Skip Navigation


PEDS Advance Access originally published online on October 9, 2007
Protein Engineering Design and Selection 2007 20(9):461-472; doi:10.1093/protein/gzm044
This Article
Right arrow Abstract Freely available
Right arrow FREE Full Text (PDF) Freely available
Right arrow Supplementary Data
Right arrow All Versions of this Article:
20/9/461    most recent
gzm044v1
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in ISI Web of Science
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Add to My Personal Archive
Right arrow Download to citation manager
Right arrow Search for citing articles in:
ISI Web of Science (1)
Right arrowRequest Permissions
Google Scholar
Right arrow Articles by Løset, G.A.
Right arrow Articles by Sandlie, I.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Løset, G.A.
Right arrow Articles by Sandlie, I.
Related Collections
Right arrow 2007
Social Bookmarking
Add to CiteULike   Add to Connotea   Add to Del.icio.us  
What's this?

© The Author 2007. Published by Oxford University Press. All rights reserved. For Permissions, please e-mail: journals.permissions@oxfordjournals.org

Functional phage display of two murine {alpha}/ß T-cell receptors is strongly dependent on fusion format, mode and periplasmic folding assistance

G.Å. Løset1,4, E. Lunde1, B. Bogen2, O.H. Brekke3 and I. Sandlie1,4

1Department of Molecular Biosciences, University of Oslo, N-0316 Oslo, Norway 2Institute of Immunology, University of Oslo, N-0316 Oslo, Norway 3 Birkeland Innovation, N-0316 Oslo, Norway

4 To whom correspondence should be addressed. E-mail: g.a.loset{at}imbv.uio.no or inger.sandlie{at}imbv.uio.no


   Abstract
Top
 Abstract
Introduction
 Materials and Methods
 Results
 Discussion
 Supplementary data
 Funding
 Acknowledgements
 References
 
Phage display has been instrumental for the success of antibody (Ab) technology. The aim of the present study was to explore phage display of soluble T-cell receptors (TCRs). A library platform that supports engineering and selection of improved TCRs to be used as detection reagents for specific antigen presentation will be very useful. In such applications, high, equal and clone independent display levels are a prerequisite for ‘fair’ selection. Therefore, we explored how different pIII fusion formats and modes affected the display levels of two murine {alpha}/ß TCRs. Both are derived from T-cell clones associated with the MOPC315 myeloma model. The results show that the design of the pIII fusion particle significantly affects the subsequent display levels. Furthermore, successful display may be obtained both in phagemid and phage versions. Importantly, improvement of poor display can be achieved by over-expressing the periplasmic chaperone FkpA.

Keywords: FkpA/phage display/T-cell receptor


   Introduction
Top
 Abstract
 Introduction
Materials and Methods
 Results
 Discussion
 Supplementary data
 Funding
 Acknowledgements
 References
 
The antigen (Ag) specific receptor of the T-cell lineage, the TCR, is a transmembrane heterodimer of a covalently coupled {alpha}- and ß-polypeptide chain. Each chain exposes two extracellular immunoglobulin (Ig) domains [a variable (V) and a constant (C) domain], and the two V domains comprise the ligand binding portion that specifically interacts with a peptide/major histocompatibility complex (pMHC) (Clements et al., 2006Go). Although the last decade has provided fundamental insight into the TCR-pMHC interaction (Rudolph et al., 2006Go), few reagents capable of specific pMHC detection are as of yet available.

A few monoclonal antibodies (MAbs) have been raised by hybridoma technology that show both specificity and sensitivity in their pMHC binding (Murphy et al., 1989Go; Aharoni et al., 1991Go; Duc et al., 1993Go; Dadaglio et al., 1997Go; Porgador et al., 1997Go; Zhong et al., 1997Go; Reay et al., 2000Go). As an alternative, in vitro selection of such Abs from libraries of specificities displayed on phage offers an attractive solution. However, such selection has so far been reported for pMHC class I binders only and not class II (Noy et al., 2005Go).

The TCR has evolved to confer fine specificity for pMHC complexes, but the mode of T-cell activation ensure that the interaction most often fall within a relatively narrow window of affinities in the lower micro-molar range (Davis et al., 2003Go; Davis and van der Merwe, 2006Go). Given this low binding strength, it has become clear that without affinity maturation, the use of native soluble TCRs as pMHC detection reagents may have limited utility. Alternatively, raising the functional affinity of the TCR–pMHC interaction by e.g. TCR-avidin tetramerization, might offer a gain in sensitivity (Subbramanian et al., 2004Go; Laugel et al., 2005Go). However, TCRs seem to be intrinsically unstable and hence difficult to produce as soluble molecules. Consequently, most protocols rely on extensive engineering and purification to obtain functional protein in reasonable quantities [e.g. Boulter et al. (2003)Go and references therein]. For these reasons, tools for engineering TCRs to obtain improved structures, both with respect to stability, solubility and affinity would be very useful.

Molecular evolution has been extremely important in Ab technology, and indeed successful TCR engineering has been reported using yeast display (Kieke et al., 1999Go; Shusta et al., 2000Go; Weber et al., 2005Go). However, in contrast to the success with Ab fragments (Hoogenboom, 2005Go), only a few reports describe TCR phage display (Onda et al., 1995Go; Weidanz et al., 1998Go), and only very recently has affinity maturation of phage displayed human TCR ectodomains been reported (Li et al., 2005Go; Dunn et al., 2006Go). Thus, our current understanding of parameters that determine successful TCR phage display is only rudimentary.

To this end, we have engineered two different MHC class II restricted murine {alpha}/ß TCRs from the MOPC315 myeloma model that recognize the same tumor-specific peptide amino acid (aa) 91–101 of the myeloma protein {lambda}2315 light chain in complex with MHC class II, I-Ed (Bogen et al., 1986Go; Bogen and Lambris, 1989Go), and surveyed the effects of (i) fusion format: scTCR versus cFab versus dsTCR, (ii) fusion mode: alternation of which TCR chain that is fused to the viral coat, and (iii) the phagemid versus phage system in TCR-pIII display. In addition, the effect of chaperone co-expression was explored. The results show that several routes may lead to successful display, but that this is critically dependent upon the mode of fusion. For single chain TCR (scTCR), the V{alpha}ß fusion was superior when both TCRs were considered. Likewise, for the four domain chimeric TCR/Ig (cFab) variant, the H{alpha}Lß was consistently superior for both TCRs, but this was dependent on over-expression of the periplasmic chaperone FkpA. Moreover, display differences observed between the TCR clones also decreased as a consequence of this FkpA over-expression. Finally, as both successful phagemid and phage mediated TCR-pIII display are shown, this adds a flexibility that may be of crucial importance when mapping low affinity interactions and for designing future selection regimes.


   Materials and Methods
Top
 Abstract
 Introduction
 Materials and Methods
Results
 Discussion
 Supplementary data
 Funding
 Acknowledgements
 References
 
Gene segment nomenclature

The official IMGT (http://imgt.cines.fr/) gene segment nomenclature is used throughout (Glusman et al., 2001Go; Lefranc et al., 2005Go). Thus, the following TCR gene segment re-designation of the murine T-cell clones 4B2A1 [V{alpha}1, J{alpha}19/Vß8.2, Dß, Jß1.2] and 7A10B2 [V{alpha}3, J{alpha}1/Vß6, Dß, Jß1.1] is [TRAV7D-3*01, TRAJ40*01/TRBV13-2*01, TRBD1*01, TRBJ1-2*01] and [TRAV9-3*01, TRAJ58*01/TRBV19*01, TRBD1*01, TRBJ1-1*01], respectively (Snodgrass et al., 1992Go).

Plasmids, bacterial strains and phage

The pSEX-Dummy (Loset et al., 2005Go), pSEX81 (Welschof et al., 1997Go) and pFAB-Display (unpublished) phagemid vectors were kindly provided by Affitech AS (Oslo, Norway). Both pSEX81 and pFAB-Display harbour the same Ab V regions with specificity against 2-phenyloxazol-5-one (phOx) coupled to bovine serum albumin, which originally were isolated from a human Ab phage library (Marks et al., 1991Go). The phage vector fUSE5 (Scott and Smith, 1990Go), and both Escherichia coli strains K91K (thi, lacZ–:Tn{lambda}NK1105 (Kanr)) and MC1061 (hsdR mcrB araD139 {Delta}(araABC-leu)7679 {Delta}(lac)174 galU galK strA thi) were obtained from Dr G. P. Smith (Division of Biological Sciences, University of Missouri, USA). The fUSE5-pSC construct, as well as the pLNOH2 and pLNOK vectors have been described previously (Lauvrak et al., 2004Go; Norderhaug et al., 1997Go). The E. coli XL1-Blue strain (recA1 endA1 gyrA96 thi-1 hsdR17 supE44 relA1 lac [F' proAB lacIqZ{Delta}M15 Tn10 (Tetr)]), M13K07 helper phage and HyperPhageTM were purchased from Stratagene (LaJolla, CA, USA), Amersham Biosciences (Uppsala, Sweden) and Progen Biotechnik GmbH (Heidelberg, Germany), respectively.

Abs and additional reagents

The mouse anti-pIII, rabbit anti-fd and sheep anti-M13-HRP Abs were purchased from MoBiTec (Goettingen, Germany), Sigma-Aldrich (Oslo, Norway) and Amersham Biosciences (Uppsala, Sweden), respectively. Purified SEC2 was obtained from Toxin Technology, Inc. (Sarasota, FL, USA). The murine MAb K13 specific for the human C{kappa} domain (Fjeld et al., 1992Go) was a kind gift from Dr Terje E. Michaelsen (Norwegian Institute of Public Health, Oslo, Norway). The anti-FLAG M2-HRP Ab was purchased from Sigma-Aldrich (Oslo, Norway). All restriction enzymes were purchased from New England Biolabs (Ipswich, MA, USA). Synthetic peptide and DNA oligos were purchased from NeoMPS SA (Strasbourg, France) and MWG Biotech AG (Ebersberg, Germany), respectively. For TCR-specific Ab reagents, see Table I.


View this table:
[in this window]
[in a new window]

  		Table I. TCR-specific Abs used in the current study

 
Cloning of the TRA and TRB genes

mRNA was isolated from 2 x 106 cells of the murine T-cell clones 4B2A1 and 7A10B2 (Bogen et al., 1986Go), respectively, using the Dynabeads mRNA DIRECT kit (Invitrogen, Oslo, Norway). cDNA was prepared with First-Strand cDNA Synthesis kit (Amersham Biosciences, Uppsala, Sweden) as described by the manufacturers' protocol, using 0.2 µg NotI-d(T)18 primer. Gene specific PCR primers were designed based on published sequences (Hayday et al., 1985Go; Winoto et al., 1985Go; Chou et al., 1987Go; Snodgrass et al., 1992Go) and tagged with RE sites (Supplementary material, Table SI). They were used to amplify the TCR ectodomains from cDNA by the use of Herculase polymerase (Stratagene, LaJolla, CA, USA). The TCR {alpha}- and ß-chain PCR products were subsequently cloned into the pLNOH2 and pLNOK expression vectors, respectively, and confirmed by sequencing.

Construction of the scTCR- and cFab-pIII fusions

The 4B2A1 and 7A10B2 TCR V genes were PCR amplified (Supplementary material, Table SI) from the corresponding pLNOH2 and pLNOK plasmids. Pfu Ultra DNA polymerase (Stratagene, LaJolla, CA, USA) was used for primer extension and the amplified DNA fragments purified with a QIAquick® PCR Purification Kit (Qiagen GmbH, Hilden, Germany) before RE digestion. The resulting digests were separated by agarose gel electrophoresis, purified with a QIAquickTM Gel Extraction Kit (Qiagen GmbH, Hilden, Germany) and ligated into the pSEX-Dummy vector using the NcoI/HindIII and MluI/NotI RE sites, before transformation into E. coli XL1-Blue cells. All scTCR constructs were confirmed by sequencing. The TCR 4B2A1-based pSEX constructs were used as template in a subsequent PCR to amplify the scTCR cassettes using primers tagged with SfiI RE sites (Supplementary material, Table SI) for cloning into the phage vector fUSE5. The amplified DNA fragments were processed as described above and the resulting fUSE5 constructs transformed into E. coli MC1061 cells. cFab-pIII fusions were constructed with standard techniques by moving the TCR V gene segments from pSEX into the compatible RE sites of pFAB-Display followed by transformation into E. coli XL1-Blue cells.

Construction of the dsTCR-pIII fusions

The human TRAC*01 (GenBank accession no. X02883 [GenBank] ) and TRBC2*01 (GenBank accession no. M12888 [GenBank] ) aa sequences were aligned (Supplementary material, Figure S1) against the corresponding murine segments, and the homologues to the reported TRAC Thr48 and TRBC Ser57 residues identified (Boulter et al., 2003Go). The 4B2A1 TRA and TRB segments were amplified by PCR (Supplementary material, Table SI). Primer extension was done with Phusion DNA polymerase (Sigma-Aldrich, Oslo, Norway) with the pLNOH2 and pLNOK vectors described above as template. After a PCR clean-up as described above, the DNA fragments were spliced by overlap extension in a second PCR yielding 4B2A1 TCR ectodomains with TRA Thr161Cys and TRB Ser171Cys (TCR 4B2A1 numbering). These DNA fragment were RE digested, purified and ligated into pFAB-Display on NcoI/SpeI and MluI/SfiI RE sites, creating the pTCRD {alpha}ß-dsTCR phagemid. Finally, the unpaired TRBC Cys185 was replaced with an Ala residue using QuikChangeTM mutagenesis according to the manufacturers' instructions (Stratagene, LaJolla, CA, USA). The pTCRD ß{alpha}-dsTCR variant was made by PCR re-amplifying these mutant TRA- and TRB-segments with primers tagged with the appropriate RE sites (Supplementary material, Table SI), followed by cloning into pFAB-Display. The resulting dsTCR constructs were confirmed by sequencing.

Cloning and expression analysis of the E. coli chaperone FkpA

The genomic E. coli K12 sequence (GenBank accession no. NC_000913 [GenBank] ) was used for primer design (Supplementary material, Table SI) to isolate the fkpA gene from the 3-end of the slyX open reading frame (ORF) to the stop codon of the fkpA ORF. Both primers included 5'-end SalI RE sites and the 3-end primer was also tagged with a FLAG tag encoding sequence (N-DYKDDDDK-C). PCR was applied to 1 µl of E. coli XL1-Blue overnight culture using Vent DNA polymerase (New England Biolabs, Ipswich, MA, USA). The resulting DNA fragment was ligated into pSEX-Dummy on the corresponding SalI RE sites creating the new pFKPDN-Dummy phagemid, which was transformed into E. coli XL1-Blue. Clones with the correct orientation were identified by RE analysis and subsequently confirmed by sequencing. rFkpA expression level and subcellular localization were analyzed essentially as described (Kipriyanov et al., 1997Go). The rFkpA cassette was subsequently shuffled between the various TCR-pIII containing phagemids on compatible ScaI/NheI RE sites using standard techniques.

Preparation of TCR-displaying bacteriophage

Phagemid rescue from E. coli XL1-Blue using M13K07 or HyperPhageTM helper phage was done essentially as described (Welschof et al., 1997Go; Rondot et al., 2001Go). Prior to analysis, the various supernatants were normalized according to the averaged virion titers obtained in several independent titration experiments. Recombinant phages were amplified from E. coli MC1061 transformed with fUSE5 essentially as described (Smith and Scott, 1993Go). Virion assembly was monitored either by spot titration as described (Koch et al., 2000Go), or by an enzyme-linked immunosorbent assay (ELISA) as follows: Polyclonal anti-fd Abs were coated into MaxiSorpTM microtiter plate wells (Nunc, Roskilde, Denmark) at 5 µg/ml in PBS overnight at 4°C. Residual protein binding capacity was blocked with 4% PBSTM (PBS, pH 7.4 with 4% v/w skim milk and 0.05% v/v Tween 20) for 1 h at room temperature (RT). Serial dilutions of virion-containing supernatant in 2x YT medium were then allowed to react for 1 h at RT before captured virions were detected with anti-M13-HRP (1:5000) for 1 h at RT. The plates were developed with ABTS substrate and the absorbance read at A405nm after 30 min. The sample titers were computed with GraphPad Prism version 3.03 for Windows (GraphPad Software, San Diego, CA, USA) from a standard curve made from phage with known titers. For the cellular panning (see below), the virions were purified and concentrated by PEG/NaCl precipitation as described (Marks et al., 1991Go), and re-suspended in PBS.

Phage capture ELISAs

The various MAbs were absorbed to MaxiSorpTM microtiter plate wells (Nunc, Roskilde, Denmark) in concentrations from 2.5 to 5 µg/ml in PBS, pH 7.4 overnight at 4°C. The wells were blocked with PBSTM for 1 h at RT, virion preparations were then added and allowed to react for 1 to 2 h at RT before captured virions were detected with anti-M13-HRP (1:5000) for 1 h at RT. The wells were developed with ABTS substrate and the absorbance read at A405nm after 30 min.

Detection of phagemid-derived scTCR- or cFAb-displaying virion binding to SEC2 was achieved through an inhibition ELISA. The MAbs F23.1, F23.2 and GB113 were immobilized to MaxiSorpTM microtiter plates (Nunc, Roskilde, Denmark) at concentrations previously determined to yield low signal, yet specific phage detection. Dilutions of SEC2 ranging from 30- to 850-fold molar surplus binding sites (assuming one and two binding sites for SEC2 and the MAb, respectively) of SEC2 when compared with the MAbs were then allowed to react with the various virion preparations in PBSTM using pre-blocked tubes. After 1 h at RT, these mixtures were added to the MAb-containing wells followed by incubation for 1 h at RT. Captured virions were detected with anti-M13-HRP (1:5000) for 1 h at RT and the wells developed by adding ABTS substrate and the absorbance read at A405nm after 30 min.

ELISA analysis of differences in TCR valence

A 2-fold serial dilution of MAb GB113, starting at 2 µg/well, was absorbed to MaxiSorpTM microtiter plate wells (Nunc, Roskilde, Denmark) in PBS, pH 7.4 overnight at 4°C. The wells were blocked with PBSTM for 1 h at RT, the various virion preparations [M13K07 versus HyperPhageTM helper phage packaging, as well as phagemid versus phage format (pSEX versus fUSE5)] added and allowed to react for 1.5 h at RT before bound virions were detected with anti-M13-HRP (1:5000) for 1 h at RT. The wells were developed by adding ABTS substrate and the absorbance read at A405nm after 30 min.

SDS–PAGE and western blotting

Virion-containing supernatants were separated by non-reducing 8% SDS–PAGE essentially as described (Laemmli, 1970Go), or by 4–12% Bis/Tris XT Criterion precast gels (Bio-Rad, Hercules, CA, USA), and blotted onto a polyvinylidene fluoride membrane (Millipore, Madison, USA) in Tris/glycine buffer (25 mM Tris, 192 mM glycine, and 20% methanol, pH 8.3) at 25 V for 30 min using a semi-dry blotting apparatus (Bio-Rad, Hercules, CA, USA). The membrane was blocked in PBSTM before pIII-fusions were detected with mouse anti-pIII MAb (1:4000) followed by sheep anti-mouse-HRP (1:10 000). The membrane was washed and developed with SuperSignalTM West Pico substrate (Pierce, Rockford, IL, USA) and exposed to BioMaxTM MR film (Kodak, Fernwald, Germany). For the rFkpA expression analysis, identical volumes of periplasmic and cytosolic fractions isolated in parallel from A600nm normalized amounts of E. coli cell cultures were run on 12% SDS–PAGE precast gels (Bio-Rad, Hercules, CA, USA). The gels were either stained with Bio-SafeTM Coomassie Brilliant Blue (Bio-Rad, Hercules, CA, USA), or the proteins transferred to a polyvinylidene fluoride membrane as described above. The membrane was developed as described above using an anti-FLAG M2-HRP Ab (1:4000) for protein detection.

Cell cultures

F9A15.3.19 (herein termed F9) and A20HA (the latter was a generous gift from Dr Hyam I. Levitsky, Department of Oncology, Johns Hopkins University, Baltimore) are A20 BALB/c B lymphoma cells transfected with {lambda}2315 Ig light chain and viral hemaglutinine, respectively (Weiss and Bogen, 1989Go; Levitsky et al., 1996Go). Both cell lines were cultured in RPMI1640 (BioWhittaker) supplemented with 10% fetal calf serum, penicillin (100 IU/ml), streptomycin (100 µg/ml), 1 mM sodium pyruvate, 0.1 mM non-essential aa and 50 µM monothioglycerol (all from Invitrogen) at 37°C under 5% CO2.

Cellular panning

The cellular panning protocol applied was a modified version of BRASIL (Giordano et al., 2001Go) performed as follows: Day 1, both F9 and A20HA were split to a cell density of 5 x 105 cells/ml. The F9 cells were then incubated with either 25 or 50 µM of the MOPC315-specific peptide {lambda}2315, aa 89 to 105 (N-FAALWFRNHFVFGGGTL-C) for 16 h. Day 2, the F9 cells were again pulsed with 50 µM peptide and incubated for 1.5 h (phagemid panning only). The cells were then pelleted by centrifugation at 300 g for 10 min at RT, re-suspended in either RMPI1640/1% v/w BSA, or PBS/3% v/w BSA to a final cell density of 3.3 x 106 cells, and portions of 300 µl transferred to pre-blocked 1.5-ml tubes. The tubes were then kept for 1 h at 4°C with agitation. Appropriate volumes of the various virion preparations were also blocked accordingly in parallel. Thereafter, volumes of 200 µl of virions corresponding to either 1 x 109 kanamycin resistant (kanR) colony forming units (cfu) (fUSE5) or 1 x 106 ampicilin resistant (ampR) cfu (phagemids) were mixed with the cells (total volume 500 µl) and allowed to interact for 2 h at 4°C with agitation. The organic phase used in the subsequent BRASIL step was prepared by mixing dibutyl phthalate (Sigma-Aldrich) with cyclohexane (Aldrich) at a 9:1 ratio (v/v). Volumes of 200 µl of this mixture were transferred to 0.2 ml thin-walled PCR tubes. Volumes of 100 µl of each of the virion:cell mixtures were gently laid onto the organic phase and the tubes centrifuged at 10,000 g for 10 min at RT. The tubes were then snap-frozen in liquid nitrogen and immediately put on dry ice. The bottom part of the tubes (containing the cell pellets) was cut of using a sterile scalpel and transferred to fresh 1.5 ml tubes. Each pellet was then dissolved using 50 µl 2x YT. The pellet-recovered virions were rescued by adding either log-phase E. coli K91K (fUSE5) or XL1-Blue (phagemids) and the output titer determined (Day 3) as cfu resulting from serial dilutions plated on LB agar dishes containing either 30 µg/ml tetracycline (fUSE5), or 100 µg/ml ampicilin (phagemids).


   Results
Top
 Abstract
 Introduction
 Materials and Methods
 Results
Discussion
 Supplementary data
 Funding
 Acknowledgements
 References
 
Design of the TCR-pIII fusions

A panel of TCR variants was made as N-terminal fusions to the minor coat protein pIII using both phage and phagemid vectors (Fig. 1). The three formats scTCR, cFab and dsTCR were made as follows: (i) the scTCR was constructed by genetically fusing the TRAV and TRBV segments by a synthetic linker, creating a single polypeptide (Fig. 1A–D). (ii) Similarly, a cFab was constructed by genetically fusing the TRAV and TRBV segments to the human IGKC (which is fused to pIII) and IGHG1 CH1 segments, hence making two separate polypeptides that only form a functional covalently coupled unit upon correct chain pairing in the periplasm (Fig. 1E and F). The same is true for the (iii) dsTCR, which was constructed by cloning of the TCR ectodomains (4B2A1 only), ending C-terminally at the third last ({alpha}-chain) or last (ß-chain) residue that precedes the native inter-chain disulphide bond (Fig. 1G and H). To increase the stability of the hetereodimer, an artificial inter-chain disulphide bond was engineered into the dsTCR structure as described (Boulter et al., 2003Go). To gain insight into how domain orientation and fusion mode affect display level, all three formats were made in two versions with either the TCR {alpha}- or ß-derived subunit as the pIII fusion partner (Fig. 1J).


Figure 1
View larger version (53K):
[in this window]
[in a new window]
[Download PowerPoint slide]
  		Fig. 1. Schematic drawing of the TCR-pIII fusion constructs. (A) and (B) scTCR V{alpha}ß and Vß{alpha} in pSEX. (C) and (D), scTCR V{alpha}ß and Vß{alpha} in fUSE5. (E) and (F) cFab H{alpha}Lß and HßL{alpha} in pFAB-Display. (G) and (H) dsTCR {alpha}ß and ß{alpha} in pFAB-Display. (I) The fkpA ORF cloned into pSEX and pFAB-Display. The rFkpA cassette was inserted with the transcriptional direction in the opposite orientation as to the lacPO controlled cassette, resulting in transcriptional termination of both cassettes into the same terminator, but on the opposite DNA strands. (J) Illustration of the resulting pIII fusions obtained from the DNA constructs described above. Abbreviations: lacPO, lac promoter, sd, Shine-Dalgarno sequence, t, T7 transcriptional terminator, X0.25, native pIII promoter, xPO, native fkpA promoter.

 
Analysis of phagemid-derived scTCR-, cFab- and dsTCR-pIII display

Supernatants containing virions encoded by pSEX (scTCR format) or pFAB-Display (cFab format) were prepared with M13K07 phagemid rescue as described. Then, several MAbs specific for the TCR V domains were coated on microtiter plates and supernatants allowed reacting before detection of bound virions with a peroxidase-coupled anti-M13 MAb. In the case of the 4B2A1 TCR formats, a clear difference in MAb reactivity was observed in this ELISA (Fig. 2A). What was immediately apparent, was the stronger MAb F23.1 and F23.2 binding to the scTCR with the V{alpha} domain N-terminally (scTCR V{alpha}ß), as well as the stronger MAb F23.2 and GB113 binding to the cFab with the V{alpha}domain fused to the C{kappa} domain (cFab HßL{alpha}). In contrast, neither of the constructs derived from the 7A10B2 TCR exhibited any appreciable MAb (44-22-1 and RR4-7) reactivity (data not shown). Table I lists the specificities of the TCR-specific Abs employed.


Figure 2
View larger version (29K):
[in this window]
[in a new window]
[Download PowerPoint slide]
  		Fig. 2. Strucural analysis of TCR 4B2A1-derived phagemids encoding scTCR-, cFab- and dsTCR-pIII fusions. (A) ELISA analysis in which equal amounts of virions (5 x 109 cfuampR/well) were captured on immobilized TCR V gene-specific MAbs (500 ng/well) as described. The anti-phOx scFv and Fab were included as negative controls. One of three independent experiments performed in duplicates is shown with the SD indicated by error bars. (B)Virions corresponding to 3 x 108 cfuampR/lane were separated by non-reducing SDS–PAGE (8%) followed by western blotting and pIII detection as described. The M13K07 was included as a positive control. One out of three representative gels is shown. (C) ELISA analysis in which equal amounts of virions (1.5 x 1010 cfuampR/well) were captured on immobilized MAb (250 ng/well, except for MAB F23.1, which was coated at 500 ng/well) as described. One of three independent experiments performed in duplicates is shown with the SD indicated by error bars.

 
To investigate if the discrepancy in MAb reactivity could be due to varying display levels, whole virions were separated by SDS–PAGE and pIII analyzed by western blotting. The TCR-pIII fusion levels did indeed mirror the MAb reactivity profile of the constructs, confirming that the 4B2A1-derived scTCR V{alpha}ß and cFab HßL{alpha} were displayed at a higher level than their counterparts with the alternative domain orientation (Fig. 2B). Both exhibited strong bands at a MW of ~62 kDa. This corresponds to pIII fused to two Ig domains (either scTCR or V{alpha}C{kappa}). In the case of the scTCR, this was the expected MW, whereas the four Ig domain cFab was revealed only after over-exposure of the film (data not shown).

Screening for heterodimeric dsTCR display on the virions was performed by ELISA essentially as described above, using an extended panel of TCR-specific MAbs (Table I). Only the {alpha}ß-dsTCR (TCR ß-chain fused to pIII) was detected and only by F23.1 and H57-597 (Fig. 2C).

TCR-virions bind to the superantigen SEC2

To further assess the folding of the scTCR V{alpha}ß and cFab HßL{alpha} constructs, we analyzed whether or not they could bind the bacterial superantigen Staphylococcus aureus enterotoxin C2 (SEC2). SEC2 binds to a large conformational epitope spanning the complementarity-determining region (CDR) 2 and to a lesser extent CDR1 and hypervariable region 4 of the murine TRBV13-2 segment (Fields et al., 1996Go). Thus, we hypothesized that upon TCR binding, SEC2 would mask the epitopes of all three MAbs F23.1, F23.2 and GB113. The SEC2-TCR interaction in the absence of MHC is weak with a Kd in the low 10–6 M range (Malchiodi et al., 1995Go). The binding strength of MAb F23.1 has been determined to Kd ~ 10–8 M (Rojo and Janeway, 1988Go), whereas those of MAbs F23.2 and GB113 are unknown. However, in the ELISAs above, the three MAbs bound in a hierarchy of F23.1<F23.2<GB113. Thus, assuming Kd values in the range of 10–8 to 10–9 M, we adjusted the amounts of SEC2 added to obtain ~50% inhibition. SEC2 was pre-incubated with the virion preparations before unbound virions were captured on immobilized MAbs and detected as described above. The results showed that SEC2 inhibited virion binding to all three MAbs (Fig. 3). The maximum inhibition was 45% (F23.1), 50% (F23.2) and 39% (GB113) in the case of the scTCR, and 42% (F23.1), 47% (F23.2) and 43% (GB113) in the case of the cFab. This clearly indicated that both the scTCR- and the cFab-pIII fusions had retained the overall native structural integrity of at least the Vß portion.


Figure 3
View larger version (23K):
[in this window]
[in a new window]
[Download PowerPoint slide]
  		Fig. 3. SEC2 inhibition ELISA of TCR 4B2A1-derived phagemids encoding scTCR V{alpha}ß- and cFab HßL{alpha}-pIII. Virions corresponding to 5 x 109 cfuampR/well below, were mixed with amounts of SEC2 that corresponded to either 2 (6 µg) or 9 (25 µg) µM/well below. Following incubation, these mixtures were added to microtiter plates coated with MAbs as described (F23.1: 500 ng/well, F23.2: 170 ng/well and GB113: 80 ng/well). Subsequently, bound virions were detected as described. The anti-phOx scFv and Fab were included as negative controls. One representative of three independent experiments is shown.

 
Multivalent TCR-pIII display increases the sensitivity of Ag detection

As the affinity of most TCRs towards pMHC is low, we investigated how the functional affinity could be raised by increasing the valence of the TCR on the virions. As a source of multivalent TCR display (3–5 copies per virion), we used the phage genome-based vector fUSE5, which ensures that all pIII displayed on the virions are fusions, and constructed the scTCR format in both domain orientations employing the 4B2A1 TRAV and TRBV segments. Virion binding to the MAbs in ELISA were investigated as before and in all cases scTCR V{alpha}ß was the superior orientation (Fig. 4A). These scTCR V{alpha}ß4B2A1-pIII multivalent virions were then tested in cellular panning experiments against an A20 transfectoma (F9) known to endogenously present the cognate {lambda}2315 T-cell epitope on I-Ed to 4B2A1 T cells (Weiss and Bogen, 1989Go; Lauritzsen et al., 1994Go). As a negative control, we used a different A20 transfectoma (A20HA). Indeed, a 5-fold enrichment favouring the target cell line F9 was seen in comparison with the negative control cell line A20HA, but only when the F9 cells were exogenously boosted with specific synthetic peptide (Fig. 4B). Hence, it appears that the endogenous {lambda}2315 T-cell epitope/I-Ed presentation on F9 cells is sufficient for sensitive T-cell assays, but not for recombinant TCR binding to be observed. Nevertheless, the results indicated that the scTCR V{alpha}ß4B2A1 was displayed in a functional form.


Figure 4
View larger version (16K):
[in this window]
[in a new window]
[Download PowerPoint slide]
  		Fig. 4. Structural and functional analysis of phage (fUSE5)-encoded scTCR 4B2A1. (A) ELISA analysis in which equal amounts of virions (1 x 108 cfutetR/well) were captured on immobilized TCR V gene-specific MAbs (500 ng/well) as described. The anti-phOx scFv served as negative control. One of the two independent experiments performed in duplicates is shown with the SD indicated by error bars. (B) Cellular panning of scTCR V{alpha}ß4B2A1 (1 x 109 cfutetR/sample) on F9 and A20HA cells (1 x 106 cells/sample) as described in Materials and methods. The ratios depicted were computed as the output cfu obtained from F9 divided by those obtained from A20HA. The graph must be interpreted as: (i) ratio >1 indicates preferential binding to F9, (ii) ratio <1 indicates preferential binding to A20HA and (iii) ratio = 1 indicates no preferential binding to either cell type. The pSC clone was included as negative control and one of two independent experiments is shown.

 
To assess multivalent TCR-pIII display of phagemid-encoded TCR-pIII, we prepared virions rescued with either HyperPhage or M13K07 helper phages. HyperPhage is a modified M13K07 helper phage without endogenous pIII expression (Rondot et al., 2001Go), hence all pIII molecules displayed are of recombinant origin leading to multivalent display comparable to phage genome-based vectors. In contrast, regular M13K07 helper phage encodes wild-type pIII and therefore directs display of varying, but lower copy numbers of recombinant pIII-TCR. In initial experiments determining virion titers, we found that the samples prepared with HyperPhage had 100- to 1000-fold reduced titers when compared with their M13K07 counterparts when based on infectivity. When titrations were done by ELISA, it became apparent that this reduction did not reflect the actual virion numbers, as the ELISA showed HyperPhage titers only 10-fold lower than with M13K07 (data not shown). As the virion size varies with the genome size (Marzec and Day, 1983Go), deviations in actual titers between the different constructs will also occur with ELISA titration, but not to the extent seen when based on infectiousness. Hence, to normalize the virion numbers, all subsequent titrations relevant to the assay shown in Fig. 5 were performed using ELISA.


Figure 5
View larger version (15K):
[in this window]
[in a new window]
[Download PowerPoint slide]
  		Fig. 5. Functional affinity ELISA of TCR 4B2A1-derived scTCR- and cFab-pIII fusions. MAb GB113 was coated as capturing ligand in a serial dilution series (2,000–10 ng/well). Equal amounts of scTCR V{alpha}ß ({blacktriangledown}/{triangledown}) and cFab HßL{alpha} ({blacktriangleup}/{triangleup}) displaying virions packaged using M13K07 ({triangledown}/{triangleup}) or HyperPhageTM ({blacktriangledown}/{blacktriangleup}) were added (5 x 107/well) and captured virions detected as described. For comparison, the fUSE5-derived scTCR V{alpha}ß was also included (•). The M13K07 packaged anti-phOx Fab ({blacksquare}) served as negative control (NC). The specific Ag detection limit was set to A405nm 0.2 (dotted line). One of the two independent experiments performed in triplicates is shown with the SD indicated by error bars.

 
MAb GB113 binding was then used to measure the binding sensitivity of the scTCR V{alpha}ß and cFab HßL{alpha} constructs. The MAb was coated as capturing agent on microtiter plates in 2-fold serial dilutions. Virions displaying either scTCR or cFab were then allowed to react, and the amounts of captured virions detected. The result clearly showed that higher valence of the TCR displayed resulted in a substantial increase in sensitivity (Fig. 5). The highest sensitivity was obtained with the multivalent phagemid-based scTCR format, exhibiting a difference (compared with the low valence counterpart) in signal strength at high Ag load of about 5-fold (A405nm of 1.67 ± 0.04 versus 0.32 ± 0.01) and in specific Ag detection of about 8-fold (500 versus 63 ng MAb GB113). The corresponding signal for the cFab format was 3-fold stronger at high Ag load (A405nm of 0.94 ± 0.05 versus 0.28 ± 0.01) and 5-fold lower in specific Ag detection (500 versus 100 ng MAb GB113) with the multivalent virions. As expected, the analysis also revealed that the two multivalent scTCRs forms, either as phagemid or phage, yielded comparable results.

Over-expression of the periplasmic chaperone FkpA increases productive TCR-pIII display

The initial ELISAs and western blot analyses showed that there was large heterogeneity with respect to display level and MAb reactivity. Most prominent was the fact that neither of the 7A10B2 TCR constructs were displayed. But also in the case of the 4B2A1 TCR, display level varied with format and mode of fusion, such that scTCR{alpha}ß > scTCRß{alpha} and cFabHßL{alpha} > cFabH{alpha}Lß.

As eukaryotic proteins produced in a prokaryotic host, TCRs often show poor functional expression. In an attempt to enhance the display, we cloned and co-expressed the periplasmic chaperone FkpA on the same phagemid containing the TCR format in question. The gene construct was made such that the recombinant FkpA (rFkpA) protein was expressed from its native promoter, hence only increasing the fkpA gene dosage in the E. coli host harboring the phagemid (Fig. 1I). The expression of the resulting rFkpA protein was analyzed by SDS–PAGE and western blotting, which showed strong over-expression in the periplasmic compartment (Fig. 6).


Figure 6
View larger version (68K):
[in this window]
[in a new window]
[Download PowerPoint slide]
  		Fig. 6. rFkpA expression analysis. (A) Normalized amounts of the periplasmic (P) and cytosolic (L) compartments fractionated from E. coli XL1-Blue, harboring phagemid without (pSEX), or with (pFKPDN) the rFkpA cassette were separated by SDS–PAGE (12%) and coomassie stained. (B) Western blot of the corresponding samples as in A, detected by anti-FLAG.

 
The effect of rFkpA co-expression on 4B2A1 TCR display levels was assessed by ELISA as described. Increased MAb reactivity was observed for both scTCR V domain orientations, but the relative effect was stronger for the construct initially found to be inferior (scTCR Vß{alpha}) (Fig. 7A). This effect was even more pronounced with respect to the cFab format. Here, the H{alpha}Lß variant that previously had shown poor MAb reactivity, now clearly exhibited the most favorable binding profile to the TCR-specific MAbs (Fig. 7B). Moreover, as opposed to being virtually non-reactive in MAb K13 binding in initial experiments (data not shown), both cFab variants now exhibited a strong and similar increase in binding to this Ab, which was comparable to the binding of the Ab anti-phOx Fab control. When analyzed by SDS–PAGE/western blotting, this increased 4B2A1 cFab MAb binding could be directly attributed to increased recombinant pIII display levels (data not shown).


Figure 7
View larger version (25K):
[in this window]
[in a new window]
[Download PowerPoint slide]
  		Fig. 7. Analysis of putative effects caused by rFkpA co-expression during phagemid-derived TCR 4B2A1-pIII containing virion assembly. (A) Equal amounts of virions (1 x 109 cfuampR/well) were captured on immobilized TCR V gene-specific MAbs (500 ng/well) as described. The anti-phOx scFv served as negative control. One of three independent experiments performed in duplicates is shown with the SD indicated by error bars. (B) Equal amounts of virions (1 x 1010 cfuampR/well) were captured on immobilized TCR V gene-specific MAbs [250 ng/well (F23.2 and GB113) and 500 ng/well (F23.1 and K13)] as described. The anti-phOx Fab served as control. One of three independent experiments performed in duplicates is shown with the SD indicated by error bars. (C) Equal amounts of virions (1 x 1010 cfuampR/well) were captured on immobilized MAb (250 ng/well, except for MAB F23.1, which was coated at 500 ng/well) as described. The cFab and anti-phOx Fab were included as positive and negative controls, respectively. One of three independent experiments performed in duplicates is shown with the SD indicated by error bars.

 
Also the dsTCR format exhibited an increase in MAb reactivity as a consequence of the rFkpA co-expression (Fig. 7C). However, only MAbs directed towards the ß-chain, namely F23.1 and H57-597, exhibited significant binding and only in the case of the {alpha}ß-dsTCR variant. A modified packaging protocol with low temperature and increased packaging time (Li et al., 2005Go) did not lead to improved dsTCR detection (data not shown).

Taken together, both scTCR and cFab 4B2A1 exhibited high display levels in combination with co-expression of rFkpA. scTCR V{alpha}ß and cFab H{alpha}Lß had the strongest MAb recognition, but the constructs with the opposite domain orientation were almost as good.

To further address the functionality of scTCR V{alpha}ß4B2A1 and cFab H{alpha}Lß4B2A1, virions were prepared in combination with over-expression of rFkpA and phagemid rescue using either M13K07 or HyperPhageTM. These virions were then tested in cellular panning experiments as described. We observed increased binding to the target cell line F9 relative to the negative control cell line A20HA (Fig. 8). The increase in F9 binding was seen for both the scTCR and cFab formats, indicating that both are displayed in a functional form. However, it was apparent that the increase in functional affinity obtained by multivalent display was required, as the scTCR prepared with M13K07 did not show any differential enrichment. It was also apparent that the cFab exhibited a slightly stronger preferential F9 binding than the scTCR.


Figure 8
View larger version (10K):
[in this window]
[in a new window]
[Download PowerPoint slide]
  		Fig. 8. Cellular panning on F9 and A20HA (1 x 106 cells/sample) using TCR 4B2A1-derived scTCR V{alpha}ß– and cFab H{alpha}Lß-containing virions (1 x 106 cfuampR/sample) prepared with rFkpA co-expression as described in Materials and methods. The ratios depicted were computed as the output cfu obtained from F9 divided by those obtained from A20HA and graph interpretation is as described in Fig. 4B. Open bar: virions prepared with M13K07, black bars: virions prepared with HyperPhageTM. The scFv anti-phOx clone was included as negative control and one of two independent experiments is shown.

 
Encouraged by the 4B2A1 results, we hypothesized that rFkpA co-expression would have a similar positive effect on 7A10B2 folding and display. Indeed, the results showed improved recognition of the scTCR 7A10B2 in ELISA with two different MAbs (RR4-7 and 44-22-1) (Fig. 9A). This beneficial effect was stronger with the scTCR V{alpha}ß variant, which has the same V domain orientation as was most productive for 4B2A1. The increased MAb reactivity was mirrored by a strong increase in display levels as assessed by SDS–PAGE/western blotting (Fig. 9B). Similarly, rFkpA co-expression affected the MAb binding pattern of the 7A10B2 cFab variants in a manner very similar to that of the 4B2A1 cFabs, and again was the H{alpha}Lß version the most favorable (Fig. 9C).


Figure 9
View larger version (27K):
[in this window]
[in a new window]
[Download PowerPoint slide]
  		Fig. 9. 7A10B2 TCR-pIII display analysis of effects caused by rFkpA co-expression during phagemid-derived virion assembly. (A) Equal amounts of TCR 7A10B2-derived scTCR-pIII containing virions (1 x 1010 cfuampR/well) were captured on immobilized TCR V gene-specific MAbs (500 ng/well) as described. The anti-phOx scFv and scTCR V{alpha}ß4B2A1 served as negative controls. One of four independent experiments performed in duplicates is shown with the SD indicated by error bars. (B) Virions corresponding to 4 x 109 cfuampR/lane were separated by SDS–PAGE (4–12%) followed by western blotting and pIII detection as described. The M13K07 and scFv anti-phOx were included as controls. One out of three representative gels is shown. (C) Equal amounts of TCR 7A10B2-derived cFab-pIII containing virions (2 x 109 cfuampR/well) were captured on immobilized MAbs (500 ng/well) as described. The anti-phOx Fab and cFab H{alpha}Lß4B2A1 served as controls. One of two independent experiments performed in duplicates is shown with the SD indicated by error bars.

 

   Discussion
Top
 Abstract
 Introduction
 Materials and Methods
 Results
 Discussion
Supplementary data
 Funding
 Acknowledgements
 References
 
The aim of the current study was to elucidate some of the intrinsic properties of TCRs that affect successful pIII phage display. We have previously shown for a large Ab scFv phage library that a high initial display level that reflects a high functionality of the library indeed is important for efficient downstream selection (Loset et al., 2005Go). As a model system, we utilized the {alpha}/ß TCRs of the murine T-cell clones 4B2A1 and 7A10B2. Phage display of Ab-derived scFv and Fab is now standard procedure. In the current study, we utilized phagemid vectors designed for phage display of Ab fragments, and show that they also support display of TCRs using standard phagemid rescue with the M13K07 helper phage. As standard phagemid rescue renders a mixture of phagemid- and helper phage-derived pIII on each virion leading to low valence display, we also investigated the use of multivalent pIII display using both a phage genome-based vector that ensures that all pIII on the virions are fusions, as well as phagemid rescue by use of the HyperPhageTM helper phage that lacks endogenous pIII and hence ensures that all pIII on the virions are of phagemid origin. As the TCR is a heterodimer of an {alpha}- and a ß-chain, we investigated which of the two chains, {alpha} or ß, may preferentially be fused to pIII. Furthermore, since a dsTCR format recently published holds great promise for human TCR phage display (Li et al., 2005Go), we generated a similar dsTCR of the murine 4B2A1 TCR.

Clearly, the only assay that irrevocably reveals functionality is one that includes cognate ligand binding. In the absence of specific soluble pMHC ligand in our system, we used binding to a panel of well characterized MAbs and also a superantigen as an indication of structural integrity. Importantly, the 4B2A1 scTCR and cFab were both recognized by the clone specific MAb GB113. To address functionality, we used binding to F9, a cell line with endogenous expression of both I-Ed and the myeloma {lambda}2315 light chain, that was also loaded exogenously with a T-cell epitope derived from {lambda}2315 (aa 89–105). This was compared to binding the same cell line without {lambda}2315 light chain expression and peptide loading, namely A20HA.

The results presented clearly show that format and domain orientation play key roles in successful display. Firstly, it was evident that the scTCR with the Vß rather than the V{alpha} domain located proximal to pIII exhibited stronger MAb reactivity. This was the case for both 4B2A1 and 7A10B2. The folding of these fusion constructs might be closer to the native state than that of its counterparts, or the epitopes more readily available for Mab binding. The latter is less likely though, as the SDS–PAGE/western blot analysis revealed a notable difference in display level between the two 4B2A1 scTCRs-pIII fusion molecules. Since all analyses performed in this study utilized normalized amounts of virions, the differences in MAb reactivity rather reflect differential incorporation of TCR-pIII fusion proteins into the virions, a phenomenon previously observed e.g. in the case of aggregation-prone Ab fragments (Bothmann and Pluckthun, 1998Go).

Secondly, in cFab display the V domains were also recognized by MAbs, but in this case GB113 and F23.2 detection was initially obtained only when the 4B2A1 Vß domain was fused to the IgG1 CH1 domain and the V{alpha} domain was fused to C{kappa} (the cFab HßL{alpha} variant). Accordingly, only MAb F23.1 detection was seen when the same Vß domain was fused to the C{kappa} domain. Therefore, MAb F23.1 does not seem to be well fit to sense global domain architecture, a notion supported by the fact that it also readily binds free ß-chains containing the TRBV13 segments (Brodnicki et al., 1996Go; Weidanz et al., 1998Go). The epitope recognized by MAb F23.2, which has been shown to involve both the CDR1 and CDR2 loops of the TRBV13-2 segment, may be a more stringent sensor of correct folding (Manning et al., 1998Go). Whether the MAb GB113 epitope is linear or conformational is unknown, but the Ab seems to require the clonotypic 4B2A1 {alpha}-chain only, either alone or in combination with any ß-chain, which implies that key residues of the epitope most likely map to the unique CDR3 loop of the V{alpha} domain (Bogen et al., 1986Go, 1992Go; Munthe et al., 1995Go, 1996Go).

The human dsTCR format initially described by Boulter et al. (2003)Go has recently been successfully introduced into a phagemid phage display context (Li et al., 2005Go), and has also been reported to work with a soluble murine TCR (Gronski et al., 2004Go). We therefore investigated if this could be further extended to murine TCR phage display. However, the initial results from binding studies involving five different MAbs, only showed recognition by two, both ß-chain specific, and only when the ß-chain was the pIII fusion partner. Notably, the dsTCR exhibiting partial MAb binding has the same domain orientation as that reported by Li et al. (2005)Go. Thus, the {alpha}-chain seems to be the source of the dsTCR malfunction, an observation also made by others when expressing soluble murine TCRs (Weber et al., 1992Go). Furthermore, the fact that the scTCR and cFab were bound by both MAbs and SEC2, as well as they bound F9, makes it tempting to focus on the C{alpha} domain. This domain diverges from the typical Ig-fold topology (Halaby et al., 1999Go), and in the murine 2C TCR, it has a large carbohydrate moiety attached to Asn{alpha}185 that makes extended stabilizing non-covalent interactions with the Cß domain (Garcia et al., 1996Go). A comparison between the murine 2C and human B7 {alpha}/ß TCR crystal structures (bacterially expressed B7) indicates that the C{alpha} domain folds are nearly identical (Ding et al., 1998Go). A search through the PROSITE database (Hulo et al., 2004Go) using the 2C and B7 PDB files readily identified the homologous Asn{alpha}185 and Asn{alpha}183 sites in 2C and B7, respectively (data not shown). However, there is to the best of our knowledge no report that Asn{alpha}183 in human C{alpha} is actually used for glycosylation. Moreover, the aa sequence identity and similarity between the human and murine C{alpha} domains are only 52.7 and 73.1%, respectively. Such differences could easily translate into differences in stability and folding, which in turn would affect heterologous expression (Roodveldt et al., 2005Go).

Two previous reports point to the advantage of pVIII as fusion partner in TCR phage display due to avidity effects (Onda et al., 1995Go; Weidanz et al., 1998Go). Such display may be applicable for single clone evaluation under certain circumstances, but it effectively undermines affinity maturation protocols, and as a general tool, pIII display is highly preferable (Kretzschmar and Geiser, 1995Go; Li et al., 2005Go). In pIII-facilitated phage display employing phagemids, it is generally believed that 1–10% of the virions carry one single copy of the recombinant protein, whereas the remaining pIII is wt derived from the helper phage (Bradbury and Marks, 2004Go). In contrast, phage genome-based facilitated pIII display renders three to five recombinant protein copies on each virion. The latter would therefore be expected to increase the sensitivity of the TCR-pIII/ligand interaction in comparison with phagemid display, as has been shown by e.g. avidin tetramerization of native, soluble TCRs (Subbramanian et al., 2004Go; Laugel et al., 2005Go). Indeed, the fUSE5-scTCR constructed and tested here exhibited preferential binding to the F9 cell line relative to a negative control cell line, whereas using standard phagemid rescue of a phagemid-encoded scTCR did not, indicating that increased functional affinity was needed to reveal cognate ligand binding. From a combinatorial library point of view however, it is most often advantageous to use a phagemid instead of phage (Bradbury and Marks, 2004Go). Our results also show that phagemid-derived multivalent TCR-pIII display may be achieved when rescued using HyperPhageTM and that such increase in valence increases the reactivity with polymeric ligands (in the form of arrayed MAb GB113 on a solid phase). The construct exhibiting the strongest increase in ligand binding was the scTCR format, whereas a smaller, yet significant, effect was seen with the cFab version. Rather than difference in folding, this phenomenon probably reflects the difference in ability of the scTCR-containing virion to incorporate more recombinant pIII per particle in comparison with the more sterically challenging cFab (de Haard et al., 1999Go).

Our results so far clearly indicate that certain fusion versions and formats yield a superior result with respect to folding and functionality. However, it is clear that successful functional, heterologous protein expression may depend on the primary sequence of the TCR clones studied (Maynard et al., 2005Go). Our findings may therefore not be applicable to all TCR sequences. For a molecular diversity platform to be of general applicability, it should ideally be able to accommodate a large sequence space, which inevitably will be of unknown identity and hence unknown physiochemical behavior. It is well documented that over-expressing certain chaperones may have a major impact on the functional expression of proteins (Thomas et al., 1997Go). The periplasmic E. coli chaperone FkpA has in this regard been shown to have a highly beneficial effect on the functional expression of toxic Ab scFvs, whereas already well-expressing clones were unaffected (Bothmann and Pluckthun, 2000Go). This property seems to be a result of the ability of the chaperone to (i) prevent premature aggregation of unstable proteins, which in turn allows efficient polypeptide folding into a stable, native fold and (ii) reactivate inactive protein (Ramm and Pluckthun, 2000Go; Ramm and Pluckthun, 2001Go). Also the periplasmic E. coli chaperone Skp has been shown to influence functional expression of toxic Ab scFvs in a positive manner similar to FkpA (Bothmann and Pluckthun, 1998Go). In initial analyses, we therefore cloned Skp into our system in parallel with FkpA. However, these Skp co-expression analyses showed only a minor effect in the context of scTCR 4B2A1-pIII display (data not shown), hence we chose to focus on the effect of FkpA only.

Our results clearly show that the beneficial effect of FkpA extends to increased productive display of both scTCR- and cFab-pIII. Both scTCR 4B2A1 V{alpha}ß and Vß{alpha} were affected by the FkpA over-expression, but in such a way that the display levels of the two variants became roughly equal. The effect of elevated FkpA levels during virion assembly was even more dramatic with the cFab 4B2A1. Here, display of the previously inferior H{alpha}Lß version increased several folds, and thus, this version became the most favorable, both regarding fusion mode and format. However, also in this case, the two domain orientations gave roughly similar display. In case of the dsTCR 4B2A1 format, there was also an enhancement in MAb reactivity, even if {alpha}-chain detection was still not achieved. In light of our negative findings, it is striking that the phage displayed human dsTCR reported by Li, et al. suffers from host down regulation of expression, which strongly indicates toxicity (Jakobsen et al., 2004Go; Li et al., 2005Go; Dunn et al., 2006Go).

Thus, it is apparent that even FkpA over-expression cannot restore productive TCR-display of all formats. Importantly, however, rFkpA over-expression resulted in improved display of both the scTCR and cFab based on the 7B2A1 TCR and the most productive display was as for 4B2A1, namely scTCR V{alpha}ß and cFab H{alpha}Lß. These two TCRs utilize completely different gene segments, which indicates that this strategy may be viable for display of many different TCRs. Differential binding to F9 versus A20HA control cells strongly indicates that these constructs are functional and can be selectively enriched even in a complex cellular panning context. This was true both with the scTCR and the cFab format (4B2A1), which gives flexibility with respect to further engineering strategies. Moreover, FkpA over-expression also greatly facilitates soluble periplasmic expression of both scTCR 4B2A1 and 7A10B2 (K.S.Gunnarsen et al., in preparation).

In summary, our results show for the first time that functional TCR-pIII display is feasible both in two-domain scTCR and four-domain cFab formats, but the display levels are highly dependent on domain orientation. Furthermore, both phage and phagemid systems may be compatible with scTCR-pIII display. Both productive scTCR and cFab-pIII display is enhanced in the presence of elevated FkpA levels, which may overcome display differences caused by variations in primary sequence. Our aim in the current analysis was to accommodate the TCR to the phage display platform in a functional form and the findings may have broad applicability in future use of TCR phage display.


   Supplementary data
Top
 Abstract
 Introduction
 Materials and Methods
 Results
 Discussion
 Supplementary data
Funding
 Acknowledgements
 References
 
Supplementary data are available at PEDS online


   Funding
Top
 Abstract
 Introduction
 Materials and Methods
 Results
 Discussion
 Supplementary data
 Funding
Acknowledgements
 References
 
This work was supported by the Norwegian Research Council (grant no. 155373/310).


   Footnotes
 
Edited by Jane Osbourne


   Acknowledgements
Top
 Abstract
 Introduction
 Materials and Methods
 Results
 Discussion
 Supplementary data
 Funding
 Acknowledgements
References
 
The authors would like to thank Sathiyaruby Manikam Vadivelu and Tom Ole Løvås for excellent technical assistance.


   References
Top
 Abstract
 Introduction
 Materials and Methods
 Results
 Discussion
 Supplementary data
 Funding
 Acknowledgements
 References
 
Acha-Orbea H., Zinkernagel R.M., Hengartner H. Eur. J. Immunol (1985) 15:31–36.[Web of Science][Medline]

Aharoni R., Teftelbaum D., Arnon R., Puri J. Nature (1991) 351:147–150.[CrossRef][Medline]

Bogen B., Lambris J.D. EMBO J (1989) 8:1947–1952.[Web of Science][Medline]

Bogen B., Malissen B., Haas W. Eur. J. Immunol (1986) 16:1373–1378.[Web of Science][Medline]

Bogen B., Lauritzsen G.F., Weiss S. Eur. J. Immunol (1990) 20:2359–2362.[Web of Science][Medline]

Bogen B., Gleditsch L., Weiss S., Dembic Z. Eur. J. Immunol (1992) 22:703–709.[Web of Science][Medline]

Bothmann H., Pluckthun A. Nat. Biotechnol (1998) 16:376–380.[CrossRef][Web of Science][Medline]

Bothmann H., Pluckthun A. J. Biol. Chem (2000) 275:17100–17105.[Abstract/Free Full Text]

Boulter J.M., Glick M., Todorov P.T., Baston E., Sami M., Rizkallah P., Jakobsen B.K. Protein Eng (2003) 16:707–711.[Abstract/Free Full Text]

Bradbury A.R., Marks J.D. J. Immunol. Methods (2004) 290:29–49.[CrossRef][Web of Science][Medline]

Brodnicki T.C., Holman P.O., Kranz D.M. Mol. Immunol (1996) 33:253–263.[CrossRef][Web of Science][Medline]

Chou H.S., Anderson S.J., Louie M.C., Godambe S.A., Pozzi M.R., Behlke M.A., Huppi K., Loh D.Y. PNAS (1987) 84:1992–1996.[Abstract/Free Full Text]

Clements C.S., Dunstone M.A., Macdonald W.A., McCluskey J., Rossjohn J. Curr. Opin. Struct. Biol (2006) 16:787–795.[CrossRef][Web of Science][Medline]

Dadaglio G., Nelson C.A., Deck M.B., Petzold S.J., Unanue E.R. Immunity (1997) 6:727–738.[CrossRef][Web of Science][Medline]

Davis S.J., van der Merwe P.A. Nat. Immunol (2006) 7:803–809.[CrossRef][Web of Science][Medline]

Davis S.J., Ikemizu S., Evans E.J., Fugger L., Bakker T.R., van der Merwe P.A. Nat. Immunol (2003) 4:217–224.[CrossRef][Web of Science][Medline]

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. J. Biol. Chem (1999) 274:18218–18230.[Abstract/Free Full Text]

Ding Y.H., Smith K.J., Garboczi D.N., Utz U., Biddison W.E., Wiley D.C. Immunity (1998) 8:403–411.[CrossRef][Web of Science][Medline]

Duc H.T., Rucay P., Righenzi S., Halle-Pannenko O., Kourilsky P. Int. Immunol (1993) 5:427–431.[Abstract/Free Full Text]

Dunn S.M., et al. Protein Sci (2006) 15:710–721.[CrossRef][Web of Science][Medline]

Fields B.A., Malchiodi E.L., Li H., Ysern X., Stauffacher C.V., Schlievert P.M., Karjalainen K., Mariuzza R.A. Nature (1996) 384:188–192.[CrossRef][Medline]

Fjeld J.G., Michaelsen T.E., Nustad K. Eur. J. Nucl. Med (1992) 19:402–408.[Web of Science][Medline]

Garcia K.C., Degano M., Stanfield R.L., Brunmark A., Jackson M.R., Peterson P.A., Teyton L., Wilson I.A. Science (1996) 274:209–219.[Abstract/Free Full Text]

Giordano R.J., Cardo-Vila M., Lahdenranta J., Pasqualini R., Arap W. Nat. Med (2001) 7:1249–1253.[CrossRef][Web of Science][Medline]

Glusman G., Rowen L., Lee I., Boysen C., Roach J.C., Smit A.F.A., Wang K., Koop B.F., Hood L. Immunity (2001) 15:337–349.[CrossRef][Web of Science][Medline]

Gronski M.A., et al. Nat. Med (2004) 10:1234–1239.[CrossRef][Web of Science][Medline]

Halaby D.M., Poupon A., Mornon J. Protein Eng (1999) 12:563–571.[Abstract/Free Full Text]

Hayday A.C., Diamond D.J., Tanigawa G., Heilig J.S., Folsom V., Saito H., Tonegawa S. Nature (1985) 316:828–832.[CrossRef][Medline]

Hoogenboom H.R. Nat. Biotechnol (2005) 23:1105–1116.[CrossRef][Web of Science][Medline]

Hulo N., Sigrist C.J.A., Le Saux V., Langendijk-Genevaux P.S., Bordoli L., Gattiker A., De Castro E., Bucher P., Bairoch A. Nucl. Acids Res. (2004) 32:D134–D137.[Abstract/Free Full Text]

Jakobsen B.K., Andersen T.B., Molloy P.E., Li Y., Boulter J.M. (2004) WO 2004/044004-A2.

Kanagawa O., Palmer E., Bill J. Cell Immunol (1989) 119:412–426.[CrossRef][Web of Science][Medline]

Kieke M.C., Shusta E.V., Boder E.T., Teyton L., Wittrup K.D., Kranz D.M. Proc. Natl Acad. Sci. USA (1999) 96:5651–5656.[Abstract/Free Full Text]

Kipriyanov S.M., Moldenhauer G., Little M. J. Immunol. Methods (1997) 200:69–77.[CrossRef][Web of Science][Medline]

Koch J., Breitling F., Dubel S. Biotechniques (2000) 29:1196–1198.[Web of Science][Medline]

Kretzschmar T., Geiser M. Gene (1995) 155:61–65.[CrossRef][Web of Science][Medline]

Kubo R.T., Born W., Kappler J.W., Marrack P., Pigeon M. J. Immunol (1989) 142:2736–2742.[Abstract]

Laemmli U.K. Nature (1970) 227:680–685.[CrossRef][Medline]

Laugel B., et al. J. Biol. Chem (2005) 280:1882–1892.[Abstract/Free Full Text]

Lauritzsen G.F., Weiss S., Dembic Z., Bogen B. PNAS (1994) 91:5700–5704.[Abstract/Free Full Text]

Lauvrak V., Berntzen G., Heggelund U., Herstad T.K., Sandin R.H., Dalseg R., Rosenqvist E., Sandlie I., Michaelsen T.E. Scand J. Immunol (2004) 59:373–384.[CrossRef][Web of Science][Medline]

Lefranc M.-P., Giudicelli V., Kaas Q., Duprat E., Jabado-Michaloud J., Scaviner D., Ginestoux C., Clement O., Chaume D., Lefranc G. Nucleic Acids Res (2005) 33:D593–D597.[Abstract/Free Full Text]

Levitsky H.I., Montgomery J., Ahmadzadeh M., Staveley-O'Carroll K., Guarnieri F., Longo D.L., Kwak L.W. J. Immunol (1996) 156:3858–3865.[Abstract]

Li Y., et al. Nat. Biotechnol (2005) 23:349–354.[CrossRef][Web of Science][Medline]

Loset G.A., Lobersli I., Kavlie A., Stacy J.E., Borgen T., Kausmally L., Hvattum E., Simonsen B., Hovda M.B., Brekke O.H. J. Immunol. Methods (2005) 299:47–62.[CrossRef][Web of Science][Medline]

Malchiodi E.L., Eisenstein E., Fields B.A., Ohlendorf D.H., Schlievert P.M., Karjalainen K., Mariuzza R.A. J. Exp. Med (1995) 182:1833–1845.[Abstract/Free Full Text]

Manning T.C., Schlueter C.J., Brodnicki T.C., Parke E.A., Speir J.A., Garcia K.C., Teyton L., Wilson I.A., Kranz D.M. Immunity (1998) 8:413–425.[CrossRef][Web of Science][Medline]

Marks J.D., Hoogenboom H.R., Bonnert T.P., McCafferty J., Griffiths A.D., Winter G. J. Mol. Biol (1991) 222:581–597.[CrossRef][Web of Science][Medline]

Marzec C.J., Day L.A. Biophys. J (1983) 42:171–180.[Web of Science][Medline]

Maynard J., Adams E.J., Krogsgaard M., Petersson K., Liu C.W., Garcia K.C. J. Immunol. Methods (2005) 306:51–67.[CrossRef][Web of Science][Medline]

Munthe L.A., Sollien A., Dembic Z., Bogen B. Scand. J. Immunol (1995) 42:651–661.[CrossRef][Web of Science][Medline]

Munthe L.A., Blichfeldt E., Sollien A., Dembic Z., Bogen B. Cell Immunol (1996) 170:283–290.[CrossRef][Web of Science][Medline]

Murphy D.B., Lo D., Rath S., Brinster R.L., Flavell R.A., Slanetz A., Janeway C.A. Nature (1989) 338:765–768.[CrossRef][Medline]

Norderhaug L., Olafsen T., Michaelsen T.E., Sandlie I. J. Immunol. Methods (1997) 204:77–87.[CrossRef][Web of Science][Medline]

Noy R., Eppel M., Haus-Cohen M., Klechevsky E., Mekler O., Michaeli Y., Denkberg G., Reiter Y. Expert Rev. Anticancer Ther (2005) 5:523–536.[CrossRef][Web of Science][Medline]

Onda T., LaFace D., Baier G., Brunner T., Honma N., Mikayama T., Altman A., Green D.R. Mol. Immunol (1995) 32:1387–1397.[CrossRef][Web of Science][Medline]

Porgador A., Yewdell J.W., Deng Y., Bennink J.R., Germain R.N. Immunity (1997) 6:715–726.[CrossRef][Web of Science][Medline]

Ramm K., Pluckthun A. J. Biol. Chem (2000) 275:17106–17113.[Abstract/Free Full Text]

Ramm K., Pluckthun A. J. Mol. Biol (2001) 310:485–498.[CrossRef][Web of Science][Medline]

Reay P.A., Matsui K., Haase K., Wulfing C., Chien Y.-H., Davis M.M. J. Immunol (2000) 164:5626–5634.[Abstract/Free Full Text]

Rojo J.M., Janeway C.A. Jr. J. Immunol (1988) 140:1081–1088.[Abstract]

Rondot S., Koch J., Breitling F., Dubel S. Nat. Biotechnol (2001) 19:75–78.[CrossRef][Web of Science][Medline]

Roodveldt C., Aharoni A., Tawfik D.S. Curr. Opin. Struct. Biol (2005) 15:50–56.[CrossRef][Web of Science][Medline]

Rudolph M.G., Stanfield R.L., Wilson I.A. Annu. Rev. Immunol (2006) 24:419–466.[CrossRef][Web of Science][Medline]

Scott J.K., Smith G.P. Science (1990) 249:386–390.[Abstract/Free Full Text]

Shusta E.V., Holler P.D., Kieke M.C., Kranz D.M., Wittrup K.D. Nat. Biotechnol (2000) 18:754–759.[CrossRef][Web of Science][Medline]

Smith G.P., Scott J.K. Methods Enzymol (1993) 217:228–257.[Web of Science][Medline]

Snodgrass H.R., Fisher A.M., Bruyns E., Bogen B. Eur. J. Immunol (1992) 22:2169–2172.[CrossRef][Web of Science][Medline]

Staerz U., Rammensee H., Benedetto J., Bevan M. J. Immunol (1985) 134:3994–4000.[Abstract]

Subbramanian R.A., et al. Nat. Biotechnol (2004) 22:1429–1434.[CrossRef][Web of Science][Medline]

Thomas J.G., Ayling A., Baneyx F. Appl. Biochem. Biotechnol (1997) 66:197–238.[Web of Science][Medline]

Weber S., Traunecker A., Oliveri F., Gerhard W., Karjalainen K. Nature (1992) 356:793–796.[CrossRef][Medline]

Weber K.S., Donermeyer D.L., Allen P.M., Kranz D.M. PNAS (2005) 102:19033–19038.[Abstract/Free Full Text]

Weidanz J.A., Card K.F., Edwards A., Perlstein E., Wong H.C. J. Immunol. Methods (1998) 221:59–76.[CrossRef][Web of Science][Medline]

Weiss S., Bogen B. PNAS (1989) 86:282–286.[Abstract/Free Full Text]

Welschof M., Terness P., Kipriyanov S.M., Stanescu D., Breitling F., Dorsam H., Dubel S., Little M., Opelz G. PNAS (1997) 94:1902–1907.[Abstract/Free Full Text]

Winoto A., Mjolsness S., Hood L. Nature (1985) 316:832–836.[CrossRef][Medline]

Zhong G., Reis e Sousa C., Germain R.N. PNAS (1997) 94:13856–13861.[Abstract/Free Full Text]

Received January 16, 2007; revised May 30, 2007; accepted July 6, 2007.


Add to CiteULike CiteULike   Add to Connotea Connotea   Add to Del.icio.us Del.icio.us    What's this?



This Article
Right arrow Abstract Freely available
Right arrow FREE Full Text (PDF) Freely available
Right arrow Supplementary Data
Right arrow All Versions of this Article:
20/9/461    most recent
gzm044v1
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in ISI Web of Science
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Add to My Personal Archive
Right arrow Download to citation manager
Right arrow Search for citing articles in:
ISI Web of Science (1)
Right arrowRequest Permissions
Google Scholar
Right arrow Articles by Løset, G.A.
Right arrow Articles by Sandlie, I.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Løset, G.A.
Right arrow Articles by Sandlie, I.
Related Collections
Right arrow 2007
Social Bookmarking
Add to CiteULike   Add to Connotea   Add to Del.icio.us  
What's this?