Bespoke design of whole‐cell microbial machines

Abstract

Six years ago, I wrote a perspective article with Lars Blank (now at Aachen) and Jens Kromer (The University of Queensland) describing a grand challenge of developing chassis cells: tailorable cells that can be used to rapidly engineer production of industrially useful biochemicals (Vickers et al., 2010). We listed four requirements to develop chassis cells: a minimal cell, with associated reduction of complexity; the ability to tightly and predictably control overall cellular behaviour; the ability to precisely direct cellular carbon flux towards desired products; and a toolbox of technologies that enable high-level microbial engineering. Here, I will examine progress in each of these areas. I will also speculate on where we might go with this technology in the future. The minimal genome developed recently by the J. Craig Venter Institute (JCVI; (Hutchison et al., 2016) is the smallest known genome capable of sustaining self-replication of a free-living organism – albeit one that grows relatively slowly and that requires fairly complex nutritional support. Significant work is required to develop an industrially useful chassis cell using this technology, including improved growth rate and the ability to grow well under stresses typical of an industrial bioprocess (Vickers, 2016). However, it does demonstrate proof of concept for extreme genome minimization – one of the two approaches to construct a chassis cell. The other is greenfield genome design, a considerably more challenging approach requiring both the capacity to synthesize complete genomes and a full understanding of minimal metabolic requirements. While writing DNA has notably lagged behind reading DNA, we are now at the stage where, with a reasonable amount of resources and infrastructure, one can write entire microbial genomes from templates. However, of the 473 genes encoded on the minimal genome, the function of 149 is currently unknown (Hutchison et al., 2016) – indicating that we still have some way to go to achieve the sufficiently detailed understanding of cellular requirements that would enable true greenfield design. Notwithstanding this, it is fair to say that we are starting to move towards the point where we can seriously consider the ground-up construction of chassis cells. This will be accelerated by the ability to interrogate in detail what it takes to make a functional genome through the kind of genome minimization experiments pioneered at JVCI. The second requirement for chassis cell design is the ability to tightly and predictably control overall cellular behaviour. This goes hand in glove with the detailed understanding of metabolic behaviour required to build a cell from the ground-up. Understanding how regulatory circuits control native cellular behaviour is of course only the first step; an ability to exploit those regulatory circuits will be required to fully realize the potential of a chassis cell system. Synthetic biology approaches for the construction of discrete genetic circuitry are paving the way towards more complex and broad-reaching regulatory control, and orthologous methods – which are developing rapidly – will undoubtedly be required to obtain full control over cellular behaviour. Our ability to precisely direct cellular carbon flux towards desired products has come a long way over the last handful of years. In particular, metabolic modelling approaches – including an improved ability to incorporate kinetic information into models (Saa and Nielsen, 2016) – have progressed significantly. In the future, such models might start to provide a foundation for greenfield genome design. Experimentally, the ability to control carbon flux at specific metabolic nodes using a variety of different approaches and, perhaps more importantly, the ability to balance metabolic flux across metabolic pathways and within the entire metabolic network (aided by modelling) are improving rapidly. Key to this is understanding metabolic fluxes on a global cel