As promised here are a few notes and links on the key points of today’s lecture: Synthetic Biology and its Application (28th November 2013). Follow the links for further information.
Synthetic Biology – it’s not what you do, it’s the way that you do it.
Three key aims of synthetic biology were highlighted: the first was ‘making biology easier to engineer’. From this flow the two other aims, ‘learning by building’ and ‘commercialisation of product’ (being an equally applied and fundamental science).
Making Biology Easier to Engineer
For the last 30-40 years, genetic engineering has been based on PCR and recombinant DNA technologies for writing, and automated sequencing for reading DNA. Automated DNA synthesis and new assembly methods (e.g. Biobrick, Gibson, Golden Gate etc.) and an industry (e.g. DNA2.0, IDT etc.) that can supply these products cheaply and quickly mean that the emphasis is now on what you write, not how you write it.
This synthesis and assembly capacity allows well characterised parts, engineered to behave in a predictable, standard way. This is recognised in recent high profile papers such as Mutalik et al. 2013a, Mutalik et al. 2013b and Kosuri et al 2013 in which these methods have been used to build and assess different combinations of regulatory units (primarily promoters and 5’UTRs). The Nat. Methods papers are behind paywall and not open access (which kind of defeats the object of communicating science) so you can get an overview here and here. The strategies were different but the outcome the same: defined, quantifiable functions relating to strength and quality of the part. Also, see the Synthetic Biology Open Language standard (SBOL) and part characterisation performed for example, during iGEM.
Parts like this can be assembled into more complex devices that perform predictable functions. In turn these can be represented in an abstract manner (abstraction) allowing the designer to focus on a few key design parameters at a time. You don’t need to go to the DNA level if you don’t want to. Computer-aided design (e.g. Clotho, GeneDesigner) is becoming more and more common. Also, as we create more synthetic, engineered sequences and learn the rules of DNA-coding, we are likely to move further and further away from ‘natural’ to truly ‘synthetic’ genes.
As an example the two papers on a genetically-encoded oscillator and toggle switch were briefly introduced. Taken together this encourages people to think of biosciences as the place to be for driving economic growth and the next industrial revolution.
Learning By Building
Firstly, taking parts from nature and finding out they don’t do what we thought they did is a pretty quick way of testing the literature!
Secondly, building synthetic diversity (based on rational, statistical choices) is preferable to natural diversity for understanding biology. In the artificial situation we can deliberately maximise our search of the design space, but in natural systems we are restricted to the places evolution has taken biology. See here and here. This point was returned to later.
Commercialisation of product
Production of artemisinin via metabolic engineering of S. cerevisiae used 14 upregulated/introduced genes and 2 repressed/omitted. This approach highlights the importance of taking a holistic approach – in this instance it is better to produce a chemical precursor (artemisinic acid) rather than the final product – chemical and biological engineering working in tandem – and learning how to ferment these strains has the biggest impact on improving yield.
Biofuels are a different prospect. The recent elucidation of the genetic basis for alkane biosynthesis in cyanobacteria and the higher plant Arabidopsis thaliana has resulted in a series of reports in which alkane biosynthesis has been tailored towards the production of desirable retail fuel molecules. These include the biosynthesis of n-alkanes and -alkenes of different chain lengths (here here and here) and of the structural complexity (iso-alkanes, here) required to supply the ability to blend fuel.
However, if you work on therapeutics, drugs or cosmetics sugar is cheap. If you work on biofuels, sugar is expensive.
Strategies to exploit cheaper (and preferably non-food) feed stocks (esp. lignocellulosic material) are therefore underway. These follow similar strategies to the artemisinin and biofuels work above. A bigger challenge is maximising conversion efficiency (I didn’t get to the trade-off between this % conversion and concentration (mg/L) and yield (mg/h) – but it is a key processing problem – which do you want more of? It is unlikely one solution will give you everything). So, how do we engineer a system whose complexity so far escapes us? Taking a leaf out of computer science we can look to couple machine learning approaches to our new automated DNA writing technologies and the creation of synthetic diversity described previously. Biology as a computer science problem. But that’s not for now.
The final question was: that sounds all well and good, but,
How easy is it to engineer biology?
How do you even test this? The answer is, give it to a bunch of amateurs….
The results (of course) can be amazing. Here are a couple of posts I wrote some time ago about iGEM projects, (Grand Prizes and Materials Science), but the best thing to do is to go and browse around yourself: iGEM main page, and Exeter 2012 and Exeter 2013.
Here are some other links you may find interesting:
Video: Synthetic Biology Explained
Any questions, feel free to ask. The best thing to do if any of this interests you is to get involved.