One of the main challenges of industrial (white) biotechnology today is the production of fuel from biomass at a cost that ultimately must be competitive with fossil fuels - but potentially more sustainable in the long term as a renewable, carbon-neutral energy source providing energy security. "Green" (i.e. environmentally-friendly) industrial production lines - characterised by reduced energy consumption, waste and CO2-emissions - are clearly attractive, but are crucially reliant on the discovery, improvement and adaptation of robust and efficient biocatalysts. This means that methods and strategies have to be developed that allow identification of suitable catalysts. While the utility of enzymes for biocatalysis is clear, it is still not trivial to find or make such efficient, useful catalysts for a wide range of purposes by enzyme engineering (now arguably based more often on Darwinian cylces of 'directed evolution, rather than design - although the two approaches are not exclusive).

We tackle in this proposal the challenge that each step in the 'bioenergy pipeline' (from growing biomass to fermentation for biofuels) can potentially become rate- or cost-limiting and that enzymes for those purposes are clearly needed. The natural resistance of plant cell walls to microbial and enzymatic deconstruction is largely responsible for the high cost of lignocellulosic biomass conversion. To date, only a small proportion (approximately 40%) of the energy content available from lignocellulose feedstocks (unusable portions of plant materials in the form of agricultural, industrial, domestic, and forest residues) is convertible to ethanol.

We address this problem at a number of fronts: using a ultra-high throughput screening system we identify new protein catalysts (from metagenomic libraries) and we improve these and already characterised catalysts by multiple rounds of directed evolution. Our directed evolution is propelled by access to a new type of libraries (mimicking natural mechanisms - involving insertion and deletions) and by the availability of ultrahigh-throughput screens of very large libraries (>10e7 members). We hope that the analysis of the selected novel enzymes will unravel evolutionary relationships between cellulases/hemicellulases and the structural basis for substrate recognition as a basis to improve the engineering of new useful biocatalysts enabling efficient plant cell wall hydrolysis. Our final goal is to to make robust and multispecific biocatalysts that should be useful for the hydrolysis of recalcitrant lignocellulosic components available as well as general rules and experimental approaches to make this process more controllable and enable it to be widely used.

Technical Abstract:
We propose to create new glycosidases by directed evolution starting from either known glycoside hydrolases (with promiscuous side activities) or from metagenomic libraries. Libraries will be generated by a method that generates 3-9 base pair deletions and insertions. The ambitious objective of identifying catalysts from indel and metagenomic libraries is enabled by our use of a ultrahigh-throughput screening system, in which single clones are compartmentalised in picolitre water-in-oil emulsion droplets that are generated and optically sorted in microfluidic lab-on-a chip devices. The chances of identifying catalysts in structurally highly disrupted libraries (or in metagenomic libraries with low potential hit rates) are enhanced by a turnover of 10e7 droplets per day in this format. We hope that the quantitative analysis of mutants along the observed evolutionary trajectories will inform us about the determinants of specificity and efficiency, the strategies necessary to bring about generalist or specialist enzymes and provide insight into the molecular recognition mechanisms employed by this class of enzymes.

Potential Impact:
The skills, methods and results generated in this project will first be important for companies and research institutions that engage in protein or metabolic engineering ranging from small biotech start-ups to large pharmaceutical companies. Many protein engineering companies will be interested in the generic methods and strategies for protein engineering that we hope to explore (e.g. Novozymes, GSK, Johnson-Matthey, DSM) and some of these are already collaborating with our group. For example, such companies may directly use the fundamentally new methodologies (indel libraries, ultrahigh-throughput screening in microfluidic droplets), the strategies for exploring sequence space (systematic analysis of catalytic promiscuity, analysis of fitness landscapes, the management of mutational load vs the chance of finding new functions). Finally there will be specific interest in the actually evolved or newly identified glycoside hydrolase enzymes, especially in the many companies (mainly in the USA) that are specifically focused on the biofuel pipeline: these include, for example, Verenium, Solazyme, Amyris, Iogen, LanzTech, Genecor, GEM Biofuels, Sustainable Power Corp, Lignol and many others. In addition there is also a growing market for analytical instruments (such as the microfluidiic devices that will be used in tshi project) in companies such as Speher Fluidics, Dolomite, Raindance or Drop-Tech.
All these companies will be looking for skilled staff for introducing the novel approach of evolution in droplets using microfluidic devices. The coworkers involved in this project will cross traditional boundaries between microengineering and directed evolution and this interdisciplinary aspect will distinguish the staff development in this project from more conventional (but in itself also highly important) training in protein engineering. Thus the postdoctoral workers to be employed in this project will receive training that will give them an excellent position to join smaller biotech start-ups or larger biotechnology. Several members of the Hollfelder group are already working in biotech and pharma companies.
The project will be accompanied by interactions with the university's technology transfer office (Cambridge Enterprise) and interactions with industrial stakeholders, so that information is initially protected and then shared and commercialised.