Difference between revisions of "Team:Amsterdam/Project/emulsions"

Stable Romances tend to evolve when there is a perfect understanding between the individuals. Whether a match is a perfect fit (or not!) can be tested when partners are forced to interact for instance, when people move in together. So imagine applying the same principle to consortia were we deduced that an underlying property of all consortia is that ultimately they will display a so-called “high-yield strategy”. By High-yield strategy, it is meant that a more efficient use of natural resources leading to a greater number of individuals will be observed. This is even more so the case when instability is a serious threat. ,

What if there was a way to select these high-yield strategy consortia and actually make them interact better? How would one do this? Well… just like with people, this is possible to test by making microbes move in together.

We decided to tackle this problem by looking at studies done using emulsion based techniques selecting for high-yield strategies (Bachmann et al.). In this paper, the authors discuss the tradeoff that occurs between growth rate and growth yield in microbes. They demonstrate that compartmentalizing the availability of public goods while serial propagating leads to the selection of organisms with increased yield strategy, i.e. able to generate ever higher cell numbers (biomass) from the same portion of substrate. We apply the same logic to consortia by developing a dedicated emulsion-based protocol. The basic principle is that by randomly combining different individuals of each side of the partnership in isolated compartments (micro-droplets), perfect matches will lead to increased numbers of offspring of both photo- and chemotroph. If propagated in time, ultimately this will enrich the mixture in both photo- and chemotrophs that do well when placed in a consortium while purging the population of the ones that do not.

This system allow us to reduce the time of observing a mutation but now it comes the second problem. How do we know if a mutation leading to a change in the parameters has happened? The first option that comes to our synthetic biologist mind is by sequencing periodically the genome of the population and screen it for mutations. The major drawback of sequencing is that it does not inform whether the mutation affects the parameters or not. The remaining solution is then to estimate the physiological parameters with high time definition to spot when the mutation has occurred. Using the turbidostat it would be possible to store all the values recorded by the spectrophotometer and calculate the growth rate from these data. This is done by fitting a linear model to the logarithm of the OD over the time. The production rate can obtained by periodically measuring the amount of product in the medium (link to protocols for enzymatic assays and HPLC).

Luckily for us when we came to the lab there was a master student, Joeri Jongbloets{link}, who had transformed a multicultivator (MC) like this{link} in a 8 channel turbidostat. For his internship he wrote the program that connect with the MC hardware and control the pumps to dilute the culture when necessary. In addition the software stores the measurements from the MC spectrophotometers in a database and analyzes it to obtain the growth rate. Thanks to this impressive piece of software engineering we were able to run long term experiments and observe evolution.

The strains producing glycerol, lactate and acetate (Δacs) were grown in turbidostats for 900 hours under constant LED lights providing 20 μE/ m²* s * OD.

The Pta1, Pta2, Δacs and Ack strains were cultured on photostats. This systems, also implemented by Joeri Jongbloets in the MC, is basically a batch culture where the light intensity per OD provided to the cells is maintained constant. In this case the light intensity per OD was set to 20 μE/ m²* s.

In both cases the culture medium was BG-11 supplemented with NO_3- and TES buffer {link to recipe}.

Samples were taken from the cultures and centrifuged (5 min, 14.500 rpm) to obtain the supernatant. Product concentration was then estimated by enzymatic assays {links to protocols} in the case of lactate and acetate and by HPLC for glycerol {links to protocols}.

In Our turbidostat experiments we observed the results we expected. In the lactate strain (figure 1), a non growth coupled carbon producer, we started to observe an increase in growth rate about 300 hours after the beginning of the experiment. At the same time point we recorded a drop in the production of lactate. This result confirms our hypothesis about how mutations leading to the loss of production can be quickly fixated inasmuch as it releases the burden imposed by the carbon production increasing the fitness of the mutated cells.

On the other hand the growth coupled acetate producer, Δacs, shown a constant production and growth rate over the length of the experiment (figure 2) demonstrating that this strategy is more stable than the classical engineering approach. The Q_p estimated around 400 hours shows a decrease in the production but it is probably due to experimental failure. In the turbidostat cells were grown at very low OD (threshold was set to 0.35) therefore the concentration of acetate was below the detection limit of the enzymatic assay method. To overcome this problem we dehydrate the samples by lyophilization, also called freeze-drying, then dilute them in an smaller volume therefore increasing the concentration above the detection limit. Although this method worked (we obtained similar Q_p values using different culture systems), the sample processing could have introduce noise in this measurement.

In table 2 the physiological parameters obtained for six strains producing glycerol, lactate and acetate are presented. The lactate strain shown the highest Q_p followed by Δacs and Syn1413. An intriguing results is the differences in Q_p and growth rate of Δacs between cultivation methods. In the turbidostat the length of the exponential phase from where the growth rate is calculated is smaller than in photostat which could be influencing the estimation. For the Q_p the previously discussed argument could also be an explanation for this difference. In addition it is possible that the photostat conditions are more favourable for the growth of Synechocystis.