Introduction of Modelling

Our team has done a modeling based on the model iFF708 whose authors Forster, Famili, et al. are. This has been validated experimentally [1].

The metabolic network in the yeast Saccharomyces cerevisiae was reconstructed using currently available genomic, biochemical, and physiological information.

The metabolic reactions were compartmentalized between the cytosol and the mitochondria, and transport steps between the compartments and the environment were included. A total of 708 structural open reading frames (from now on ORFs) were accounted for in the reconstructed network, corresponding to 1035 metabolic reactions. Further, 140 reactions were included on the basis of biochemical evidence resulting in a genome-scale reconstructed metabolic network containing 1175 metabolic reactions and 584 metabolites [1]. The number of gene functions included in the reconstructed network corresponds to aproximately 16% of all characterized ORFs in S. cerevisiae.

The resulting metabolic network, iFF708, has been used in this project, together with the Flux Balance Analysis tools for assessing the productive capacity of this yeast in the production of 1,8-cineole y S-linalool (products of interest in our project), and geranyl diphosphate (precursor of these products). We also compared the results obtained for the wild type, with those obtained for the mutant ERG20_2 [2], a mutant that increases production of geranyl diphosphate (from now on GPP).

Simulation and Results

  1. Growth
    First we have done a simulation to estimate the growth of the organism, adjusting glucose entry to 3mmol(gDCW*h) because the model was validated against experimental data obtained under these conditions. Also allowed free entry other inorganic nutrients essential for growth (salts of nitrogen, sulfur, phosphorus).

    The simulation objective function was growth function (biomass equation).

    With this first simulation we show that the model works as expected: it is able to simulate normal growth. Flow value obtained for the growth equation was 0.283344 (mmol Biomass)/(gDCW * h), which is adjusted to the expected value according to the experimental validation.

  2. GPP production
    In this section we simulate GPP production, for it, we add to the initial model transport the following reaction:

    GPP : Geranyl diphosphate -> GPPxt.

    It allows to excrete GPP. Since the FBA algorithm simulates steady state metabolism, this reaction is necessary to simulate the accumulation of this product.

    The model run was conducted as a series of simulations that allow us to observe the possible combinations of production growth and optimize the use of inputs.

    With these slogans got the answer shown below:

    Graphic 1. GPP face to Growth.

    This graphic shows the production flow and the flow of GPP growth equation (on the vertical axis) against growth (on the horizontal axis). Shows that the growth and production of GPP are linked by an inverse relationship, if it is able to slow the growth of GPP will be a surplus. At the point of optimal growth of net production value of GPP is zero, since all the GPP produced is used to grow.

    The metabolic cost of production of GPP, that is the reduction in growth caused by the increase in production (∆Growth/∆GPP) gives a value of -0.32 (per unit of production increases GPP growth slows by 0.32 units). The negative value makes sense, as both streams are fed by the entry of glucose, that means are competitive.

    While we must add that the points near the optimum have higher accuracy, since they are closest to the point validated experimentally.

  3. iGEM mutants
    In this section we simulate the production of 1,8-cineole and S-linalool, both products as mentioned above are of interest in our project. Reactions were included on one side corresponding synthesis reactions:

    #18CS : Geranyl diphosphate -> 1,8-cineole + Pyrophosphate

    #SLS : Geranyl diphosphate -> S-linalool + Pyrophosphate

    And the following extraction reactions

    #18C_Out : 1,8-cineole -> 18Cxt

    #SL_Out : S-linalool -> SLxt

    The model run was conducted in three phases. In the first step, production was maximized in 1,8-cineole without the inclusion of S-linalool, the second step was maximized S-linalool production without the addition of 1,8-cineole. And finally a third simulation was made in which both products are maximized, which is constructed for the next reaction.

    #des_obj : 1,8-cineole + S-linalool -> obj

    The data obtained in the simulations showed us, first, that these products could be synthesized in yeast with only the inclusion of the corresponding enzymes. In all these simulations the maximum flow growth reaction remained, without being affected by the modifications added, and remained being 0.283344(mmol of Biomass)(gDCW*h).

    Furthermore, the net GPP flow remained constant since the reaction stoichiometry was not affected by the mutations changes. It is also observed that, both when the flow is optimized for each individual product as when optimizing the sum of both the productive flow coincides with the maximum peak obtained in the previous section for the production of GPP. This is because all of the GPP is produced and is not consumed for the growth is available for synthesis of these products.

  4. ERG20_2 Mutants
    The ERG20 gene encodes an enzyme which shows, on one hand, dimetilaliltransferasa activity, catalyzing the following reactions:

    Isopentenyl diphosphate (IPP) -> Geranyl diphosphate (GPP)

    Dimethylallyl diphosphate (DMAPP) -> Geranyl diphosphate (GPP)

    And on the other hand, geranyl transferasa activity:
    Geranyl diphosphate (GPP) -> Farnesyl diphosphate (FPP)

    This enzyme, in the wild, has a 75% geranyltransferasa activity, and 25% dimethylallyl transferase activity, so that in the wild type all GPP generated immediately becomes FPP. A change in certain amino acids of the enzyme (such as the site 197 of an amino acid for E K, K197E) causes changes in the activity, so that the activities are reversed (now has 25% geranyltransferasa activity and 75% dimethylallyl transferase activity) [2]. This makes more GPP is accumulated and which may occur more compounds derived from this compound, instead of primarily derived FPP as before. Thus, any enzyme that clones in the yeast and to use GPP as substrate, present a markedly improved productivity.

    Since the algorithm FBA is based on steady-state simulations, we cannot directly simulate the effect of a change in activity, since it is a kinetic effect. Therefore, in order to simulate this, there has been a change in the stoichiometry of the reaction GPP production, thus trying to simulate the accumulative effect of the mutation.

    ERG20_1 : Dimethylallyl diphosphate + Isopentenyl diphosphate -> 3 Geranyl diphosphate + Pyrophosphate

    ERG20_2 : Geranyl diphosphate + Isopentenyl diphosphate -> trans,trans-Farnesyl diphosphate + Pyrophosphate

    Since now there is triple affinity by GPP precursors, the production reaction (ERG20_1) will get triple times the consumer (ERG20_2) and this causes it to accumulate. This effect was simulated by changing the ratio 1 to 3 accompanying stoichiometry Geranyl diphosphate.

    With these changes we see:

    Graphic 2.

    This shows that we have maximized evaluating growth and productive capacities shows that the mutant always produces more than the wild type, what means the production of interesting products (1,8-cineole and S-linalool) has increased significantly . As exception we see the optimal point where all the resources are used to grow.

  5. Conclusions
    To conclude, we must emphasize that the wild type organism already has the precursor in the metabolic network, it allows to add the corresponding enzymes to product 1,8-cineole and S-linalool.

    If further we employ the mutant ERG20_2 we get GPP increase production, this implies an increase and hence an improvement in the production of 1,8-cineole and S-linalool.


  1. Förster J., Famili I., Fu P., Palsson B. Ø., and Nielsen J., 2003, “Genome-scale reconstruction of the Saccharomyces cerevisiae metabolic network.,” Genome Res., 13(2), pp. 244–53.

  2. Fischer, M. J., Meyer, S., Claudel, P., Bergdoll, M., & Karst, F. (2011). Metabolic engineering of monoterpene synthesis in yeast. Biotechnology and bioengineering, 108(8), 1883-1892.