A Modular, DNA-Based Beacon for Single-Step Fluorescence Detection of Antibodies and Other Proteins

Abstract

A versatile platform for the one-step fluorescence detection of both monovalent and multivalent proteins has been developed. This system is based on a conformation-switching stem-loop DNA scaffold that presents a small-molecule, polypeptide, or nucleic-acid recognition element on each of its two stem strands. The steric strain associated with the binding of one (multivalent) or two (monovalent) target molecules to these elements opens the stem, enhancing the emission of an attached fluorophore/quencher pair. The sensors respond rapidly (<10 min) and selectively, enabling the facile detection of specific proteins even in complex samples, such as blood serum. The versatility of the platform was demonstrated by detecting five bivalent proteins (four antibodies and the chemokine platelet-derived growth factor) and two monovalent proteins (a Fab fragment and the transcription factor TBP) with low nanomolar detection limits and no detectable cross-reactivity.

Keywords: DNA nanotechnology; antibodies; aptamers; fluorescence; molecular devices.

Figure 1

Our protein-targeting sensor is composed of a fluorophore/quencher-modified DNA stem-loop containing two single-stranded tails. To create a target-responsive sensor, these tails are hybridized with DNAs conjugated to the appropriate recognition element (red hexagons). A frame inversion at one tail-stem junction ensures that the two tails meet “head-to-head” (3′-end-to-3′-end), thus allowing a single recognition-element modified strand sequence to populate both recognition sites. The binding of a bivalent macromolecule (here an antibody) to the two recognition elements opens the stem, allowing for rapid and sensitive protein detection. As shown later, the binding of two copies of a target is also sufficient to cause stem opening, allowing for the design of switches that respond to monovalent targets.

Figure 2

(A) As proof-of-principle we employed digoxigenin (Dig) as a recognition element for the detection of anti-Dig antibodies. (B) We initially tested sensors varying in stem stability due to variations in stem GC content (2, 3 and 4 GC) and/or loop length (13 to 20 bases). (C) Melting curves obtained in the absence of the target illustrate their varying stabilities. (D) Comparison with curves obtained at 100 nM target (one is shown) provides a means of measuring the gain of each variant as a function of temperature. (Under these conditions antibody binding is effectively temperature independent; Figure S7.) (E) Sensors of intermediate stability exhibit the best compromise between gain and affinity, with a 3GC stem and a 15-base poly-T loop (variant #4) proving optimal. (F) The optimal sensor detects anti-Dig antibodies at low nanomolar concentrations. (G) K_1/2 changes with varying sensors’s concentration in precisely the manner expected for a 1:1 binding stoichiometry, thus supporting the proposed sensing principle. The experiments shown here and in the following figures represent averages of three measurements; error bars reflect standard deviations. The binding and melting curves here and in the following figures were performed in 50 mM Na₂HPO₄, 150 mM NaCl, pH 7.0, with the nanoswitch at 10 nM unless otherwise noted.

Figure 3

Our modular platform is versatile and can easily be adapted to the detection of new targets via the expedient of changing the recognition element employed. Here we demonstrate this by using three different antigens recognized by specific antibodies: (A) dinitrophenol (DNP), (B) the 8-residue FLAG peptide, and (C) a 13-residue epitope excised from the HIV protein p17. (D) Using a 35-base aptamer as recognition element the platform can also be used to detect the bivalent chemokine PDGF. All four sensors respond to their specific target at low nanomolar concentrations whilst exhibiting no significant response to high concentrations of the other switches’ targets

Figure 4

(Top) Our platform can be also used for the detection of monovalent targets. In this case the signal change arises because the simultaneous binding of two copies of the target causes steric hindrance, opening the stem. Shown are sensors detecting a monovalent anti-Dig Fab fragment (bottom left) and the monovalent DNA-binding transcription factor TBP (bottom right). Both sensors readily detect their specific target at nanomolar concentrations.

Figure 5

Modular sensors can serve as a molecular AND-logic gate that signals only in the simultaneous presence of two different macromolecular targets. (Top) To demonstrate this we modified a sensor with the recognition elements Dig and DNP. (Bottom) Only in the simultaneous presence of both anti-Dig and anti-DNP antibodies we observe any significant signal increase.