A microfluidic DNA computing processor for gene expression analysis and gene drug synthesis

Abstract

Boolean logic performs a logical operation on one or more logic input and produces a single logic output. Here, we describe a microfluidic DNA computing processor performing Boolean logic operations for gene expression analysis and gene drug synthesis. Multiple cancer-related genes were used as input molecules. Their expression levels were identified by interacting with the computing related DNA strands, which were designed according to the sequences of cancer-related genes and the suicide gene. When all the expressions of the cancer-related genes fit in with the diagnostic criteria, positive diagnosis would be confirmed and then a complete suicide gene (gene drug) could be synthesized as an output molecule. Microfluidic chip was employed as an effective platform to realize the computing process by integrating multistep biochemical reactions involving hybridization, displacement, denaturalization, and ligation. By combining the specific design of the computing related molecules and the integrated functions of the microfluidics, the microfluidic DNA computing processor is able to analyze the multiple gene expressions simultaneously and realize the corresponding gene drug synthesis with simplicity and fast speed, which demonstrates the potential of this platform for DNA computing in biomedical applications.

Figure 1

Illustration of gene expression analysis and corresponding gene drug synthesis performing Boolean calculations.

Figure 2

(a) Schematic illustration of the relationship of cancer-related gene, capture molecule, and drug segment. The drug segment (light green and blue) is the sequence of suicide gene which is most similar to that of cancer-related gene. The capture molecule (dark green and pink) is complementary to both the part of cancer-related gene (light green and purple) and the part of drug segment (light green). (b) Schematic illustration of the computing process at molecular level. Initially, the capture molecule partially hybridizes with the drug segment. When the cancer-related gene appears, the capture molecule is inclined to hybridize with the cancer-related gene and the drug segment will be displaced. The displaced drug segment can be transferred to fill the gap of the incomplete suicide gene.

Figure 3

(a) Schematic of the microfluidic chip. (b) The photography of the microfluidic chip.

Figure 4

(a) Illustration of the computing process of C-erbB-2 over-expression. (b) Illustration of the computing process of nm23 under-expression. Illumination: During the rinse step, when the electrophoresis conditions are changed, both nm23 (red) and drug segment B (green) in chamber 2 can be detached from the capture molecule 2, migrated into reservoir 8, and transferred into reservoir 13 finally. As nm23 is not complementary completely to the drug template strand and cannot ligate with drug segment A to synthesize complete gene drug AB, it has no effect on the final drug synthesis, so it is not shown in the rinse process or the transfer process in Fig. 4b.

Figure 5

(a) Fluorescent images of 2 μM C-erbB-2 migrating to the right edge of chamber 1a. (b) Fluorescent images of drug segment B in chamber 2 rinsing off when changed the electrophoresis conditions. (c) Fluorescent images of computing process of C-erbB-2 (green) in three concentrations and transferring of drug segment A (red). Drug segment A could be detected in chamber 4 only when C-erbB-2 was over-expressed (10 μM). (d) Fluorescent images of computing process of nm23 (red) in three concentrations and transferring of drug segment B (green). Drug segment B could be detected in chamber 4 only when nm23 was under-expressed (0.2 μM). Because hydrogel has concentrated function to oligonucleotides (Ref. 31), we showed the pictures of drug segments A and B in chamber 4, which were clearer than that in reservoir 13.

Figure 6

Electropherograms of the output molecule in reservoir 13. B represents drug segment B and AB represents complete suicide gene (drug AB). (a) Electropherogram of the standard sample of drug segments A, B, and AB. (b) Electropherogram of the output molecule when C-erbB-2 ↑ and nm23 in normal level. (c) Electropherogram of the output molecule when both C-erbB-2 and nm23 in normal level. (d) Electropherogram of the output molecule when C-erbB-2 in normal level and nm23 ↓. (e) Electropherogram of the output molecule when C-erbB-2 ↑ and nm23 ↓. (f) The reproducibility of three runs of computing on the same chip, with the conditions that c-erbB-2 was 10 μM and nm23 was 0.2 μM.