Modeling

Bacteria such as E. coli are very well characterized, and we know the features and the typical nucleotide patterns that has to be modified in order to tune gene expression. However, L. jensenii is a little-known bacterium and we only recently started to engineer its genome for biomedical applications [1]. With this study we propose a general pipeline to identify promoters found in the genome of L. jensenii, a first step towards the characterisation of this organism.

The aim of our modeling was to predict which natural pre-gene sequences from L. jensenii were most likely to be to be strong promoters. For this we used bioinformatic tools to identify recognizable patterns in those pre-gene sequences such as the Shine Dalgarno sequence or minus 10 and minus 35 sequences from promoters typical of this bacteria. Then we selected the pre-gene sequences having those patterns to test their promoter force with a signal of RFP. We wanted to find natural sequences with different promoter forces, in the goal to build a toolbox of promoters sequences for L. jensenii.

First, we needed to extract those pre-gene sequences from the full genome. We downloaded it from NCBI (L. jensenii JV-V16). For this we followed the protocol of the following article that had the goal to identify promoter sequences on L. plantarum : "Genome-wide prediction and validation of sigma70 promoters in Lactobacillus plantarum WCFS1”. We developed a python script to do it. Selecting a maximum length of 100bp, a minimum of 25bp and no overlap, upstream each gene sequence.

Then we used the “MEME SUITE: tools for motif discovery and searching” to identify relevant patterns on those sequences. With the results we were able to identify different known patterns present on multiples pre-gene sequences (number of hits):

The RBS pattern of the bacteria, a Shine Dalgarno sequence specific of L. jensenii:

The pattern of L. jensenii promoters:

We can clearly see the -10 pattern: TATAAT which resembles to some of other Lactobacilli patterns.

Furthermore we identified the -35 pattern by selecting the sequences having a promoter pattern for an other MEME run:

We also found pattern looking like terminators tails:

With those informations on what RBS, promoters and terminator tails looked like on L. jensenii we then ran a MAST (MEME SUITE) to found those patterns on our pre-genes sequences. We then selected some to test their promoters forces with a RFP signal, we took 4 with RBS and promoter patterns, 3 with only promoter patterns, 2 with only RBS pattern, and finally one without any. Plus, we created an artificial sequence with promoter and RBS with the most common letter from the power weight matrice of each patterns. We also took a characterized promoter from jensenii as positive control, and a sequence with only RFP as negative control.

We then cloned those sequences on the plasmid for L. jensenii (pLEM415) and prepared them to be transformed on L.jensenii. Sadly their were some issue with the transformation with L. jensenii and we were not able to have results on this part of the project.