Digital Maskless Photolithographic Patterning of DNA-Functionalized Poly(ethylene glycol) Diacrylate Hydrogels with Visible Light Enabling Photodirected Release of Oligonucleotides

Abstract

Soft biomaterials possessing structural hierarchy have growing applications in lab-on-chip devices, artificial tissues, and micromechanical and chemomechanical systems. The ability to integrate sets of biomolecules, specifically DNA, within hydrogel substrates at precise locations could offer the potential to form and modulate complex biochemical processes with DNA-based molecular switches in such materials and provide a means of creating dynamic spatial patterns, thus enabling spatiotemporal control of a wide array of reaction-diffusion phenomena prevalent in biological systems. Here we develop a means of photopatterning two-dimensional DNA-functionalized poly­(ethylene glycol) diacrylate (PEGDA) hydrogel architectures with an aim toward these applications. While PEGDA photopatterning methods are well-established for the fabrication of hydrogels, including those containing oligonucleotides, the photoinitiators typically used have significant crosstalk with many UV-photoswitchable chemistries including nitrobenzyl derivatives. We demonstrate the digital photopatterning of PEGDA-co-DNA hydrogels using a blue light-absorbing (470 nm peak) photoinitiator system and macromer comprised of camphorquinone, triethanolamine, and poly­(ethylene glycol) diacrylate (M n = 575) that minimizes absorption in the UV-A wavelength range commonly used to trigger photoswitchable chemistries. We demonstrate this method using digital maskless photolithography within microfluidic devices that allows for the reliable construction of multidomain structures. The method achieves feature resolutions as small as 25 μm, and the resulting materials allow for lateral isotropic bulk diffusion of short single-stranded (ss) DNA oligonucleotides. Finally, we show how the use of these photoinitiators allows for orthogonal control of photopolymerization and UV-photoscission of acrylate-modified DNA containing a 1-(2-nitrophenyl) ethyl spacer to selectively cleave DNA from regions of a PEGDA substrate.

DNA Reaction–Diffusion Attractor Patterns

“…Using the fit parameters and literature rate constants, the model recapitulated the shape and formation timescale of the hill gradients. The fitted rate constants were within one order of magnitude of literature values for rates of standard strand displacement reactions with the respective toehold lengths; [29] fitted diffusion coefficients were also within an order of magnitude of previously measured DNA diffusion coefficients in 30 % (v/v) PEGDA hydrogels [28] (SI: Table S5). The predicted timescale of formation was 4.3 h, which was within a factor of 2 of the experimental timescale of 6.7 h. We simulated pattern recovery using the rate constants and diffusion coefficients as model inputs, and the normalized fluorescence intensity profile measured after perturbation as an initial condition.…”

“…TheD NA reaction-diffusion processes that we design operate within hydrogels ( Figure 2);atthe hydrogel boundaries the concentrations of input species are actively maintained at constant levels.T hese active boundaries serve as as ource of chemical energy,a llowing reaction-diffusion processes to remain far from equilibrium for arbitrarily long times.W ea chieved these conditions within am icrofluidic device containing diffusion cells of 1500 mml ength, 50 mm width, and 20 mmh eight. [28] Reactants were supplied and waste was removed by pressure-balanced flows perpendicular to the diffusion cells.P erturbations were induced via UV photocleavage of a1 -(2-nitrophenyl) ethyl linker within the phosphodiester backbone of ad ouble stranded (ds) competitor complex, exposing previously occluded toeholds that allow binding and degradation of the patterned species.30% (v/v) poly(ethylene glycol) diacrylate (PEGDAM n = 575) was photopatterned with ac amphorquinone (470 nm excitation) photoinitiator [28] within each diffusion cell as amedium for reaction-diffusion that minimizes convection. Camphorquinone does not strongly absorb UV light, making it possible to trigger the UV-photosensitive reaction.…”

“…These active boundaries serve as a source of chemical energy, allowing reaction‐diffusion processes to remain far from equilibrium for arbitrarily long times. We achieved these conditions within a microfluidic device containing diffusion cells of 1500 μm length, 50 μm width, and 20 μm height [28] . Reactants were supplied and waste was removed by pressure‐balanced flows perpendicular to the diffusion cells.…”

“…Perturbations were induced via UV photocleavage of a 1‐(2‐nitrophenyl) ethyl linker within the phosphodiester backbone of a double stranded (ds) competitor complex, exposing previously occluded toeholds that allow binding and degradation of the patterned species. 30 % (v/v) poly(ethylene glycol) diacrylate (PEGDA Mn=575) was photopatterned with a camphorquinone (470 nm excitation) photoinitiator [28] within each diffusion cell as a medium for reaction‐diffusion that minimizes convection. Camphorquinone does not strongly absorb UV light, making it possible to trigger the UV‐photosensitive reaction.…”

DNA Reaction–Diffusion Attractor Patterns

“…Using the fit parameters and literature rate constants,the model recapitulated the shape and formation timescale of the hill gradients.The fitted rate constants were within one order of magnitude of literature values for rates of standard strand displacement reactions with the respective toehold lengths; [29] fitted diffusion coefficients were also within an order of magnitude of previously measured DNAd iffusion coefficients in 30 %(v/v) PEGDAhydrogels [28] (SI:T able S5). The predicted timescale of formation was 4.3 h, which was within af actor of 2o ft he experimental timescale of 6.7 h. We simulated pattern recovery using the rate constants and diffusion coefficients as model inputs,a nd the normalized fluorescence intensity profile measured after perturbation as an initial condition.…”