High-affinity target binding engineered via fusion of a single-domain antibody fragment with a ligand-tailored SH3 domain

Abstract

Monoclonal and recombinant antibodies are ubiquitous tools in diagnostics, therapeutics, and biotechnology. However, their biochemical properties lack optimal robustness, their bacterial production is not easy, and possibilities to create multifunctional fusion proteins based on them are limited. Moreover, the binding affinities of antibodies towards their antigens are suboptimal for many applications where they are commonly used. To address these issues we have made use of the concept of creating high binding affinity based on multivalent target recognition via exploiting some of the best features of immunoglobulins (Ig) and non-Ig-derived ligand-binding domains. We have constructed a small protein, named Neffin, comprised of a 118 aa llama Ig heavy chain variable domain fragment (VHH) fused to a ligand-tailored 57 aa SH3 domain. Neffin could be readily produced in large amounts (>18 mg/L) in the cytoplasm of E. coli, and bound with a subpicomolar affinity (K(d) 0.54 pM) to its target, the HIV-1 Nef protein. When expressed in human cells Neffin could potently inhibit Nef function. Similar VHH-SH3 fusion proteins could be targeted against many other proteins of interest and could have widespread use in diverse medical and biotechnology applications where biochemical robustness and strong binding affinity are required.

Figure 1. Design of a VHH-SH3 fusion protein (Neffin) targeted against HIV-1 Nef.

(A) Size exclusion chromatography analysis confirming the expected molecular sizes for the monomeric Nef, sdAb19, and SH3-B6 proteins and for the dimeric sdAb19/Nef and the trimeric sdAb19/Nef/SH3-B6 complexes. (B) Domain organization of Neffin and the amino acid sequences of the different linkers tested for joining of sdAb19 and SH3-B6.

Figure 2. Bacterial expression of Neffin.

(A) sdAb19 and Neffin were expressed in the cytoplasm of E. coli Origami cells in 50 ml flask cultures and captured to 0.2 ml of nickel-agarose resin. A Coomassie blue–stained gel containing 10 µl aliquots of the first three 0.5 ml fractions (F1–F3) of sdAb19 and Neffin eluated from the resin is shown. (B) Comparison of Nef-binding capacity of Neffin produced in BL21 or Origami cells. 10 µg (lanes 1 and 3) or 5 µg (lanes 2 and 4) of BL21- or Origami-derived Neffin were incubated with 10 µg GST-Nef or plain GST. Equal fraction of proteins captured to glutathione-resin well as input material were analyzed by SDS-PAGE and Coomassie blue–staining.

Figure 3. Estimation of binding affinity and kinetics for the Nef-Neffin interaction.

(A) Biacore sensorgrams used for calculating the on-rates 1.41×10⁶ M⁻¹s⁻¹ and 1.51×10⁶ M⁻¹s⁻¹ for binding of sdAb19 and Neffin to Nef. The off-rates for both interaction were too slow to be reliably estimated by this method. (B). Competitive ELISA for estimation of the off-rates of 2.17×10⁻⁵ s⁻¹ and 8.12×10⁻⁷ s⁻¹for binding of sdAb19 and Neffin to Nef.

Figure 4. Efficient association of Neffin and Nef expressed in human cells.

GFP-tagged Nef was co-expressed in 293 T cells together with sdAb19, Neffin-B6 or Neffin-C1, and the amount of sdAb19, Neffin-B6, or Neffin-C1 associated with anti-GFP immunocomplexes was determined by Western blotting (top panel). Even precipitation of GFP-Nef as well as equal expression of Nef, sdAb19, Neffin-B6, and Neffin-C1 in the lysates of the transfected cell cultures is shown as indicated. Blotting of the lysates with an antibody for the endogenous α-tubulin is used as a loading control. (bottom panel).

Figure 5. Potent inhibition of Nef-induced Hck activation by Neffin.

An expression vector for the tyrosine kinase Hck was co-transfected to 293 T cell with an empty control vector (left lane) or with an expression vector for HIV-1 Nef (other lanes). In Nef-transfected cells a vector for sdAb19, Neffin-B6, or Neffin-C1 was included as indicated. Lysates of these cells were subjected to Hck-pulldown and the amount of activated (pHck; top panel) and total Hck (second panel) these precipitates determined by Western blotting. Corresponding amounts of Nef, sdAb19, Neffin-B6, and Neffin-C1 in the total lysates of the transfected cells was confirmed (third and bottom panels). A quantitation of relative Hck activation in the transfected cells normalized to the total amount of precipitated Hck is shown on the right, where the amount of autophosphorylated Hck in cells lacking Nef is set to one, and the phospho-Hck blotting signals from the other cells are graphed relative to this.