Team:TAU Israel/Experiments
Experiments
Evolution experiment – what does it mean?
In an evolution experiment, a particular genetically modified strain (or several strains in parallel) that expresses an inserted construct is grown over time. An expression measure is defined beforehand, for example, a fluorescent reporter gene in the construct. The evolutionary stability of that construct is measured at different time points, as individuals/populations adapt to new environmental conditions by natural selection. If the pre-defined measure indicates a significant decrease or a total disappearance of the construct from the population (depending on its location, viability, and other features), the experiment is aborted.
Based on the defined parameter, one can conclude the evolutionary half-life: the amount of time that the construct's frequency in the population is reduced by half.
We have conducted two evolution experiments to prove our hypothesis and gather mutational data.
Our experiments included measuring genetic stability of constructs composed of an exogenous gene (that adds no value to the fitness) attached to the "N" terminus of an endogenous open reading frame (ORF) (figure 1). We conducted two separate experiments: The Gene-SEQ experiment and the POC experiment (10 gene experiment).
Libraries
For each experiment we used two unique Saccharomyces cerevisiae libraries, a GFP- and an RFP-based, provided by Prof. Maya Schuldiner from the Weizmann Institute. These libraries contain about 6000 yeast variants, one for each ORF in the yeast genome, with a distinct promotor per library. Each variant has a fluorescent gene (either GFP or RFP) fused to the N-terminus of a different gene[1]. 
This means the library contains a variant with a fluorescent-tagged gene for every known yeast gene. The fluorescent protein-ORF conjugate is translated together, resulting a fused protein.
We used these libraries as a basis for both our experiments since they allowed us to conveniently test our hypothesis and produce our Gene-SEQ measurement protocol. This was possible due to the unique design of the library constructs - As you can see in figure 1, each variant contains a construct composed of GFP/RFP, a linker, and a specific ORF. This design fits our proposed idea - to interlock a target gene (in this case, GFP or RFP), to an essential gene of the host organism. Thus, we can use the libraries to collect empirical data for our models and understand the behavior of such constructs in evolution experiments.
Proof of Concept (POC) Experiment
This experiment also referred to as the "10-gene experiment", was conducted as a preliminary proof of concept.
Our hypothesis hypothesis is that by interlocking a target gene upstream to an essential gene will prolong the evolutionary half-life of the target gene. However, each target gene can be linked with a limited number of essential genes, regarding the conjugate’s folding and translation. Those protein conjugates need to be as stable as possible and this could be achieved through tailoring target and essential linkage.
This experiment had two main goals:
- Validate that fusion to an essential gene likely prolongs a target gene’s evolutionary half-life, in comparison to the target gene alone (as the negative control for the stability).
- Demonstrate that different essential genes provide varying stability levels.
For this purpose, we compared the evolutionary half-life of ten diverse constructs against the half-life of the unattached target gene, by measuring their relative preservation of fluorescence, which is used as a measure of stability.. Each construct from the ten constructs that were tested is derived from the GFP library described above, thus it is composed of GFP, a fixed linker, and an essential gene. The GFP serves as the target gene for this experiment.
We chose the 10 essential genes based on a hierarchical clustering algorithm and many descriptive features we created to characterize the genes and define their "diversity". The selection procedure is described in the Engineering Success Page.
Measurement Considerations
We grew triplicates of each construct separately, where the unattached GFP was the negative control for stability. Every day, we repeatedly transferred them to fresh media for overnight growth, measuring fluorescence and optical density (OD) for 180 generations.
In every ongoing evolution experiment, which spans over days, weeks and even months, the problem of sterility arises. Therefore, in the 10-gene experiment that lasted for months, we had to work in a very rigorous and sterile method in a biological hood at all times.
Handling contaminations: To make sure that our biological repeats did not become contaminated, we transferred samples from the previous day to another plate of selective media (lacking amino acids). This media would not allow the library’s yeast growth. If high turbidity was observed the next day, it was indication that the previous day's experiment went wrong, and we would have to abort the experiment, to avoid dragging the mistake into the next day. To prevent starting the experiment all over again in case of contamination, we froze samples from each day that would serve as a backup when necessary.
Throughout the experiment, we measured fluorescence values and turbidity accurately and uniformly. This was achieved by eliminating possible variables that could disrupt the results, such as:
- Calibration of the plate reader.
- Sterile and uniform work by one staff member throughout the whole experiment.
- Precipitate the yeasts and replacing the old substrate with transparent liquid (PBS) before measuring the fluorescence values, to prevent GFP residues and other proteins and waste in the substrate from interfering with the reading.
- Perform the experiment at a constant time each day.
- Use of two controls for the experiment: wild-type (WT) control of the library's background strain (i.e. yeast) to serve as a negative review for the fluorescence test. On the other hand, we have transformed the same species with a GFP that is not fused with any gene in order to serve as a positive control for the fluorescence test.
The protocols and results can be seen in their respective pages.
Protocols Results
Raw Data Analysis
The triplicates of each of the ten variants were grown in 96-well plates, with three positive controls for fluorescence (GFP alone), and two negative controls for fluorescence (WT). As mentioned, we measured fluorescence and OD for 180 generations. At the same plates, there were also 2 wells with media only, as described in the protocol.
The analyzed raw data consisted of fluorescence and OD measurements for the start (day one) and the end of the experiment, after 180 generations.
The analysis included the following steps:
- 1. Averaging the fluorescence values of the two media-only wells and reducing the average value from the fluorescence levels of the other wells. This way we take into consideration differences between media in the triplicates since sometimes the liquid itself appears to be turbid regardless of growth or contaminations.
- Normalization of the reduced values with the OD values of the same strains. The OD reflects the amount of yeast in the media. By normalizing it, we consider the possibility that some strains are reproducing faster than others, thus they will show higher fluorescence without it indicating actual stability.
- Calibrating the normalized values according to the plate-reader calibration protocol.
- Normalization of the fluorescence values of the fluorescent variants (10 construct's triplicates and the negative control) with the fluorescence value of the WT. The reason behind this normalization is to take into account inaccuracies of the plate reader. The difference between the WT and the fluorescent variants should remain the same regardless of variations resulting from the plate reader errors.
- Averaging the reads from the triplicates of each strain, resulting in average fluorescence value for the 10 constructs and the negative control for the first and last day of the experiment.
- 6. Normalization of the average fluorescence value from the last day with the values from day one. Thus, we receive relative values that indicate the preservation of fluorescence – if this value is high, it means that this variant kept expressing the fluorescent protein.
You can see that the results matched our hypothesis in the Proof of Concept Page!
Proof of Concept
Gene-SEQ experiment
This experiment is the first test of our large-scale stability measurement technique, called Gene-SEQ (Stability Enhancing Quantifier).
It has two main goals:
- Large-scale testing of our hypothesis, that the target-essential gene linkage prolongs the half-life of the target gene.
- Analyzing the mutational footprints of each construct, in order to eventually predict the preservation of the target gene.
As an evolution experiment that is meant to test the genomic stability of any gene coupling over time, we had to grow the yeast variants for many generations. We could not grow thousands of different variants separately because of time and logistics limitations – this kind of procedure would have required the use of thousands of test tubes, plates, tens of thousands of tips, tens of liters of media, and perishable equipment every day.
Therefore, we decided to grow all the strains together as a "co-culture" – a mix of all strains – in a reactor/chemostat. But the existing devices on the market are expensive, large, and cumbersome. Fortunately, Chi.Bio was recently developed for exactly this purpose. It is an autonomous, compact, simple-to-use, and even programmable and upgradeable turbidostat (robotic growth platform). This machine monitors OD and Fluorescence (if needed) and transmits the data directly to the computer. It constantly dilutes the culture with fresh media, to always maintain an OD of choice.
The device is made of three units:
- A reactor into which a glass test tube is inserted. That is capable of mixing and heating the solution at a fixed rate and a constant temperature. In addition, the reactor contains sensors for measuring turbidity, fluorescence, temperature and more.
- A set of peristaltic pumps that attach to the reactor and pump a new substrate into it and an old substrate out of it.
- A control panel that is connected to the pumps and the reactor and gives commands through a computer.
You can read more about chi-bio in the following link.
Since it is a DIY device, it had to be adapted to our experiment. To avoid problems of sterility and leaks, we designed caps with nozzles for both test-tubes and media bottles inspired by the optional cap offered on the chi-bio website. Regarding this design, we would like to thank Ben Oz who designed the caps fully voluntarily according to our needs. In fact, the new caps allowed us total autonomy without having to regularly test the experiment except to change the media every few days.
The yeast cultures were monitored continuously and were grown in constant turbidity while monitoring fluorescence levels for about 300 generations. Then, we extracted and prepared sequencing samples with our novel Gene-SEQ protocol that is aimed to measure stability in a larger-scale. The protocol and a detailed explanation of our planned data-analysis can be found in our Measurement Page.
Measurement
More protocols and the results of the Gene-SEQ preliminary experiment can be observed in the respective pages.
Protocols Results
references
[1] Yofe I, Weill U, Meurer M, Chuartzman S, Zalckvar E, Goldman O, Ben-Dor S, Schütze C, Wiedemann N, Knop M, Khmelinskii A, Schuldiner M. (2016) One library to make them all: streamlining the creation of yeast libraries via a SWAp-Tag strategy. Nature methods, DOI: doi.org/10.1038/nmeth.3795. Pdf.
