Computationally Optimised DNA Assembly of synthetic genes

Abstract

Gene synthesis is hampered by two obstacles: improper assembly of oligonucleotides; oligonucleotide defects incurred during chemical synthesis. To overcome the first problem, we describe the employment of a Computationally Optimised DNA Assembly (CODA) algorithm that uses the degeneracy of the genetic code to design overlapping oligonucleotides with thermodynamic properties for self-assembly into a single, linear, DNA product. To address the second problem, we describe a hierarchical assembly strategy that reduces the incorporation of defective oligonucleotides into full-length gene constructs. The CODA algorithm and these biological methods enable fast, simple and reliable assemblies of sequence-correct full-length genes.

Figure 1

Gene Assembly Schematics. Self-assembled overlapping oligonucleotides are extended by overlap extension and PCR to generate intermediate DNA fragments. The intermediate DNA fragments are cloned to be sequenced and PCR amplified from the selected vectors. Intermediate DNA fragments with correct DNA sequences are mixed together and extended by overlap extension and PCR amplification to generate the full-length gene

Figure 2

Hybridisation melting temperatures and assembly of the integrase (IN) gene of the S. cerevisiae transposable element Ty3. Computational melting temperatures without (A) and with (B) CODA melting temperature optimisations. Solid and dashed lines represent correct hybridisations of oligonucleotides and intermediate DNA fragments, respectively. Dot-dash and dotted lines represent corresponding incorrect hybridisations. Lanes 1-10, IN intermediate DNA fragments 1 through 10 assembled from oligonucleotides without (C) and with (D) CODA melting temperature optimisations. Lanes 11 and 13, molecular weight markers. Lane 12, corresponding assembly of full-length IN gene (1,640 bp)

Figure 3

Codon usage in the CODA designed Ty3 IN gene vs. codon usage in E. coli highly expressed genes. A. X (solid regression line), CODA designed IN gene. O (dashed regression line), E. coli highly expressed genes. Each codon x-coordinate is usage per 1,000 in E. coli average genomic DNA; y-coordinate is in the CODA gene (X, solid line) or in highly expressed genes (O, dashed line). B. 10% SDS coomassie stained gel. Protein Marker (Lane 1). Protein products from uninduced (Lane 2) and induced (Lane 3) cells

Figure 4

Probability of Obtaining at Least One Correct Sequence for each of the Intermediate DNA Fragments. Pg (light bars) is the probability that every one of the M intermediate DNA fragments has at least one clone with the correct sequence. P[g|b] (dark bars) is the probability that every one of the M intermediate DNA fragments has at least one clone with the correct sequence given that only pCODAblue screened blue clones are considered