Nanobody-Displaying Flagellar Nanotubes

Abstract

In this work we addressed the problem how to fabricate self-assembling tubular nanostructures displaying target recognition functionalities. Bacterial flagellar filaments, composed of thousands of flagellin subunits, were used as scaffolds to display single-domain antibodies (nanobodies) on their surface. As a representative example, an anti-GFP nanobody was successfully inserted into the middle part of flagellin replacing the hypervariable surface-exposed D3 domain. A novel procedure was developed to select appropriate linkers required for functional internal insertion. Linkers of various lengths and conformational properties were chosen from a linker database and they were randomly attached to both ends of an anti-GFP nanobody to facilitate insertion. Functional fusion constructs capable of forming filaments on the surface of flagellin-deficient host cells were selected by magnetic microparticles covered by target GFP molecules and appropriate linkers were identified. TEM studies revealed that short filaments of 2-900 nm were formed on the cell surface. ITC and fluorescent measurements demonstrated that the fusion protein exhibited high binding affinity towards GFP. Our approach allows the development of functionalized flagellar nanotubes against a variety of important target molecules offering potential applications in biosensorics and bio-nanotechnology.

Figure 1

Construction of flagellar nanotubes displaying single-domain antibodies. (a) Arrangement of flagellin subunits within the flagellar filament. The outer hypervariable D3 portions (circled) of flagellin subunits, situated on the filament surface, were replaced by single-domain antibodies. The distance between the solvent-exposed nanobodies on the filament surface is about 5.5 nm. Solid surface representation of a longitudinal section of the filament according to Mimori-Kiyosue et al. . (Copyright (1996) National Academy of Sciences, USA.) (b) The Ca backbone trace of flagellin (PDB code: 1UCU) and aGFP_ENH nanobody (PDB code: 3K1K). The aGFP_ENH nanobody with various oligopeptide linkers attached to both ends was inserted into the middle part of flagellin to replace the D3 domain. (c) Representation of the domain arrangement of flagellin. D0-D2 are discontinuous domains formed by segments from both the N- and C-termini. D3 is constructed by the middle portion of the polypeptide chain consisting of residues 190–283. This internal segment was removed and replaced by a single-domain antibody. (d) Oligonucleotide segments coding for the chosen oligopeptide linkers (L₁, L₂, …) were joined via segments encoding the cleavage sites of XhoI, SacI or AgeI restriction enzymes to form a multilinker gene. Arrows show the local direction of the reading frame. In the regions indicated by white and black arrows linker segments are coded by the opposite strands of DNA. The multilinker gene was synthesized and fragmented by a mixture of restriction enzymes (XhoI, SacI, AgeI) to obtain linker-encoding segments (shown in the same orientation relative to the reading frame) with sticky ends facilitating construction of the FliC-aGFP_ENH fusion genes (see text).

Figure 2

SDS-PAGE analysis of the filament-forming and secreted monomeric fractions of the FliC-aGFP_ENH fusion proteins, overexpressed in flagellin-deficient SJW2536 Salmonella host cells. For comparison, similar fractions from cells expressing the D3-deleted flagellin variant (FliC-delD3) were also run on the gel. Lanes 2, 5, 7 are the secreted monomeric fractions for FliC-delD3, FliC-aGFP_ENH_V1 and FliC-aGFP_ENH_V2, respectively. Lanes 1, 4, 6 show the amount of protein subunits for FliC-delD3, FliC-aGFP_ENH_V1 and FliC-aGFP_ENH_V2, respectively, forming filaments on the cell surface. The intensity of the bands suggests that the fusion proteins were mainly secreted into the culture medium and only a smaller fraction formed filaments. The calculated molecular masses of FliC-delD3, FliC-aGFP_ENH_V1 and FliC-aGFP_ENH_V2 are 42.8 kDa, 58.6 kDa and 57.7 kDa, respectively.

Figure 3

Flagellar nanotube formation from FliC-aGFP_ENH_V2 subunits in vivo and in vitro. (a) Bright-field TEM image of Salmonella cells that possess mutant flagellar filaments. Samples were stained with 2% phosphotungstate to enhance image contrast. (b) Filaments polymerized from purified subunits as visualized by dark-field optical microscopy. Polymerization experiments were carried out in PBS buffer (pH 7.4) at 2 mg/ml protein concentration and 4 M AS was added to 0.6 M final concentration to initiate filament formation.

Figure 4

Interaction of the FliC-aGFP_ENH fusion protein with sfGFP. (a) The fluorescence emission spectrum of sfGFP (solid) showed an emission maximum at 508 nm. Addition of the FliC-aGFP_ENH fusion protein (dashed) resulted in a nearly 60% increase of fluorescence intensity. As a control experiment, wild-type flagellin was also added (dotted) but it produced only a small effect on sfGFP fluorescence. Measurements were taken in PBS buffer (pH 7.4) at an excitation wavelength of 488 nm. (b) Isothermal calorimetric titration of FliC-aGFP_ENH with sfGFP at 25 °C. The FliC-aGFP_ENH sample (c = 0.006 mM) was loaded into the cell and the sfGFP (c = 0.065 mM) solution was injected in 10 μl portions. Changes in binding enthalpy (▪) of the injections are shown as a function of the molar sfGFP to FliC-aGFP_ENH ratio. The solid line is the least-squares fit to the data by using a one-binding-site model, resulting in the following parameters: stoichiometry N = 0.86, dissociation constant K = 1.1·10⁻⁹ M, binding enthalpy ΔH = −9.3 kcal/mol. Titrations were done in 20 mM Tris-HCl 150 mM NaCl, pH 8.5, buffer.