DNA computing on surfaces: DESTROY and READOUT operations

This thesis is focused on the development of the DESTROY and READOUT operations for surface-based DNA computation.; In order to solve larger and more complex computational problems using DNA computing, our strategy has switched from single-word DNA computing to multiple-word DNA computing. A new DESTROY operation has been developed for multiple-word DNA computing, as the previous E. coli Exonuclease I-based DESTROY operation for single-word DNA computing is not compatible with multiple-word DNA computing. The new DESTROY operation, which selectively removes unmarked DNA strands from surfaces, consists of primer extension followed by restriction enzyme cleavage. DNA polymerase is used to extend DNA strands immobilized on a chemically modified gold thin film. Complete extension of the DNA strands creates a Dpn II restriction enzyme site in the duplex DNA; these molecules may then be cleaved from the surface by addition of Dpn II, with an overall destruction efficiency exceeding 90%. DNA molecules may be protected from such destruction by hybridization of a peptide nucleic acid (PNA) oligomer to one of the words. The hybridized PNA blocks primer extension, thereby preventing formation of the restriction site and consequent strand cleavage. The utility of this DESTROY operation for DNA computing is demonstrated by solving a small (2-bit) Satisfiability problem in which information was encoded in two tandem words.; The READOUT step in surface-based DNA computing identifies the DNA molecules present at the end of the computational process. These DNA molecules encode the solutions to the computational problem. A new structure-specific cleavage-based READOUT strategy for surface-based DNA computing has been developed and demonstrated in the solution of a 4-variable-3-Satisfiability problem. The specificity of the sequence detection method utilized here derives from the sequence specificity of DNA hybridization coupled with the structure specificity of the enzymatic cleavage. The process is linear, with an increased uniformity of detection of the DNA computing products compared to that obtained using PCR amplification. The structure-specific cleavage-based readout is simple, accurate, and compatible with multiple-word DNA computing.