Protein analysis by time-resolved measurements with an electro-switchable DNA chip

Abstract

Measurements in stationary or mobile phases are fundamental principles in protein analysis. Although the immobilization of molecules on solid supports allows for the parallel analysis of interactions, properties like size or shape are usually inferred from the molecular mobility under the influence of external forces. However, as these principles are mutually exclusive, a comprehensive characterization of proteins usually involves a multi-step workflow. Here we show how these measurement modalities can be reconciled by tethering proteins to a surface via dynamically actuated nanolevers. Short DNA strands, which are switched by alternating electric fields, are employed as capture probes to bind target proteins. By swaying the proteins over nanometre amplitudes and comparing their motional dynamics to a theoretical model, the protein diameter can be quantified with Angström accuracy. Alterations in the tertiary protein structure (folding) and conformational changes are readily detected, and even post-translational modifications are revealed by time-resolved molecular dynamics measurements.

Figure 1. Time-resolved DNA switching measurement.

(a) The biochip features 24 circular Au microelectrodes (∅=120 μm), which are individually addressable with different receptors for target molecules via DNA-directed assembly. Six microelectrodes are arranged together with one Pt counter electrode in four flow channels (not shown), which are connected via eight holes in the glass substrate. White scale bars, 10 mm and 1 mm (inset). (b) An epifluorescence setup is used for the optical detection of DNA orientation in real time. A frequency generator supplies alternating voltages to the electrodes. This actuates the DNA levers and simultaneously triggers an event timer that records single photons emitted by the fluorescently labelled DNA. (c) Schematic of the electrically induced DNA switching. The fluorophore emission (depicted as a star) is quenched in close proximity to the surface. (d) Square wave voltage, which is applied to the Au electrodes, versus Pt (10 kHz) and (e) time-resolved fluorescence response of the 48 bp Cy3-labelled DNA layer.

Figure 2. Protein detection and size analysis.

(a) Schematic of a DNA lever with a protein (effective protein diameter D_H) bound via a capture molecule (purple), which is covalently attached to one single strand. Dashed lines indicate the ‘Lollipop’ silhouette used for the calculation. (b,c) Binding of the enzyme DHFR (21 kDa) to the anticancer drug MTX, which is attached to the DNA’s top end. (b) Fluorescence response during upward switching for MTX-modified DNA without and with bound DHFR. Areas under the curves denote the DR between 0 and 3 μs, which is a parameter gauging the swiftness of the motion. (c) Real-time binding and dissociation of 10 nM DHFR to MTX-modified DNA levers. The solid lines are single exponential fits to the data. Kinetic rate constants of k_on=(1.25±0.09) · 10⁵ M⁻¹ s⁻¹ and k_off=(5.7±1.0) · 10⁻⁴ s⁻¹ are obtained, yielding a dissociation constant of K_D=(4.6±0.8) nM. (d) Upward switching fluorescence response for bare DNA and three DNA–protein complexes. The His₆-tagged proteins ubiquitin (Ub, 8.5 kDa), chloramphenicol acetyl transferase (CAT, 29 kDa) and protein L (43 kDa) were captured via NTA₃-tagged DNA. Dashed lines are calculated with the theoretical Lollipop model. (e) Comparison of protein diameters analysed from switching dynamics experiments using the Lollipop model with DLS (circles) or literature values from X-ray diffraction structure data (squares). Cytochrome-C (Cyt-C, 12 kDa), interferon-α (IFNα, 18 kDa), carbonic anhydrase (CA, 29 kDa), IgG-binding recombinant proteins A (42 kDa) and G (26 kDa), protein kinase ERK2 (ref. 45) (42 kDa), Fab fragment (50 kDa). Representative structures of four proteins with D_H≈2–6 nm are drawn to scale on the right. Black scale bar, 2 nm.

Figure 3. Mixture and conformation analyses.

(a,b) Mixture analysis of antibody solutions. (a) Time-resolved upward motions of biotinylated DNA levers before and after binding antibiotin full IgGs and IgG fragments (Fab) from solutions containing different mixing ratios. (b) Comparison of measured and preadjusted mixing ratios. Data were analysed from DR values evaluated between 3 and 9 μs. (c–e) Unfolding of protein A. (c) Relative DR values for protein A-modified DNA levers in urea (which acts as a denaturation agent) and glycerin (as a negative control) as a function of solution viscosity. Values are normalized by the DR in urea- and glycerin-free buffer. Lines are parallel linear fits. All errors for glycerin are smaller than the symbols. (d) Differential analysis of protein A-modified samples reveals the disintegration of tertiary protein structure (shaded region). (e) Circular dichroism measurements at 220 nm indicate unfolding of secondary structures of protein A at urea concentrations above 2 M. (f,g) Conformation analysis of CaM. (f) 1 μM His₆-tagged CaM is bound to NTA₃-modified DNA. Ion-free CaM (blue circles) and CaM mixed with 100 μM MgCl₂ (cyan triangles) show identical binding curves, whereas the addition of 100 μM CaCl₂ (black arrow, purple diamonds) causes an additional decrease in DR. (g) Comparison of the fluorescence response during the upward switching cycle for saturated layers of Ca²⁺-free CaM and CaM with bound Ca²⁺. All error bars represent s.d. derived from measurements on three electrodes.

Figure 4. Post-translational modifications.

(a,b) Detection of the deglycosylated state of the protein hCGβ. (a) Upward switching fluorescence response for conjugates of DNA with deglycosylated (+PNGase) and glycosylated forms (−PNGase, untreated) of hCGβ. (b) Average upward DR values for the conjugates (n=3). (c,d) Detection of the phosphorylation state of the protein kinase ERK2. (c) Downward switching fluorescence response of NTA₃-tagged DNA levers before and after binding His₆-tagged ERK2 and phosphorylated ERK2-P. (d) Average downward DR values of unmodified (negative control, NC) DNA levers and NTA₃-tagged DNA levers after exposure to ERK2 and ERK2-P, respectively (n=6). Values are normalized by the DR of bare DNA levers.