Single-chain Fv phage display propensity exhibits strong positive correlation with overall expression levels

Abstract

BackgroundSingle chain Fvs (scFvs) are widely applied in research, diagnostics and therapeutic settings. Display and selection from combinatorial libraries is the main route to their discovery and many factors influence the success of this process. They exhibit low thermodynamic stability, resulting in low levels of premature cytosolic folding or aggregation which facilitates sec YEG-mediated translocation and phage in E. coli. However, there is little data analysing how this is related to and influenced by scFv protein expression.ResultsWe characterised the relationship between overall scFv expression and display propensity for a panel of 15 anti-tetanus toxin scFvs and found a strong positive correlation (Rho = 0.88, p < 0.005) between the two parameters. Display propensity, overall expression and soluble localisation to the periplasm and extracellular fractions were clone specific characteristics which varied despite high levels of sequence homology. There was no correlation between display of scFv or its expression in non-fused (free) form with soluble scFv localisation to the periplasm or culture supernatant. This suggests that divergence in the fate of scFv-pIII and non-fused scFv after translocation to the periplasm accounts for the observed disparity. Differential degrees of periplasmic aggregation of non-fused scFv between clones may affect the partitioning of scFv in the periplasm and culture supernatant abrogating any correlation. We suggest that these factors do not apply to the scFv-pIII fusion since it remains anchored to the bacterial inner membrane as part of the innate phage packaging and budding process.ConclusionWe conclude that in the absence of premature cytosolic aggregation or folding, the propensity of a scFv to be displayed on phage is directly related to its overall expression level and is thus indirectly influenced by factors such as codon bias, mRNA abundance or putative DNA motifs affecting expression. This suggests that scFvs capable of high overall expression and display levels may not produce high yields of non phage-fused soluble protein in either the periplasmic or extracellular fractions of E. coli. This should be considered when screening clones selected from combinatorial libraries for further study.The nucleotide and amino acid sequences of the anti-tetanus toxin scFvs have been deposited in the EMBL data base: accession numbers-C1: AM749134, C2: AM749135, C3: AM749136, C4: AM749137, C5: AM749138, N1: AM749139, N2: AM749140, N3: AM749141, N4: AM749142, N5: AM749143 J1; AM749144, J2: AM749145, J3: AM749146, J4: AM749147, J5: AM749148.

Antibody isolation from immunized animals: comparison of phage display and antibody discovery via V gene repertoire mining

“…To name a few: biases begin already during library construction (Sblattero and Bradbury, 2000;Holt et al, 2000) as a result of PCR-based bias in amplification of different templates, during the subsequent PCR amplifications used for assembly of scFvs and due to the fact that actual library size falls short of covering all possible combinations of V H and V L domains present in the cDNA. Second, some scFvs display better on phage than others (Malone and Sullivan, 1996;Pavoni et al, 2007;Scott et al, 2008), as is shown in our evaluation of the three HTS-chosen clones that displayed very poorly on phage. Third, the affinity selection process itself is very sensitive to varying conditions and experimental variations (Hoogenboom et al, 1999;de Wildt et al, 2000;Smith et al, 2005), frequently leading to different antibodies being isolated for a particular antigen by different individuals or even by the same individual under slightly varying experimental conditions.…”

Characterisation of a panel of anti-tetanus toxin single-chain Fvs reveals cooperative binding

“…They were also compared in terms of overall expression levels, soluble localisation to the periplasm and culture supernatant as well as their propensity for display on phage, as reported separately (Scott et al, 2008). …”

Virus-based assay for antigen detection using infective growth as signal transduction mechanism

“…The zz-phage were constructed based on a phagemid/helper phage system utilizing a modified pCantab6 amp r phagemid (McCafferty et al 1996) and were rescued using M13KO7 helper phage. The pCantab6 phagemid p3 gene was genetically modified to contain the zz domain near the amino terminus after the leader signal sequence for SecYEG mediated periplasmic translocation (Scott et al 2008). The insert containing the zz domain was obtained by PCR of a TAP tag insertion cassette derived from pFA6a-TAP-His3MX (Ghaemmaghami et al 2003) using the following primers (zz for: ACAACTG-CAGGTGGACAACAAATTCAACAAAGAACAAC; zz rev: TATTGC GGCCGCCTGATGATTCGCGTCTACTTTCG) to generate PstI and NotI restriction sites, respectively.…”