Programmable Magnetic Tweezers and Droplet Microfluidic Device for High-Throughput Nanoliter Multi-Step Assays
Programmable Magnetic Tweezers and Droplet Microfluidic Device for High-Throughput Nanoliter Multi-Step Assays
复制标题
DOI:
10.1002/anie.201203862
复制
发表时间:
2012-01-01
影响因子:
16.6
通讯作者:
Malaquin, Laurent
中科院分区:
文献类型:
--
作者:
Ali-Cherif, Anais;Begolo, Stefano;Malaquin, Laurent
Droplet microfluidics offer unique capabilities for the development of high-throughput analytical systems.[1] It allows fluids to be aliquoted into volumes in the nanoliter or picoliter range, and then transported over arbitrary distances without dispersion or cross-contamination. Several droplet-based functions, such as merging,[2, 3] splitting,[2] sorting,[2, 4] or cell encapsulation,[3–5] have already been demonstrated. Moreover, thanks to their “pipeline” architecture, in which different samples follow the same track in a row, droplet microfluidics enable the implementation of high-throughput assays with relatively simple microfluidic designs, as compared for example, to systems in which the aliquoting is performed by valves integrated in the microfluidic system itself.[7] Unfortunately, the ability to efficiently purify or extract molecules of interest from a complex matrix, a key component of most biochemical methods, is still missing from the functions currently available for droplet microfluidics. In macroscopic methods, the use of superparamagnetic beads as a solid state support has become very popular: they can bind an analyte of interest, be retained with a magnet while the supernatant fluid is removed, and release the analyte in an elution buffer. This process can be multiplexed, for example, using multiple magnets at the bottom of microtiter plates, but it still suffers from significant constraints, notably owing to 1) mass transfer limitations, 2) the need for relatively large volumes, and 3) poor mixing and washing efficiencies. We already proposed an alternative approach, using selfassembled magnetic microcolumns in microfluidic format.[8, 9] A large reduction in analysis time and increased automation was achieved in this way, but as with other microcolumnbased systems, this device does not allow for a high level of multiplexing and is not adapted to low-volume sample handling (that is less than 10 μL). In this respect, the combination of magnetic solid-phase extraction with droplet microfluidics is appealing. Magnetic-bead transfer between two droplets of several microliters was first introduced by Shikida et al.[10] Results were achieved by the displacement of a permanent magnet along a millimeter-sized fluidic channel, interconnecting different reservoirs. A similar method, using multiple wetting valves to allow for more complex designs, was recently presented,[11] but it still requires rather large volumes and incubation times. To increase the flexibility of these assays, Sista et al. combined this strategy with droplets manipulated on an array of electrodes by electrowetting (EWOD).[12, 13] This device, generally named a digital microfluidic device, is very flexible, but it requires complex microfabrication steps to integrate the array of electrodes onto the device surface. A variant of this method, in which drops are immobilized on hydrophilic patches and droplets containing magnetic particles are magnetically actuated through microfabricated coils, was also proposed.[14] The drop size, however, is large (approximately 10 μL) and microfabrication of the system is almost as demanding as that of the EWOD systems. Another strategy was proposed recently, in which droplets containing magnetic particles are hydrodynamically split in the presence of an asymmetric field.[15, 16] This approach shares with conventional droplet microfluidic devices the possibility of processing droplets at a high throughput of tens to hundreds of drops per second. However, even if effective particle separation in a single daughter droplet is achieved, using this method for a purification process is not efficient, since the removal of the supernatant fluid is inefficient. More recently, Gu et al. attempted to …