When initially designing the yeast and lactic acid bacteria (LAB) implementations the creation of a symbiotic relationship between the two organisms was explored, with the goal of preventing direct competition between the two and maintaining stable populations of both production strains. The design that was eventually decided on used a modified yeast using the B12 dependent methionine synthase MetH instead of the more commonly found Met6. This made growth in media where B12 was not present unfeasible for the yeast and thus made the already negligible risk of the yeast outcompeting the LAB even less of an issue. Engineering the LAB to be dependent on the yeast was early found to be outside of the scope of what was deemed reasonable, but the naturally strong color of vitamin A would at least provide a way of confirming that both organisms were alive. Vitamin A production would therefore act as an indicator for if either the yeast died, the LAB died, B12 production stopped or yeast stopped producing vitamin A. As a proof of concept a genome scale model of vitamin A production using the Yeast8 model [1] was also established at this point where the model indicated that β-carotene production was possible without sacrificing too much in the way of cell growth.
The assembly process for the β-carotene construct started off well with the level 0 plasmids used for the MoClo pYTK [2] Golden Gate assembly being done relatively quickly. The level 1 plasmids on the other hand kept being either assembled incorrectly or weren’t being amplified sufficiently once transformed into E.coli. Several different changes were made to the methodology to try and find some success such as using higher quality enzymes, doing the procedures in triplets or ordering new reagents in case they were contaminated. Sadly none of this had the desired result and the finished vitamin A constructs were never produced.
When the process of combining and amplifying the parts required for vitamin A production via Golden Gate Assembly started to stall, the decision was made to transition the vitamin B12 project over to using Gibson assembly of larger, ordered fragments. The rationale here was that either the methodologies or reagents used for Golden Gate assembly may be one of the causes of the repeated failures, and a transition away from type IIs enzyme assembly may be advantageous.. The full B12 construct was designed to be ordered as three separate fragments which could then be combined to form the full construct. Around the same time it was also decided that since time was starting to be an issue, the B12 construct would not be stably integrated into the LAB genome but kept as transient plasmids and work as a proof of concept instead. At this point time was running out and there was only enough time for one attempt at assembling the construct, which then proceeded to yield a non-viable final product.
So while the results were a bit disappointing we did at the very least follow the standard “Design, Build, Test, Learn”-cycle, even though most of the learning boiled down to concluding that something does not work and never finding the true source of the error.
References
References
- [1]: Lu, H., et al. "A consensus S. cerevisiae metabolic model Yeast8 and its ecosystem for comprehensively probing cellular metabolism. Nat Commun10: 3586." (2019).
- [2]: Michael E. Lee, William C. DeLoache, Bernardo Cervantes, and John E. Dueber. “A Highly Characterized Yeast Toolkit for Modular, Multipart Assembly” ACS Synthetic Biology” (2015) DOI: 10.1021/sb500366v
