Lignin degrading microbial enzymes are capable of degrading plastics. Some previous iGEM teams have exploited this ability. We have built on their work and extended the plastic degradation capabilities of the synthetic biology community by improving polyurethane (PUR) degradation, as this has not been successfully achieved before. We have identified 5 PUR-esterase enzymes from the literature that are capable of catalyzing the below reaction.

Secretion Strategy:

E.coli is commonly used as a chassis in innovative iGEM projects that aim to prove a concept and make the case for a novel function in a biologically engineered machine. In our case, we aim to degrade and synthesise plastic and the degradation part of our system needs to be extracellular. There are many strategies for secretion. You can read about these on our Secretion Guide page and find the corresponding biobricks and projects. If you are currently in the process of designing a Genetically Engineered Machine in E.coli, this page provides a good starting point to look for information.

We chose the pelB secretion tag as it has been demonstrated to work in many cases, sometimes with very high transport efficiency (1)(2). pelB has been used in iGEM projects for many years and is part of 50+ constructs. The UC-Davis team last year used it to secrete LC-Cutinase, a PET plastic degrading enzyme [http://parts.igem.org/Part:BBa_K936013 (BBa_K936013)] successfully, which is somewhat similar to our plastic-degrading enzymes.

Safety:

We considered the safety aspects of using the PudA enzyme, since its sequences were originally from a Risk Group 2 (RG2) organism. We submitted our safety information to iGEM for review. Safety forms were approved on October 2nd, 2013 by the iGEM Safety Committee. We did not use the organism only the part itself and the final sequences were codon optimised for expression in E. coli, which included the elimination of the forbidden restriction sites.

References

1. Thanassi DG, Hultgren SJ. Multiple pathways allow protein secretion across the bacterial outer membrane. Current opinion in cell biology 2000;12(4) 420-430.
2. Sletta H, Tondervik A, Hakvag S, Aune TEV, Nedal A, Aune R, et al. The presence of N-terminal secretion signal sequences leads to strong stimulation of the total expression levels of three tested medically important proteins during high-cell-density cultivations of E.coli. Applied and Environmental Microbiology 2007;73(3) 906-912.

Poly-3-hydroxybutyrate P(3HB) is a polyester which is naturally produced inside bacteria such as Ralstonia eutropha, where it accumulates as granules inside the cell. In its native bacteria it is produced as an energy store (1).It's properties as a plastic has gained it the attention of researchers and industrialists. In our system, the bacteria will take up organic molecules from waste and use them for P(3HB) production.

We have made P(3HB) in E.coli, transferring three genes naturally found in Ralstonia eutropha into E.coli MG1655. These encode the three enzymes necessary for P(3HB) production; polyhydroxyalkanoate synthase(phaC), 3-ketothiolase(phaA) and acetoacetyl coenzyme A reductase(phaB). These are encoded by the pha CAB operon. We have altered the expression of these three genes to maximise the production of P(3HB) as high yields are required for it to be economically viable.

3-ketothiolase:

Acetoacetyl coenzyme A reductase:

Polyhydroxyalkanoate synthase:

We used the pha CAB biobrick [http://parts.igem.org/Part:BBa_K934001 BBa_K934001] to produce P(3HB) which contains the native operon of Ralstonia eutropha. Inthe future we would like to see our system implemented at an industrial scale. This requires high yields to be economically viable. We created P(3HB) synthesis models to inform our experimental attempts to increase P(3HB) yield. We were especially interested to find out if up or down-regulation of any of the enzymes involved in the biosynthetic pathway could increase yields. To analyse this, we needed to know the reactions in the metabolic pathway and also the relative parameters of the enzymes and their dynamic relationships in order to identify bottlenecks in the flux of metabolites. The results from the metabolic model suggest that the amount of pha B is especially critical and increasing its level should give us more P(3HB). (Please see our modelling section for details.) To do this, we designed constructs with stronger promoters to up-regulate the expression of the operon.

We designed and constructed two new constructs for increased expression of the operon (including pha B). The first construct uses the constitutive promoter [http://parts.igem.org/Part:BBa_J23104 J23104] and the [http://parts.igem.org/Part:BBa_B0034 RBS 0034] with a following scar site TACTAGAG in front of the ATG of the pha C gene. The second construct was designed as a hybrid promoter incorporating BBa_J23104 and the original promoter first, followed by the native promoter and RBS, due to recent results of high expression from a hybrid promoter[http://www.plosone.org/article/info%3Adoi%2F10.1371%2Fjournal.pone.0056321]. The region containing the native promoter and RBS is 352 nucleotides long and might contain important regulatory elements and therefore we were interested to test the hybrid construct in comparison to both the native and constitutive promoter configurations.</

We chose the MG1655 E.coli strain as chassis because it constitutively expresses genes that make it resistant to toxic waste degradation products, such as Ethylene glycol. The [http://parts.igem.org/Part:BBa_K892010 aldA] and [http://parts.igem.org/Part:BBa_K892009 fucO] genes have an important role in decreasing the toxic effects of ethylene glycol by converting it into glycolaldehyde which is a link to the cell`s central metabolism. We have received these genes from the registry and future work could be to express these in Bacillus or cellulose-degrading organisms to make them tolerant as well.

E.coli is commonly used as a chassis in innovative iGEM projects that aim to prove a concept and make the case for a novel function in a biologically engineered machine. In our case, we aim to degrade and synthesise plastic and the degradation part of our system needs to be extracellular. There are many strategies for secretion. You can read about these on our Secretion Guide page and find the corresponding biobricks and projects. If you are currently in the process of designing a Genetically Engineered Machine in E.coli, this page provides a good starting point to look for information.

We chose the pelB secretion tag as it has been demonstrated to work in many cases, sometimes with very high transport efficiency (3). pelB has been used in iGEM projects for many years and is part of 50+ constructs. The UC-Davis team last year used it to secrete LC-Cutinase, a PET plastic degrading enzyme [http://parts.igem.org/Part:BBa_K936013(BBa_K936013)] successfully, which is somewhat similar to our plastic-degrading enzymes.

References

1. Economou A. Following the leader: bacterial protein export through the Sec pathway. Trends in microbiology 1999;7(8) 315-320.
2. Thanassi DG, Hultgren SJ. Multiple pathways allow protein secretion across the bacterial outer membrane. Current opinion in cell biology 2000;12(4) 420-430.
3. Sletta H, Tondervik A, Hakvag S, Aune TEV, Nedal A, Aune R, et al. The presence of N-terminal secretion signal sequences leads to strong stimulation of the total expression levels of three tested medically important proteins during high-cell-density cultivations of E.coli. Applied and Environmental Microbiology 2007;73(3) 906-912.

Safety issues regarding our chassis can be found here.

Introduction

The efficiencies for polyurethane (PUR) degradation and ethylene glycol production are important for the performance of our system. We built a mathematical and deterministic model that is based on MATLAB extension Simbiology for polyurethane degradation. The model contains the kinetic properties of degradation enzymes that are helpful for the design of assays. As we scaled up the initial concentrations of all substrates to meet the conditions for a bio-reactor, the model can provide preliminary simulations and predictions for the MAPLE system.

Design

Objective

Here are some specific objectives for the model to achieve:

1. The model should contain the gene expression model of the degradation enzymes because the enzyme concentration determines the rate of plastic degradation. In our case for PUR degradation, we used [http://parts.igem.org/Part:BBa_K206000 pBAD strong promoter K206000] for most enzymes. We built the gene expression model based on inducible pBAD promoter, which gene expression rate can be regulated by inducer concentration.

2. The model should show the efficiency of the enzyme secretion to the culture from the cells. It's also important because the enzyme concentration in the culture depends on it. Here we used pelB secretion tag for most enzymes in order to achieve a high efficiency.

3. The model basically predicts how long will take to degrade a known concentration of soluble polyurethane. It is assumed that the enzyme in our assays has the same kinetic properties as the enzyme used in the literature. The model can suggest a suitable concentration of the plastic to use in order to get good results from the assays.

4. It is known that ethylene glycol is toxic to E.coli. It has, however, no clear effect on the growth of our MG1655 strain when the concentration of ethylene glycol is below 200 mM. The model, therefore, should suggest a safe range of polyurethane concentration to avoid a high concentration (>200mM) of ethylene glycol produced.

The Model

Polyurethane (PUR) degradation involves 5 different degradation enzymes:

4 of the 5 enzymes are not well characterised before, so we could't find enough kinetic data from the literature. The only well-characterised PUR degradation enzyme PudA is used in the model as an illustration of all PUR degradation enzymes. The model will be more complete when the kinetic data of the other enzymes are defined. The finished PUR degradation model is shown as below:

There are two compartments which represents cells and the culture from left to right. The "cell" compartment contains the gene expression module whereas the "culture" compartment contains the degradation module. The "secretion" block that connects two compartments is the secretion module.

Parameters and assumptions

1.Derivation of the maximal expression rate,β

• Average molecular weight (Mw) of a base pair = 660g/mol[http://www.geneinfinity.org/sp/sp_dnaprop.html][http://www.lifetechnologies.com/uk/en/home/references/ambion-tech-support/rna-tools-and-calculators/dna-and-rna-molecular-weights-and-conversions.html]
• Average mass of a base pair = 660g/mol x 1.66x10^-24 = 1.1x10^-21g
• Volume of an E.coli cell = 1µm³[http://kirschner.med.harvard.edu/files/bionumbers/fundamentalBioNumbersHandout.pdf] = 1x10^-15L
  • ∴Mass concentration =
  • ∴Molar concentration of 1 base pair in the volume of E.coli = = 1.66x10^-6 mM
• BioBrick assembly plasmid pSB1C3 is a high copy number plasmid (100-300 copies per cell)[http://parts.igem.org/Part:pSB1C3?title=Part:pSB1C3]
  • assume 200 copies per cell
• ∴ concentration of the gene per cell = N x 200 x 1.66x10^-6mM, where N = number of base pairs
  • ∴ concentration of the gene PudA (N = 1644) in the volume of an E.coli cell is = 0.55mM
• Transcription rate in E.coli = 80bp/s[http://kirschner.med.harvard.edu/files/bionumbers/fundamentalBioNumbersHandout.pdf] = 80 x 1.66x10^-6mM/s = 80 x 1.66x10^-6 x 60mM/min = 7.97x10^-3mM/min
• ∴ Rate of mRNA_pudA production under the control of pBAD = 7.97x10^-3 ÷ 0.25 = 0.015/min

• Average molecular weight(Mw) of an amino acid(aa)= 110g/mol[http://www.genscript.com/conversion.html][http://www.promega.com/~/media/Files/Resources/Technical%20References/Amino%20Acid%20Abbreviations%20and%20Molecular%20Weights.pdf]
• Average mass of an amino acid = 110g/mol x 1.66x10^-24=1.83x10^-22g/L
  • ∴Mass concentration of one aa in the volume of an E.coli = = 1.83x10^-6g/L
  • ∴Molar concentration of one aa = = 1.66x10^-5mM
• Translation rate = 20aa/s = (20 x 1.66x10^-5 x 60)mM/min = 0.020mM/min
• PudA comprises of 548aa[http://www.sciencedirect.com/science/article/pii/S0964830598000663]
  • ∴concentration of PudA's aa in the volume of an E.coli = 1.66x10^-5mM x 548 = 9.10x10^-3mM
• ∴ Rate of protein production = 0.020 ÷ 4.25x10^-3 = 2.2/min

PUR degradation module

The reaction equation of the PUR degradation is:

[Polyurethane]+[PudA]= 4 [ethylene glycol] + 4 [polyisocyanate] + [PudA]

Assumptions:

We assumed 1 mole of polyurethane dispersion can produce 4 moles of ethylene glycol.

The molecular weight of a single polyurethane monomer is 470 g/mol whereas the molecular weight of the polyurethane dispersion is around 2000 g/mol.[http://www.polyurethanes.basf.de/pu/solutions/en/content/group/Arbeitsgebiete_und_Produkte/Grundprodukte/Lupraphen-Produktuebersicht_Gewicht]Therefore, the short-chain polyurethane consists approximately 4 monomers. 4 molecules of ethylene glycol will be produced by degrading one chain of polymer.

We also assumed a simple Michaelis-Menten mechanism for PudA

"[PudA]"is the concentration of the PudA enyme whereas "[Polyurethane]" is the concentration of the polyurethane. The kinetic parameters are defined as below:

Assumptions:

1. The kinetic parameters of PudA are not from the strain E.coli K12 MG1665 we are using. So the assumption here is that the kinetic property of pudA in E.coli is close to that in Comamonasacidovorans

2. Considerations of temperature change and PH change are not included in the model. Therefore, the assumption is that they remain constant during the degradation

Secretion module

The efficiency of Secretion is assumed to be 90% secretion over 2 hours.[http://www.ncbi.nlm.nih.gov/pmc/articles/PMC1251600/] The rate of secretion in the model is therefore:

rate of secretion = 0.9[concentration of PudA]/120 (mM/mins)

Simulation Results and Industrial Implementation

The simulation result we worked out is especially related to the industrial degradation SRF which we aim to achieve a high degradation efficiency. According to the data we estimated for the percentage Polyurethane in SRF, we have about 5% Polyurethane contained in the waste. Therefore, in the model, we assumed 50g of polyurethane in 1kg of SRF based on the estimation. If we put 1kg of waste into a 5L bioreactor, the approximate concentration of polyurethane would be 5mM. Here is the simulation result:

The modelling result shows that we can efficiently degrade Polyurethane in SRF even in industrial scale. 5mM of polyurethane dispersion can be efficiently degraded in 15 mins. The final concentration of ethylene glycol we produced would be 20 mM that is under the toxicity limitation, this proves that our bio-reactor can maintain the efficiency without the toxicity effects.

As for the implementation of MAPLE system, the polyurethane degradation enzymes in our system are very efficient. Therefore, we need an efficient filter which removing ethylene glycol from the system.

PUR degradation model:Download

This section is about modelling the plastic-synthesising E.coli developed in our project. The plastic that this module synthesises is the bioplastic poly-3-hydroxybutyrate, also known as P(3HB). The content below shows the process of building, designing and optimising the model. Also, the considerations, assumptions and limitations will be discussed.

The model was built and simulated in Simbiology, a Matlab package designed for modelling biological systems. In our model we have a compartment (labelled "E.coli_1") that represents our engineered cell. The compartment contains the reactions and species that interact with the polymer synthesis pathway.

Here is our model for P(3HB) synthesis. It is part of module 1, this model uses glucose from waste as the input. Here we have the phaCAB operon, which convertsthe metabolite acetyl-coA into poly-3-hydroxybutyrate. We integrated our synthetic pathway into the metabolic pathway. This allowed us to quantitatively predict the P(3HB) production over time. More details about the metabolic model can be found in Module 2.

Firstly, consider the model concerning the synthetic pathway. This model does not take into account the existence of the metabolic pathway present within the bacteria. The synthetic pathway is mainly used for a sensitivity analysis by assuming minimum interference to the cell metabolism. The pathway is rearranged for clarity.

The pathways on the right half of the cell are genetic expressions of the 3 enzymes involved in the plastic synthesis pathway. The pathways on the left represent the actual P(3HB) production pathway.

As can be seen in the Key (above), the yellow, solid circle represents a reaction object and for each object parameter the following needs to be specified for the simulation to work:

• Rate equation or kinetic law (e.g. mass action, Michaelis-Menten etc.)
• Parameters in the rate equation
• Species involved in the reaction
• Value and units for each parameter

Simbiology will then use an ODE solver (ode15 or sundial) to solve these ODEs and give a plot of the specified output(s) (concentration level of a species over time, for instance).

• Note that 3HB deg, Acetoacetate deg, AcetoacetylCoA deg and 3HB-CoA deg are degradation rates (all = 0.035/min, please see assumptions below) of the corresponding species formed along the P(3HB) synthesis pathway.

Genetic regulations and expressions

BDH2 (3-hydroxybutyrate dehydrogenase)

Schematic of the gene expression of BDH2 as seen in the model diagram

PhaB (Acetoacetyl-CoA reductase)and PhaC (P(3HB) synthase)

Schematic of the gene expression of PhaB and PhaC as seen in the model diagram

• Average molecular weight (Mw) of a base pair = 660g/mol[http://www.geneinfinity.org/sp/sp_dnaprop.html][http://www.lifetechnologies.com/uk/en/home/references/ambion-tech-support/rna-tools-and-calculators/dna-and-rna-molecular-weights-and-conversions.html]
• Average mass of a base pair = 660g/mol x 1.66x10^-24 = 1.1x10^-21g
• Volume of an E.coli cell = 1µm³[http://kirschner.med.harvard.edu/files/bionumbers/fundamentalBioNumbersHandout.pdf] = 1x10^-15L
  • ∴Mass concentration =
  • ∴Molar concentration of 1 base pair in the volume of E.coli = = 1.66x10^-6 mM
• BioBrick assembly plasmid pSB1C3 is a high copy number plasmid (100-300 copies per cell)[http://parts.igem.org/Part:pSB1C3?title=Part:pSB1C3]
  • assume 200 copies per cell
• ∴ concentration of the gene per cell = N x 200 x 1.66x10^-6mM, where N = number of base pairs
  • ∴ concentration of the gene BDH2 (N = 768) in the volume of an E.coli cell is = 0.25mM
• Transcription rate in E.coli= 80bp/s[http://kirschner.med.harvard.edu/files/bionumbers/fundamentalBioNumbersHandout.pdf] = 80 x 1.66x10^-6mM/s = 80 x 1.66x10^-6 x 60mM/min = 7.97x10^-3mM/min
• ∴ Rate of mRNA_BDH2 production under the control of pBAD = 7.97x10^-3 ÷ 0.25 = 0.032/min

2.Protein production rate of BDH2, k₂

• Average molecular weight(Mw) of an amino acid(aa)= 110g/mol[http://www.genscript.com/conversion.html][http://www.promega.com/~/media/Files/Resources/Technical%20References/Amino%20Acid%20Abbreviations%20and%20Molecular%20Weights.pdf]
• Average mass of an amino acid = 110g/mol x 1.66x10^-24=1.83x10^-22g/L
  • ∴Mass concentration of one aa in the volume of an E.coli = = 1.83x10^-6g/L
  • ∴Molar concentration of one aa = = 1.66x10^-5mM
• Translation rate = 20aa/s = (20 x 1.66x10^-5 x 60)mM/min = 0.020mM/min
• BDH2 comprises of 256aa[http://www.uniprot.org/uniprot/Q2PEN2&format=html]
  • ∴concentration of BDH2's aa in the volume of an E.coli= 1.66x10^-5mM x 256 = 4.25x10^-3mM
• ∴ Rate of protein production = 0.020 ÷ 4.25x10^-3 = 4.7/min

3.Protein production rate for PhaB and PhaC, k₃

• Relative promoter strengths: J23104 = 1.3RPU, J23101 = 1.0RPU.[http://sb6.biobricks.org/poster/automated-bioparts-characterisation-for-synthetic-biology/]
  • ∴ 104 is 1.3x stronger than 101.
• In absolute units: take GFP synthesis rate (molecules per min per cell) and approximate that as a generic protein synthesis rate for the promoter.
  • GFP synthesis rate of 101 = 2232 molecules per min per cell.[http://sb6.biobricks.org/poster/automated-bioparts-characterisation-for-synthetic-biology/]
  • ∴ GFP synthesis rate of 104 = 1.3 x 2232 = 2902 molecules per min per cell
• Assume 1 molecule in an E.coli cell gives a concentration of 1nM.
  • ∴ GFP synthesis rate of 104 = 2902 x 1nM = 2.9x10^-6nM/min per cell
• Plasmid copy number assumed as 200 (as in derivation 1)
  • ∴ GFP synthesis rate of 104 in our E.coli = 200 x 2.9x10^-6 = 0.00058nM/min = 0.58mM/min

• Note:
  • v = velocity of reaction
  • Ac-CoA = acetyl-CoA
  • AcAc-CoA = Acetoacetyl-CoA
  • AcAc = acetoacetate

BDH2 (3-hydroxybutyrate dehydrogenase)

Reaction R1:

Reaction R1r:

Values, sources and assumptions

atoAD (Acetyl-CoA:acetoacetyl-CoA transferase (α and β subunits))

Reaction R2:

Reaction R2r:

Values, sources and assumptions

phaB (Acetoacetyl-CoA reductase)

Reaction R3:

Reaction R3r:

Values, sources and assumptions

phaC (P(3HB) synthase)

Reaction R4:

Values, sources and assumptions

atoB (acetyl-CoA acetyltransferase)

Reaction R5:

Reaction R5r:

Values, sources and assumptions

List of enzymes and the corresponding mechanisms by which they work in the synthesis pathway

• BDH2: Ordered-sequential bi-bi mechanism[http://www.brenda-enzymes.org/php/result_flat.php4?ecno=1.1.1.30]
• atoAD: Ping-pong bi-bi mechanism[http://www.sciencedirect.com/science/article/pii/000398617590003X]
• atoB: Ping-pong bi-bi mechanism[http://www.sciencedirect.com/science/article/pii/000398617590003X]
• PhaB: Ordered-sequential bi-bi mechanism[http://digitalcommons.usu.edu/engineering_datasets/1/]
• PhaC: Michaelis-Menten[http://connection.ebscohost.com/c/articles/85133798/new-insights-activation-substrate-recognition-polyhydroxyalkanoate-synthase-from-ralstonia-eutropha]

Using the values presented in the tables above, the deterministic model was simulated to see how much P(3HB) could be formed inside the cell. However, with the single cell model, it would be difficult to reflect how much of the plastic could actually be formed on a macroscopic scale. Therefore, modifications to the simulation algorithm were needed to give scaled-up simulation results.

In the scaled-up regime, though, new limitations were encountered. For example, growth, cell division and the concentration gradient that each cell encounters and how much can be taken up by the cells. Given the data available, what could be done was to assume that the bacteria were in a bioreactor, with constant nutrient and oxygen supply, and that they were in the phase of stationary growth (please see assumptions and explanations under the graphs below).

Here, three graphs are presented to show P(3HB)formation over time. The first two graphs are (i)single-cell model results, concentration in mM and (ii) single-cell model results, concentration in g/L. The first graph was the original graph produced from Simbiology as all concentrations were expressed in mM in the model. However, to gain a more intuitive sense of how much could be produced g/L was used. The third graph is the scaled-up simulation result.

(i) please see assumptions and explanations below

(ii) please see assumptions and explanations below

(iii) please see assumptions and explanations below

Graphs (i) and (ii): assumptions and explanations

• To convert from mM to g/L:
  • Molecular weight(Mw) of P(3HB) = 1.8x10¹²g/mol[http://webpages.charter.net/hvalentin/PDF-files/18.pdf]

(Please note that this value was taken from a paper about the "Production of poly(3-hydroxybutyrate-co-4-hydroxybutyrate) in recombinant Escherichia coli grown on glucose" instead of the production of poly(3-hydroxybutyrate). However, this is the closest we managed to get as far as species, substrate and type of plastic are concerned.

Also, it was noticed that the molecular weight would change depending on the cultivation time (as conditions such as growth and pH can also affect it) and it was hard to find a paper that contains the exact same conditions and the range of cultivation times for our simulations. Furthermore, in reality there wouldn't be a single value of molecular weight obtained, but a range of values (known as polydispersity).[http://aem.asm.org/content/78/9/3177.full] Therefore, it was assumed that the molecular weight would remain constant during the time course of the simulation and that the value should be considered as an average.)

• ∴P(3HB)concentration in mM x 10^-3 x 1.8x10¹²g/mol = concentration in g/L

Graph (iii): assumptions and explanations

• Assume 1.5x10¹²cells in a 1L broth, calculated from value obtained from BioNumbers. A range was given: 1-2x10⁹ cells/ml, so 1.5x10⁹ was taken. [http://bionumbers.hms.harvard.edu/bionumber.aspx?s=n&id=104831&ver=5]
  • Conditions and considerations accompanying this value are:
    • growth medium: LB broth
    • Cultivation temperature: 37°C
    • wild-type E. coli K-12 strain MG1655 (NB: though not exactly our engineered one, it is the correct strain)
• Assume stationary growth phase.
• Assume that the amount of P(3HB) accumulated during the time course of simulation doesn't exceed the cell's capacity or ability to contain it.
• Same metabolic considerations as above (single-cell results)
• Simbiology single-cell simulation results multiplied by the number of cells stated above to give the red curve.

Sensitivity analysis: species concentrations

What is sensitivity analysis and why do it?

In order to look at how we could increase the production of P(3HB) we decided to run a sensitivity analysis (also in Simbiology) to identify what species in the model P(3HB) is sensitive to, given the specific conditions with which we set the model. In other words, we calculated the time-dependent sensitivity of P(3HB) with respect to the initial conditions of other species (species formed along the synthesis pathway and enzymes involved).[http://www.mathworks.co.uk/help/simbio/ug/calculating-sensitivities.html#brbumlr|ref]

For a species x, whose concentration depends on time, it can be expressed as x(t). To calculate the sensitivity of x(t) with respect to another species y(t), the sensitivity with normalisation relative to the numerator x(t) can be calculated as:

Therefore, the underlying algorithm for this sensitivity analysis of P(3HB) is:

Sensitivity analysis: enzyme concentrations

The sensitivity analysis of the enzymes involved in the synthetic pathway is carried out, which determine the most sensitive enzyme in the pathway

We plot a time integral of the sensitivity analysis which has the algorithm as:

The senstivity result is:

According to the simulation result, we found that the PHB production is highly sensitive to the concentration of PhaB and PhaC. In contrast, the system is not sensitive to the concentration of atoAD and atoB. In theory, increase the concentration of any of the enzyme will increase the flux of the system. However, for the strain E.coli K12 MG1655, both atoAD (acetoacetate:acetoacetyl-coA transferase) and atoB （acetoacetyl-coA thiolase) favour the reverse reactions, it becomes the rate limiting step in our pathway. Acetoacetyl-coA intends to be converted to acetoacetate instead of 3HB-coA. [http://www.sciencedirect.com/science/article/pii/0003986177904969]Therefore, the concentration of PhaB become really critical which push the system to flow in the forward direction.

Scan with different levels of PhaB

The plot below is to show that increasing PhaB concentration within the engineered organism would have a positive effect on the concentration of the P(3HB) synthesised. As so many parameters within the model were taken from different sources, it would be difficult to give a simulation plot that is accurate quantitatively. It was decided to show the single cell simulation for this because the scans were carried out up to the doubling time of 20mins, hence avoided the issue of scaling-up which would have otherwise subjected the simulation to more errors. Therefore, this plot should be interpreted from a qualitative perspective and the trend should be observed.

Difference between promoter expressions after optimisation

In order to choose the strongest promoter that is available to increase the gene expression. We ran a parameter scan of enzyme expression rates of a range of Anderson's constitutive promoters. The simulations shows the PHB production under different Anderson's promoters. The promoters we tested are:

The derivations of the GFP synthesis rate are:

*Relative promoter strengths: J23104 = 1.3RPU, J23101 = 1.0RPU, J23100 = 0.92RPU, J23118 = 0.76RPU[http://sb6.biobricks.org/poster/automated-bioparts-characterisation-for-synthetic-biology/]

• In absolute units: take GFP synthesis rate (molecules per min per cell) and approximate that as a generic protein synthesis rate for the promoter.
  • GFP synthesis rate of 101 = 2232 molecules per min per cell.[http://sb6.biobricks.org/poster/automated-bioparts-characterisation-for-synthetic-biology/]
• Assume 1 molecule in an E.coli cell gives a concentration of 1nM.
  • ∴ GFP synthesis rate of 101 = 2232 x 1nM = 2.2x10^-6nM/min per cell
• Plasmid copy number assumed as 200 (as in derivation 1)
  • ∴ GFP synthesis rate of 101 in our E.coli = 200 x 2.2x10^-6 = 0.00045nM/min = 0.45mM/min
  • ∴ The GFP synthesis rates of other promoters can be calculated by multiplying the RPU value with the GFP synthesis rate of J23101 promoter.

The simulation result:

Although there is just a small increase the gene expression rate, J23104 promoter achieved a much higher PHB production than other constitutive promoters. Therefore, we decided to use J23104 promoter for our improved biobrick [http://parts.igem.org/Part:BBa_K1149051 BBa_K1149051].

Here is our experimental results that show we have 11 times more PHB after the change of promoter.

Comparison of P3HB production (left) 1.5ml tube, natural phaCAB (BBa_K934001) (right) 5ml tube, phaCAB expressed from the hybrid promoter, (BBa_K1149051).

P(3HB) synthesis model:Download

Additional script:Download

In Module 1: Resourceful Plastic, glucose from the degradation is used as the input for the system that produces P(3HB). The P(3HB) synthetic pathway hence needs to be adapted for this purpose. To do so, 3HB monomer would no longer be the source input to the engineered system. So, BDH2 should be removed from the pathway. Acetyl-coA would then become the key resource to produce P(3HB). Also, a metabolic model was needed to show the production of Acetyl-coA from glucose. Furthermore, our synthetic pathway involves several key metabolites from the pathway such as NADPH, NADP+, NADH and NAD+. The initial concentration of those metabolites can be determined from the metabolic model by assuming they are in their steady states. More features and results of the metabolic model can be found in Module 2.

Initial concentrations of metabolites

The model presented in Section 2.0 contains a synthesis pathway that interact metabolic species involved in so many other metabolic reactions elsewhere in the cell. It was, therefore, difficult to determine how much of these metabolites would be left available for our plastic synthesis pathway to utilise. So, a simple metabolic model of the cell (more detail in next paragraph) was needed to simulate and show insight into this problem. We decided to use the steady state levels of these metabolites (without the engineered pathway) and plug the values into the P(3HB) synthesis model in Section 2.0 as the initial concentrations.

The metabolic model involves two pathways - glycolysis pathway and the TCA cycle, where glucose is used as the sole carbon source. The reason is that glucose is directly taken by the glycolysis pathway and there are interactions between Acetyl-coA and TCA cycle. This Simbiology model is based on the one built by Angela Dixon in Predictive Mathematical Model for Polyhydroxybutyrate Synthesis in Escherichia coli.[http://digitalcommons.usu.edu/engineering_datasets/1/]. All reactions and pathways are based on ODEs and kinetics data presented in the original model. It can dynamically predict the concentration of all metabolites involved over a defined simulation time course.

The metabolic model for determining steady state concentrations is shown as below:

Keys:

The initial concentrations of all metabolites and kinetic data are referenced in the paper.[http://digitalcommons.usu.edu/engineering_datasets/1/] The simulation of the metabolites is:

Table of initial and steady-state concentrations of the metabolites:

<p align="justify"> It should also be noted that acetate, acetoacetate and acetoacetyl-CoA cannot be predicted by the metabolic model. Therefore, their initial concentrations are rough estimates (value similar to other metabolites) such that a reasonable P(3HB) output can be produced.

These assays were designed to test whether our chassis, E.coli (MG1655) could grow directly with waste over a long period of time.

Waste media

LB-waste assay(A) Waste media (B) E.coli containing mCherry stress biosensor (BBa_K639003) were grown in mixed waste (A) over 3 days, then streaked in a qualitative assay to check for growth. (C) mCherry stress biosensor (BBa_K639003) transformed E.coli were streaked again after 7 days growth in SRF. Figure made by Imperial College London 2013 iGEM.

PBS-waste assay(A) waste media made up in PBS (phosphate buffered saline). (B) E. coli expressing mCherry stress biosensor (BBa_K639003) grown in waste media (A) over 3 days, then streaked onto an antibiotic containing plate to qualitatively assess whether the E.coli had survived. (C) Streaked again after 6 days growth in SRF. Figure made by Imperial College London 2013 iGEM.

Conclusion: MG1655 E.coli are viable and grow on mixed waste alone. Therefore we have established that our chassis could survive in a mixed waste bio-reactor context, which is validation of our concept to industrially implement our system.

Waste conditioned media

These assays were designed to test whether our chassis, E.coli MG1655 strain could grow with waste conditioned media (WCM) over a period of 24-48 hours. Waste conditioned media is a filter sterilised version of the waste media and was designed for several reasons; Firstly we were unsure whether mixed waste would be toxic to E.coli and hence a less concentrated version may be more suitable and secondly large chunks of waste would prevent accurate OD600 measurements and therefore we decided to filter out the largest chunks.

Photograph of waste conditioned media cultures mCherry stress biosensor (BBa_K639003) transformed MG1655 were grown with LB-WCM at 37ºC, with shaking for 48 hours. Figure made by Imperial College London 2013 iGEM.
Growth assay in waste conditioned media (WCM).E.coli MG1655 strain transformed with [http://parts.igem.org/Part:BBa_K639003 mCherry stress biosensor.] were grown with LB-based waste conditioned media at 37ºC, with shaking. Error bars represent S.E.M., n=4. Figure made by Imperial College London 2013 iGEM.	Growth assay in waste conditioned media (WCM). E.coli MG1655 strain transformed with either empty vector control (EV) or mCherry stress biosensor (BBa_K639003) were grown in WCM for 5 hours, with shaking at 37ºC. Error bars are SEM, n=4. Figure made by Imperial College London 2013 iGEM.

Conclusion: MG1655 transformed with either empty vector (EV) control or mCherry stress biosensor (BBa_K639003) vector are viable and can grow in waste conditioned media. Therefore waste conditioned media is an appropriate and novel experimental media with which to characterise biobricks within a mixed waste/landfill context. These data are also characterisation of an existing biobrick (BBa_K639003)

PUR Esterase enzyme activity

Cell lysate assay

Since the PUR Esterases were not secreted, initially the cells were lysed to obtain crude cell extracts in order to test whether the enzymes are active. The Western Blot results showed that the constructs EstCS2 [http://parts.igem.org/Part:BBa_K1149002 BBa_K1149002], PueB [http://parts.igem.org/Part:BBa_K1149004 BBa_K1149004] and PulA [http://parts.igem.org/Part:BBa_K1149006 BBa_K1149006] were being expressed. The cultures expressing these three constructs were grown, lysed by sonication and utilised in a colourimetric assay with the substrate analog para-Nitrophenyl butyrate. The data shows that PUR Esterase EstCS2 [http://parts.igem.org/Part:BBa_K1149002 BBa_K1149002] is definitely active.

para-Nitrophenyl butyrate (p-NP) is commonly used to indicate an enzyme’s esterase activity. The enzyme cleaves the ester bond and releases the 4-nitrophenol (4-NP), thus causing a colour change from colourless to yellow and an increased absorbance at wavelength 405 nm.

para-Nitrophenyl butyrate is cleaved by Esterases to release 4-Nitrophenol. This is accompanied by an increase in absorbance at 405 nm. Figure adapted from TU Darmstadt 2012 iGEM

The assay was run in the [http://www.eppendorf.com/int/index.php?sitemap=2.1&action=products&contentid=1&catalognode=87236 Eppendorf BioSpectrometer] was used to automatically read the absorbance of the reaction mixture every 30 seconds. The concentration of 4-Nitrophenol produced from the reaction was calculated using the Beer-Lambert Law; the extinction coefficient of 4-NP is 18,000 M-1 cm-1 at 405 nm.

The results below show the concentration of 4-NP produced by the three different PUR esterases and compare them to the Empty Vector and Substrate alone as negative controls.

The increase in absorbance that accompanies the cleavage of para-Nitrophenyl butyrate by PUR Esterase EstCS2. Figure by Imperial College London iGEM 2013

The concentration of 4-Nitrophenol released by PUR Esterase EstCS2 activity. Empty Vector and Substrate alone were used as negative controls. Figure by Imperial College London iGEM 2013

The above graphs clearly show that PUR Esterase EstCS2 is active and cleaving p-NP. The recorded concentrations of 4-NP, in the presence of this PUR Esterase, are much greater than with the Empty Vector or the Substrate alone. Is PUR Esterase EstCS2 active? Yes!

The enzyme activity of PUR Esterase PueB. The graph shows the concentration of the 4-Nitrophenol released during 10 minutes incubation of para-Nitrophenyl butyrate with PueB and Empty Vector, as well as the p-NP substrate by itself. Figure made by Imperial College London 2013 iGEM

The enzyme activity of PUR Esterase PulA. The graph shows the concentration of the 4-Nitrophenol released during 10 minutes incubation of para-Nitrophenyl butyrate with PulA and Empty Vector, as well as the p-NP substrate by itself.Figure made by imperial College London 2013 iGEM

The enzymes expressed by both PueB and PulA do not appear to have esterase activity for this substrate. There is a tiny increase in 4-NP concentration for PueB, PulA, Empty Vector and Substrate alone. This probably indicates that the substrate is slowly degrading by itself. Other constituents of the cell lysates are likely to be causing a slight increase in p-NP degradation, as PueB, PulA and Empty Vector show higher concentrations of 4-NP than the Substrate alone.

Conclusion: PUR Esterase EstCS2 is active!

Ethylene glycol

Ethylene glycol in LB with Stress Response cells. MG1655 were grown in ethylene glycol, a byproduct of polyurethane degradation. Cells were grown in 0mM, 100mM or 200mM Ethylene Glycol. At 37°C, the concentrations of ethylene glycol used do not affect growth, however at 30°C, the increasing concentration results in halved growth. A two-tailed t-test addressing the null hypothesis, temperature does not affect growth with ethylene glycol shows that the null hypothesis must be rejected as p = 0.001. Data points show final time point after 6h growth for each concentration. Growth was at 37°C and 30°C with shaking over 6h. Error bars are SEM, n=4. Figure made by Imperial College London 2013 iGEM.

Reduced growth at 30°C likely due to decreased efficiency of MG1655 ethylene glycol break down enzymes. These enzymes (see UC Davis 2012) are endogenously expressed and detoxify Ethylene Glycol.

E.coli MG1655 containing the PhaCAB construct Part:BBa_K934001[http://parts.igem.org/wiki/index.php?title=Part:BBa_K934001] accumulates the bioplastic P3HB inside its cells. P3HB can be visualised through staining with Nile Red, which fluoresces strongly in a hydrophobic environment, such as when it binds to the P3HB. However staining will be much lower in cells where P3HB is not present. To see whether P3HB was being produced we grew E.coli containing the phaCAB operon under the native promoter from Ralstonia eutropha. This produced the results we hoped for as seen in the image below.

The P3HB production model indicated that upregulating atoB would lead to an increase in bioplastic production. We changed the promoter before the phaCAB operon to the strong promoter J23104 and a hybrid promoter we created.

The Nile Red plates also provide a preliminary way in which to investigate the effect of changing the promoter to a stronger version. The darker staining of the streaks of the constitutive and hybrid promoter containing constructs suggests the production of higher levels of P3HB in these strains.

E.coli transformed with the phaCAB operon(native, hybrid and constitutive are different promoters expressing the operon) fluoresce when grown on LB-Agar plates with Nile Red stain. The P3HB produced within them is stained. Those without the operon do not(EV- transformed with empty pSB1C3). When the promoter is changed from the native form to the hybrid or constitutive J23104 then the fluorescence is more intense, indicating increased P3HB production.

We imaged the cells using the fluorescent microscope to qualitatively understand how much P3HB is produced per cell. As indicated by other authors(1), we saw that P3HB is found in localised granules within the bacteria. Empty vector containing cells show little fluorescence. In those that do, the staining is qualitatively different than in the phaCAB containing strains. Nile Red is also used as a lipid stain and this is indicated by its staining of the membranes of EV cells. In contrast, in phaCAB containing cells, staining is restricted to points within the cells.

Each fluorescent image was created by exciting the stain at 530nm for 100ms. Despite this standard treatment of samples the intensity of fluorescence produced by the cells containing the hybrid promoter is greater than from the constitutive ones. This could indicate a difference in the strength of the promoters which lead to phaCAB expression and therefore the amount of P3HB produced. The images also suggest that not every single E.coli is stained. This may be due to the conditions in which the cells were grown or may be dependent upon the stage in the lifecycle of the E.coli.

Fluorescent microscope images of E.coli stained with Nile Red. 1 -Empty vector, 2 - constitutive promoter phaCAB, 3 - hybrid promoter phaCAB. Images labelled a are fluorescent images excited at 530nm ovelaid on the bright field images shown in the b set of images. Images by Imperial College iGEM Team

We extracted the P3HB using a simple technique which maintains the chain length of the P3HB while releasing it from the cells. The amount of plastic extracted from the hybrid promoter-phaCAB(BBa_K1149051[http://parts.igem.org/Part:BBa_K1149051]) was much larger compared to the native promoter-phaCAB. All other conditions were kept the same.

A comparison between the amount of plastic produced by bacteria containing an empty plasmid(EV), native promoter-phaCAB and hybrid promoter-phaCAB after plastic extraction. The EV does not produce P3HB while the hybrid produces much more than the native promoter.

We converted the data into the format with which industry compares P3HB production. These are mass of plastic per dry biomass and mass of plastic per litre of culture media. The strain containing the hybrid promoter had a ten fold higher ratio of plastic to biomass than that produced by the original part BBa_K934001

A summary of the improved production of P3HB by our hybrid promoter-phaCAB construct(BBa_K1149051) over the native promoter-phaCAB.

Comparing the amount of plastic produced by the two constructs is visually striking.

Comparison of P3HB production (left) 1.5ml tube, 90mg produced by native promoter-phaCAB (BBa_K934001), (right) 5ml tube, hybrid promoter-phaCAB produced 2.05g P3HB, (BBa_K1149051).

P(3HB) From Mixed Waste

E. coli (MG1655) transformed with native phaCAB (BBa_K934001) were grown with 600ml waste media (12g mixed waste in 600ml M9M- Minimal Media) overnight (16h) at 37oC with shaking. Prior to centrifugation large mixed waste chunks were filtered out. P(HB) was purified as described in Protocols.

P(3HB) purification: The diagram summaries the process of P(3HB) purification from E. coli (MG1655) transformed with native phaCAB (BBa_K934001)

See here a video of the process.

3HB Assay: Confirming production of 3HB

We have adapted a medical kit for 3HB detection for chemically analysing our samples. This provided a simple but accurate mean to measure the 3HB liberated from the polymer after phaZ1 depolymerase enzyme treatment. The monomer 3HB indicated the presence of P(3HB) in the material we purified from cells grown on waste as a sole carbon source. Please see the protocol here and our data and calculations here.

The chemical analysis of the produced bioplastic. The samples break break down to 3HB monomers after treatment with our PhaZ1 enzyme (BBa_K1149010). We synthesised P(3HB) using our improved Biobrick part (hybrid promoter phaCAB, BBa_K1149051). Our engineered bioplastic producing E.coli synthesised P(3HB) directly from waste. (Imperial iGEM data here.)

The bioplastic produced by our bacteria contains PHB and can be broken down into monomers by PhaZ1 enzyme (expressed by our bacteria)Imperial iGEM data.

We repeated the previously described experiment with E.coli cells containing all three P(3HB) production operons, grown on waste media. We also also tested the material left after the extraction protocol from cells containing the Empty vector only, grown on waste media. The yellow colour indicates presence of 3HB.

Production of PHB from waste confirmed. Imperial iGEM data.