The transcriptome ‒ the total collection of every RNA transcript in a cell ‒ provides a unique readout of a cell’s commands as it executes its own genetic code. Studying the transcriptomes of individual cells is a powerful way to identify new cell types, building a better understanding of complex tissues from the ground up. Furthermore, transcriptomic characterization of tumor samples is already leading to tangible advancements in personalized cancer care. However, the process of capturing a transcriptome requires careful isolation of mRNA from a single cell that is potentially entrenched in an entangled, three-dimensional tissue structure. Our lab has previously developed a method to successfully isolate mRNA from single cells still contextualized within living tissue. This method, Transcriptome In Vivo Analysis (TIVA), utilizes a light-activatable oligonucleotide probe to offer fine spatio-temporal control of mRNA capture. TIVA probes are highly modified RNA hairpins or loops incorporating a poly(U) “capture” sequence complementary to the poly(A) tail of mRNA, as well as a biotin moiety to enable pull-down of bound mRNA. To enable spatio-temporal control of mRNA binding, the probes are locked into an inactive “caged” conformation. Laser excitation of the target cell severs photoactivable o-nitrobenzyl- or Ru(II) polypyridyl-based linkages built into the probe, freeing the poly(U) capture sequence to effectively biotinylate mRNA. After photolysis, a small tissue region containing the target cell is aspirated out and lysed so that the target mRNA can be isolated by streptavidin-biotin affinity purification. In this dissertation I present the synthesis, characterization, and application of next-generation TIVA constructs that aim to address various limitations of the original probe. (I) I demonstrate that phosphorothioation extends the serum stability of the TIVA probe from a few minutes to over 24 hours, and mediates uptake into cells without the need for a cell-penetrating peptide. (II) I also present versions of the probe with hairpin-terminating GC pairs, longer blocking strands, and a pegylated hairpin turn which dramatically reduce the probe’s pre-photolysis background binding of mRNA. (III) Finally, I show that incorporation of a Ru(II) polypyridyl-based photocleavable linker extends the probe’s activation response from 1-photon near-UV light to 2-photon near-IR light. Together, these advancements move TIVA towards a broader range of applications – including deeper regions of challenging, nuclease-abundant tissues – with greater confidence that our construct will remain stable and generate a low-background transcriptome.