Cambridge iBRAIN 2008Engineering self-organisation and electrical signalling in bacteria

IntroductionWe are attempting to engineer multicellularity in single-cel-lular organisms. iBRAIN (integrated Bacterial Recombinant Artificial Intelligence Network) involves engineering or-dered pattern formation and simulating action potentials, the two hallmarks of CNS function. Our team designed a reaction-diffusion system that couples the Agr system of S. aureus, and the Lux system of V. fischeri. We developed a mathematical model which considers and optimises param-eters to produce ideal Turing patterns. To achieve efficient peptide signalling, we developed Bacillus subtilis as an iGEM chassis. Our novel Bacillus compatible vectors also comply with iGEM standards. We also demonstrated the potential of the PCR-based In-FusionTM, which we used to construct these vectors. Finally, we assembled an in vivo bacterial sys-tem that produces a measurable voltage output in response to glutamate.This project sets the foundation for future research with the aim of creating self-organizing multicellular bacte-rial systems capable of performing complex tasks.

ModellingA self-organising system starts off in a homogeneous state with undifferentiated and genetically identical cells, and equal distribution of signalling molecules. However, small stochastic events will break the symmetry, result-ing in differentiation and formation of spatial patterns.Our model of the Agr quorum sensing system of S.aureus was verified by comparing predic-tions with quasi-static behav-iours observed in vivo. Several potential genetic circuits that would generate spatial patterning were investigated in sili-co, and we arrived at two parallel quasi-static systems, both self-activating and cross-inhibiting. The overall model of this system consists of two instances of our agr-signalling model, with added cross-inhibitory and spatial diffusion terms. Thus, from the almost uniform state, we arrived at the following patterns of gene expression:

SignallingOur goal was to implement a simple two-component Reaction-Diffusion system in Bacillus subtilis. Two well-characterized bacterial communication systems can be used to generate this behaviour: Agr peptide signalling from S. aureus is an ideal activatory signal, while the clas-sic Lux system from V. fischeri is a suitable inhibitor.

This project will focus on a tight integration of modelling and experiment; different promoter strengths and other variables were tested, and these parameters were then fed into multicellular models. Next step: use these models to tweak the regulatory ma-chinery that will control signal production.

ReferencesBallal et. al. (2007) ‘The Kdp-ATPase system and its regulation,’ J.Biosci. 32, 559-568Canton et al (2008) “Refinement and standardization of synthetic biological parts and devices’, Nature Biotech 26, 787 - 793Cheng et al (1999) ‘Functional characterization of a potassium-selective prokaryotic glutamate receptor’, Nature 402, 817-821 Meinhard (1992) ‘Pattern formation in biology: a comparison of models and experi-ments’, Rep. Prog. Phys. 55, 797-849Pédelacq et al (2006) ‘Engineering and characterization of a superfolder green fluores-cent protein’, Nature Biotech 24, 79-88Waters & Bassler (2005) ‘Quorum Sensing: Cell-to-Cell Communication in Bacteria‘, Annu. Rev. Cell Dev. Biol. 21 319-346

Our TeamStudents: Linda Boettger, Marie Chapart, Kevin Cheng, Kathryn Chung, Daniel Goodman, Chris Hill, Rebecca Koenigsberg-Miles, Arjun Mehta, Dmitriy Myelnikov, Ellis O’Neill, Ian (Yee Yen) Tang, James Wu, Xiao-Hu Yan

Advisors: Stefan Milde, James Brown, Hugo Schmidt

Faculty: Jim Ajioka, Jim Haseloff, Duncan Rowe, Gos Micklem

VoltageIn order to simulate neural activity in bacteria, a mech-anism resembling a synaptic function is necessary. At-tempting to mimic this in a prokaryotic chassis is partic-ularly attractive as, in a more general sense, it provides an interface between chemical, biological and electrical systems.

The aim was to engineer E.coli to respond to a specific neu-rotransmitter (glutamate) by inducing an ion flux, measur-able as a small change in field p.d. of the growth medium. GluR0, a glutamate-gated transmembrane K+ channel from Synechocystis wPCC 6803, was identified as a candidate pro-tein. The GluR0 BioBrick (BBa_K090002) was then synthe-sised by DNA2.0 after the necessary codon optimisation for expression in both E.coli and B.subtilis.

To assess the effect of GluR0, a sensitive microelectrode was used to measure minuscule p.d. changes in cell suspensions, these were amplified and visualised using an oscilloscope:

These results show a dose dependent voltage change. The initial drop in voltage is probably due to the addition of a dilute solution of an organic anion. The overall response of adding glutamate is a rise in voltage, and is related to amount of glutamate. The 20μl addition rose to above the starting value after about 1 minute, and continued to rise.Next step: design a similar system in Bacillus, and to investi-gate other CNS-like elements, e.g. voltage-gated channels.

DH5αZ1 was used as an expression strain for GluR0. Cells were grown in high (100mM) K+, then resuspended in a zero-potassium osmotic buffer before gluta-mate (agonist) was added. Aspartate was used as a negative control.

Response with different amounts of Glutamate added

0.2V

/div

.

Neutral relative to medium Negative relative to medium

K+ K+K+

GlutamateConstitutive channelInducible

pump

Novel Techniques1. Parts for Bacillus subtillisTo build complex cellular systems, we generated standard-ized parts and techniques for the gram-positive B. subtillis. This chassis offers many advantages to E. coli, including effi-cient protein secretion and accurate integration of DNA into the chromosome. We have submitted 2 gram-positive RBSes (BBa_K090505-6) and 4 promoters (BBa_K090501-4), two inducible and two constitutive.2. Bacillus vectors and InFusionTM

We also designed and built a pair of Bacillus- and E.coli-com-patible vectors: an episomal BBa_K090402 that exists as a 3-5 copy plasmid, and integrational BBa_K090401 that transfers the insert into the chromosome through selecta-ble site-specific recombination.

To construct these, we used the InFusionTM method from ClonTech, which joins PCR fragments using short regions of homology between their ends. This has many advantages over traditional digestion/ligation, including the ability to assemble 4 or 5 pieces together in a single step.3. Superfolder GFPTo enable visualisation of single-copy chromosome inte-gration in B. subtilis, we created a codon-optimised and Bio-Bricked version of the ‘Superfolder’ GFP (BBa_I746916) en-gineered by Pédelacq et al (2006). This protein folds faster and glows brighter than other GFP variants:

I0500 pBAD promoter

B0034RBS

GFP variantcoding region

B0015terminator

J23116promoter

J61117RBS

K090000GluR0 coding region

B0012terminator

Superfolder GFP: (a) E. coli BW27783 growing on LB agar plate containing 10mM arabinose. The three streaks are expressing the GFP variant indicated in the picture under the control of pBAD promoter. (b) Graph showing change in fluorescence levels relative to mut3 GFP with in-creasing concentration of inducer. (c) BioBrick construct for GFP expression studies.

P2 agrB agrD

agrC agrApLac

P

AgrA

AgrA

AgrC

luxI

AgrB

AHL

luxR lacI

LuxR

pLux

AIP