Molecular Assemblies

J. William Efcavitch

We are a startup. We opened in January 2014. Bootstrapping with SBIRs, NSFs, we have raised over a million dollars from small seed investors. We are trying to develop next gen DNA synthesis technology. All of the DNA synthesis companies use a chemical process invented in 1981. There's been a lot of physics and cost minimization. Nothing much has changed in the fundamental process. We wanted to use a de novo DNA synthesis method using enzyme.

Because fedex became the largest suppliers of oligos, there was a company called ABI that commercialized the phosphoramidite chemistry method and it has become a billion dollar industry in supplying oligos. It's based on the chemistry invented at U of Colorado and the young man in the picture is showing the first development of a commercial DNA synthesizer in 1981. Phosphoramidite chemistry is wonderful, but it has reached a plateau in its tech development, even though there has been miniaturization, inkjet delivery etc, the chemistry is essentially the same as the young man was using in 1981.

The problem is yield. Nobody seems to understand why the limitations in the chemistry exists. There's a group of four reactions that must occur at high yield, so it becomes simple to say that if all of the sum of the chemistry steps are 99.9% something, that's going to tell you the max length of the oligo that you can make. In 1986, I ssynthesized a 176 bp length fragment and that's still the same max length we can do.

The length has not really increased much since then. There's been slow improvement in cost. The DNA molecules almost always require error correction. It's all based on organic chemistry so there's hazardous waste generation which is not really sustainable.

At molecular assemblies, we wanted to approach the problem not from physics, automation or miniaturization, but to develop a remarkable enzyme called terminal deoxyribonucleotide transferase. We don't use a copy process. We're doing de novo synthesis.

Some of the hallmarks are that TdT can make very long single stranded DNA molecules. It's sustainable. Aqeous environment. It's scalable. It's not just small synthesis, it can be large scale synthesis.

Why TdT? It's a remarkable enzyme. It's very promiscous. It will accept a large wide range of modified oligonucleotides. It can make 10k bp strands of holopolymers in solution. The basic premise we have is that if one uses an aqeous base synthesis with this enzyme, one could theoretically make very long polymers. Yesterday someone asked about the limits of phosphoramidites-- well it's in acetonitrile in anhydrous system, well let's try water.

TdT properties are well known. It's available.

What's our approach? We're doing this solid phase, reversible terminator with TdT, add one nucleotide, wash, then remove the terminator, and just repeat that cycle. Because it's aqeuous based chemistry, we should be able to make long molecules. Some people say hey block the 3 prime hydroxy but we decided to use another approach and use what we call virtual terminators that attached to the exocycli amines. We cleave it with mod reducing agents; the scar eliminates and we are leaving behind natural DNA at every step. Because we have no 3 prime hydroxyl, we can have a large degree of control over modifications, we can leave scars on if we want, and we think this could be a useful synthetic tool for making highly modified nucleic acids.

I realize that this conference is devoted to synthetic biology, we consider DNA to be the industrial polymer of the 21st century. We're not going to be producing suits out of DNA but we think that with our scalable technology we could do synthetic biology and crispr-cas9 applications or the recently emerging immunotherapeutic cancer applications, precision diagnostics, and beyond biology like DNA memory, nanotechnology, these are fields that are being hindered by the expense of making rather long strands of DNA and hihgly modified DNAs and we think our technology will be applicable there.

Our team is down in San Diego. Prem K Sinha is our newest member. Still in development, hope to show you data in the future.

Q&A

Q: I am associated with some igem teams that tried this technique. The cost of the modified oligos was prohibitive. We were using heat to cleave the modification off. And then the enzyme, ... how are you.

A: cost of modified nucleotide is just simply related to scale, ... the only reason why phosphoramidites are inexpensive is due to the large scale. Eventually that cost goes down. We are trying to use as minimal modification as we can. We are exploring both exocylcic amine and 3 prime block modification. With reference to heat reomvables, that's one of the reasons we didn't pursue 3 prime blocks, because TdT will not accept modifications of the 3 prime end. We can hang ... on the bases.. and the enzyme will incorporate it, but any modification of the 3 prime end, will inhibit the use of the wild type protein. So we didn't want to engineer the protein and we decided to go that way. Our molecules come off with mild reducing agents and we can leave the scar until the end or take it off with the mild application of heat. But no heat required.

Q: A big part of the contribution to error rates in oligo synthesis comes from reactions that don't go to completion. Whether enzymatically or with chemistry, you still have to push a reaction to completion. I don't see how enzymes are going to do a better job of pushing the reaction to completion than traditional chemistry.

A: With some respect, you're right. We can... TdT quantitative additions. I had a discussion yesterday-- what are the limitations of phosphoramidite is a coupling reaction or is it subsequent degradation of strands formed during synthesis? I don't think that has been well defined. TdT has, I don't think it has an equilibrium point, we've been able to drive it whith low micromolar concentrations of heavily modified triphosphates to what looks like quantitative addition. We were able to implement sequencing based assay to see how many 9s we can put tin that.

Q: Secondary structure effects.

A: Synthesis in solution. Secondary structure effects, sure. We can leave the scar on. We are thinking minimalistic scar on the 3 of G or the n 4 of C will vanish secondary structure effects, we hope.