust strip blot assayto detect the presence of anti-SNV antibodies in rodentserum (Hjelle et al., 1997), allowing detection ofinfection in rodents under field conditions. However, ittakes two to three weeks after infection to detect anti-SNV serum antibodies, indicating a need to detect SNVcomponents directly, rather than the immune response tothem. Monoclonal antibody based diagnostic kits areavailable to detect viral antigen in other infections, suchas measles (Olding-Stenkvist and Bjorvatn, 1976),CMV (Rawlinson and Scott, 2003), HIV (Iweala,2004), Hepatitis B (Hatzakis et al., 2006) and C (Semeet al., 2005), and some tissue culture isolated viruses.Such kits are increasingly used for diagnostic purposesdue to their relatively low cost, rapid results and abilityto detect infection at early time points (Grandien, 1996).Given the strong response to SNV-N, and the fact thatrecombinant SNV-N prepared in E. coli is effectivelyrecognized by serum antibodies (Jenison et al., 1994),the nucleocapsid protein is a good target for directantigen detection. This is in contrast to the use ofrecombinant G1 surface glycoprotein, which beingglycosylated, is more difficult to prepare in largequantities.

Traditional methods to derive antibodies rely on theimmunization of laboratory animals, with either theharvesting of polyclonal antibodies, or the creation ofmonoclonal antibodies. Phage display is an alternative

in vitro method to develop antibodies which relies onthe creation of large phage antibody libraries, fromwhich monoclonal antibodies binding to a target ofinterest can be selected. In general two kinds of librarycan be used: naïve or immune. Naïve libraries aregenerally derived from natural unimmunized humanrearranged V genes (Marks et al., 1991; Vaughan et al.,1996; Sheets et al., 1998; de Haard et al., 1999; Sblatteroand Bradbury, 2000) or synthetic human V genes(Griffiths et al., 1994; Nissim et al., 1994; de Kruif et al.,1995; Knappik et al., 2000), while immune libraries arecreated from the V genes of immunized humans (Burtonet al., 1991; Zebedee et al., 1992; Williamson et al.,1993; Amersdorfer and Marks, 2000; Amersdorferet al., 2002) or mice (Orum et al., 1993; Ames et al.,1994, 1995b). Immune libraries have the advantage thatthey are composed of antibodies with a strong biastowards the antigen/organism used for immunization,resulting in higher affinities, but the disadvantage thatnew libraries need to be made for each immunogen. Asantibody genes are cloned simultaneously with selectionin these in vitro systems, antibody fragments can besubjected to downstream genetic engineering to increaseaffinity (Schier and Marks, 1996; Daugherty et al.,1998; Boder et al., 2000; Hanes et al., 2000; Coia et al.,2001) and multimericity (de Kruif and Logtenberg,1996; Krebber et al., 1997; Kortt et al., 2001; Zhanget al., 2004), as well as being linked to desired effecterfunctions for downstream applications (Griep et al.,1999; Muller et al., 1999; Casey et al., 2000; Hink et al.,2000; Han et al., 2004), including recloning into the fulllength immunoglobulin format (Ames et al., 1995a;Persic et al., 1997; Nowakowski et al., 2002). Recently,both naïve and immune phage antibody libraries havebeen used to select specific high-affinity humanantibodies against a number of infectious disease agentsand toxins, including botulinum toxin (Amersdorferet al., 2002), crotoxin (Cardoso et al., 2000), HIV(Burton et al., 1991; Barbas et al., 1994; Moulard et al.,2002; Zhang et al., 2003), herpes virus (Sanna et al.,1995), SARS (van den Brink et al., 2005), rabies(Kramer et al., 2005) and hepatitis B (Zebedee et al.,1992; Park et al., 2005), with the potential for bothdiagnosis and therapy.

In the work described here, we have expanded theuse of large naïve phage antibody libraries to selectsingle chain Fvs (scFvs) against recombinant SNV-Nprotein. A number of different antibodies were isolated,some of which were shown to be specific for the SNVnucleocapsid protein, and not recognizing a panel ofrelated hantavirus nucleocapsid proteins. This confirmsthe utility of these in vitro methods in the study of

Fig. 1. Expression and purification of recombinant SNV-N. SNV-Ncloned into a pET23 based vector was expressed in E. coli BL21 DE3by induction with IPTG. After expression, the protein was purified byimmobilized metal affinity chromatography. The samples were M)marker/standard 1) sample before purification 2) flow through from thecolumn 3) Wash I 4) Wash II 5) Wash III 6) protein elution 7) stripbuffer 8) SNV-N after dialysis.

62 N. Velappan et al. / Journal of Immunological Methods 321 (2007) 60–69

infectious disease, and indicates their potential forbroader use.

2. Materials and methods

2.1. Plasmids and bacteria

pET vectors were obtained from Novagen, and thealkaline phosphatase gene from A. Schots (Griep et al.,1999). DH5αF′ E. coli (Gibco BRL, Rockville, MD): F′/endA1 hsdR17 (rK

− mK+) supE44 thi-1 recA1 gyrA (Nalr)

relA1 Δ (lacZYA-argF)U169 deoR (ϕ80dlacΔ(lacZ)M15); BL21 DE3 E. coli (Novagen): E. coli B F− dcmompT hsdS(rB

− mB−) gal λ(DE3).

2.2. Constructs made

After selection, scFvs 3A, B2, C3 and F4 weresubcloned from the pDAN5 phage display vector intopET-28 based protein expression vectors which containeda pelB based leader sequence upstream and either thealkaline phosphatase (Griep et al., 1999) genewith theSV5and His tags downstream, or only the SV5 and His tags.

2.3. Antigen preparation

The SNV-N coding sequence was cloned into pET-23(Novagen) and transformed into BL21 DE3 cells.Transformed cells were grown at 37 °C in LB ampicillin(50 μg/ml), and protein expression induced with 0.4 mMIPTG at OD600 0.6–0.8. Bacterial cells were pelleted bycentrifugation, resuspended in the binding buffer provid-ed in the His-Bind purification kit (Novagen) andsonicated to release inclusion bodies. The centrifugationand sonication was repeated twice more, and the finalinclusion body pellet was resuspended in 5 ml of bindingbuffer containing 6 M urea for solubilization. After 1 hincubation on ice, the cell debris was pelleted bycentrifugation at 39,000 g for 20 min and the recombinantHis-tagged SNV-N purified using the His-Bind purifica-tion kit (Novagen) according to the manufacturer'sinstructions. All washing and elution buffers contained6 M urea to maintain SNV-N in the denatured solubilizedstate. The different elution fractions were analyzed bypolyacrylamide gel electrophoresis and coomassie stainedto assess the quantity and quality of SNV-N in eachfraction. The best fractions were combined and dialyzedovernight into 1× phosphate buffered saline (PBS) at 4 °Cto remove the urea and refold the protein. Quantificationwas carried out by comparison to bovine serum albumin(BSA) standards. Antigen thus prepared (Fig. 1) was usedin all subsequent analyses.

2.4. Phage display selection of scFv antibodies

MaxiSorp® immunotubes (Nunc) were coated with4 ml SNV-N (10 μg/4 ml) at 4 °C overnight, washedtwice with PBS and blocked with 4% milk PBS(MPBS). In the meantime the phage scFv library(Sblattero and Bradbury, 2000) was blocked byincubation with 2% MPBS for 30 min prior to addingto the antigen-coated immunotubes. After incubation for2 h at room temperature, unbound phage were removedby washing once with PBS+Tween (0.1%) (PBST) andtwice with PBS. Bound phage were eluted with 0.1 MHCl for 4 min and neutralized with 1 M Tris–HCl pH7.4. Eluted phage were infected into E. coli DH5αF′,plated onto 2×YT carbenicillin (50 μg/ml) agar platesand incubated at 30 °C overnight. The following daybacteria were scraped from the plate, resuspended in1 ml 2×YT and inoculated into 50 ml of 2×YTcontaining 100 μg/ml ampicillin and 3% (w/v) glucose(amp/glu 2×YT) and grown to 0.5 OD600 at 37 °C, priorto infection with helper phage (M13K07). Bacteria werepelleted by centrifugation and resuspended in 50 ml of2×YT media containing 50 μg/ml ampicillin and100 μg/ml kanamycin (kan 2×YT) and allowed toproduce phage at 30 °C overnight. These were used forthe next round of selection after PEG precipitation. Thesecond and third rounds of panning were identical to thefirst except that washing was more stringent (5× PBSTand 5× PBS). Single colonies, containing monoclonalscFvs, were picked after the second and third rounds ofselection and analyzed by phage ELISA for their ability

Fig. 2. Analyses of SNV-N binding clones. a: scFv ELISA signals, withbackground subtracted, for each of the ten scFvs chosen for furtheranalysis. b: DNA fingerprints for the different scFvs. c: showscomparisons between the phage and scFv ELISA signals for the 4 scFv(3A, B2, F4, C3) chosen for more detailed study. Alcohol dehydro-genase (ADH) is an antigen used to assess non-specific binding.

63N. Velappan et al. / Journal of Immunological Methods 321 (2007) 60–69

to recognize antigen. scFv genes from positive cloneswere analyzed by PCR, using pDPH5′ (GCA GCC GCTGGA TTG TTA TTA) and pDpH 3′ (TTG TCG TCTTTC CAG ACG TTA) primers followed by BstN1restriction enzyme fingerprinting.

2.5. Phage ELISA

Single colonies picked after two or three rounds ofselection were grown overnight at 30 °C in 96 well U-bottom plates (Nunc) in 100 μl of amp/glu 2×YT media.The overnight culturewas diluted 1/10 into 100μl of amp/glu 2×YT and grown for 2 h at 37 °C. The culture wasinfected with helper phage (M13K07), at a multiplicity ofinfection of 20:1, for 30min at 37 °C, the bacteria pelletedby centrifugation and resuspended in 150 μl of amp/kan2×YT prior to growth overnight at 30 °C to producephage. The overnight culture was centrifuged and thesupernatant used as the source of monoclonal phages.MaxiSorp® plates (Nunc) were used for all ELISAs.Antigenwas coated by incubating 1 μg SNV-N per well at4 °C overnight or at 37 °C for 1 h. The antigen bound platewas washed with PBS and blocked with 200 μl of 4.5%fish gelatin for 1 h at room temperature. The blocked platewas washed with PBS and incubated with 75 μl of phagesupernatant and 25 μl 4.5% fish gelatin for 1.5 h at roomtemperature. Plates were washed with PBST (three times)and PBS (three times). Bound phage were revealed usingan anti-M13 horse radish peroxidase labeled antibody(Amersham Pharmacia) detected with TMB (Sigma). Thereaction was terminated with 1 M sulfuric acid andabsorbancewas read at 450 nm using a Spectrafluor plus®spectrophotometer (Tecan US, Inc.).

2.6. scFv ELISA

scFv ELISAs were carried out in a fashion similar tophage ELISA. Soluble scFvs were produced by inductionof bacteria with IPTG (250 μM) and growth overnight at25 °C in 2×YT.Under these conditions, scFvs are releasedinto the growth medium where they can be used fordetection. Bound scFvs were detected using the SV5monoclonal antibody, which recognizes the C-terminalSV5 tag (Hanke et al., 1992), and horse radish peroxidaselabeled anti-mouse antibody (Dako Corp). Washesbetween incubations with PBST and PBS were carriedout as described above. ELISAs with scFv-alkalinephosphatase (scFv-AP) fusions were carried out in asimilar fashion except that there was no need forsecondary antibodies after the incubation with scFv-AP.The alkaline phosphatase activity was detected using thePhosphatase substrate kit (Pierce).

2.7. scFv expression and purification

scFvs 3A, B2, C3 and F4 in the pET28 vector wereexpressed in 250 ml kan/glu 2×YT grown to 0.9–1OD600. Bacteria were pelleted, resuspended in 250 mlkan/IPTG (250 μM) 2×YTand incubated at 30 °C for 3 hor 18 °C overnight for protein production. Bacterial cellsexpressing scFvs were then pelleted at 9000 g for 10 minand stored at 4 °C. scFv purification from the periplasm,by osmotic shock, was carried out by incubating the cellpellet in 50 ml PPB buffer [50 mM Tris–HCl pH 7.4,1 mMEDTA, 20% sucrose] for 20min on ice followed by

64 N. Velappan et al. / Journal of Immunological Methods 321 (2007) 60–69

centrifugation at 19,000 g for 5 min. The supernatant wascollected and the cell pellet resuspended in 5 mMMgSO4

for 10 min at room temperature. The bacterial cells wereagain centrifuged at 6000 g for 6 min. The supernatantsfrom both osmotic shocks were combined and filteredbefore purification using the C-terminal His-tag on an Ni-NTA agarose (Qiagen) column using the Biorad biologicpurification system according to the manufacturer'sinstructions. Purified scFvs were dialyzed into 1× PBS,and 30% glycerol and 0.1% tween-20was added for long-term storage at −20 °C. Quantification was carried out bycomparing the intensity of purified SNV-N with titratedBSA standards of known concentrations.

2.8. Western and dot blotting analysis

Antigens were separated by polyacrylamide gelelectrophoresis using a 4–12% gradient Novex acryl-amide gel (Invitrogen), and electro-transferred ontonitrocellulose using a semidry electroblotter. Prior to

Fig. 3. Determination of sensitivity by western blotting. a: The sensitivity ofwestern blot containing a titration of purified SNV-N. Two hundred microgramponce staining of the transferred proteins, while those below show antigen dwestern blotting. The ability of the scFvs to recognize viral antigen was tesinfected cells. The recombinant SNV-N was used as a positive control. Theantiserum, followed by anti-rabbit AP. In all blots, the recombinant SNV-N iSNV-N is indicated with the long arrow. The gels are cropped to show all p

analysis, the blot was blocked using MBT blockingsolution (4% milk, 4% BSA, 0.25% tween-20) for atleast 30 min. Three–five micrograms of scFv wasdiluted in 10 ml of 2% fish gelatin and incubated withthe transferred blot for 1.5 h. The blot was then washedfor 10 min with PBST followed by 10 min with PBS.Similar washings were carried out after incubation withthe SV5 monoclonal and alkaline phosphatase labeledanti-mouse (Dako) antibodies. Alkaline phosphataseactivity was detected using NBT/BCIP (Pierce). Anti-SNV-N rabbit polyclonal and alkaline phosphataselabeled anti-rabbit (Cell Signaling) antibodies wereused as positive controls when detecting viral SNV-N.In the case of scFvs B2 and F4 the alkaline phosphataseactivity was detected using NBT/BCIP substrate, whilefor scFv 3A and the positive control the AP activity wasdetermined using lumiphos® WB substrate. For dotblots, 300 ng of recombinant viral nucleocapsid proteins(from Puumala, Prospect hill, Rio Mamore and Seoulviruses) were spotted onto a nitrocellulose membrane.

the individual scFvs to detect recombinant SNV-N was analyzed on as of lysozyme was used as negative control (NC). The top panel showsetection by each of the scFvs. b: Detection of natural viral antigen byted by western blot analysis on vero E6 lysate from infected and non-positive control experiment was carried out using a rabbit polyclonals indicated with the short arrow, while the expected size of the naturalroteins recognized by the different antibodies.

Fig. 4. Analyses of functionality of scFv-alkaline phosphatase fusionproteins. a: shows the AP activity and the SNV-N detection ability ofthe four scFvs in an ELISA format, with green fluorescent proteinfused to alkaline phosphatase (GFP-AP), and lysozyme (lys) used asnegative controls. In b:, the ability of scFv B2-AP to recognize relatednucleocapsid proteins was tested against 300 ng of the differentproteins spotted on a nitrocellulose filter.

65N. Velappan et al. / Journal of Immunological Methods 321 (2007) 60–69

The membrane was blocked with 4.5% fish gelatin for1 h and incubated with 1 μg purified scFv-AP fusion for1.5 h. The membrane was washed with PBST and PBS,and the alkaline phosphatase signal detected using NBT/BCIP.

2.9. Vero E6 cell lysate preparation

In order to test antibody reactivity on natural SNV-N,Vero E6 cells were infected with SNV and extractsprepared. Cells were resuspended at 25,000 cells/μl inlysis buffer [1% Triton X-100, 0.5 M Tris–HCl, 0.1 MEDTA, 2%glycerol, 10%B-mercaptoethanol, 2% sodiumdodecyl sulfate (SDS), pH 6.8] with a protease inhibitorcocktail (Boehringer Mannheim), transferred to beadbeater tubes filled to 1/3 volume with 2.5 mm zirconia/silica beads (Biospec Products), and homogenized in abead beater at maximum speed for 15 s. The homogenateswere then transferred to 1.5 ml eppendorf tubes andheated to 100 °C for 5 min prior to centrifugation at16,000 g for 15 min at 4 °C. The supernatants weretransferred to new tubes, flash frozen in a dry ice/ethanolbath, and stored at −70 °C. Non-infected cells were usedas negative controls.

2.10. scFv-alkaline phosphatase (AP) fusion protein

scFv-AP fusions were expressed in E. coli BL21 DE3using 250 ml autoinduction media (Studier, 2005)containing kanamycin grown at 18 °C for 2 days.Periplasmic extraction and purification of the scFv-APwas carried out as described above.

Surface plasmon resonance analysis of scFvs:purified recombinant SNV-N or anti-SNV-N scFvclones, prepared as described above, were coupled to aBiacore CM-5 surfaces, following the manufacturer'sinstructions, to give RU values between 200 and 1000.Purified scFvs or SNV-N, respectively, were then passedover the coupled chips. Although binding curves wereobtained, the multivalent nature of SNV-N (Alfadhliet al., 2001, 2002; Kaukinen et al., 2004; Alminaiteet al., 2006) made it difficult to accurately determinebinding affinities for the scFv. However, from differ-ences in off rates, we determined the relative bindingaffinities between the clones.

3. Results

3.1. Phage display selection

Recombinant SNV-N was expressed from a pET23based vector and purified using the His6 tag, as

described in the Materials and methods, and shown inFig. 1. Three rounds of phage display panning againstSNV-N were carried out and ∼400 single colonies wereassayed for antigen recognition by ELISA. One hundredand five ELISA positive scFv phage were obtained, ofwhich twenty-five clones were found to be unique byDNA fingerprinting. These were retested for antigenreactivity as soluble scFvs and ten different scFvs wereidentified as recognizing the nucleocapsid protein.Fig. 2a shows the ELISA signals for these ten clones,while their DNA restriction fingerprint patterns areshown in Fig. 2b. Four of the scFvs (3A, B2, C3, andF4) which gave consistently strong ELISA signals andproduced high scFv levels, were selected for furthercharacterization. The ELISA values for these clones asboth phage and soluble scFv are shown in Fig. 2c.

66 N. Velappan et al. / Journal of Immunological Methods 321 (2007) 60–69

3.2. scFv sensitivity determination and detection ofnative SNV-N

scFvswere expressed and purified using osmotic shockand the His6 tag, as described in the Materials andmethods. The scFvs were tested for their ability torecognize the recombinant antigen in a Western blot inwhich antigen was titrated downwards from 200 to 0.8 ng(Fig. 3a) and 3–5 μg of each purified scFv used fordetection. Detectionwas defined as the lowest amount thatprovided a clearly visible signal stronger than background.scFv 3A had a detection limit of 7.3 ng antigen, B2 and F4had detection limits of 66 ng, whereas C3 only recognized1 μg of antigen (data not shown). Although these scFvswere able to recognize the recombinant antigen, this didnot indicate they could recognize the natural viral antigen,which was tested in a similar fashion by Western blottingof denatured vero-E6 cell lysate derived from hantavirusinfected and non-infected cells. scFvs B2 and F4 clearlyshowed recognition of both recombinant SNV-N andnative SNV-N, while 3A only showed recognition of therecombinant SNV-N, indicating that its recognition motifwas probably composed of peptides encoded by thevector. scFv B2 recognized additional faint bands in boththe infected and non-infected lysates, while F4 identified asingle band. The rabbit polyclonal positive controlrecognized additional bands in the infected lysate,probably representing SNV-N degradation products(Fig. 3b).

3.3. Evaluation of cross-reactivity with other hantavirusspecies

In order to simplify antigen detection, the scFvs werefused to alkaline phosphatase. This multifunctionalfusion (Griep et al., 1999; Muller et al., 1999; Han et al.,2004) has been shown to stabilize and dimerize the scFv(alkaline phosphatase is a dimer), as well as provide aneasily assayed enzymatic activity allowing the scFv-alkaline phosphatase fusion to be used directly, withoutthe need for secondary reagents. The activities of thesedifferent fusions, as well as their ability to recognizeSNV-N or a negative control (lys) are shown in Fig. 4a.GFP fused to alkaline phosphatase (GFP-AP) was usedas a negative control. scFvs 3A and B2 showedrecognition of recombinant SNV-N, while scFv F4 didnot, indicating that not all scFvs are functional in the APfusion format. scFv-B2 which recognizes viral SNV-Nwas further characterized, and its sensitivity found to besimilar whether it was fused to alkaline phosphatase ornot. B2-AP was tested for its ability to recognize SNV-Nin a dot blot, under which conditions, the protein is not

denatured. The positive signal (Fig. 4b) indicates thatB2-AP is capable of recognizing both the denatured andnative forms of the antigen. Similar testing of B2-APagainst a panel of additional nucleocapsid proteins fromPuumala virus (PUU), Prospect hill virus (PHV), RioMamore virus (RMV) and Seoul virus (SEO), showedno cross-reactivity, indicating that this antibody isspecific for the Sin Nombre Virus nucleocapsid protein.

4. Discussion

Currently diagnosis of SNV infection relies on thedetection of anti-SNVantibodies (Schmidt et al., 2005) inpatient serum.However, this requires time for the immuneresponse to develop, and could be improved, orcomplemented, by direct antigen detection. Althoughboth polyclonal (Green et al., 1998; Hjelle, 2002) andmonoclonal (Zvirbliene et al., 2006) (and www.austral-biologicals.com) antibodies have been developed againstSNV nucleocapsid protein, where testing has beendescribed, these antibodies also tend to react with thenucleocapsid proteins of other species, complicatingidentification. Using the phage display method describedhere, we were able to select two different scFvs thatclearly recognize not only the recombinant but also thenative nucleocapsid protein in the presence of cell extractas background. One of the scFvs, B2, was also fullyfunctional as an AP fusion protein, which allowed it to beused in a one step assay in both ELISAs and western/dotblots. This scFv did not cross react with the other testednucleocapsid proteins derived from Puumala, Prospecthill, Rio Mamore or Seoul viruses.

This use of a large naïve library to select antibodiesrecognizing products of an infectious agent is similar tothe use others have made of naïve libraries to select scFvsagainst different pathogens (Sanna et al., 1995; Cardosoet al., 2000; Amersdorfer et al., 2002; Park et al., 2005;van den Brink et al., 2005), with the advantage that only asingle large library needs to be made. Although immunelibraries often provide higher affinity neutralizing anti-bodies (Zebedee et al., 1992; Amersdorfer et al., 2002;Moulard et al., 2002; Zhang et al., 2003; Kramer et al.,2005), the affinities of antibodies selected from naïvelibraries can usually be increased by affinity maturation.Furthermore, it is frequently difficult to obtain lympho-cytes from infected patients.

The nucleocapsid protein binds to membranes, viralRNAs and associates with transcription and replicationcomplexes, oligomerizing during the process of virusassembly (Alfadhli et al., 2001, 2002). This oligomeriza-tion also occurs when the SNV-N protein is producedrecombinantly in either bacterial, mammalian or yeast cells

67N. Velappan et al. / Journal of Immunological Methods 321 (2007) 60–69

(Alfadhli et al., 2001, 2002; Kaukinen et al., 2004;Alminaite et al., 2006). Although this precluded an accuratedetermination of affinity using surface plasmon resonance,we were able to obtain an off-rate ranking of3AbB2bF4bC3, which reflected the signals obtained inboth ELISA and western blots. The off-rate analysissuggests that of the selected scFvs, 3A and B2 have thehighest affinity, since off rate is usually the most variablecomponent of affinity. The oligomeric nature of the antigenalso made it an unsuitable target for yeast display basedaffinity determination and affinity maturation (Boder et al.,2000; VanAntwerp and Wittrup, 2000).

The different reactivity patterns given by the differentscFvs in Fig. 3b suggests that they all recognize differentepitopes, and the fact that they all recognize denaturedrecombinant SNV-N further suggests that these epitopesmay be linear. However, the oligomeric nature of theantigen prevented confirmation of this and further epitopecharacterization will require additional experiments withpeptides or truncated SNV-N, which does not oligomerize(Yoshimatsu et al., 2003; Kaukinen et al., 2004). One ofthe scFvs selected (B2) showed exquisite specificity forSNV-N, not recognizing the nucleocapsid protein fromfour other hantaviruses (Fig. 4b). A sequence comparisonof these five nucleocapsid proteins, revealed that theyonly share 52% amino acid homology. Of these, the SNVnucleocapsid sequence is closest to the Rio Mamorenucleocapsid, with which it shares 85% homology.However, most of the differences are scattered throughoutthe sequence, with the exception of one region of 65amino acids (234–301) where the homology is reduced to59% between these two sequences, and 12% for all fivesequences, suggesting that B2 probably recognizes alinear epitope localized to this area.

The conversion of scFvs to full length IgGs (Persicet al., 1997) is known to improve stability, avidity andsensitivity in detection assays. Further modification ofthis antibody to the IgG format will render it more usefulfor direct antigen detection in patient samples, as has beencarried out with other infectious agents (Grandien, 1996).

Acknowledgments

This research was funded by the LDRD-DR programat Los Alamos National Laboratory. We would like tothank Ms. Priya Dighe for her technical assistance.

References

Alfadhli, A., Love, Z., Arvidson, B., Seeds, J., Willey, J., Barklis, E.,2001. Hantavirus nucleocapsid protein oligomerization. J. Virol.75, 2019.

Alfadhli, A., Steel, E., Finlay, L., Bachinger, H.P., Barklis, E., 2002.Hantavirus nucleocapsid protein coiled-coil domains. J. Biol.Chem. 277, 27103.

Alminaite, A., Halttunen, V., Kumar, V., Vaheri, A., Holm, L., Plyusnin,A., 2006. Oligomerization of hantavirus nucleocapsid protein:analysis of the N-terminal coiled-coil domain. J. Virol. 80, 9073.

Amersdorfer, P., Marks, J.D., 2000. Phage libraries for generation ofanti-botulinum scFv antibodies. Methods Mol. Biol. 145, 219.

Amersdorfer, P., Wong, C., Smith, T., Chen, S., Deshpande, S.,Sheridan, R., Marks, J.D., 2002. Genetic and immunologicalcomparison of anti-botulinum type A antibodies from immune andnon-immune human phage libraries. Vaccine 20, 1640.

Ames, R.S., Tornetta, M.A., Jones, C.S., Tsui, P., 1994. Isolation ofneutralizing anti-C5a monoclonal antibodies from a filamentousphage monovalent Fab display library. J. Immunol. 152, 4572.

Ames, R.S., Tornetta, M.A., Deen, K., Jones, C.S., Swift, A.M.,Ganguly, S., 1995a. Conversion of murine Fabs isolated from acombinatorial Phage display library to full length immunoglobu-lins. J. Immunol. Methods 184, 177.

Ames, R.S., Tornetta, M.A., McMillan, L.J., Kaiser, K.F., Holmes, S.D.,Appelbaum, E., Cusimano, D.M., Theisen, T.W., Gross, M.S., Jones,C.S., et al., 1995b. Neutralizing murine monoclonal antibodies tohuman IL-5 isolated from hybridomas and a filamentous phage Fabdisplay library. J. Immunol. 154, 6355.

Barbas, C.F.R., Hu, D., Dunlop, N., Sawyer, L., Cababa, D., Hendry,R.M., Nara, P.L., Burton, D.R., 1994. In vitro evolution of aneutralizing human antibody to human immunodeficiency virustype 1 to enhance affinity and broaden strain cross-reactivity. Proc.Natl. Acad. Sci. U. S. A. 91, 3809.

Bharadwaj, M., Nofchissey, R., Goade, D., Koster, F., Hjelle, B., 2000.Humoral immune responses in the hantavirus cardiopulmonarysyndrome. J. Infect. Dis. 182, 43.

Boder, E.T., Midelfort, K.S., Wittrup, K.D., 2000. Directed evolutionof antibody fragments with monovalent femtomolar antigen-binding affinity. Proc. Natl. Acad. Sci. U. S. A. 97, 10701.

Burton, D.R., Barbas, C.F., Persson,M.A.A., Koenig, S., Chanock, R.M.,Lerner, R.A., 1991. A large array of human monoclonal antibodies totype 1 human immunodeficiency virus from combinatorial libraries ofasymptomatic individuals. Proc. Natl. Acad. Sci. U. S. A. 88, 10134.

Cardoso,D.F.,Nato, F., England, P., Ferreira,M.L.,Vaughan, T.J.,Mota, I.,Mazie, J.C., Choumet, V., Lafaye, P., 2000. Neutralizing human anticrotoxin scFv isolated from a nonimmunized phage library. Scand.J. Immunol. 51, 337.

Casey, J.L., Coley, A.M., Tilley, L.M., Foley, M., 2000. Greenfluorescent antibodies: novel in vitro tools. Protein Eng. 13, 445.

Coia, G., Hudson, P.J., Irving, R.A., 2001. Protein affinity maturationin vivo using E. coli mutator cells. J. Immunol. Methods 251, 187.

Daugherty, P.S., Chen, G., Olsen, M.J., Iverson, B.L., Georgiou, G.,1998. Antibody affinity maturation using bacterial surface display.Protein Eng. 11, 825.

de Haard, H.J., van Neer, N., Reurs, A., Hufton, S.E., Roovers, R.C.,Henderikx, P., de Bruine, A.P., Arends, J.W., Hoogenboom, H.R.,1999. A large non-immunized human Fab fragment phage librarythat permits rapid isolation and kinetic analysis of high affinityantibodies. J. Biol. Chem. 274, 18218.

de Kruif, J., Logtenberg, T., 1996. Leucine zipper dimerized bivalentand bispecific scFv antibodies from a semi-synthetic antibodyphage display library. J. Biol. Chem. 271, 7630.

de Kruif, J., Boel, E., Logtenberg, T., 1995. Selection and applicationof human single chain Fv antibody fragments from a semi-synthetic phage antibody display library with designed CDR3regions. J. Mol. Biol. 248, 97.

68 N. Velappan et al. / Journal of Immunological Methods 321 (2007) 60–69

Grandien, M., 1996. Viral diagnosis by antigen detection techniques.Clin. Diagn. Virol. 5, 81.

Green, W., Feddersen, R., Yousef, O., Behr, M., Smith, K., Nestler, J.,Jenison, S., Yamada, T., Hjelle, B., 1998. Tissue distribution ofhantavirus antigen in naturally infected humans and deer mice.J. Infect. Dis. 177, 1696.

Griep, R.A., vanTwisk, C., Kerschbaumer, R.J., Harper, K., Torrance, L.,Himmler, G., van der Wolf, J.M., Schots, A., 1999. pSKAP/S: anexpression vector for the production of single-chain Fv alkalinephosphatase fusion proteins. Protein Expr. Purif. 16, 63.

Griffiths, A.D., Williams, S.C., Hartley, O., Tomlinson, I.M., Water-house, P., Crosby,W.L., Kontermann, R.E., Jones, P.T., Low, N.M.,Alison, T.J., Prospero, T.D., Hoogenboom, H.R., Nissim, A., Cox,J.P.L., Harrison, J.L., Zaccolo, M., Gherardi, E., Winter, G., 1994.Isolation of high affinity human antibodies directly from largesynthetic repertoires. EMBO J. 13, 3245.

Han, Z., Karatan, E., Scholle, M.D., McCafferty, J., Kay, B.K., 2004.Accelerated screening of phage-display output with alkalinephosphatase fusions. Comb. Chem. High Throughput Screen 7,55.

Hanes, J., Schaffitzel, C., Knappik, A., Pluckthun, A., 2000. Picomolaraffinity antibodies from a fully synthetic naive library selected andevolved by ribosome display. Nat. Biotechnol. 18, 1287.

Hanke, T., Szawlowski, P., Randall, R.E., 1992. Construction ofsolid matrix-antibody–antigen complexes containing simianimmunodeficiency virus p27 using tag-specific monoclonal anti-body and tag-linked antigen. J. Gen. Virol. 73, 653.

Hatzakis, A., Magiorkinis, E., Haida, C., 2006. HBV virologicalassessment. J. Hepatol. 44, S71.

Hink, M.A., Griep, R.A., Borst, J.W., van Hoek, A., Eppink, M.H.,Schots, A., Visser, A.J., 2000. Structural dynamics of greenfluorescent protein alone and fused with a single chain Fv protein.J. Biol. Chem. 275, 17556.

Hjelle, B., 2002.Vaccines against hantaviruses. Expert Rev. Vaccines 1,373.

Hjelle, B., Jenison, S., Torrez-Martinez, N., Herring, B., Quan, S.,Polito, A., Pichuantes, S., Yamada, T., Morris, C., Elgh, F., Lee,H.W., Artsob, H., Dinello, R., 1997. Rapid and specific detectionof Sin Nombre virus antibodies in patients with hantaviruspulmonary syndrome by a strip immunoblot assay suitable forfield diagnosis. J. Clin. Microbiol. 35, 600.

Iweala, O.I., 2004. HIV diagnostic tests: an overview. Contraception70, 141.

Jenison, S., Yamada, T., Morris, C., Anderson, B., Torrez-Martinez,N., Keller, N., Hjelle, B., 1994. Characterization of humanantibody responses to four corners hantavirus infections amongpatients with hantavirus pulmonary syndrome. J. Virol. 68, 3000.

Kaukinen, P., Kumar, V., Tulimaki, K., Engelhardt, P., Vaheri, A.,Plyusnin, A., 2004. Oligomerization of Hantavirus N protein:C-terminal alpha-helices interact to form a shared hydrophobicspace. J. Virol. 78, 13669.

Knappik, A., Ge, L., Honegger, A., Pack, P., Fischer, M., Wellnhofer,G., Hoess, A., Wolle, J., Pluckthun, A., Virnekas, B., 2000. Fullysynthetic human combinatorial antibody libraries (HuCAL) basedon modular consensus frameworks and CDRs randomized withtrinucleotides. J. Mol. Biol. 296, 57.

Kortt, A.A., Dolezal, O., Power, B.E., Hudson, P.J., 2001. Dimeric andtrimeric antibodies: high avidity scFvs for cancer targeting.Biomol. Eng. 18, 95.

Kramer, R.A., Marissen, W.E., Goudsmit, J., Visser, T.J., Clijsters-Vander Horst, M., Bakker, A.Q., de Jong, M., Jongeneelen, M.,Thijsse, S., Backus, H.H., Rice, A.B., Weldon, W.C., Rupprecht,

C.E., Dietzschold, B., Bakker, A.B., de Kruif, J., 2005. The humanantibody repertoire specific for rabies virus glycoprotein asselected from immune libraries. Eur. J. Immunol. 35, 2131.

Krebber, A., Bornhauser, S., Burmester, J., Honeggar, A., Willuda, J., H.R.,B., Pluckthun, A., 1997. Reliable cloning of functional antibodyvariable domains from hybridomas and spleen cell repertoiresemploying a reengineered phage display system. J. Immunol.Methods 201, 35.

Marks, J.D., Hoogenboom, H.R., Bonnert, T.P., McCafferty, J., Griffiths,A.D., Winter, G., 1991. By-passing immunization. Human anti-bodies from V-gene libraries displayed on phage. J. Mol. Biol. 222,581.

Moulard,M., Phogat, S.K., Shu, Y., Labrijn, A.F., Xiao, X., Binley, J.M.,Zhang, M.Y., Sidorov, I.A., Broder, C.C., Robinson, J., Parren, P.W.,Burton, D.R., Dimitrov, D.S., 2002. Broadly cross-reactive HIV-1-neutralizing human monoclonal Fab selected for binding to gp120–CD4–CCR5 complexes. Proc. Natl. Acad. Sci. U. S. A. 99, 6913.

Muller, B.H., Chevrier, D., Boulain, J.C., Guesdon, J.L., 1999.Recombinant single-chain Fv antibody fragment-alkaline phos-phatase conjugate for one-step immunodetection in molecularhybridization. J. Immunol. Methods 227, 177.

Nichol, S.T., Spiropoulou, C.F.,Morzunov, S., Rollin, P.E., Ksiazek, T.G.,Feldmann, H., Sanchez, A., Childs, J., Zaki, S., Peters, C.J., 1993.Genetic identification of a hantavirus associated with an outbreak ofacute respiratory illness. Science 262, 914.

Nissim, A., Hoogenboom,H.R., Tomlinson, I.M., Flynn, G.,Midgley, C.,Lane, D., Winter, G., 1994. Antibody fragments from a 'single pot'phage display library as immunochemical reagents. EMBOJ. 13, 692.

Nowakowski, A.,Wang, C., Powers, D.B., Amersdorfer, P., Smith, T.J.,Montgomery, V.A., Sheridan, R., Blake, R., Smith, L.A.,Marks, J.D.,2002. Potent neutralization of botulinum neurotoxin by recombinantoligoclonal antibody. Proc. Natl. Acad. Sci. U. S. A. 99, 11346.

Olding-Stenkvist, E., Bjorvatn, B., 1976. Rapid detection of measlesvirus in skin rashes by immunofluorescence. J. Infect. Dis. 134,463.

Orum, H., Andersen, P.S., Oster, A., Johansen, L.K., Riise, E.,Bjornvad, M., Svendsen, I., Engberg, J., 1993. Efficient method forconstructing comprehensive murine Fab antibody libraries dis-played on phage. Nucleic Acids Res. 21, 4491.

Park, S.G., Jeong, Y.J., Lee, Y.Y., Kim, I.J., Seo, S.K., Kim, E.J.,Jung, H.C., Pan, J.G., Park, S.J., Lee, Y.J., Kim, I.S., Choi, I.H.,2005. Hepatitis B virus-neutralizing anti-pre-S1 human antibodyfragments from large naive antibody phage library. Antivir. Res.68, 109.

Persic, L., Roberts, A., Wilton, J., Cattaneo, A., Bradbury, A.,Hoogenboom, H., 1997. An integrated vector system for theeukaryotic expression of antibodies or their fragments afterselection from phage display libraries. Gene 187, 9.

Rawlinson, W., Scott, G., 2003. Cytomegalovirus. A common viruscausing serious disease. Aust. Fam. Physician 32, 789.

Sanna, P.P.,Williamson, R.A., De Logu, A., Bloom, F.E., Burton, D.R.,1995. Directed selection of recombinant human monoclonalantibodies to herpes simplex virus glycoproteins from phagedisplay libraries. Proc. Natl. Acad. Sci. U. S. A. 92, 6439.

Sblattero, D., Bradbury, A., 2000. Exploiting recombination in singlebacteria to make large phage antibody libraries. Nat. Biotechnol.18, 75.

Schier, R., Marks, J.D., 1996. Efficient in vitro affinity maturation ofphage antibodies using BIAcore guided selections. Hum. Antib.Hybridomas 7, 97.

Schmaljohn, C., Hjelle, B., 1997. Hantaviruses: a global diseaseproblem. Emerg. Infect. Dis. 3, 95.

69N. Velappan et al. / Journal of Immunological Methods 321 (2007) 60–69

Schmidt, J., Meisel, H., Hjelle, B., Kruger, D.H., Ulrich, R., 2005.Development and evaluation of serological assays for detection ofhuman hantavirus infections caused by Sin Nombre virus. J. Clin.Virol. 33, 247.

Seme, K., Poljak, M., Babic, D.Z., Mocilnik, T., Vince, A., 2005. Therole of core antigen detection in management of hepatitis C: acritical review. J. Clin. Virol. 32, 92.

Sheets, M.D., Amersdorfer, P., Finnern, R., Sargent, P., Lindqvist, E.,Schier, R., Hemmingsen, G., Wong, C., Gerhart, J.C., Marks, J.D.,1998. Efficient construction of a large nonimmune phage antibodylibrary; the production of panels of high affinity human single-chainantibodies to protein antigens. Proc. Natl. Acad. Sci. U. S. A. 95, 6157.

Studier, F.W., 2005. Protein production by auto-induction in highdensity shaking cultures. Protein Expr. Purif. 41, 207.

VanAntwerp, J.J., Wittrup, K.D., 2000. Fine affinity discriminationby yeast surface display and flow cytometry. Biotechnol. Prog.16, 31.

van den Brink, E.N., Ter Meulen, J., Cox, F., Jongeneelen, M.A.,Thijsse, A., Throsby, M., Marissen, W.E., Rood, P.M., Bakker, A.B.,Gelderblom, H.R., Martina, B.E., Osterhaus, A.D., Preiser, W.,Doerr, H.W., de Kruif, J., Goudsmit, J., 2005. Molecular andbiological characterization of human monoclonal antibodies bindingto the spike and nucleocapsid proteins of severe acute respiratorysyndrome coronavirus. J. Virol. 79, 1635.

Vapalahti, O., Lundkvist, A., Kallio-Kokko, H., Paukku, K., Julkunen, I.,Lankinen, H., Vaheri, A., 1996. Antigenic properties and diagnosticpotential of puumala virus nucleocapsid protein expressed in insectcells. J. Clin. Microbiol. 34, 119.

Vaughan, T.J., Williams, A.J., Pritchard, K., Osbourn, J.K., Pope, A.R.,Earnshaw, J.C., McCafferty, J., Hodits, R.A., Wilton, J., Johnson,

K.S., 1996. Human antibodies with sub-nanomolar affinitiesisolated from a large non-immunized phage display library. Nat.Biotechnol. 14, 309.

Williamson, R.A., Burioni, R., Sanna, P.P., Partridge, L.J., Barbas, C.F.,Burton, D.R., 1993. Human monoclonal antibodies against aplethora of viral pathogens from single combinatorial libraries.Proc. Natl. Acad. Sci. U. S. A. 90, 4141.

Yoshimatsu, K., Lee, B.H., Araki, K., Morimatsu, M., Ogino, M.,Ebihara, H., Arikawa, J., 2003. The multimerization of hantavirusnucleocapsid protein depends on type-specific epitopes. J. Virol.77, 943.

Zebedee, S.L., Barbas, C.F., Hom, Y., Caothien, R.H., Graff, R.,Degraw, J., Pyati, J., LaPolla, R., Burton, D.R., Lerner, R.A.,Thronton, G.B., 1992. Human combinatorial antibody libraries tohepatitis B surface antigen. Proc. Natl. Acad. Sci. U. S. A. 89, 3175.

Zhang, M.Y., Shu, Y., Phogat, S., Xiao, X., Cham, F., Bouma, P.,Choudhary, A., Feng, Y.R., Sanz, I., Rybak, S., Broder, C.C.,Quinnan, G.V., Evans, T., Dimitrov, D.S., 2003. Broadly cross-reactive HIV neutralizing human monoclonal antibody Fabselected by sequential antigen panning of a phage display library.J. Immunol. Methods 283, 17.

Zhang, J., Tanha, J., Hirama, T., Khieu, N.H., To, R., Tong-Sevinc, H.,Stone, E., Brisson, J.R., MacKenzie, C.R., 2004. Pentamerizationof single-domain antibodies from phage libraries: a novel strategyfor the rapid generation of high-avidity antibody reagents. J. Mol.Biol. 335, 49.

Zvirbliene,A., Samonskyte, L., Gedvilaite, A.,Voronkova, T., Ulrich, R.,Sasnauskas, K., 2006. Generation of monoclonal antibodies ofdesired specificity using chimeric polyomavirus-derived virus-likeparticles. J. Immunol. Methods 311, 57.