INFORMATION MOVEMENT IN A BACTERIAL NETWORK

Overview

The goal of this project is to identify a probabilistic model for propagation of information (mobilizable plasmids transmission) in a dynamic population (batch culture). The motivation for this goal is identifying movement of the transmissible plasmid as to determine the distribution of E. coli having received the information. Accordingly, the transient response of information must be interpreted as it spreads throughout the system via conjugation. This model serves as a generalized case of bacterial growth and conjugation that may be utilized for a specific system in which conjugation acts a means of forming a bacterial communication network.

Problem Assumptions

The population is well mixed; trajectories of cells are therefore ignored
The carrying capacity of the system is determined by the size of the vessel
The cells of one population type are identical in cell size to all others
Individual cell growth is ignored; rather we assume instantaneous fission after which daughter cells are of equal size as previous existing parent cell
Assume cells have different replication rates depending upon population type
Assume the replication rates vary with time dependent on age of culture
All donor cells are capable of conjugating with any recipient cell
All transfers result in recipient cell becoming a donor cell (capable of transmission of plasmid)
There is no latent period between moment of transmission and moment of becoming capable of transmission.
No cell death (purely birth process); not interested in system in death phase
Cell replication rates increase with number of plasmids contained in the cell
Only two cells can conjugate
Transfer of only one plasmid at a time during conjugation event
Transfer is unidirectional from donor to recipient
Rolling replication mechanism for conjugation ensures one copy will remain in the donor after transfer.

One can consider the model design smilar in principle to an epidemiological system. For instance, the plasmid (representative of the disease in the epidemiological model) possesses multiple transmission modes. It is capable of vertical transmittance from parent to offspring and also horizontal transmittance from one cell type (donor) to another cell type (recipient). Upon transmission of plasmid from donor to recipient, the "infected" cell changes phenotype to now becoming capable of transmitting the mobilizable plasmid as well (or becoming infectious).

Growth Model

Growth phases of the system are modeled based on the concentration of the system and time of initial inoculation of culture. The replication is considered to occur instantaneously at the generation times which are determined by the amount of plasmid in each cell. The model incorporates three states of either “open”, “replicating”, or “conjugating”. The open state represents the time during which either conjugation or replication can occur. During the lag phase, the transition to a replication state has a probability of approximately zero as the cells are adapting to their environment. After 2.5 hours, the cells have entered the log phase to experience a doubling at there specified generation time +/- 1% as ln(xt)=ln(xo) + u*t, where u=max replication rate and x is the number of cells. The rate of replication decreases as the culture approaches the carrying capacity designated. As the carry capacity is reached and stationary phase is initiated, the rate of replication, u, again approaches zero.

Conjugation Model

The conjugation events were modeled using the Gillespie algorithm. The Gillespie algorithm determined the waiting time for the next conjugation event and the type of conjugation event which occurred. The two types of conjugation events, u, in the system included: a1) recipient cells and donor cells with mobilizable plasmids incorporated into their chromosome; a2) recipient cells and donor cells with separate mobilizable plasmids. At this point, the conjugation event was either determined to be of transfer state or no-transfer state. The system allows the Gillespie algorithm to be simplified to au=cu*xD*xR given:

• au represents the conjugation rate

• xD represents number of donor cells

• xR represents the number of recipient cells

• Cu represent stochastic transfer rate constant

• Cu=transfer rate * concentration

The value of au is then used to determine the waiting time for the next reaction as t = (1/ao)ln(1/r) where ao is the total conjugation rate of the system (ao=a1+a2) and r is a uniformly distributed variable from (0,1). The algorithm given a larger conjugation rate or concentration of cells will results in a greater chance the reaction will happen in the next step of the simulation.

Computational Approach

Initialization

• Set values for system parameters

• Set initial total number of cells in culture at inoculation

• Set seed percentage of donor cells

• Set simulation time to 0

Iteration

• Calculate values for a1 & a2

• Generate uniformly distributed random variables

• Calculate wait time for next event, dt

• Calculate index of the next event

• Calculate the transfer state

• Adjust the number of cells in the system (due to replication or conjugation event)

• Calculate the fraction of cells with information

• Set the time of the simulation to t + dt

Termination

• End when time reaches 48 hours or concentration of recipient cells equals zero

Interpretation of Results

Simulation has provided us with a look at the nature of the response times for plasmid transmission through a bacterial network. Results indicate faster response times, noted by the lower time constants, for larger seed percentages of donor cells. This is intuitive as the transfer rate is concentration dependent. These values for the time constant were shown to decrease almost linearly for an increase in seed percentage. One can also see a faster response time for an increase in the total initial concentration of cells in the system. This represents a faster spreading of plasmid throughout the bacterial network by inoculating an environment with a higher concentration of cells. We would like to see a faster spread of plasmid throughout the system though initial total concentrations are preferred to be lower as this is the purpose of bacterial growth culture (ie. creating many cells from few). If one did not care about how much bacteria was being utilized to transfer plasmid, the largest initial concentration would be used. This is seen from results indicating a gradual increase in the abundance of F plasmid bearing cells after a 48 hour culture period for larger total initial concentrations. It should be noted that the initial concentration would necessarily be less than 10^9 cells/mL as this is the assigned maximum concentration to allow sustained stationary phase growth, and there would be no purpose of such a culture in practice. To recapitulate, a higher initial concentration will provide a higher fraction of cells containing the mobilizable plasmid at stationary phase. Lower initial concentrations may be used, though in order to attain the maximum information propagation by the end of stationary phase a higher seed percentage must be used. The result of the lower initial concentration will be an increase in the final fraction of cells without carrying information (mobilizable plasmids) representing the error in a perfect broadcast in a bacterial communication network. The simulation results have been successful in demonstrating the information movement and have identified the competition for mobilizable plasmids apparent by the sigmoid transient response

Validation

Validation of the model is identifiable through flow cytometry experimentation in which the plasmid information will represent a gene for a fluorescent protein. The fluorescent protein will act as a marker for identification of the cell type as donor. Those cells not exhibiting this fluorescent marker are counted as recipient. Counting of these cells in cultures flowed into analysis system at various time points for various initial concentration and seed percentages provide a histogram of the distribution of these cell types in the culture. In this sense, the information content can be assayed providing an accurate method for comparing model generated transient responses of mobilizable plasmid transmission in the bacterial network.

References

Yin et al. Computers and Chemical Engineering 27 (2003) 235/249
Biemont et al. Genetics 169: 467–474 (January 2005)
Brickner et al. Infection and Immunity, Jan. 1983, p. 60-84
Verstraete et al. Journal of Bacteriology, Sept. 1992, p. 5953-5960
Taddei et al. Genetics 162: 1525–1532 (December 2002)
Charlesworth et al. Genetics 112: 359-383 February, 1986
Seidler et al. Applied and Environmental Microbiology, Feb. 1988, p. 343-347
Henson et al. Current Opinion in Biotechnology 2003, 14:460–467
Dubnau et al. Science. 310 (2 DECEMBER 2005) 1456
Berryman, A. Principles of population dynamics and their application. 1999 Stanley
Zwietering et al. Applied Environmental Microbiology. 60:1 (Jan. 1994) 204–213.
Molin et al. Microbiology 145 (1999) 2615–2622
Kierzek, A. Bioinformatics. 18:3 (2002) 470-481
Starlinger P, Saedler H. Current Topics in Microbio. & Immunology. 75 (1976) 111

Comments

I	Attachment	Action	Size	Date	Who	Comment
ppt	BIOEpresentation.ppt	manage	525.0 K	2005-12-12 - 09:25	TWikiGuest	PPT Presentation (Monday)
pdf	JustynJaworskiProjectReportBacterialNetworkModel.pdf	manage	278.2 K	2005-12-20 - 07:32	TWikiGuest	Paper (PDF)

Parents: StudentDesignedGraduateProjects