Immobilization of proteolytic enzymes on replica-molded thiol-ene micropillar reactors via thiol-gold interaction

We introduce rapid replica molding of ordered, high-aspect-ratio, thiol-ene micropillar arrays for implementation of microfluidic immobilized enzyme reactors (IMERs). By exploiting the abundance of free surface thiols of off-stoichiometric thiol-ene compositions, we were able to functionalize the native thiol-ene micropillars with gold nanoparticles (GNPs) and these with proteolytic α-chymotrypsin (CHT) via thiol-gold interaction. The micropillar arrays were replicated via PDMS soft lithography, which facilitated thiol-ene curing without the photoinitiators, and thus straightforward bonding and good control over the surface chemistry (number of free surface thiols). The specificity of thiol-gold interaction was demonstrated over allyl-rich thiol-ene surfaces and the robustness of the CHT-IMERs at different flow rates and reaction temperatures using bradykinin hydrolysis as the model reaction. The product conversion rate was shown to increase as a function of decreasing flow rate (increasing residence time) and upon heating of the IMER to physiological temperature. Owing to the effective enzyme immobilization onto the micropillar array by GNPs, no further purification of the reaction solution was required prior to mass spectrometric detection of the bradykinin hydrolysis products and no clogging problems, commonly associated with conventional capillary packings, were observed. The activity of the IMER remained stable for at least 1.5 h (continuous use), suggesting that the developed protocol may provide a robust, new approach to implementation of IMER technology for proteomics research. Graphical abstract Electronic supplementary material The online version of this article (10.1007/s00216-019-01674-9) contains supplementary material, which is available to authorized users.


Fabrication of SU-8 masters
Silicon wafers used for master fabrication were first dipped into hydrofluoric acid to remove any native oxide and dehydrated overnight at 120°C. SU-8 100 (Micro Resist Technology, Germany) was spin coated (1500 rpm, 30 s) and soft baked on a hotplate (65°C for 25 min followed by 95°C for 3.5 h) to yield a 200-m-thick SU-8 layer. The layer was then UV exposed with a dose of 1.35 J/cm 2 on the MA-6 mask aligner (Süss Microtec, Garching, Germany). Post exposure bake was done on the hotplate (65°C for 1 h followed by a slow 4-hour ramp back to room temperature). The slow ramp was used to minimize the thermal stress that can lead to adhesion loss of the thick SU-8 structures. The development was done in propylene glycol methyl ether acetate for 75 min. After development, the master was hard baked on the hotplate at 150°C for 3 h to improve the adhesion.
Finally, the master was coated by a nominally 40-nm-thick fluoropolymer layer for anti-adhesion.

Optimization of replica-molding of high aspect ratio thiol-ene micropillars
To ensure good uniformity of the thiol-ene micropillar array (height), a vacuum step was

Solvent compatibility
The effects of five different organic solvents on two different thiol-rich (50 mol-% excess) thiol-ene compositions, fabricated from either a trithiol or a tetrathiol monomer and cured for 5 or 10 min, were examined. The solvent compatibilities of the materials were evaluated by visual monitoring of the quality of the cured thiol-ene plates (thickness 0.5 mm, A=1 cm 2 ) immersed in 1 mL of each solvent for 1 h. If no visual damage was observed with 1h, the solvent exposure was continued for 4 days. The tetrathiol-rich composition was shown to be more stable toward toluene and acetone compared with the trithiol-rich composition, whereas dimethyl sulfoxide, tetrahydrofuran, and dichoromethane caused visible degradation of both compositions.

Effect of gold nanoparticle deposition on the wetting properties
The efficiency of gold nanoparticles (GNPs, 10 nm, in phosphate buffer saline) and dodecanethiolfunctionalized particles (d-GNPs, 3-6 nm, in toluene) immobilization on thiol-rich thiol-ene plates was examined by advancing and receding water contact angle measurement (Fig S-2) and compared to that of allyl-rich thiol-ene plates. As expected, the d-GNPs were shown to attach on both thiolrich and allyl-rich thiol-ene surfaces, as evidenced by the clear reduction in the hysteresis (as the result of increased receding contact angle). Instead, binding of GNPs did not significantly affect the contact angles compared with native surfaces and thus, AFM and XPS analyses were required to distinguish between GNP binding on thiol-rich and allyl-rich thiol-ene surfaces.

Effect of crosslinking degree on the amount of free surface thiols
The amount of free surface thiols was determined by using the Ellman's reagent [1] as described in the main text. The effect of photoinitiator and an additional UV exposure on the amount of free surface thiols on a thiol-rich (+50 mol-%) surface is illustrated in Figure S3-a, and the effect of curing time on the number of free surface thiols on a stoichiometric surface in Figure S3

Specificity of enzyme immobilization
IMERs prepared via incubation with only GNPs (no CHT) or only CHT (no GNPs) were tested for their catalytic efficiency toward bradykinin hydrolysis. As expected, the IMERs lacking the enzyme showed no reaction product (Fig S-4a), whereas IMERs lacking the GNPs contributed to bradykinin hydrolysis suggesting nonspecific adsorption of CHT to the native thiol-ene surface (Fig S-4b). (a)