Create Ultralong DNA Constructs in One Assembly Step
Firebird Biomolecular Sciences LLC
Steven A. Benner
Foundation for Applied Molecular Evolution
Shuichi Hoshika
Abstract
Frost & Sullivan found a 2014 global market for DNA oligos at $241 million, $137 million for genes. Private
investment in DNA synthesis companies like Twist, Ginkgo, and DNA Script give collective valuations of
several billion dollars. Federal public investment by the NIH, DARPA, and others in "synthetic biology" that
depends on DNA synthesis exceeds $100 million annually. These numbers stand behind this project to develop
two innovations to (a) deliver, under a custom synthesis model, long DNA (L-DNA) assemblies (b) secure a
licensing platform, and (c) create collaboration and buyout opportunities. These technologies are:
1. Artificially expanded genetic information systems (AEGIS), which add 4 nucleotides forming 2 additional
orthogonally binding nucleobase pairs (Z:P and S:B) to the pairs (C:G and T:A) found in natural DNA. Eight-
letter DNA increases the number of sequence accurate fragments that can be autonomously assembled.
2. Transliteration, which converts Z, P, S and B to C, G, T and A respectively, giving an entirely natural L-
DNA construct by removing the AEGIS components after they have done their job assembling fragments.
Highlights of Phase I results include:
(a) OLIGARCHTM software predicting stability of 8-letter GACTZPSB DNA duplexes.
(b) Fidelity of DNA products made by AEGIS + transliteration as good as in commercial G-blocks.
(c) Constructed genes for kanamycin resistance and green fluorescence protein were active in E. coli cells.
These successes shift the cost/quality burden for L-DNA synthesis towards residual error management.
Aim 1. Manage residual error using, as experiments suggest:
1.1 C-glycosides to eliminate depurination and depyrimidinylation, should these cause residual error.
1.2 Enzymatically removable protecting groups to eliminate chemical damage during deprotection.
1.3. Capturable capping groups to achieve simple >99.999% removal of truncated species.
1.4 Enzymatic DNA synthesis to eliminate all "chemical" degradation in "harsh" phosphoramidite synthesis.
Residual error will be further managed using MutS and Surveyor nuclease error correction.
Aim 2. Create synthetic pipelines to prepare the building blocks and reagents used to manage residual error.
Aim 3. Develop array-based phosphoramidite synthesis of fragments with continuous error evaluation.
Reproducibility will be ensured by making the reagents themselves available for sale. This is a source of
immediate revenue as well as a major part of our marketing strategy. Already, reagent sales to satisfied
customers have yielded licensing deals worth over $2.5 MM. For commercialization, Firebird just executed an
agreement with DNA Script, a pioneer for non-templated enzymatic DNA synthesis and its automation, to go
forward after Phase 2, should enzyme-based DNA synthesis be preferred to manage residual error. This
includes licensing Firebird's patents for enzymatic cyclic reversibly terminated untemplated DNA synthesis.