Cells at disease onset are often associated with subtle
changes
in the expression level of a single or few molecular components, making
traditionally used biomarker-driven clinical diagnosis a challenging
task. We demonstrate here the design of a DNA nanosensor array with
multichannel output that identifies the normal or pathological state
of a cell based on the alteration of its global proteomic signature.
Fluorophore-encoded single-stranded DNA (ssDNA) strands were coupled
via supramolecular interaction with a surface-functionalized gold
nanoparticle quencher to generate this integrated sensor array. In
this design, ssDNA sequences exhibit dual roles, where they provide
differential affinities with the receptor gold nanoparticle as well
as act as transducer elements. The unique interaction mode of the
analyte molecules disrupts the noncovalent supramolecular complexation,
generating simultaneous multichannel fluorescence output to enable
signature-based analyte identification via a linear discriminant analysis-based
machine learning algorithm. Different cell types, particularly normal
and cancerous cells, were effectively distinguished using their fluorescent
fingerprints. Additionally, this DNA sensor array displayed excellent
sensitivity to identify cellular alterations associated with chemical
modulation of catabolic processes. Importantly, pharmacological effectors,
which could modulate autophagic flux, have been effectively distinguished
by generating responses from their global protein signatures. Taken
together, these studies demonstrate that our multichannel DNA nanosensor
is well suited for rapid identification of subtle changes in a complex
mixture and thus can be readily expanded for point-of-care clinical
diagnosis, high-throughput drug screening, or predicting the therapeutic
outcome from a limited sample volume.