Final Report Summary - BASYNTHEC (Bacterial Synthetic Minimal Genomes for Biotechnology)
Executive Summary:
The main goal of BaSynthec is to develop the computational and experimental methods, which are necessary to modify the non-pathogenic and industrially-important bacterium Bacillus subtilis. These methods are applied to modify B. subtilis in a predetermined and rational way in order to increase substantially the natural capacity of the bacterial cell to synthesize chosen products. We focused on applying our rational method to create better producing strains for vitamin B5 and for the extracellular enzyme α-amylase. To achieve this goal, many scientific problems need to be tackled.
First, we hypothesized that by eliminating genes that encode for proteins acting in dispensable cellular processes, we would obtain less wasteful and better performing strains. We established a repertoire of the dispensable regions of the chromosome and reduced the genome of B. subtilis up to 35% by accumulating deletions. Altogether, BaSynthec has generated 1234 deletion mutant strains, which were all assayed for phenotypes and/or production. Better performing strains were identified, and new knowledge on important pathways for production was generated. However, this study also revealed the strong tendency of strains to lose performance as their genome grew smaller, rendering strains with a large genome reduction unfit for use in biotechnology.
To better understand the phenotypes of deleted strains, a large scale metabolic model was used to predict the phenotypes of each deletion and predictions were compared to experimental observations. Initially, many model predictions were false. The model was then corrected and tested in many cycles to obtain a highly accurate model. Also, a new model, called RBA, taking into account the allocation of resources between the different cell processes was validated experimentally by quantifying several cell components at the whole cell scale. This model was shown to predict accurately the amount of a protein in a cell under a given growth condition. This is extremely useful because the computer model can be set to simulate the production a lot of a desired protein (e.g. α-amylase). To achieve this objective, the model will determine what needs to be changed in the cell. This information will guide the biologists to build a better producing strain.
To store data sets generated by experiments and by models, BaSynthec created an entire bioinformatics infrastructure. Importantly, the components of the BaSynthec infrastructure were linked together so that model predictions and phenotype data could be fully integrated into a single operating environment (KBase).
High-throughput test systems for vitamin B5 and secreted α-amylase production have been developed. Synthetic vitamin B5 plug-in modules containing B5 biosynthetic genes were constructed and introduced into 300 strains with reduced genomes. B5 production in these strains was tested and four strains were shown to have improved production capabilities. HTP secretion assays and synthetic plug-in secretion machinery modules, including synthetic sec and tat operons, have been developed. Sec- and Tat-dependent protein secretion was assessed in strains with reduced genomes. Synthetic Sec machineries were shown to perform equally well as the authentic Sec machinery. Analyses on the Tat machinery identified novel substrates, novel components, novel functionalities and novel biological roles. Novel concepts for the construction of super-secreting Bacillus cell factories were developed.
In summary, BaSynthec has successfully achieved most of its tasks and has created foundational tools for future chromosome engineering in B. subtilis. Chromosome engineering is now effectively “à la carte”. Reduced strains have been extremely useful to identify new genes important for vitamin B5 production and protein secretion, and two patent applications (WO2012/175574 and WO2012/175575) were filed. We discovered that reduced strains are not suitable cell factories for industrial applications. In other words, the removal of expansive dispensable cell processes affects in an unknown fashion some cellular function important for cell fitness. To circumvent this problem, we have successfully designed, built and validated a chassis strain suitable for industrial applications with much smaller genome reduction. Also, we have validated the RBA method using high quality standardized omics datasets generated in the project, effectively opening the way to model-driven (rational) optimization of strains. Because the scope of RBA validation was expanded, and because unexpected problems were encountered in the construction of tester strains, the formal experimental validation of model-driven global strain optimization was postponed after the end of the project. However, the construction of synthetic modules with various composite expression signals attests that the developed computational and experimental tools were applied successfully for module engineering.
Project Context and Objectives:
The overarching goal of BaSynthec is to develop a model-based approach for engineering the Gram-positive bacterium B. subtilis. We want to create a biotechnological platform and to develop synthetic modules for the production of metabolites and proteins of interest. To this end, BaSynthec will build upon the major progress in systems biology achieved in a previous EU-funded project (BaSysBio). BaSynthec has in particular the ambition to leverage the integrated computational-experimental approaches and the innovative tools developed in European projects to position itself at the centre of the worldwide efforts made in these areas by establishing a strong collaboration with a leading laboratory in bioinformatics and computer science in the USA.
Our scientific and technological objectives are to develop a model-based approach for engineering the Gram-positive bacterium B. subtilis and to develop synthetic modules for the production of metabolites and proteins of interest. The feasibility of the proposed approach will be demonstrated with vitamin B5, as well as a synthetic module for the generic production of extracellular enzymes using the alpha-amylase AmyQ as a general reporter for protein production. Iterative cycles of model refinement and experimental validation of predictions will be focused on engineering (i) B. subtilis strains with slow growth in the feeding phase of the fermentation and unrestricted metabolic activity, and (ii) improved synthetic modules for the production of a small molecule (vitamin B5) and of an extracellular protein (AmyQ), including biosynthesis and regulatory components, will be designed using DNA parts from B. subtilis and closely related Bacilli. Such modules will be transplanted into the cellular network of the best suited minimal cell used as a biotechnological platform (as well as in B. subtilis 168), and their operation will be monitored. The developed integrated theoretical-experimental approach will be used to predict the modifications to improve module operation. Iteration of this systems biology cycle will lead to a better understanding of the rules which underpin functional integration of modules in cellular networks, and to a better capacity to exploit functional modules for biotechnological applications. Finally, validated simpler bacterial strains together with the modelling framework will be used as generic biotechnological platforms to better control cell metabolism in industrial processes.
BaSynthec will:
-Design and generate bacterial cells of reduced complexity by a model-driven experimental strategy for iterative deletion of chromosome parts. Focus will be put on the following traits: low energy waste, global deregulation of gene expression, and cell growth control;
-Generate quantitative data sets using a few selected strains lacking potentially expensive processes such as flagella, sporulation, biofilms, etc., hereafter termed “simpler” cells, to model at a systems level how some basic biological processes functionally integrate. This includes the real time monitoring of the expression of selected genes in response to changes in the level of expression of enzymes or regulators predicted to be key in the control of growth rate or metabolic activity ;
-“Plug-in” new modules, which include entire pathways with their genetic regulations and their deregulated overexpression variants, in the “simpler” cell (experimentally) and in its cognate models (computationally). The selected modules will be used for establishing the proof-of-principle of the assembly of artificial operon/gene-clusters to generate functional synthetic modules to be “plugged-in”. Two modules of high biotechnological relevance will be constructed: i) the vitamin B5 biosynthetic pathway encoded by 5 genes, and ii) the secretion machinery for the export of extra-cellular enzymes encoded by up to 7 genes.
-Monitor quantitatively the operation of the “plugged-in” modules within the simpler host cell in order to tune and optimize the module. Input signals will be compounds inducing specifically a few promoters, each controlling the expression of a different key element of the module (or pathway). The output expression of these elements and others will be monitored in real time using fluorescent reporters. The availability of input and output values will enable the development of sophisticated control laws which can be tested and validated through online computer monitoring. This will facilitate efficient tuning of the synthetic modules, and will reduce the number of experimental trials.
-Apply both the methods and the gained knowledge about module plug-in in a synthetic cell to unravel high-level bacterial functions relevant to biotechnology (i.e. growth rate control).
To reach these objectives, BaSynthec will use recently developed technology to reduce the genome of B. subtilis, to monitor the expression of genes in a time-resolved fashion, and to reduce the genome and to design synthetic modules using computational approaches. Genome reduction will be a model-driven iterative process generating multiple strains (about 800) with progressively smaller genomes. The proposed research will investigate in a systematic way the resilience of biological networks to the progressive removal of dispensable biological parts. This will help us explore the effect of high order deletions on the cell regulation. We will investigate at a systems level what is causing some of the discrepancies between model predictions and experimental observations, and how the new knowledge can be included in the models to make them more accurate and predictive. Answering these questions will also help us pave the way for the more sophisticated bottom-up synthetic biology as it will enable a more accurate design of bacterial strains. This will allow us to harness the fantastic capabilities of bacteria for biosynthesis and biodegradation for human benefit.
Project Results:
- WP1 – Reduction and engineering of the Bacillus subtilis genome
WP1 overall objective was to construct Bacillus subtilis reduced genomes “à la carte” using efficient chromosome engineering methods, and to perform the biological implementation of synthetic modules with the aim of producing cell factories useful for biotechnological applications. WP1 reports an arrayed collection of 539 deletion mutant strains generated, culminating in the construction of i) a repertoire of dispensable chromosome regions covering 76% of the chromosome; ii) a ~35% reduced B. subtilis genome; and iii) a chassis strain (CS1) validated for industrial use. This arrayed collection was used to construct 290 derivative strains tested for vitamin B5 production and 405 derivative strains tested for α-amylase secretion. Altogether, this project has generated 1234 strains which were all assayed for phenotypes and/or production. Better performing strains were identified, and new knowledge on important pathways for production was accumulated. Massively reduced strains have reduced growth and are unfit for industrial use. However, specifically tailored strains are promising factories. A synthetic vitamin B5 plug-in module containing B5 biosynthetic genes was constructed and analyzed. The adequate gene expression of the biosynthetic genes was achieved using different promoters, mRNA stabilizing elements and RBS’s. The influence of the gene organization (operon and gene cluster) was also analyzed. Computational and experimental tools have been developed to rationally modify synthetic modules and chassis strains. However, due to a larger scope of the knowledge-based mathematical model of the transcription process, which needed incorporation of additional information on transcriptional start sites (see WP2, T2.4) and to the unanticipated delay in delivering a chassis strain fit for use in biotechnology (see WP1, T1.2) the assembly of a library of synthetic expression signals (promoter, TIR) aimed at testing the modularity of gene expression and the rational modification of industrially relevant strains could not be completed during the time of the project.
Minimal genome to sustain life
A large array of 248 mutant strains in which deletions were accumulated, was produced by Partner INRA. After 28 iterations of the deletion cycle, we obtained the strain MGP254, which has accumulated 28 large deletions representing a total reduction of ~1476 Kb relative to the 168 prototype strain. From the lineage of deleted strains, we chose MGP192, MGP229 and MGP254 for further characterization relative to the master strain. The genomes of these 4 strains were sequenced confirming that the chromosome structure of each strain was as expected. Large-scale omics and phenotyping analyses (WP4) and testing for industrial applications (WP5) were performed. Among other phenotypes tested, the most prominent one was the increased thermosensitivity as the genome gets reduced. Importantly, massively reduced strains do not appear as suitable industrial platforms for the production of vitamins and proteins because of their decreased capacity to produce biomass (see WP4). To circumvent this problem, we combined modeling and experimental approaches to add back selected genes into the genome. Specifically, 6 genes were added back to the MGP253 genome to restore its capacity to grow on minimal medium supplemented with glucose and glutamine. Also, a completely phototrophic strain was generated from MGP192 (1.1 Mb deleted) by combining restoration of chromosome regions and positive selection.
Reduced genomes for biotechnological processes
The design of a chassis strain suitable for biotechnological applications was undertaken by the whole consortium. The constraints for chassis strain design, which were industrial, computational, and methodological (strain construction), were described in the mid-term report. The final design of “Chassis 1” (CS1) strain involved the combination of 12 deletions and the exchange of the tryptophan biosynthesis pathway by that from a strain used in biotechnology. CS1 construction was achieved by INRA but the resulting strain had lost competence for genetic transformation, making them unfit for use. This problem was fixed by combining the deletions in different orders. The CS1 strain was distributed to partners, it genome was sequenced by NOVOZYMES, and CS1 used in WP5.
Fine tuning of strains with reduced genomes
Various vitamin B5 synthesis modules were constructed and tested for B5 production. The best module was introduced into 290 deleted strains and these strains were assayed for B5 production. Several better and worst performers were identified and the genes responsible for the effects were identified, providing new information on important players in B5 production. However, the better performers showed no increase in B5 production under conditions mimicking industrial conditions. Production of the BAN amylase was tested in 405 strains including single and multiple deletions. No strain with enhanced performance was identified. A miniDR strain was constructed to specifically characterize antibiotic resistance efflux pumps from Gram-positive bacteria.
Controlled expression of selected genes in strains with Live cell array (LCA) reporters
Because expression of genetic circuits is dependent on the physiological state of bacteria, the design of genetic sequences leading to the “optimal” protein/module abundance (relative to the objective) has to take into account the given growth condition. We therefore decomposed the major processes of gene expression in order to verify the assumption of gene expression modularity, which entails that the combination of promoters and translation initiation regions harboring specific expression profiles as a function of the rate of growth will result in a mathematically derivable outcome of protein abundance. Under the assumption of gene expression modularity, one is able to set the abundance of a given protein by combining transcription and translation processes to modulate differentially and predictably protein production across growth conditions. To verify this assumption, we combined 8 natural, well-characterized promoters (-50 position to +1 relative to the +1 of transcription; Nicolas et al., in preparation), 5 well-characterized synthetic and natural TIRs (Borkowski et al., submitted) and 8 short sequences known to enhance transcription (+1 to +8 relative to the +1 of transcription). Altogether this led to 47 synthetic sequences that were inserted upstream of the gfpmut3 coding sequence within a shuttle vector for E. coli and B. subtilis. The 47 plasmids were then transformed by single cross-over in 2 insertion loci (10’ and 90’ relative to the chromosome) of Bacillus subtilis.
We now have in hand a library of 76 GFP producing strains that were quickly screened (by Live Cell Array, LCA) for their GFP production using several growth media giving rise to a wide range of growth rates (from 0.2 h-1 to 1.6 h-1). This first screen is currently being analyzed in order to choose specific constructs that will be further investigated. The second round of LCA will allow concluding on the gene expression modularity and estimate the modeling parameters. Afterwards, a rational design of synthetic modules will be facilitated and systematic approaches will become possible.
- WP2 – Computational design and modelling of reduced genomes
The WP2 objectives were organized around two main lines led by UC and INRA respectively:
1) The design of computational web tools permitting the management and the characterization of reduced and minimal strain phenotypes.
2) The development of a rational computational pipeline for the design of synthetic modules.
3) The experimental validation of the model-driven design of synthetic module by plugging a synthetic module in the reduced strain for biotechnological purposes designed by WP1.
Progress in axis 1 (Task 2.1 and 2.2 leader UC): At the beginning of the reporting period, the database managing the reduced strains (all designed intervals, all implemented strains, and all observed and predicted phenotypes) and a web-API for accessing this database and predicting phenotypes were already implemented. Additionally, a web-tool prototype called the “StrainManager” had been developed and enables the design of new intervals and strains, and the automated prediction of strain viability on a variety of media conditions during the design phase. This web-tool has been achieved during the second reporting period and is now operational and accessible, which completes Task 2.1.
For Task 2.2 the objective was the integration of genome-scale metabolic models and regulatory network, the simulation of all observed phenotypes using the model, and when simulations and observations conflicted, the correction of the model to eliminate these conflicts. Significant results on the simulation of observed phenotypes and the correction of the metabolic model were already obtained during the first reporting period. During this reporting period, INRA and UC collaborated to complete the reconciliation of metabolic and regulatory models using a new stimulon-based approach (defined as the set of genes responding to the same set of stimuli). Regulatory models and the latest metabolic reconstruction for B. subtilis have first been updated. Then, the latest metabolic reconstruction for B. subtilis, the iBsu1103V2 model and the regulatory model have been integrated. UC applied this integrated metabolic regulatory model to the simulation of all growth phenotype data currently available for B. subtilis, demonstrating how the addition of regulatory constraints improved consistency of model predictions with experimentally observed phenotype data. The addition of regulatory constraints greatly improved consistency of model predictions with experimentally observed phenotype data. Altogether, these tasks should lead to publications including one joint paper between UC and INRA on the integration of genome-scale metabolic models and regulatory network.
Progress in axis 2 (Task 2.3 2.4 and 2.5 leader INRA): At the beginning of the reporting period, RBA method had already been deeply improved through enlargement of the metabolic network and the refinement of the bacterium description (Task 2.3). Some datasets necessary for RBA calibration and the models of transcription and translation were already acquired while other datasets remained to be produced. Preliminary tests performed with the available data had indicated that the data quality was high and suitable for the project purpose.
During the second reporting period, all the standardized omics datasets that were initially planned have been acquired and analyzed. The analyses confirmed the high quality and consistency of the whole dataset for model purposes. Complementary biological experiments (RNA-seq) dedicated to the fine characterization of +1 of transcription for the model of transcription appeared to be necessary. These data are currently under production and should be available in the next months.
Thanks to this large set of standardized omics data dedicated to growth rate management and modeling purposes, RBA prediction capability has been greatly improved especially through the accurate estimation of enzyme activities (Task 2.3) while dynamical models of protein production (transcription/translation) revealing the key elements managing the growth rate have been consolidated. The bank of 76 synthetic native promoters/TIR sequences has been built on the basis of these elements and is currently under biological validation. The second step is to exhaustively decompose for each gene the contribution of promoter/TIR sequences with respect to the model of transcription and translation (Task 2.4). Finally, a computational pipeline combining RBA predictions and the bank of native promoters/TIR sequences is currently under development (Task 2.5). Task 2.3 is now complete. Task 2.4 and 2.5 will be completed as soon as RNA-seq data will be available. Altogether, Task 2.3 2.4 and 2.5 should lead to 3 major joint publications.
Progress in the experimental validation of the model-driven design of synthetic module (Task 2.6): The purpose of Task 2.6 consists in developing a proof of principle methodology combining expression reporter strain, Live Cell Array technology and dedicated algorithms to control specific entitie(s) (growth rate) in bacteria in real-time. More precisely, the objective was to control the growth rate of bacterial cells by the Biotech, typically through the tracking of sinusoids for which period and amplitude are defined by reference signals. The achievement of such objective requires (a) the biological system permitting the control of growth rate in real-time, (b) the algorithms estimating the promoter activities of reporter strains from OD and fluorescence data and (c) the algorithms computing the injection characteristics from the growth rate target value (reference signals) and from the state of the cell (measured signals). Significant results on the estimation of promoter activities have been obtained. User-friendly software integrating the sets of algorithms estimating the promoter activities of reporter strains from OD and fluorescence data have been developed and published (http://genome.jouy.inra.fr/basylica). However, the biological system permitting the control of growth rate in real-time is still under development, which postponed the achievement of Task 2.6. In view of the strong interest to dispose of such biological devices, this project will be pursued after Basynthec ending.
Development of web tools to facilitate the use of the genome-scale model
Towards this end, we have made significant progress. We completed design of a database that includes all designed intervals, all implemented strains, and all observed and predicted phenotypes. We also created a web-API for accessing this database, which is located here: http://pubseed.theseed.org/model-prod/StrainServer.cgi and we created documentation for using this web-API here: https://bionet.mcs.anl.gov/index.php/Strain_Server. Project participants at ETHZ successfully tested the API and applied the system to download strain data.
We created a web-tool called the “StrainManager”. This web-tool enables the viewing of all strains, intervals, and observed/predicted phenotypes. It also enables the design of new intervals and strains, and the automated prediction of strain viability on a variety of media conditions during the design phase.
Refinement of the predictive capability of the RBA method
The refinement of prediction capability of RBA method has been obtained by (a) the refinement of the cell description; (b) the integration of additional structural constraints such as membrane occupancy and protein localization and (c) the identification of parameters from omics data produced in WP4 (mainly absolute protein quantification and uptake and excretion rates of nutrients) (see Illustration 8). These steps have been described in details in the second period report. RBA prediction capability has been investigated through (a) the prediction of growth rate, flux/protein distribution in media PYR, S, TS and CHG with identified enzyme efficiencies; (b) the prediction of growth rate, flux/protein distribution in CH medium by estimating enzyme activities from the ones of PYR, S, TS and CHG media; (c) the prediction of growth rate, flux/protein distribution in a strain disrupted from flagella/motility (sigD KO). For this last set of validation, growth of strains was performed in microplates (96 wells) and growth rate values were determined from 15 replicates. We found that RBA prediction were very accurate. Discrepancies mainly occurred in rich media for which uncertainties of omics datasets were larger and at the level of TCA cycle. Some enzymes in the TCA cycle are clearly over-abundant with respect to optimal resource allocation for growth. RBA then reveals missing constraints or trade-off acting on system design.
Definition of a list of elements with respect to the growth rate effect
Refinement in transcription model should be achieved with the integration of Transcription Start Site data (RNAseq) which will allow the fine characterization the +1 transcription, and thus the nature of the first nucleotides. The impact of the first nucleotide incorporated in mRNA became fully appreciated during the course of the project, and corrective action was undertaken. Experimental protocols to map the +1 by RNAseq had to be adapted, validated and applied to samples prepared from the 5 growth conditions. Although this new endeavour is causing some delay in task completion, it is clearly augmenting its scope. No other major obstacle has been identified. Objectives of this task should be achieved in the next months.
The bank of synthetic promoter/TIR sequence is under progress. At the moment, 76 clones have been successfully obtained and preliminary high throughput expression analyses sound very promising. Altogether, this work should lead to two joint publications.
Development of a computer based approach
Computational pipeline for the rational design of synthetic modules is presented on Illustration 11 and is divided in two steps:
1. The optimal protein configuration of the cell which permits the maximal production of the compound of interest is first computed by the RBA method [Goelzer et al. 2011]. A list of target genes with protein outputs (cellular concentrations) that have to be modified is then identified and enters into the second step.
2. The suitable combination of promoters and Translation Initiation Region (TIR) sequences for these genes can be deduced from (a) a model of the protein production and (b) a bank of native promoter/TIR sequence at the genome-scale (see previous Task), which by combination allows us to obtain the desired protein concentration.
- WP3 – Bioinformatics & information system
The annotation of the B. subtilis genome has been enormously enriched by structural information about 3000 transcription units being annotated. We have also built a comprehensive repertoire of dispensable chromosome regions in Bacillus subtilis and have used the phenotypes of deletion mutant strains to carry out a major refinement of the iBsu1103V2 metabolic model. All servers have been setup and are available to all consortium members. A consortium wiki, an exchange application for large unstructured data and a server for hosting the BaSynthec database for experimental data and metadata has been setup by Partner ETHZ, a server for hosting the strain database has been setup by Partner UC. The experimental result data set formats and result metadata have been successfully defined in an iterative process between all partners. A relational schema for experiment metadata and clean ASCII formats for automated retrieval and processing are available. The formats are documented in the BaSynthec wiki. Many components of the BaSynthec infrastructure were built into the DOE Systems Biology Knowledgebase (www.kbase.us). One such component is the “Workspace” data-store (http://140.221.84.142/browse.html) which provides a mechanism for storing, sharing, and analyzing a wide variety of data products. All growth-phenotype data produced by the BaSynthec project was loaded into a KBase workspace, called BaSynthECPhenotypes. The KBase workspace also serves as a repository for all metabolic models produced for all strains generated by the BaSynthEC project, and all model data is stored in the BaSynthECModels. All data is private, but readily accessible to project personnel. In this way, modeling output and phenotype data is fully integrated into a single operating environment (KBase). Based on conceptual work being performed in prior periods, the integration of the database for storing experiments and data sets with the databases for storing the mutant strains and phenotypes was finalized in this term.
Commission and operate infrastructure
A wiki and a biology information system has been setup for BaSynthec and all members have been given to the ETH file exchanger (http://cifex.ethz.ch). The BaSynthec wiki is available at
https://wiki-bsse.ethz.ch/display/BST the information system at https://basynthec.ethz.ch. The strain database is available at http://pubseed.theseed.org/model-prod/StrainServer.cgi and we created documentation for using this web-API here: https://bionet.mcs.anl.gov/index.php/Strain_Server.
Document experimental result data set formats and result metadata
The model and formats have been refined iteratively by ETHZ based on feedback by the other consortium partners. The result of this work is documented in the BaSynthec wiki at https://wikibasynthec.ethz.ch/display/BST/Building+principles+of+the+data+sheets and See mid-term report
Partner UC constructed a JSON based data model for strain and phenotype metadata. This format is used to integrate databases based on web services and is documented at
https://bionet.mcs.anl.gov/index.php/Strain_Server.
Management of phenotypes
Many components proposed in BaSynthec were built into the DOE Systems Biology Knowledgebase (www.kbase.us). One such component is the “Workspace” data-store (http://140.221.84.142/browse.html) which provides a mechanism for storing, sharing, and analyzing a wide variety of data products. All growth-phenotype data produced by the BaSynthec project was loaded into a KBase workspace, called BaSynthECPhenotypes. This includes data for over 300 strains, with growth tested in four media conditions: NMS, LB, M9, and Spizizen media. Details of all four media conditions are also stored in the BaSynthECPhenotypes workspace. All data is private, but readily accessible to project personnel through the “basynthec2” KBase account (password is “basynthec”), and all data is readily available for analysis by metabolic models that are also stored in a KBase workspace (see next section).
For this period, Partner ETHZ integrated the BaSynthec database for experimental data with the strain database provided by partner UC.
Integrating modeling output into a phenotype management engine
The KBase workspace also serves as a repository for all metabolic models produced for all strains generated by the BaSynthEC project. All phenotype data is stored in the BaSynthECPhenotypes workspace, and all model data is stored in the BaSynthECModels. All data is private, but readily accessible to project personnel through the “basynthec2” KBase account (password is “basynthec”). All models in the BaSynthECModels workspace may be viewed using the sophisticated model viewer available in KBase (Figure 2), and models may be applied to simulate and reconcile phenotypes using tools built into the IRIS tool of KBase (http://www.kbase.us/services/docs/invocation/Iris/iris.html). IRIS permits the application of all KBase tools on data stored in the KBase workspace. Tools include flux balance analysis, phenotype simulation, phenotype reconciliation, gapfilling, model curation, and model reconstruction. All output from analysis tools is also stored in the workspace. In this way, modeling output and phenotype data is fully integrated into a single operating environment (KBase).
Integrate experiment, genome annotation, mutant and phenotype databases
A browser has been built for the BaSynthec datasets and models, based on the BaSynthec database run in Basel by ETHZ. All data uploaded to the database by the partners have been immediately available to all other consortium members. Strain information and predications from the Strain DB are synchronised to the BaSynthec database regularly and can be viewed in the BaSynthec browser. Strain information (including details on how a strain was constructed) and predications from the Strain DB and experimental datasets can be browsed side by side.
Many components of the BaSynthec infrastructure were built by Partner UC into the DOE Systems Biology Knowledgebase (www.kbase.us). One such component is the “Workspace” data-store (http://140.221.84.142/browse.html) which provides a mechanism for storing, sharing, and analyzing a wide variety of data products. All growth-phenotype data produced by the BaSynthec project was loaded into a KBase workspace, called BaSynthECPhenotypes. The KBase workspace also serves as a repository for all metabolic models produced for all strains generated by the BaSynthEC project, and all model data is stored in the BaSynthECModels. All data is private, but readily accessible to project personnel. In this way, modeling output and phenotype data is fully integrated into a single operating environment (KBase). Based on conceptual work being performed in prior periods, the integration of the database for storing experiments and data sets with the databases for storing the mutant strains and phenotypes was finalized in this term.
- WP4 – Phenotyping and Omics analyses
As the experimental technology application arm, this WP had four major aims.
i) Phenotypic screening of approximately 800 strains to identify candidate chassis strains and industrial phenotypes for use in WP5 (biological engineering for industrial application);
ii) Provide basic phenotypic data for a large number of strains to guide and validate the modelling efforts in WP2;
iii) Characterize the 4 initially selected strains with respect to their omics content.
iv) Develop novel HTP phenotypic screens
Note that essentially no information was available for the above objectives at the beginning of this period.
The first and second objectives were met by testing approximately 1100 strains. 500 of them for general phenotypes of growth rate and yield, and 300 each for protein secretion and B5 production. The third objective on detailed characterization of the omics content of the 4 selected strains has been achieved and the data was made available for modeling. Several novel HTP assays were developed in the 4th objective, leading to the discovery of many novel protein functions and several industrially relevant phenotypes, in particular increased B5 production.
In-depth analysis of selected genome-reduced strains:
Based on the criteria outlined in WP 1 and 2, the consortium decided to interrogated the genome-reduced strains MGP 192, MGP 229, and MGP 254 as well as the master strain. All four strains were quantitatively analyzed in biological triplicates with the following methods:
• Physiology
• Proteome
o Absolute quantification of the cytosolic fraction revealing absolute quantities for > 1500 proteins
o Separation of the cytosolic fraction with 2D-gels. The ongoing analysis of these gels will identify post-translational modifications (e.g. phosphorylation)
o Relative quantification of the membrane (1200 proteins), cytosolic, and extracellular fraction (15N metabolic labelling)
• mRNA levels (tiling arrays)
• mRNA half-lives (DNA microarrays)
• mRNA levels in response to growth rate variation of the master strain (tiling arrays at 5 different growth rates)
• Metabolome
• Genome re-sequencing, revealing several SNPs mainly in the flanking regions of the chromosome deletions.
While the three genome-reduced strains grew significantly slower and to lower optical densities than the master strain, their metabolism exhibited a remarkable robustness. Most of the 50 detected metabolites showed only minor or no changes relative to the master strain. A notable, albeit small (2-fold) change included decreased TCA cycle metabolite abundances in all three of the MGP strains, and increased GxP pools in MGP192 and MGP229. A small number of metabolites showed significant (over 10-fold) differences between the strains that were mostly attributed directly to deleted enzymes in the MGP strains. Consistently, both the relative and absolute protein quantification revealed no significant changes in the central metabolic pathways.
Most probably due to the deletion of AhpC and AhpF, proteins involved in oxidative stress response (KatA, SodA, TrxA) accumulated significantly. Relative quantification of the membrane proteome fraction shows significant alterations in 155 proteins in at least one mutant strain in relation to the master strain. Differences in mRNA levels were mostly related to the deleted genes with no particular consistent pattern between the strains.
For transcriptomics, cultures from the MS, and 3 minimal strains were grown in NMS medium to mid-exponential growth (OD600 0.3) and RNA samples were prepared as previously described (Nicolas, Mäder, Dervyn et al. Science 2012). For each sample, two biological replicates were obtained. To parallel the proteomics analyses for relative quantification, the same growth conditions were used. Because the mutant strains used in this study carry genome reductions ranging from ~32 to ~35%, the usual normalization procedure, which assumes that only a small number of mRNA abundance levels will change, cannot be applied. Instead, a normalization procedure based on the remaining genes was used. A statistical model was developed by P. Nicolas to take into account the fact that deletions are oriented (towards accumulation of deletions along the strain lineage), allowing to compare MS/MGP192, MGP192/MGP229, MGP229/MGP254 in M9 Bioexpress and to derive interaction terms between the other medium (NMS) and each strain.
The results revealed that massive genome deletions had large and complex effects on global transcription, as could be expected. In the 4 strains representing the accumulation of deletions, 52% of the 3839 annotated RNA features remaining in MGP254 did not show any significant change. Therefore, 48% of the genes exhibited a differential expression in at least one mutant condition, suggesting a progressive deregulation of global transcription in the minimal strain.
HTP phenotyping
General physiological characterization revealed that essentially all deletants were physiologically inferior to the control strain as expected. The potentially industrially interesting phenotype of sustained high metabolic activity in the absence of growth was entirely related to genetic inactivation of sporulation that is already inactive in industry strains. To test biotechnological potential, the 300 strains were transformed with B5 and protein secretion modules.
The present studies on sublancin resistance, motility, pellicle formation, protein secretion and antibiotic resistance show that the collection of strains with reduced genomes represents a formidable tool for systematic HTP phenotyping and subsequent pathway analysis. Specifically, we identified many genes involved in sublancin resistance, motility and pellicle formation.
Proteomics and transcriptomics
The proteome was analysed in 3 biological triplicates with three different approaches:
1. Absolute quantification of the cytosolic fraction using the label free Hi3 technique, revealing absolute quantities for more than 1300 proteins in total.
2. Relative quantification based on 15N metabolic labelling was performed for 3 different subcellular fractions. The analyses of the membrane fraction led to relative quantities of nearly 1200 proteins. In the cytosolic fraction 1100 proteins could be relatively quantified. The analyses of the extracellular fraction provided relative quantities for almost 1000 proteins.
3. Separation of the cytosolic and extracellular fraction with 2Dgels. The analysis of these gels led to only a few proteins with differential expressed isoforms pointing to changes in the yet unknown post-translational status of these protein forms.
Using DNAmicroarrays, EMAU analyzed mRNA half-lives and mRNA levels in response to different growth rates linked to T2.4 “Definition of a list of elements with respect to the growth rate effect”. Because global changes in the cellular components in response to growth rate variation do not allow the use of standard normalization approaches, the normalization procedure developed for mRNA half-life measurements was applied to determine the proportion of mRNA to total RNA and individual mRNA levelsin response to 4 different growth rates. Using NimbleGen tiling arrays, EMAU analyzed mRNA half-lives and mRNA levels in response to 3 different growth rates linked to T2.4 “Definition of a list of elements with respect to the growth rate effect”. Because of major changes in gene expression in response to growth rate variation, global normalization approaches are not appropriate for determining mRNA concentrations. Therefore, partner INRA determined the total RNA amount per OD unit for different growth rates, and RNA samples were applied to microarray analysis using external RNA standards for normalization.
Metabolomics and fluxomics
Most metabolites showed some minor (less than 2-fold) differences between the master strain and the large genome deletions. The more notable of these included decreased TCA cycle metabolite abundances in all three of the MGP strains, and increased GxP pools in MGP192 and MGP229. A small number of metabolites showed significant (over 10-fold) differences between the strains. While some of these may be attributed directly to deleted enzymes in the MGP strains, several cases do not have immediate explanation. One such example is the detection of fructose-1-phosphate in MGP254, a metabolite usually only present when cells use fructose as a carbon source. While the entire fructose operon is deleted in MGP254, this operon is not expected to be expressed under glucose. This phenotype may be explored further in upcoming work. A second example is the absence of the lysine biosynthesis intermediate diaminopimelate in MGP192 and MGP229. These strains are missing the aspartate kinase yclM, one of the isoenzymes that catalyzes the first step of the pathway, however, the pathway is not expected to be active with lysine present externally. Exploration of this phenomenon may lead to a better understanding of amino acid biosynthesis.
Novel phenotypic assays including live cell array
A range of physiology-based assays was developed for sustained metabolic activity in the absence of growth (ETHZ), B5 production (DNP), protein secretion (NOV), sublancin resistance, motility and pellicle formation (AZG). The results relating to sublancin resistance have been validated by live cell array analyses.
The many already identified genes show that the collection of strains with reduced genomes represents a formidable tool for systematic HTP phenotyping and subsequent pathway analysis. With respect to industrially relevant phenotypes, the single deletion strains proved extremely useful to identify new genes and pathways involved in production of vitamin and secreted proteins. However, strains with massively reduced genomes do not appear to be suitable biotechnological platforms, mainly because of their low biomass yield.
- WP5 – Biological engineering for industrial application (5 pages max.)
High-throughput test systems for vitamin B5 and secreted AmyQ production have been delivered on May 31st, 2011. Synthetic vitamin B5 plug-in modules containing B5 biosynthetic genes were constructed and introduced into 300 strains with reduced genomes. B5 production in these strains was tested and four strains were shown to have improved production capabilities. Deliverable 5.2 has been delivered on May 31st, 2013. HTP secretion assays and synthetic plug-in secretion machinery modules have been developed. Synthetic sec and tat operons have been developed. Sec- and Tat-dependent protein secretion were assessed in strains with reduced genomes. Synthetic Sec machineries were shown to perform equally well as the authentic Sec machinery. Analyses on the Tat machinery identified novel substrates, novel components, novel functionalities and novel biological roles. Novel concepts for the construction of super-secreting Bacillus cell factories were developed. Deliverable 5.2 has been delivered on May 31st, 2013. Existing fermentation protocols have been adapted for cells with reduced genomes and have been implemented in the testing of vitamin B5 production and secretion of the α-amylase AmyQ. The objectives were achieved and Deliverable 5.3 has been delivered on May 31st, 2013. The results of high-throughput screening were assessed under relevant fermentation conditions. Compared to the original master strain MS168, improved vitamin B5 production could be shown in strains lacking individual genomic regions in fed-batch fermentation experiments. The new Chassis I strain was shown to perform about 40% better in vitamin B5 production than the original master strain MS168, but in this strain no further improvement of vitamin B5 production was observed upon deletion of genomic regions. High-level expression of the α-amylase AmyQ was established from the chromosomally located sigF locus of the master strain and 300 strains with reduced genomes. None of the tested strains exhibited improved α-amylase secretion characteristics. The constructed strains and insights will serve as valuable starting points for further optimization of the vitamin B5 and α-amylase production by B. subtilis. The deliverable D5.4 will be delivered to the EC on 15th September, 2013.
Development of high-throughput test systems for vitamin B5 and secreted AmyQ production
High-throughput test systems for vitamin B5 and secreted AmyQ production have been developed. This involved a high throughput small-scale production assay that mimics well the vitamin B5 production in fed-batch fermentation for which two patent applications (WO2012/175574 and WO2012/175575) were filed. Furthermore, to analyse strains with reduced genomes for their capabilities in protein secretion, protocols were developed for growing cells in 96 well microtiter plates. Specifically, the bacteria were grown in 2x LB broth until they reached an OD600 between 0.3 and 0.4 in the Biotek Synergy II. Bacteria were then separated from their growth medium by centrifugation, and the medium fractions were mixed with Laemmli buffer. Proteins in the medium fractions were separated by SDS-PAGE and visualized by silver staining. Different secretion profiles have been observed for several strains, but no major secretion defects were identified.
Testing of minimal cells with synthetic vitamin B5 plug-in modules for Vitamin B5 production
Synthetic vitamin B5 plug-in modules containing vitamin B5 biosynthetic genes were constructed. The adequate gene expression of the biosynthetic genes was achieved using different promoters, ribosome-binding sites (RBS) and terminators. The influence of the gene organization (operon and gene cluster) was also analyzed. The vitamin B5 production of strains carrying these modules was determined and compared to reference strains. The best vitamin B5 plug-in module was introduced into 300 strains with single genomic deletions around the chromosome to determine the vitamin B5 production capability of these strains. Four strains with improved vitamin B5 production characteristics and nine strains with drastic decreases in vitamin B5 production were identified for further characterization. The responsible genes influencing vitamin B5 production were localized on the corresponding chromosomal regions deleted in the 13 strains with engineered genomes. Clean mutants of the responsible genes were constructed and the corresponding strains were characterized in small-scale vitamin B5 production assays. The four best candidates were tested in glucose-limited fed-batch fermentation, but they showed no significant improvement of the vitamin B5 yield compared to the parental strain.
Establishment and testing of minimal cells with synthetic plug-in secretion machinery
High-throughput secretion assays and synthetic plug-in secretion machinery modules have been developed. An amyQ expression cassette was developed for production of the secreted α-amylase AmyQ and this cassette was introduced into 300 strains with reduced genomes. Synthetic genes for the Sec and Tat secretion machineries were combined in several different synthetic operons containing one to five genes. All the constructed sec operons have been made with at least 3 different promoters. Synthetic sec genes encoding the essential Sec secretion machinery components were introduced into cells with reduced genomes carrying the amyQ expression cassette. The wild-type copies of these sec genes were deleted from the chromosome of some of these strains to confirm the expression and activity of the synthetic modules. This showed that the sec modules encode active protein translocases. In total, 105 new strains with various expression levels of Sec components were constructed, but none of the tested strains showed improved secretion of the AmyQ α-amylase. The strains expressing sec genes only from the synthetic modules are now valuable tools for optimization of the secretion machinery by further genetic modifications. Furthermore, to analyse strains with reduced genomes for their capabilities in protein secretion, cells were grown in 96 well microtiter plates and the patterns of secreted proteins were analyzed. No major effects on protein secretion were detected. Next, the function of the Tat machinery was characterized in detail. This revealed: (i) a new Tat substrate, the metallo-phosphoesterase YkuE, (ii) novel roles of the Tat machinery in iron uptake and biofilm formation, (iii) functionality and novel roles of the TatAc subunit in protein secretion, (iv) the network structure of the Tat machinery in the B. subtilis membrane including three novel factors needed for the secretion of particular Tat substrates, and (v) the existence of a Tat-specific protein quality control system. Lastly, novel concepts for the construction of super-secreting Bacillus cell factories were developed. The results of the protein secretion studies were documented in 13 publications (see WP6).
Adaptation and optimization of existing fermentation protocols for minimal cells
Existing fermentation protocols have been adapted for cells with reduced genomes to analyse AmyQ and vitamin B5 production. Specifically, a glucose-limited fed-batch fermentation protocol was established for vitamin B5 production by the B. subtilis strains with reduced genomes generated in the project. The fermentation batch medium was a mineral medium with glucose and yeast extract to rapidly generate bacterial biomass in the batch-phase. The batch phase was followed by an exponential glucose feed phase. Subsequently, a constant glucose feed phase was implemented to ensure glucose-limited fermentation conditions and efficient vitamin B5 production. During the fermentation the pH, temperature and dissolved oxygen (DO) were controlled. The supply of air and exhaust gas production were continuously analysed to calculate carbon dioxide evolution (CER) and oxygen uptake rates (OUR).
Assessment of high-throughput screening results under relevant fermentation conditions
In the context of task 5.5 high-throughput screening results were assessed under relevant fermentation conditions. For genome deletion, the starting (or master) strain was MS168. Chassis I, the platform strain for biotechnological purposes was generated in WP1 by introducing 10 deletions into the genome of MS168. The two strains were analyzed in glucose-limited fed-batch fermentations developed in task 5.4. The master strain MS168 with the synthetic vitamin B5 module produced 26% lower vitamin B5 levels in fed-batch fermentation as compared to the small scale production assay in the 24 deep well plates (DWP; task 5.1). The 10 deletions in the biotechnology platform strain Chassis I carrying the synthetic vitamin B5 plug-in module enhanced the vitamin B5 yield by 40% in fed-batch fermentation as compared to the vitamin B5 yield produced by the master strain MS168 carrying the same vitamin B5 synthetic module. In small-scale vitamin B5 production assays, four strains with different genome deletions produced 20-50% higher vitamin B5 yields than the MS168 parent strain (task 5.2). These beneficial genome deletions were transferred into the wild-type B. subtilis 168 strain since the MS168 strain did not perform optimally in fed-batch fermentations. The four new strains were tested in glucose limited fed-batch fermentations and produced similar vitamin B5 yield as the wild-type parental strain carrying the same synthetic vitamin B5 module. The constructed strains and insights will serve as a valuable starting point for further optimization of the vitamin B5 production by B. subtilis. In parallel, 300 strains with reduced genomes were tested for their capacity to produce the AmyQ α-amylase from Bacillus amyloliquefaciens. High-level AmyQ expression was established from the chromosomally located sigF locus of the master strain and 300 strains with reduced genomes. Approximately 200 of the single interval-deleted strains and a few of the multiple interval-deleted strains were extensively tested for the α-amylase yield. All multiple interval- deleted strains turned out to perform poorly in α-amylase production compared to the parental master strain, and none of the strains with single interval deletions exhibited improved α-amylase secretion characteristics.
Potential Impact:
- WP1 – Reduction and engineering of the Bacillus subtilis genome
Potential socio-economic impact of the research activities in WP5:
The impact of constructing a collection of deletion mutant strains is in relation with the lessons learned from it by the industrial partners. The methodology has first been transferred to industrial partners, it is now public and disseminating in the scientific and biotechnology community. Our finding that the benefits of making a bacterial cell simpler and potentially more amenable to engineering is generally offset by unexpected detrimental effects on cell performances is a very important guide for future development of cell factories. But still, the discovery of new phenotypes is of high interest for biotechnology (e.g. one patent application by DNP). Also, the approach of reducing a genome still represents an attractive way to eliminate unwanted side products. Our capacity to add back specific genes and restore cell fitness represent a route to be considered for such types of applications. Through publications, this knowledge will also be disseminated to the public and other industries.
Main dissemination activities:
Researchers active in WP1 have disseminated their results through publications in peer-reviewed scientific journal, presentations in scientific conferences and workshops, and presentation of project results on the BaSynthec website (see WP6 for details). The collection of deletion mutant strains represents a formidable tool for systematic HTP phenotyping and subsequent pathway analysis, as demonstrated in WP4 and WP5. This resource will now be made available to the public, enabling other scientists to test for many more phenotypes of scientific, biotechnological and medical relevance.
Exploitation of the results:
Commercial exploitation of the results from WP1 is not directly foreseen, but their application potential is clearly underscored by one of the patent applications of partner DNP. Importantly, the strain collections created in BaSynthec form an excellent resource for further strain engineering and rational optimization through synthetic biology principles.
- WP2 – Computational design and modelling of reduced genomes
Potential socio-economic impact of the research activities:
Computational design and modelling of reduced genomes and of synthetic pathways is at the core of demonstrating efficiency of rational synthetic biology approaches (vs trial and error) for biotechnology. BaSynthec has shown that models can be greatly improved in accuracy through reconciliation of predictions with phenotypes of deletion mutant strains, up to the point where a model-driven streamlining of the metabolism is a logical step to take. The validation of Resource Balance Analysis through the highly accurate prediction of protein abundances in a given growth regime opens the way to rational design of synthetic modules and the fine tuning of the host cell to be “in synch” with the synthetic module. Although this latter step was not achieved yet, industrial partners have learned about the entire process and about what worked, what did not, and why. Again, this knowledge will guide the future development of cell factories. Through publications, this knowledge will also be disseminated to the public and other industries.
Main dissemination activities:
Researchers active in WP2 have disseminated their results through publications in peer-reviewed scientific journal, presentations in scientific conferences and workshops. Developped algorithms and software are freely accessible upon publication. The computational infrastructure developed in WP3 will be a great asset to support public release of the experimental and model datasets in a form that can be easily exploited by the scientific community.
Exploitation of the results:
Commercial exploitation of the results from WP2 is not directly foreseen. Importantly, the Kbase repository for the models corresponding to deletion mutant strains represents an excellent resource for further strain rational optimization through synthetic biology principles.
- WP3 – Bioinformatics & information system
Potential socio-economic impact of the research activities:
The impact of constructing a computational infrastructure to document in a single “virtual” place strains’genomes and genotypes, experimental omics datasets, high throughput phenotypes, mathematical models, model datasets is tighly linked with the lessons learned from the whole process by the industrial partners. This infrastructure enables data integration and facilitates its visualization and interpretation. Industrial partners have learned about the entire process and about what worked, what did not, and why. This knowledge will be extremely useful to guide the development of bioinformatics to support R&D projects in systems biology and synthetic biology. Through publications, this knowledge will also be disseminated to the public and other industries.
Main dissemination activities:
Researchers active in WP3 have disseminated their results through publications in peer-reviewed scientific journal, presentations in scientific conferences and workshops. Developped bioinformatic infrastructure are available to all project partners. Upon publication of results, these resources will support the public release of the experimental and model datasets with some analysis tools under a form which is much more accessible to the biologists.
Exploitation of the results:
Commercial exploitation of the results from WP3 is not directly foreseen. Importantly, the Kbase repository for the models corresponding to deletion mutant strains represents an excellent resource for further strain rational optimization through synthetic biology principles.
- WP4 – Phenotyping and Omics analyses
Potential socio-economic impact of the research activities:
The general socio-economic impact of the detailed characterization of few genome-reduced strains as well as the different screening results with large numbers of such strains was in education of the two industrial partners. According to their own statements, the partners profited from obtaining a clear understanding of what modern experimental analytics and computational methods can achieve, and what they can yet not deliver. Through the publications, the knowledge will also be disseminated to the public and other industries, but the two participating partners inevitably learned more because they were intimately involved in the entire process, and thus were also exposed to failures that are generally not published. The second aspect where the industrial partners profited was the novel biological understanding generated around cellular processes that are generally important for biotechnological processes such as the physiological impact of genome manipulations, sustained metabolic activity in the absence of growth, or protein secretion.
Main dissemination activities:
All scientific results are either already published in scientific journals or are in the process of being submitted. In several cases, consortium members have presented those results to different, mostly scientific audiences that are listed in section 2.1. Our results demonstrate that the collection of strains with reduced genomes represents a formidable tool for systematic HTP phenotyping and subsequent pathway analysis and – importantly – we demonstrate how to use this resource effectively. This resource will now be made available to the public, enabling others to test for many more phenotypes of scientific, biotechnological and medical relevance.
Exploitation of the results
A very tangible outcome of direct socio-economic value for one of the industry partners was the identification of 4 strains with improved vitamin B5 production and the discovery of several non-obvious genes that are important for B5 production. The identified strains and this novel knowledge will aid the further development of a commercial process for vitamin B5 production.
Another tangible result is improved understanding on the mechanisms of action for the bacteriocidal glycopeptide sublancin. Resistance to this antibiotic is, in part, mediated by proteins involved in glucose uptake, zinc homeostasis, oxidative stress management, and cysteine biogenesis. These results might be of relevance beyond the particular antibiotic tested.
- WP5 – Biological engineering for industrial application
Potential socio-economic impact of the research activities in WP5:
The research generated several novel insights for the rational construction of synthetic modules for gene and pathway expression in general, and the production of vitamin B5 and α-amylase in particular. Specifically, strategies for the design of plug-in modules, protocols for high-throughput testing of vitamin B5 and α-amylase production, and a large collection of expression modules and strains that can now be used for further improvements in the production of vitamin B5 and α-amylase were delivered. These results will be of high value to the industrial project partners DNP and NOV. This is underscored by the fact that partner DNP filed two patent applications (WO2012/175574 and WO2012/175575) on the use of high-throughput technology for screening B. subtilis strains with improved vitamin B5 production. Furthermore, special attention was attributed to the establishment of fruitful exchanges with other scientists and the communication of fundamental results to the scientific community. The success of this approach is reflected by several publications with scientists who not participated in BaSynthec, but in other FP7 projects such as the Marie Curie Initial Training Network (ITN) TranSys and the trans-national collaborative project BACELL SysMO2 (see WP6 for details on publications). Thus far, 14 publications describing results from WP5 have appeared in peer-reviewed scientific journals and a book, and results from WP5 have been highlighted on the project web site. Furthermore, we have paid close attention to technology transfer between the academic and industrial partners working in WP5, thereby contributing to improving the competitiveness of the European biotechnology sector. Lastly, WP5 has contributed to the training of researchers in synthetic biology approaches, and Standard Operating Procedures (SOPs) for synthetic biological approaches are made available to the larger scientific community through the website of BaSynthec.
Wider societal implications of the research activities in WP5:
European citizens will benefit from the societal stakes of the project as it responds to some of the “hottest” societal topics such as human health (production of vitamins), and sustainable biotransformation processes (enzyme production). Furthermore, the results of WP5 have been presented to and discussed with safety and ethics experts - Prof. Matti Sarvas and Dr. Robin Pierce - during all consecutive project meetings in order to discuss and create awareness about the societal implications of synthetic biology.
Main dissemination activities of WP5:
Researchers active in WP5 have disseminated their results through 14 publications in peer-reviewed scientific journals, a book, two patent applications, presentations in scientific conferences and workshops, and presentation of project results on the BaSynthec website (see WP6 for details).
Exploitation of the results of WP5:
Commercial exploitation of the results from WP5 is not directly foreseen, but their application potential is clearly underscored by two patent applications of partner DNP. Importantly, the presently created strain collections form an excellent starting point for further rational strain improvement through synthetic biology principles to achieve strongly enhanced vitamin and enzyme production.
- WP6 - Outreach
Communication to the general public
The BaSynthec web site is up and operational since M8 (Deliverable 6.1). It presents the project to the general public and is regularly updated. One newsletter has been posted on the web site. However, this first newsletter was seldom accessed and seemed to have a rather limited impact despite the efforts by the contributing partners to explain their work for non specialists. This observation triggered a discussion within the consortium which was led by Dr Robin L. Pierce from the BaSynthec Ethics Committee. For reasons explained in D6.5 we concluded that the communication on intermediate achievements via the newsletter was mostly likely not at the right level to raise public interest. So, the consortium decided to discontinue newsletters, brochures and communication to the mass media to instead focus on a potentially more productive communication on ethical questions directed towards the scientific community (see D6.5).
Dissemination to the scientific community
- 19 scientific papers acknowledging the EU BaSynthec program were published in peer-reviewed scientific journals between 2011 and 2013. Two pan-consortium manuscripts concernin g on one hand the characterization of strains with massively reduced genomes and on the other hand, the calibration of the Resource Balance Analysis method is currently prepared for publication. Manuscript was prepared with the title “The Call to Scientists (Response to “For Better or Worse”)” addressing societal and ethical implications of synthetic biology.
- 2 patents were filed and published in 2012.
- 38 oral and poster presentations demonstrated the achievements of the BaSynthec project in scientific conferences, meetings and workshops.
- 7 SOPs were generated and will be made available to the larger scientific community through publication of pan-consortium studies (D6.3).
- Successful BaSynthec Synthetic Biology Workshop was held on April 25, 2012 in Dublin (D6.2) with three oral presentations from partners INRA, Novozymes and DNP. BaSynthec leaflet (flyer) and poster were presented and distributed to the participants of the BACELL2012 meeting April 24-25, 2012 (see below).
Technology Transfer
To identify results worth to be protected at an early stage of development in BaSynthec each partner nominated one exploitation experts into the Intellectual Property Use & Dissemination Committee (IPUDC) in December 2010. IPUDC was responsible to review unpublished results, reports and oral and poster presentations to identify results worth to be protected by patents (Deliverable D6.4 will be delivered in September 2013). 2 patent applications were filed and published by DNP.
Ethics committee
The BaSynthec Ethics Committee is composed of Dr Robin Pierce (Delft University of Technology,Delft, Netherlands) and of Pr Matti Sarvas (Professor emeritus, National Institute for Health and Welfare, Helsinki, Finland). The Ethics committee attended our annual meetings in Leiden (June 2011), Dublin (April 2012), and Beaune (May 2013).
The Ethics Committee produced a first report on BaSynthec research activities which was annexed to the Activity Report for the first period. The second ethical report is appended to the activity report for the second period (Annex I). The committee found that “the BaSynthec Project appears to warrant little or no concern in terms of biosafety hazards” and that “there are no explicit biosecurity concerns”.
Under the guidance of our Ethics Committee, all partners were actively involved in the preparation of the manuscript with the title “The Call to Scientists (Response to “For Better or Worse”)” addressing societal and ethical implications of synthetic biology (see Task 6.4 and D6.5 - Awareness and Wider Societal Implications).
An important lesson learned from BaSynthec is about the value of the participation of ethicists to all the scientific and managerial sessions during the annual meetings, exposing them to the internal scientific debates and how problem-solving decisions were taken. We believe that this dialogue within the scientific community, observed and sometime facilitated by ethicists, is of great value to defuse the concept of ethically unaware scientists, and promote within the community of ethicists and policy makers the usefullness of an interdisciplinary (ethics, policy making, science) to ensure the development of well balanced regulations concerning synthetic biology.
List of Websites:
http://www.basynthec.eu/