In silico study of lignocellulosic biofuel processes

In theory, a genome sequence provides all of the information necessary to define the structure of the biological system of interest. For example, knowing all of the enzymes in a cell and the substrates that each one accepts and all of the products that each one can make, it is possible to formulate a bioreaction master global network that represents the complete repertoire of possible biochemical reaction systems within that cell. In this study, we plan to use a 'genome-scale' metabolic network (gsmn), reconstructed from the sequence data for a number of species with applications in biofuel production. Gsmns have already been published for a number of medically- and industrially-important species, including Streptomyces coelicolor and Mycobacterium tuberculosis (these 2 by us) facilitating novel approaches to process design and identification of antibiotic targets, respectively. We plan to use genome scale modelling to demonstrate the utility of in silico experimentation to Biorefining. The approach will link genomes, capable of carrying out lignocellulose degradation to genomes able to produce biofuels, with a view to predicting processes that will form the basis for an in vivo study. As far as we are aware, this will be the first project to link genome scale models in this way, and will therefore represent a scientific advance in addition to providing pragmatic information. This 2-stage (biomass degradation followed by a separate bioethanol production stage) is potentially more efficient than a single microbial processing step. Biomass Degradation The genomes of two 'model organisms' (the fungus Trichoderma reesei and the bacterium Clostridium thermocellum) have recently been sequenced and these species will, therefore, be included in the study. However, considering the current dependence on acid and heat pre-treatment in lignocellulose degradation, enzymes that are stable and active at low pH values and at high temperatures are of particular value. Thus, enzymes derived from thermophilic and acidophilic organisms known to degrade lignocellulose hold significant promise for industrial processes, and, for this reason Caldicellulosiruptor saccharolyticus and Acidothermus cellulolyticus (both of which have been sequenced) will be included in the biomass degradation stage. Biofuel production A significant yield limiting factor is the toxicity of ethanol to the fermenting host. Most fermenting organisms such as S. cerevisiae cannot tolerate high ethanol concentrations resulting in a product that must then be concentrated through an expensive and energy-intensive distillation step. Pichia stipitis represents one yeast species of relevance to biofuel research based on its natural ability to ferment xylose. Its recently sequenced genome revealed insights into the metabolic pathways responsible for this process and this species will be included in our second stage modelling Zymomonas mobilis, which has been described as having considerable potential for biofuel production has also been sequenced recently and will be included Finally, the use of E. coli, which has been engineered to produce isobutanol and other alcohols, using non-fermentative pathways, will be included. The feed-stocks to be examined for biomass degradation will include lignocellulose (with and without chemical pre-treatment) and co-substrates, which may enhance bioenergetic efficiency of metabolism. Each of the four species will have gsmns constructed. In each case, a 'draft' gsmn will be prepared (using the sequence annotations) which will be refined by the addition of further reactions in order to 'close' the model. Each gsmn includes numerous (often hundreds of) 'input gates', ie potential substrates, predicted by the gene sequence but never tried in the laboratory. Using in silico models, we will be able to examine the effect of combinations of substrates that would require many years of experimentation in vivo.

We will produce a web-based package that will contain in-silico cell representations (metabolic networks) of 4 potential lignocellulose-degrading micro-organisms and 4 potential bio alcohol producers. Using this package, it will be possible to simulate a process using each possible pairing of degrader/producer, varying the raw materials composition and using linear programming to optimise metabolic fluxes for bio alcohol production. The package will be capable of: screening all possible pairs of degrader/producer for optimum ethanol production for a given raw material identifying potentially valuable byproducts from each pairing optimising the raw material composition (by addition of co-substrates) for bio ethanol and/or byproduct yield. process simulation and optimisation using micro-organisms and/or pre-treatment specified by users (initially members of the consortium) Identifying mutation strategies for yield optimisation. Mass transfer modelling for raw materials degradation. The quantitative output of the package(s) will be product (including bio ethanol, byproduct and biomass) yields and all input and output rates. The users will then be able to convert these parameters into economic data. The qualitative outputs will be the optimal degrader/producer combination, potential co-substrate and the identity of byproducts. Currently, biorefining process design relies, largely on trail and error experimentation supported by limited, ad hoc process models. Genome scale modelling now gives us the opportunity to exploit gene sequencing programmes to examine the entire theoretical metabolic repertoire of potential microbial species. We will investigate a range of lignocellulosic substrates and the effect of substrate pre-treatments.

Communication and engagement Industrial colleagues designing Biorefining processes (particularly those members of the Biorefining Club). This will provide a method for carrying out numerous combinatorial experiments in silico, examining the effect of a large number of pairings of biomass degrader/bioethanol producers and the influence of co-substrates within a reasonable timeframe. It will also allow them to prioritise expensive laboratory resources. Regular meetings will be held with industrial colleagues as part of the IBTI management programme Genome scale metabolic modellers and experimental scientists who may be able to benefit from such approaches. Our project will provide an example of the use of pairs of linked genome scale metabolic models, using out-puts from the down-stream (bioethanol-producing) model as the objective function for optimising both models (including the upstream, biomass-degrading model). Applications could include the study of host-pathogen interactions as well as mixed culture bioprocesse and fuel cell development. The initial engagement will follow the presentation describing the availability of the Web site, once this has been authorised by the Industrial Club Members Track record in this area. We have implemented publically-available web-based versions of 2 genome scale metabolic models; Streptomyces (for antibiotic production) and Mycobacterium (for TB studies) which are currently used by the appropriate research communities. The cost of doing so for this programme will be born by indirect cost allocation. Collaboration and Exploitation/Application. These activities will be undertaken collectively by the IBTI consortium. The consequences of research underpinning the development of biological processes for the production of chemicals, materials and polymers to replace petro-chemical derived sources is likely to have significant impact of the quality of everyone's life and has been identified as an area where increased investment in research is required. However, the ''Biorefinery is a comparatively novel concept in the UK but is well established in other countries, particularly the USA, and, for maximum impact, should be given prominence by the BBSRC's communication office, once the programme is underway. The concept that the development of manufacturing facilities in which basic chemicals are produced from renewable resources, specifically biomass, before being converted to a broad range of chemical products, including fuels, fine chemicals and advanced materials using sustainable processes needs to be communicated to the public at a National, rather than parochial level. Of particular potential interest to the general public is the overall aim to support the shift from oil-dependant societies to recycling-oriented societies, and to contribute towards the prevention of global warming. We plan to use genome scale modelling to demonstrate the utility of in silico experimentation to Biorefining. The approach will link the genomes, capable of carrying out lignocellulose degradation to genomes able to produce biofuels, with a view to predicting processes that will form the basis for an in vivo study, in a future project. We are not aware of any published attempts to link genome scale models in this way, so this will represent a scientific advance in addition to providing pragmatic information. This is, therefore, an opportunity to place the UK community in the forefront of research capability in this approach and this also needs flagging on a wider platform. Capability These activities will be undertaken by the PI