Transient DNA binding to gapped DNA substrates links DNA sequence to the single-molecule kinetics of protein-DNA interactions

Abstract

Protein interactions with nucleic acids are central to all genetic processes and many biotechnological applications. While many sequence-dependent protein-DNA interactions have been studied in detail using single-molecule methods, there is no standard high-throughput way to link the complex single-molecule kinetics of protein-DNA interactions with the DNA sequence of a single molecule. Here we provide the missing link by introducing a single-molecule imaging method (Gap-Seq) that interrogates DNA sequences via transient binding of short fluorescent DNA to a single DNA molecule previously used to characterise a protein-DNA interaction. In Gap-Seq, we identify a base by the degree of binding of 6-9 nt-long DNAs to surface-immobilised DNA substrates featuring a short single-stranded gap. To facilitate detection, we also developed a fluorescence quenching strategy that allows single-molecule detection at up to 500 nM of unbound fluorescent DNA. We link single-base differences on single DNA molecules to the kinetics of protein-DNA interactions by studying the interaction of a transcription activator with its cognate site. Finally, we show that our assay can address mixed sequences by distinguishing between two different sequences immobilised on the same field of view, paving the way for interrogation of sequence libraries for both mechanistic work and biotechnological applications.

Bleaching-resistant, near-continuous single-molecule fluorescence and FRET based on fluorogenic and transient DNA binding

“…A standard way to reduce the fluorescence background is through the use of an evanescent excitation field in a total-internal-reflection fluorescence (TIRF) microscope; however, this mode of microscope still cannot allow detection of single immobilised molecules in the presence of >50 nM of unbound label. [20][21][22][23] To further suppress the fluorescence background, we considered a strategy for quenching the fluorescence of unbound labels. This strategy uses a short ssDNA labelled with two ATTO647N fluorophores (one on either end) that exhibit contact-mediated quenching in solution; when bound to the target, the state of quenching is lifted, leading to the appearance of fluorescence corresponding to two ATTO647N fluorophores.…”

“…A standard way to reduce the fluorescence background is through the use of an evanescent excitation field in a total-internal-reflection fluorescence (TIRF) microscope; however, this mode of microscope still cannot allow detection of single immobilised molecules in the presence of >50 nM of unbound label. 20–23…”

“…The current design operates at 100 nM r-label and 100-ms exposures. To resolve faster processes, shorter exposures can be enabled by increasing the excitation power and using shorter r-labels at higher concentrations; currently, concentrations of up to 500 nM are feasible using quenching of unbound r-labels carrying two ATTO647N fluorophores 20 . In our traces, we also see binding of a single-labelled r-label (e.g., see top trace in Figure 4), which suggests the presence of a significant proportion of non-quenched labels in the buffer, caused either by incomplete synthesis and separation of single-from double-labelled r-labels, or by bleaching of unbound r-labels whilst illuminated on the microscope.…”

“…A standard way to reduce the fluorescence background is using an evanescent excitation field in a total-internal-reflection fluorescence (TIRF) microscope; however, this mode of microscope still cannot allow detection of single immobilised molecules in the presence of >50 nM of unbound label. [24][25][26][27] To further suppress the fluorescence background, we considered a fluorogenic strategy that quenches the fluorescence of unbound labels, but allows for substantial recovery of fluorescence upon label binding.…”

“…In principle, the dark interval between the binding of two transient labels to the same target can be decreased by increasing the rate of binding, which in turn can be achieved either by increasing the transient label concentration, or by changing the properties of the transient label to increase the on-rate constant or both. However, the concentration of fluorescent transient labels cannot be increased much above 30 nΜ, as unbound labels contribute to the fluorescence background and degrade the SNR of the measurement [20][21][22][23] , and this concentration will not allow for a high enough on-rate of short ssDNAs for continuous fluorescence traces.…”