Team:UNITS Trieste/Project

From 2011.igem.org

Revision as of 21:12, 20 September 2011 by Giulio (Talk | contribs)

SYNBIOME OVERVIEW

An important challenge in the near future will be the optimization of bioreactors for the production of complex molecules. The aim of our research project is to combine different cell systems commonly used in biosynthesis through synthetic biology. To improve this system we want to use cells from different kingdoms because we believe that different cell types could cooperate and better produce complex molecules. The innovation and challenge will be to obtain a stable community of cells from different kingdoms and establish mutualism among them. This interdependence will be obtained through a metabolic and signaling pathways in which the survival and/or growth depends from the other cell types.
The project is based on a three-element system: two different bacterial strains and one eukaryotic cell type that communicate through quorum sensing (QS) signal molecules.
In order to achieve the goal in constructing this synthetic community, both the bacterial cells and the eukaryotic cell will be engineered with a genetic circuit under the regulation of the N-acyl homoserine lactone (AHL) QS signals.
More specifically, we will engineer both bacterial strains to produce the enzyme cellobiosidase, in order to convert extracellular cellobiose into glucose, while the eukaryotic cell will be engineered to produce a soluble form of beta-lactamase.
This set up will ensure interdependence among the three cell types; all cells will benefit from the free available glucose and the two bacteria will survive in an ampicillin-containing culture medium.

The mutalism between the two different bacterial strains will occur thanks to a synthetic network based on the two different AHL QS signals, namely 3-oxo-C8-AHL and 3-oxo-C12-AHL.
The inter-kingdom mutualism will be guaranteed by an eukaryotic trans-activator sensible to the AHL QS mediator 3-oxo-C8-AHL.
Importantly, this genetic circuit will be designed in such a way so that it can be adapted to different bacterial species and eukaryotic cell types.

DATA

synbiome model pTraBox P65-TraR LASnlator TRAnslator pLasI-glucosidase pTraI-glucosidase
x

 

 

pTraBox

Generation of pTraBOX-IRES-EGFP

Excision of CMV from pIRES2-EGFP and following riligation of the backbone pIRES2-EGFP supplied by Clontech has been digested in AseI and NheI (Fig.1) in order to remove the constitutive CMV promoter and then the linearized backbone has been purified using the "Wizard Gel Clean Up System" by Promega.
The extremities of the linearized backbone have been blunted in order to allow its self ligation.
XL10-GOLD competent cells have been transformed with the products of ligation and then minipreps have been done.
The colonies have been checked by enzymatic digestion with NdeI and BamHI, the positives must show only one excised fragment of 600bp (Fig.2).

Excision of TraBox-CMV from pSEAP pSEAP has been double digested with EcorI and NotI in Buffer EcoRI plus BSA in 30ul total.
The digestion has been checked on agarose Gel 0.8% W/V.
The fragment corrensponding to the TraBox/CMVmin has been purified using the "Wizard Gel Clean Up System" by Promega.

Cloning TraBox-CMVmin in pCDNA3 using NotI- EcoRI sites: pCDNA3 has been previously cut in EcorI and NotI in order to obtain the linearized backbone ready for the cloning of TraBox-CMVmin.
Different Condition of ligation has been performed looking for the best efficiency.
The colonies obtained in this way have previously been screened by colony pcr and then checked by enzymatic digestion.
All the digested colonies were positive, the fragment excised by the EcorI/XhoI double digeston is the TraBox-CMVmin. (Fig.3)
One of the positives has been chosen and then amplified by trasformation in XL10-GOLD competent cells. The plasmidic DNA has been purified using a commercial Kit supplied by Promega.
The Plasmidic DNA has been subsequently digested in EcoRI and XhoI in order to obtain the same insert previously cloned provided by the XhoI sites.
The insert TraBox-CMVmin has been purified using the "Wizard Gel Clean Up System" by Promega.

Cloning TRABOXCMVmin in pIRES2-EGFP/CMV- using EcorI/XhoI sites in order to obtain pTraBOX-IRES-EGFP TraBox-CMVmin has to be cloned in the pIRES2-EGFP/CMV- previously digested in EcoRI XhoI.(Fig.4)
The linearized backbone has been purified using the "Wizard Gel Clean Up System" by Promega and then ligated with the TraBox-CMVmin as insert.
Different Condition of ligation has been performed looking for the best efficiency.
XL10-GOLD competent cells have been transformed with the products of ligation and then minipreps has been done.
The plasmidic DNA so obtained has been screened by enzymatic digestion using EcoRI and XhoI. The positives have to show the TraboxCMVmin excised in agarose gel electrophoresis separation (Fig5).
Colony N°2 and 4 has been chosen as positive and amplified in order to obtain more plasmidic DNA.

Cloning sBLA in pTraBox-IRES-EGFP sBLA has to be cloned in the pTRABOX-IRES-EGFP previously digested in EcoRI - BamHI.
The linearized backbone has been purified using the "Wizard Gel Clean Up System" by Promega and then ligated with the sBLA as insert.
Different Condition of ligation has been performed looking for the best efficiency.
XL10-GOLD competent cells have been transformed with the products of ligation and then minipreps has been done.
The plasmidic DNA so obtained has been screened by enzymatic digestion using EcoRI and BamHI.
The positives have to show the sBLA excised in agarose gel electrophoresis separation(Fig.6).
Colony N°4 and N°5 have been chosen as positive and amplified in order to obtain more plasmidic DNA.

Checking the final constructs pTRABOX-sBLA-IRES-EGFP In order to check the final constructs both the plamidic DNA obtained by the clone N°4 and 5 has been digested with: -   EcoRI-BamHI: sBLA has to be excised
-   EcoRI-XhoI: TraBoxCMVmin has to be excised
-   NdeI-BamHI: The construct has to be linearized
All the digestions have been checked in Gel electrophoresis separation on Agarose 1% W/V (Fig.7)

P65-TraR

AHL Sensible Eukaryotic Switch

We decided to test both pTraBox-SEAP and p65-TraR (Neddermann P. et al., 2003), kindly provided by Dr. R. Cortese's group, using SEAP (Secreted alkaline phosphatase) as reporter gene, detected with the Great Escape Chemiluminescent assay kit (Clontech).
In our final system we aim to have the presence of both the OXOC8 and the OXOC12 but the eukaryotic cell has to be sensible only to OXOC8.

On this basis the assay was performed in order to test the efficiency of this inducible promoter after the induction with OXOC8 and the response to OXOC12 as unspecific ligand.
AHL has to be dissolved in a organic solvent as Ethyl-Acetate in order to prevent the lactonolysis that will occur in prolonged exposure to aqueous conditions.
2x105 cells for Hela were placed in 35mm culture dishes and transfected using the Fugene HD transfection reagent (Promega). For each transfection 2ug of DNA were transfected.
For all the experimental conditions that we tested, were performed biological triplicates and experimental triplicates.

Figure 1. SEAP activity12 hours after transfection. 2x105 cells for Hela were placed in 35mm culture dishes and transfected using the Fugene HD transfection reagent (Promega). For each transfection 1ug of transactivator plasmid (P65-TraR) and 1ug of pTraR-SEAP reporter were transfected. We decided also to test the basal activity of SEAP under the control of TraBox-CMVmin, in order to achieve this goal hela cells were transfected with 1ug of pTraR-SEAP and 1ug of pCDNA3.
After 6 hours 20uM of AHLs (OXOC8 and OXOC12 separately) were added to cell culture medium and 12 hours after the addition of ligands the medium was collected and the activity of SEAP was measured.
Hela WT were treated with a corresponding amount of Ethyl Acetate + OXOC8.

Figure 2. SEAP activity 24 hours after transfection. 2x105 cells for Hela were placed in 35mm culture dishes and transfected using the Fugene HD transfection reagent (Promega). For each transfection 1ug of transactivator plasmid (P65-TraR) and 1ug of pTraR-SEAP reporter were transfected. We decided also to test the basal activity of SEAP under the control of TraBox-CMVmin, in order to achieve this goal hela cells were transfected with 1ug of pTraR-SEAP and 1ug of pCDNA3.
After 6 hours 20uM of AHLs (OXOC8 and OXOC12 separately) were added to cell culture medium and 24 hours after the addition of ligands the medium was collected and the activity of SEAP was measured.
Hela WT were treated with a corresponding amount of Ethyl Acetate + OXOC8.

Figure 3. Luciferase activity12 hours after transfection. 2x105 cells for Hela were placed in 35mm culture dishes and transfected using the Fugene HD transfection reagent (Promega). For each transfection 1ug of Luciferase reporter plasmid and 1ug of pCDNA3 reporter were transfected as positive control of trasnfection. Cells were treated with 20 uM OXOC8 6 hours after trasnfection.

LASnlator

This part, made of two composite Biobrick, provides the continuous presence of LasR.
This trans-activator is ready to bind OC8 HLA and then it positively regulates the transcription of both the cellobiosidase and LasI, the OC8 HLA synthase present on the same plasmid.






Generation of: Constitutive Promoter – RBS – LasR – Terminator To build this plasmid we used the following parts:
    -  Constitutive Promoter BBa_J23100 - 35 bp
    -  RBS BBa_B0034 - 12 bp
    -  Las R (coding region) BBa_C0179 - 723 bp
    -  Terminator BBa_B0015 - 129 bp
These BioBricks transformed into DH5α, as suggested by the iGEM protocol. BBa_C0179 was digested with EcoRI/SpeI in order to isolate the LasR. BBa_B0015 was linearized with an EcoRI/XbaI-digestion and LasR was ligated ahead of the terminator. The ligated product was than transformed into DH5α and seeded in the presence of the appropriate antibiotic. The growing colonies were tested with a colony PCR (Vf-Vr2 primers), following this protocol: 93° 5’ | 30x (95° 30” | 50° 30” |72° 60’’) | 72° 7’| 4° ∞ The positives, highlighted through electrophoresis, have been expanded and the plasmid extracted.
The same steps were followed to verify the ligation of LasR-Terminator (XbaI/PstI digested) downstream the RBS (BBa_B0034). The positive colonies were extracted and the plasmids have been digested with XbaI/PstI. The RBS-LasR-Terminator fragments were inserted in the plasmid BBa_J23119 at first but, as seen with other constructs, this promoter didn’t work as expected. So, finally we ligated them inside a different promoter, BBa_J23100, previously digested with SpeI/PstI in order to eliminate the RFP reporter and linearize the backbone. As always, the ligation product was tested with the colony PCR and the positive colonies were inoculated to extract the plasmids.
The final BioBrick was tested with an EcoRI/PstI cut, giving our insert long 921 bp, whilst the plasmid is 2000 bp (see figure below). To make LAS- "n"lator suitable for our system, this plasmid was ligated to the PromLasR - TraI BioBrick and inserted into the kanamycin resistant vector pBBR1MCS-3.

Generation of: PromLasR – RBS – TraI – Terminator To build this plasmid we used the following parts:
    -  Promoter Las R regulated BBa_R0079 - 157 bp
    -  RBS BBa_B0034 - 12 bp
    -  TraI from A. Tumefaciens (New!) - 639 bp
    -  Terminator BBa_B0015 - 129 bp
The BioBricks BBa_R0079, B0034 and B0015 were resuspended and transformed into DH5α cells. The PromLasR was digested with EcoRI/SpeI and checked on gel electrophoresis. The purified fragments were ligated upstream of the EcoRI/XbaI-digested RBS and then transformed into DH5α. Then a colony PCR was run to check the transformed colonies (primers: Vf2 – Vr). The protocol used was the following: 93° 5’ | 25x(93° 30” | 50° 30” | 72° 40”) | 72° 7’ | 4° ∞ The positive colonies were expanded, their plasmids extracted and then linearized with a SpeI/PstI digestion.
Meanwhile, a PCR amplification of TraI was made from the gDNA of A. Tumefaciens with TaqPol, following this PCR protocol (oligos on parts page):
95° 5’ | 10x(93° 30” | 56° 30” | 72° 40”) | 23x (93° 30” | 65° 30” | 72° 40”) | 72° 7’ | 4° ∞ The TraI amplification, to which the EcoRI and PstI restriction sites were added through the PCR primers, was verified on gel electrophoresis, extracted and then digested with EcoRI/PstI.
The digested fragment was ligated into pBSIIK and transformed into DH5α. Positive white colonies growing on Xgal-LB agar plates were checked for the presence of TraI with the following colony PCR protocol (PTraI oligo on parts page):
93° 5’ | 25x(95° 30” | 65° 30” | 72° 40”) | 72° 7’ | 4° ∞ The PCR amplification was checked on gel electrophoresis, and the positive colonies selected for sequencing. When the sequencing was completed, analysed and selected the mutation-free samples.
TraI was digested with EcoRI/SpeI in order to ligate it in ahead of the terminator, previously digested with EcoRI/XbaI. The ligated product was transformed into cells and the resulting colonies were checked with a colony PCR (prmers: vf2 and vr):
93° 5’ | 25x(95° 30” | 50° 30” | 72° 60”) | 72° 7’ | 4° ∞ The positive colonies were expanded and their plasmids extracted.
The TraI-Terminator fragment was isolated with a XbaI/PstI digestion, and then ligated into the previously linearized plasmid containing the PromLasR-RBS. We thus transformed DH5α and subsequently performed a colony PCR with the same protocol used before, in order to expand the positive colonies and extract the plasmids.
The completed BioBrick was checked through an EcoRI/PstI digestion and sequencing. The digested fragment of B3 was of the appropriate length (937 bp), as confirmed with gel electrophoresis (see figure below).

TRAnslator


This part, made of two composite Biobrick, provides the continuous presence of TraR.
This trans-activator is ready to bind OC8 HLA and then it positively regulates the transcription of both the cellobiosidase and LasI, the OC12 HLA synthase present on the same plasmid.






Generation of: Constitutive promoter – RBS – TraR – Terminator To build this plasmid we used the following parts:
    -  Constitutive promoter BBa_J23100 - 35 bp
    -  RBS BBa_B0034 - 12 bp
    -  TraR from A. Tumefaciens (New!) - 705 bp
    -  Terminator BBa_B0015 - 129 bp
The BioBricks BBa_J23100, B0034 and B0015 have been resuspended and amplified through transformation into DH5α cells.
Meanwhile, a PCR amplification of TraR was made from the gDNA of A. Tumefaciens with TaqPol, following this protocol:
95° 5’ | 10x(93° 30” | 56° 30” | 72° 40”) | 23x (93° 30” | 65° 30” | 72° 40”) | 72° 7’ | 4° ∞ The amplification was verified on agarose gel and purified. The primers used for the PCR inserted an EcoRI and a PstI site at the ends of the TraR sequence, in order to cut it with the respective enzymes and then clone it into pBSIIK. DH5α were transformed and positive white colonies growing on Xgal-LB agar plates were checked with the following colony PCR protocol: 93° 5’ | 25x(95° 30” | 65° 30” | 72° 40”) | 72° 7’ | 4° ∞ The PCR amplification was checked on gel electrophoresis, and the positive colonies were selected for sequencing. When the sequencing was completed, we analysed and a selected the mutation-free samples.
The TraR obtained was digested with EcoRI/SpeI and cloned into the EcoRI/XbaI-digested BBa_B0015 vector. The ligation product was transformed into DH5α and the colonies were checked with the following colony PCR protocol (primers: vf2 and Vr):
93° 5’ | 30x (95° 30” | 50° 30” |72° 60”) | 72° 7’| 4° ∞ Some of the positive colonies, giving an amplification product of about 1149 bp, were expanded in order to amplify the DNA.
The plasmids were then digested with XbaI/PstI and cloned into the SpeI/PstI-linearized BBa_B0034 vector. As before, we transformed DH5α and we controlled the positives with a colony PCR, using the following protocol (primers: vf2 and Vr):
93° 5’ | 30x (95° 30” | 50° 30” |72° 75”) | 72° 7’| 4° ∞ Positive colonies have been expanded and their plasmids digested with XbaI/PstI, in order to ligate the RBS-TraR-Terminator to a constitutive promoter digested with SpeI/Pst. At first, we used the BioBrick BBa_J23119 but this part didn’t work as expected, so changed the promoter. The ligation was eventually made into the Biobrick Bba_J23100, and each product has been transformed into DH5α and checked with colony PCR. Positive colonies have been expanded as usual. The new construct was then verified through EcoRI/PstI digestion (its length was 903 bp as shown in the picture) as well as sequencing. To complete the final construct of the "TRA-nslator", we ligated the A2 construct downstream of A1, into the tetracycline-resistant plasmid pBBR1MCS-3.

Generation of: PromTraR – RBS – LasI – terminator To build this plasmid we used the following parts:
    -  Promoter Tra R regulated from A. Tumefaciens (New! BBa_K553002) - 151 bp
    -  RBS + LasI + Term composite BBa_K081016 - 735 bp
The BioBrick BBa_K081016 was resuspended with the standard protocol and transformed into DH5α competent cells. The plasmid was then extracted using a plasmid minipreparation commercial kit (EuroClone) and digested with EcoRI/XbaI.
PCR amplification of PromTraR from the gDNA of A. Tumefaciens was made with TaqPol, following this PCR protocol (primers: TraRFw and TraRRev):
95° 5’ | 10x(93° 30” | 56° 30” | 72° 40”) | 23x (93° 30” | 65° 30” | 72° 40”) | 72° 7’ | 4° ∞ PromTraI amplification was verified on gel electrophoresis and extracted. It was then digested with EcoRI/PstI, ligated in pBSIIK and transformed into DH5α. The positive white colonies growing on Xgal-LB agar plates were checked using the following colony PCR protocol (primers: TraRFw and TraRRev): 95° 5’ | 25x(93° 30” | 65° 30” | 72° 40”) | 72° 7’ | 4° ∞ The PCR amplification was checked on gel electrophoresis and the positive colonies selected for sequencing. The PromTraR thus obtained was digested with EcoRI/SpeI and ligated in the BBa_K081016 vector. DH5α were subsequently transformed and the colonies checked with a colony PCR, using the following protocol (primers: vf2 and vr): 93° 5’ | 30x (95° 30” | 50° 30” |72° 60’’) | 72° 7’| 4° ∞ The amplification was on gel electrophoresis and the positives selected.
The new construct was then verified with a EcoRI/PstI digestion and sequencing. As shown in the picture below, the length of the digested part was of the appropriate size (886 bp).
To complete the final construct of the “TRA-nslator”, we ligated the plasmids of A1 and A2 together. This construct was finally into pBBR1MCS-3.

pLasI-glucosidase


This plasmid hosts an OC12 HLA inducible promoter: LasI (BBa_R0079). Once OC8 HLA binds the LasR trans-activator (BBa_C0179) the cellobiosidase transcription is activated thus the bacteria itself can transform the cellobiose in glucose and use it as a source of energy.
The plasmid provides also to generate constitutively the GFP and host a kanamycin resistance.

Generation of: PromLasI – RBS – Glucosidase – terminator

To build up this plasmid we used the following parts:
-Promoter Las I BBa_R0079 - 157 bp
-RBS BBa_J15001 + Glucosidase BBa_K392008 composite Biobrick - 1681 bp
-Terminator BBa_B0015 - 129 bp

For the construction of the first part, in particular the ligation of RBS-Glucosidase inside the terminator vector, see the generation of PromTraI-Glucosidase. As seen with the construct PromTraI-Glucosidase, the first ligation made with the promoter was not successful: the DNA sequencing confirmed that. The construct was the built again from the beginning.

Unlike PromTraI-Glucosidase, the EcoRI/XbaI-digested RBS-glucosidase-terminator vector was ligated with PromLasI, previously cut with EcoRI/SpeI. Different ligation strategies were attempted and later DH5α were transformed. We then performed a colony PCR with the following protocol (PromTrFw and PromTraRev): 93° 5’ | 30x(95° 30” | 50° 30” | 72° 70”) | 72° 7’ | 4° ∞ Positive colonies were expanded and their plasmids extracted. We then proceeded with the quality control: the plasmid was control-digested with EcoRI/PstI and checked on gel electrophoresis, showing the expected pattern with the construct at about 1850 bp and the plasmid backbone at less than 3000 bp. This plasmid was sent to sequence and the results were analyzed and compared with the irregular construct we’d previously made. The composite construct is correct but we observed some mutations in the glucosidase sequence.

To complete our construct, the BBa_I13522 PTet GFP was resuspended, transformed, extracted and digested with XbaI/PstI. This fragment was then ligated into our SpeI/PstI-digested construct. The ligation product was transformed into DH5α, giving some positive green colonies expressing GFP.

Positive colonies were expanded and the extracted and digested with EcoRI/PstI. The digested fragments were finally ligated into the kanamycin-resistant vector pSB1K3, then transformed into DH5α and seeded with the appropriate antibiotic, in order to allow the selective growth only of the colonies carrying our construct B1 ligated in the new plasmid.

pTraI-glucosidase


This plasmid hosts an TraR - OC8 HLA inducible promoter (BBa_K553002). Once this OC8 HLA has bound the TraR trans-activator the beta-glucosidase (Bba_K392008) transcription is activated thus the bacteria can transform the cellobiose in glucose and use it as a source of energy.
The plasmid provides also to generate constitutively RFP and host a kanamycin resistance.
TraR and Prom TraI are new biobrick obtained by Agrobacterium Tumefacens

Generation of PromTraI – RBS – Glucosidase – terminator

To build this plasmid we used the following parts:
-Promoter TraI from A. Tumefaciens (New!) - 151 bp
-RBS BBa_J15001 + Glucosidase BBa_K392008 composite Biobrick - 1681 bp
-Terminator BBa_B0015 - 129 bp
To learn more about how we obtained PromTraI, see plasmid PromTraI – RBS – LasI – terminator

The double terminator BioBrick (BBa_B0015) has been resuspended and then amplified through transformation into DH5α competent cells. The plasmid was then extracted and digested (EcoRI/XbaI). The RFC10 compatible RBS-Glucosidase plasmid (BBa_J15001 + BBa_K392008), a gift from Dr. C. French (University of Edinburgh), was transformed into DH5α, extracted and digested (EcoRI/SpeI). The glucosidase was then ligated inside the BBa_B0015 vector and transformed into DH5α. The transformed colonies were screened with colony PCR (primers: vf2 and Vr), using the following protocol: 93° 5’ | 30x(95° 30” | 50° 30” | 72° 60”) | 72° 7’ | 4° ∞

The plasmid was extracted from the positive colonies and then digested with EcoRI/XbaI in order to linearize it and ligate it to PromTraR, previously digested with EcoRI/SpeI.
DH5α were transformed with the ligation product and the resulting colonies were PCR-screened (primers: vf2 and Vr). The first screening used the following protocol:
93° 5’ | 30x(95° 30” | 50° 30” | 72° 60”) | 72° 7’ | 4° ∞ There was no amplification: probably the sequence is too long.
An alternative protocol, which made use of different and more specific primers flanking the PromTraR sequence (primers: PromTra Fw and PromTra Rev), was employed:
93° 5’ | 30x(95° 30” | 65° 30” | 72° 30”) | 72° 7’ | 4° ∞ The second screening identified some positive colonies, giving a PromTraR amplicon of less than 200 bp. These positive colonies were then amplified to extract suitable amounts of the plasmid. The control digestion of the complete plasmid with EcoRI/PstI showed something unexpected, so we sent the DNA to sequence and in the meantime we started to build again our glucosidase construct from the beginning. We restarted from a different colony of BBa_J15001 and extracted its plasmid. We obtained a clean EcoRI/SpeI cut and ligated it to the glucosidase inside the vector containing the terminator with different strategies. For each ligation, a transformation and a colony PCR have been made (primers: vf2 and Vr).
The last colony PCR finally showed some positive colonies. The protocol used was different, in order to bypass the problem of the amplicon length:
93° 5’ | 30x(95° 30” | 50° 30” | 72° 90”) | 72° 7’ | 4° ∞ The plasmid was exracted from one positive colony, then digested with EcoRI/SpeI. Finally it was ligated to the EcoRI/SpeI-cut PromTraR.
DH5α were subsequently transformed and PCR-screened with different protocols: one amplifying the whole sequence and another specific for the PromTraR sequence. The former gave no result, but the latter revealed the incorporation of the promoter inside the glucosidase vector.
Some of the positive colonies were then chosen to amplify and extract the plasmid.
We quality-controlled the plasmid with an EcoRI/PstI digestion, which showed as expected the construct at about 2000 bp and the plasmid backbone at 3000 bp. This DNA was sent to sequence and the results have been analyzed and compared with the previous construct we had made. The composite construct is correct but we found out some mutations in the glucosidase sequence.

To complete our construct, the BBa_I13521 PTet mRFP has been resuspended, transformed, extracted and digested with XbaI/PstI. The fragment has been purified from gel and ligated into our SpeI/PstI-digested construct, and then transformed into DH5α.

The plasmid obtained from the positive colonies, which were visibly red for the expression of mRFP, was digested with EcoRI/PstI and then cloned into the kanamycine-resistant pSB1K3 expression vector.
The positive red colonies growing from this last ligation contained our completed construct in the new plasmid.

MODELING

The goal
Our goals were 1) To test if our genetic design works as planned; 2) To get preliminary ideas on the main parameters of the two-compartment reactor to be built for the lab experiments. In vivo tests would be clearly too costly and time consuming for this purpose.
The model

Principle We designed a simple model in which QS bacteria have two states. In the i) ground state, glucose intake and signal production is at a low level. If the signal concentration reaches a threshold, the cells enter into an ii) active state characterized by higher glucose intake, higher signal production and the production of glucose from cellobiose. (Figure 1).

The cells are in a perfectly mixed, closed environment. At the beginning of the simulation, a given number of ground state cells are placed into the medium, and the simulation proceeds in discrete time steps. At every step, the cells take up nutrients and carry out the signal and/or glucose production depending on their state of activation. When the cells accumulate a certain amount of energy ("glucose equivalents"), they divide. This program is repeated at every time step. As a result, the cells produce a growth curve quite similar to that seen in liquid cultures.

Figure 1. The two states of the bacterial model.

Figure 2a. Growth of wild type cells that produce a signal and respond to it by producing glucose from cellobiose.

Figure 2b. Carbon source transformation - from cellobiose to glucose - and consumption.

This is a very simple model. In which the cells do not move and the signal is not diffusing within the compartment, since the medium is perfectly mixed. At each time step we can record the number of cells, the concentration of the solutes, etc. This is a so-called agent-based model since each cell-agent executes its own program that depends on its state of activation and stored energy.

Two-compartment reactor. In this setup, two different kinds of cells are put into two equal, perfectly mixed compartments that are separated by a large pore-size semi-permeable membrane. The cells themselves can not pas through the membrane, but the dissolved materials can freely move between the compartments. As compared to the previous one-comaprtment model, the only difference is that the solutes need to equilibrate between the two compartments at every time step. If we assume perfect equilibration between the compartments, the concentration of a solute would be the average of those mesured in the two compartments at each time step. Here we employ a simple trick, we introduce a virtual diffusion coefficient that regulates the exchange between the two compartments. The diffusion coefficient D was defined in such a manner that its values be between zero and 1.0. If it is zero, the two compartments do not communicate. It if is 1.0, the compartments are fully equilibrated, i.e. the equilibrated concentration wiill be the average of the concentrations within the two compartments at every time-step.

The model was implemented using an existing Matlab code written by S. Netotea and A. Kerenyi for modelling the swarming of quorum sensing bacteria on agar plates (Netotea et al, 2009, Venturi et al, 2010), kindly provided to us by the authors. This code had to be slightly simplified, as described in Appendix [1].

Simulation results
We carried out simulation runs at a large number of parameter settings, changing the number of starting population, the ratio of the populations, the starting concentrations of the various solutes, etc. We valuated the results in a qualitatie way, i.e. groth vs. no growth, slower or faster growth. As a comparison, we used a wt model that responds to its own signal. Then we extended the experiments to the species designd to depend on each other.The main finding were as follows:
- Mutually dependent bacterial cells that respond only to the signal of the other species, can grow within the same compartment, same as the wild type. The ratio of the species is 50:50 %. This is a stable equilibrium that can be reached even if one of the populations starts with a single cell (not shown).
- When put into separate but communicating compartments, mutually dependent cell populations will grow, but the speed of growth will depend on the intensity of material exchange between the compartments. If there is no exchange, mutually dependent cells will not grow.

Figure 3a. Growth of the two bacterial populations in separated compartments. c = mixing coefficient.

Figure 3b. When the two bacterial populations are in the same compartment (c = 1) their growth is synchronized.

Figure 3c. In this last condition (3b) we can observe the usual pattern in the carbon source transformation and consumption. The concentration of the signals molecules (OC8 and OC 12 HLA) follow the cells growth patter for both spicies.
- If we put a ternary community into the two compartments, in such a way that lactamase producing partner is in one compartment, and the mutually dependent bacteria are in the other compartment, the system will start only if we jump start it by adding glucose and signal.

Figure 4a. Cell growth of the ternary system without and with the "jump start" .

Figure 4b. The bacterial signaling also switch-on the production of beta-lactamase by HeLa cells that allows the bacterial growth .

Conclusion of the simulations
We used a highly simplified qualitative model in order to have preliminary insights into the logics of our system, in hope of obtaining indications regarding the physical setup of the reaction chamber to be used in the lab experiments. The model shows that the system is in princple "viable", howeer the growth can be limited by the material exchange between the compartments. Also, the system may need to be jump-started by adding a certain amount of glucose and signal. Naturally, these predictions are qualitative and have to be checked by experiment.
Appendix [1]


The QS modelling program of S. Netotea and A. Kerenyi (Netotea et al, 2009, Venturi et al, 2010)is designed to model the swarming of QS bacteria. In this modell i) bacteria are modeled as individuals freely moving on a 2D plane, i.e. in an open environment; ii) bacteria have an individual program that activates them depending on the threshold concentration of a) signals, b) public goods, according to the known rules of QS; and iii) nutrients and solutes freely diffuse on a 2D plane which is discretized into square zones (Netotea et al, 2009). This modell was instrumental in showing that QS regulation is sufficient for a modell popultion to show density-dependent activtion, tracking of exdternal signals, co-swarming of species and community collapse.

Our system is simplified, since due to the perfect mixing within a reactor compartment, a) cells do not move by themselves; b) solutes do not diffuse (concentration is uniform throughout the compartment). So, in order to model growth within a closed compartment, the diffusion and the movement part of the original program had to be simply switched off.

In order to simulate growth in a two-compartment system, one can model the growth of bacteria in such a way, that solutes accumulated separately, i.e. the same program is executed for the two separate components. If we now want to modell the passage of solutes between the compartments, we can use a diffusion-like concentration equilibration at each time step. This setup corresponds to two compartments separated by a semipermeable membrane. The concentration of a solute can be calculated as follows:

c1(t), equilibrated = c1(t) – 0.5 * D* [(c1(t) -c2(t)]

where c1 and c2 are the concentrations in compartment 1 and 2, respectively, and D is the virtual diffusion coefficient, with a value between zero and 1.0. It is easy to see, that at D=0, the concentration remains the same, at m=1, the concentration will be the average of c1 and c2.

Based on the above design, the Matlab code was modified by A. Kerenyi and put to our disposal for modelling as a *.exe file that runs under Windows. The results are cell counts, signal and food concentrations as a function of the time steps These results were visualized by Excell.