Results

wave

Dry-Lab

Model Metabolic


The work titled "METABOLIC FLUX ANALYSIS FOR OPTIMIZING THE SYNTHESIS OF COSMOMYCIN ANTIBIOTIC BY STREPTOMYCES OLINDENSIS ICB20" (LIMA LOBATO, A. K. C., 2010) demonstrated that the production of cosmomycin can be computationally estimated with good experimental reproducibility and highlights the correlation between the fatty acid synthesis pathway and the cosmomycin pathway. This supports the hypothesis that the incorporation of malonate as a precursor of Acetyl-CoA would increase the production of cosmomycin without compromising an essential pathway for the microorganism's survival.

Model Protein and validation

Using alphafold, we predicted five potential structures for our SARP protein. Then, we did a simple validation of the protein structure using SWISS-MODEL, that displays ligands, global and local quality, and target-template alignments. The third of the five structures we predicated was the one with the best outcomes.


Validation

Figure 1. Validation of modeled protein input number 3 using Ramachandran plot of SWISS-MODEL analysis


Input Mol Prob Clash Score Ramach Fav QMEANDisCo
1 2 3 4 5
2.16 2.16 1.99 2.19 2.17
17.28 17.05 11.82 18.64 17.51
93.53% 93.53% 93.88% 93,53% 93,53%
0.70 0.70 0.71 0.70 0.70

Table 1. Results of all generated structures. The inputs 1 through 5 correspond to the built structures in Alphafold. Additionally, we have scores produced by MolProbity, ClashScore, Ramachandran Favoured, and QMEANDisCo Global. The third structure was the one that gave the best results.


It is possible to assess the precision of the predicted protein structure using the Ramachandran plot. A positive indication that our protein's structure is close to the genuine conformation is that our Ramachandran plot values were above 90%. MolProbity creates a single number by combining all geometric scores (such as clashscore, bad bonds, bad angles) to indicate quality. Higher percentiles and lower values are preferable. Finally, we have the QMEANDisCo parameter, a single model approach that combines distance constraints (DisCo) score, statistical potentials, and agreement terms. DisCo assesses the consistency of pairwise CA-CA distances obtained from a model using homologous structure-extracted constraints.


QMEANDisCo score of protein number 3

Figure 2. QMEANDisCo score of protein number 3

The SARP family has two conserved domains: the DNA-binding transcriptional activator of the SARP family (DnrI) and the Bacterial Transcriptional Activation domain (BTAD). By aligning these domains with the sequence of the Cosmomicin SARP, we were able to identify the position where these domains are located in the predicted structure.


Structure of the SARP

Figure 3: Structure of the SARP rotated by 90 degrees. In a) The SARP, in b) we have the DnrI domain highlighted, in c) we have the BTAD domain highlighted, and in d) we have both the DnrI and BTAD domains marked.


As expected, the conserved domains that we know from the literature interact with DNA are located within the protein's cavity, indicating that it is indeed the interaction site with DNA.


Wet Lab

Transformation Challenges

The transformation process with Streptomyces proved to be exceptionally challenging. This was further compounded by the issues outlined throughout our research and in our project notebook. Despite these obstacles, our team persevered in our quest to optimize antibiotic production.


Pigmentation Clues

One of the intriguing observations during our experiments was the conspicuous red pigmentation exhibited by the Streptomyces strains we worked with. This distinctive coloration led us to hypothesize that these strains were potentially producing Cosmomycin D, an important secondary metabolite of interest. However, quantifying this production proved to be a significant challenge, leaving our understanding largely speculative.


Collaboration and Insights

To advance our understanding of Cosmomycin D production, we were fortunate to receive support from Professor Padilha. He provided us with mutant Streptomyces strains known to produce Cosmomycin D. This collaboration greatly enriched our research efforts.


Notably, our observations revealed an intriguing difference in the red pigmentation between the Cosmomycin D produced by the wild-type strains we initially worked with and that produced by the mutant strains provided by Professor Padilha's students. The mutant strains exhibited a richer red coloration. This unexpected discovery suggested that the mutants produced a higher quantity of Cosmomycin D.


Building on Insights

The parallel research conducted by Professor Padilha's students, focusing on secondary metabolite production, was fortunate for our project. They successfully cloned Cos I and Cos J, providing valuable insights into future experiments.


Our observations and data analysis indicated a notably higher production of Cosmomycin D in the mutants provided by Professor Padilha's students. The culture medium used by these mutants appeared significantly redder than that of our wild-type strains. This corroborative evidence reinforced our initial hypothesis.


Furthermore, we speculated that our construct had the potential to produce approximately twice as much Cosmomycin D compared to the wild-type strains. This hypothesis aligned with a mathematical model proposed by Professor Lobato, which indicated that our construct could expel Cosmomycin D from the inner cell region at twice the rate of the wild-type strains.


Future Directions

These findings and insights have set the stage for future experimentation. Our aim is to continue exploring and optimizing the production of Cosmomycin D in Streptomyces. Through further research, we are confident that we can achieve increased production of this important secondary metabolite.


Our collaborative efforts, innovative approach, and dedication to advancing antibiotic production in Streptomyces serve as a testament to our commitment to this vital field of research.


Reference

    Studer, G., Rempfer, C., Waterhouse, A.M., Gumienny, G., Haas, J., Schwede, T. QMEANDisCo - distance constraints applied on model quality estimation. Bioinformatics 36, 1765-1771 (2020).