The diagnosis, staging and clinical management of cancer and other diseases is becoming increasingly reliant upon the identification and quantification of molecular markers as well their spatial distribution in histological samples. Yet, due to spectral overlap of dyes and the inability to remove probes without affecting marker integrity, immunohistological methods are limited by the number of markers that can be examined on a single specimen resulting in loss of information that could be vital to diagnosis or treatment. This dissertation describes the development and characterization of an erasable multi-color imaging technology capable of detecting large numbers of molecular markers on a single biological sample. The system consists of (1) 'targets', which are single or partially hybridized DNA strands conjugated to a protein of interest for biomarker recognition in cells, and (2) multi-strand, fluorophore-containing DNA 'probe complexes' that react with the DNA portion of the target in a sequence dependent fashion to create fluorescent reporting complexes. The addition of a quencher-bearing ssDNA displaces the target's DNA strand to effectively remove the dye from the marker so that the sample can be re-imaged for other markers with minimal interference from prior iii rounds of labeling. Orthogonal DNA sequences and spectrally-separated dyes can be used to create multiple, unique target/probe pairs that associate specifically and can be imaged in parallel. The overall utility of this technology depends on high specificity of targets to respective probe complexes, highly efficient labeling and erasing to ensure that fluorescent signals can be used to fully quantify target abundance without the interference of signals from previous rounds of labeling, and short reaction times to allow for multiple rounds of processing on the same sample without loss of integrity. Based on the above criteria, three classes of probes were designed and their structure-function relationships elucidated to determine the contributions of complex size, free energy differences between intermediate states, and strand displacement on labeling and erasing kinetics and efficiencies on cells. A comparison of the kinetics of the labeling and erasing reactions for the three different constructs showed that reaction efficiencies depend less on calculated net free energy change than on the engineered state of the complex during the strand displacement reaction (i.e., the type of strand displacement reaction it participates in). This new paradigm in probe design allowed the system to meet its design goals, potentially increasing the diagnostic power of individual histological specimens and opening the door to more sophisticated analyses of cell phenotype and its functional relationship to disease.