The Automated Microfluidic ChIP Essay

Wu et al. developed an
automated microfluidic ChIP ("AutoChIP")
that used 2000 cells per reaction. Figure 16.1 shows a labeled schematic of the device where different stages of ChIP are performed. In Ring 1, fixed cells were trapped and lysed and the DNA was fragmented with
micrococcal nuclease (MNase)
digestion. Antibody-bead preparation was loaded into Rings A to D and mixed with fragmented DNA using valve-actuated mixing in order to facilitate IP. After IP, the samples were washed in the columns stacked behind sieve valves SV6, SV7, and SV8 by passing washing buffer through beads, and collected from the device. Purification of DNA was conducted off-chip. The major improvement of AutoChIP when compared to conventional protocols was its short assay time. AutoChIP only took 2 h for IP instead of overnight in conventional assays. Fast mixing of the reagents and beads in ring mixers from A to D accelerated antibody-antigen binding. 400 nl microfluidic reaction chamber (compared with 1 ml in the conventional assays) created high bead density to effectively speed up ChIP. The device shown in Fig. 16.1 was also designed with four rings to accommodate four ChIP reactions at one time. As Wu et al. demonstrated in their later work, scaling up was easily accomplished by fabricating a chip with more mixing rings, resulting in even more parallel reactions. Figure 16.2 shows an upgraded device that has 16 rings to accommodate 16 ChIP reactions at one time [44]. Shen et al. also used very similar design for their ChIP-seq study on H3K4me3 epigenomic landscape using 1000 mouse early
embryonic cells
Our group (Geng et al.) designed a microfluidic device for ChIP-qPCR assay with the ultrahigh sensitivity. Optimized microfluidic and ChIP conditions permitted the analysis of histone modifications from as few as ~50 cells without losing specificity. The whole
ChIP-qPCR
could be finished within ~8.5 h. As shown in Fig. 16.3, cells were first stacked against magnetic beads that were functionalized with anti-histone antibody. They were lysed to release chromatin and chromatin was then sheared by MNase digestion. The fragmented chromatin was then incubated with the antibody-coated magnetic beads with mixing of the beads by a moving magnet and a periodically actuated valve ("flapping valve"). Once the bead surface captured profiled histone (together with the DNA that it interacts with), the beads were washed to remove nonspecifically bound chromatin/DNA, while the magnet retained the beads and washing buffer passed through beads. The beads were then collected from an outlet of the device and taken off the chip for remaining purification and qPCR procedure.
Compared to AutoChIP, several measures were taken to improve the sensitivity. First, we reduced the reaction chamber for IP down to ~50 nl and ensured IP beads to occupy a large fraction of the reaction chamber. The close proximity among beads greatly increased the efficiency and rate for chromatin adsorption on the bead surface due to the short diffusion lengths involved. The adsorption of a chromatin molecule among beads was rapid given that travel time