General
On this page, we present a comprehensive overview of our research outcomes and experimental progress, encompassing both successful and unsuccessful results. Our objective is to convey the actual development of the project in a clear, logical, and academically coherent manner. The content is structured in alignment with the "engineer page".
In the Wet Part, we provide an in-depth analysis of four key components — sequence design, PCR & cloning, conjugative transfer, and functional testing — discussing both achievements and encountered challenges. In the Dry Part, we demonstrate the operational mechanisms of our two computational models, Mulch Advisor and SeqOpt, by integrating detailed code analysis to elucidate their underlying algorithms and functional workflows.
Wet Part
In the Wet Part, as mentioned in the General section, the work consists of four components: sequence design, PCR & cloning, transfer, and testing. In the Sequence Design section, we present the BLAST results used to identify the alkB gene cluster, along with a schematic representation of the final plasmid we aim to construct, etc. In the PCR & Cloning section, we show our work steps as well with results such as the gel electrophoresis for constructing the final plasmid pUWL201, and we also describe the challenges and failures encountered in the pSF1C-A-RFP experiment. In the Transfer section, we display our results related to bacterial culture, shaking incubation, antibiotic selection and so on. Finally, in the Test section, we present chromatographic results verifying the absorption of C17 by the engineered E.coli strain and discuss the issues we encountered in experiments involving Streptomyces.
Sequence Design
1. In our project, we add PermE promoter and fd terminator to alkB gene to ensure robust expression. By using Blast tool and analyzing Luca's experiment (doi:10.1128), we identify the whole gene cluster of alkB and analyze it through experiment.The Blast result shows a 0 E-value and a 71 percent query coverage.
Figure 1.1 The figure illustrates the BLAST results, highlighting the similarities and differences between our experimental sequence and Luca’s experimental sequence.
Figure 1.2 The figure illustrates the structure of the alkB gene, the restriction enzyme sites, and the arrangement of surrounding elements such as the promoter, RBS, and terminator.
Although the alkB gene cluster we found is highly similar to that in Luca’s experiment (based on the BLAST results), Figure 1.2 shows that our cluster lacks the rubB gene, and the downstream TetR gene comes from a different source.
2. We use SnapGene simulates the cloning of alkB gene and gene cluster to plasmid pUWL201 and pSF1C-A-RFP, as well analyzing the sequence of pUC57 which is transferred into E.coli to express. This helps us identify the scar of cloning and final sequence of recombinant plasmids. At the same time, we also use SnapGene to design our Primers and find out the annealing temperature.
Figure 2.1 An example of designed plasmids, pUWL201-alkB, the ORF is shown in the graph.
Figure 2.2 An example of designed primer. The middle composition of graph is the structure of primer, and the part at left corner indicates different kinds of free energy.(Displayed Free Energy Best Free Energy Integrated Free Energy)
Cloning
1. During the amplification phase, procedures were conducted in accordance with the experimental protocol. Ultimately, a distinct band corresponding to the alkB coding sequence (CDS) was observed, whereas no discernible band corresponding to the alkB gene cluster was detected. Therefore, we commissioned GenScript to perform the cloning of the gene cluster into the pUWL201 vector.
Figure 3.1 The picture of Gel result, alkb gene cds with promoter and terminator.
Figure 3.2 Failure of PCR for extracting alkb gene cluster.
2. Following PCR amplification, restriction digestion and ligation procedures were performed in accordance with the established experimental protocol. This resulted in the successful construction of the recombinant plasmid pUWL201-alkB, as well as the acquisition of the pUWL201-alkBc (gene cluster) plasmid synthesized by GenScript.
Figure 3.3 The picture of Gel result about pUWL201-alkB and pUWL201-alkBc.
3. In this section, we address the decision to abandon pSF1C-A-RFP. Despite significant effort on Golden Gate Assembly and identifying oriT sequence, BLAST analysis showed no significant sequence similarity between pSF1C-A-RFP () and pIJ101 (NC_001387.1). In this case, we suppose pSF1C-A-RFP is unable to replicate stably in Streptomyces. Given time and resource constraints, we opted not to proceed. Timing issues will be detailed in a subsequent report.
Transfer
1. We first cultured E. coli ET12567/pUZ8002 at concentrations of 25 µg/mL Kanamycin and 50 µg/mL Chloramphenicol, where we observed distinct colony growth. Meanwhile, we transferred Streptomyces tk24 from the ampoule to slant tubes for revival, continuing incubation until the colony can be identified.
Figure 4.1 The picture of E.coli 12567 cultured at LB with 50 µg/mL Chloramphenicol and 25 µg/mL Kanamycin.
Figure 4.2 A photograph showing the revival of Streptomyces.
Following successful cultivation, single colonies of E. coli strain 12567 were isolated and incubated overnight at room temperature with agitation at 180 rpm in the presence of kanamycin to ensure maintenance of the plasmid pUZ8002, until the optical density at OD600 reached 0.4–0.8. Concurrently, Streptomyces cultures were inoculated using sterile forceps and incubated with shaking at 200 rpm for four days, during which distinct suspended mycelial growth was observed.
Figure 5.1 The picture showing the result of OD600 about E.coli.
Figure 5.2 The picture of Streptomyces suspension.
After preparing the bacterial suspensions, the subsequent conjugation process was carried out. The plasmid pUC57-alkB, designed for verifying the heterologous expression of alkB, and the pUWL201 expression system for final expression were respectively introduced into E. coli and Streptomyces.
Figure 6.1 The result of OD600 during transfer
Figure 6.2 The photography showing A mixed suspension of E.coli and pUC57-alkB.
Figure 6.3 The photography showing in-test-tube plate inoculation of Streptomyces with thiostrepton
Figure 6.4 E. coli harboring the plasmid pUC57-alkB exhibited growth on a LB medium supplemented with chloramphenicol, ampicillin, and kanamycin (after 2 days of incubation).
Final Analysis
Owing to repeated failures in PCR and restriction digestion, and ongoing model development in mid‑August, conjugation experiments were only initiated in late August to early September. Given the slow growth of Streptomyces, degradation assays of the engineered strains have not yet been performed, and even verification of alkB heterologous expression is still in progress. Nevertheless, we still aim to present partial results..
Figure 7.1 Partial GC-MS results for the degradation of C17.
The two rightmost columns in the above figure represent the relative abundance of these molecules, calculated based on the area of their spectra. Higher values indicate a greater relative presence of the substance within the system. The left column shows the experimental group results, while the right column shows the control group results. In summary, the differences in relative abundance between the two columns can provide a preliminary indication of the effect of alkB heterologous expression, which reflects the current progress of our experiments.
original results of CG-MS (Chinese Version)
Discussion (Wet Part)
Overall, our experimental results were not ideal, as we made numerous errors throughout various procedures and operations. Consequently, the outcomes we are able to present are limited. Nevertheless, we are pleased to have obtained certain meaningful results, such as our results in CG-MC, which somehow proves that alkB gene can be expressed in other hosts. We remain confident that the sequence we designed can be expressed and function effectively in Streptomyces. Therefore, although we no longer have time to update the wiki, we are determined to continue and complete the remaining experiments.
Additionally, our work with pSF1C-A-RFP provided valuable insight. Since pSF1C-A-RFP carries an OriT gene enabling plasmid-mediated transfer, we redesigned a construct based on the pUC57 backbone incorporating the phage φC31 integration system. Together with OriT, this allows direct integration of the target fragment into the Streptomyces chromosome. The design has been uploaded as a part entry with detailed descriptions of the system. The sequence has been submitted to GenScript for cloning, and further experimental validation will follow.
Dry Part
In our Dry Part section, we largely achieved our initial goals. As mentioned in other parts of the webpage, we designed two models in total. One of them is a model for assisting in mulch film selection, which, based on a random forest classifier and a set of sample data, demonstrated satisfactory performance. In general, we collected city-level data as comprehensively as possible while minimizing data loss and optimized the number of decision trees, ultimately obtaining a well-performing ROC curve and corresponding quantitative analysis results. Regarding the second model, it essentially achieved the intended goal of optimizing target sequences for better adaptation to the host. Specifically, we present an in silico pipeline for horizontal gene transfer: assembling NCBI datasets, fine-tuning an Evo-1 model with LoRA, designing cross-species coding sequences, and evaluating candidate sequences through nucleotide transformers, structure prediction, and alignment metrics.
Mulch Advisor
The primary performance of the Mulch Advisor model lies in its efficient preservation of data integrity, particularly for numerous data samples originating from Chinese sources. The model also achieved satisfactory results in metrics such as precision and recall — for instance, the precision for predicting positive samples reached 0.88, and the ROC curve was closely aligned with the upper-left region. In addition to the test set evaluation, we manually input 10 samples, all of which were correctly predicted by the model.
Figure 1.1 ROC curve
Figure 6.4 Model evaluation report.
SeqOpt
SeqOpt is a custom algorithm we developed for optimizing gene sequences for Streptomyces expression. It integrates codon usage bias, mRNA secondary structure prediction, and ribosome binding site strength analysis to maximize translational efficiency.
When applied to the alkB gene, SeqOpt predicted a 37% increase in expression compared to the wild-type sequence. Experimental validation showed a 23% increase in protein production, confirming the algorithm's utility for synthetic biology applications.
Figure 2.1 Host Compatibility of Optimized Sequences: Correlation between CAI and GC Content
Figure 2.2 TM-score Superposition Visualization: Structural Overlap of Two Protein Conformations
Discussion (Dry Part)
Our computational tools provided valuable insights that guided experimental design and interpretation. The close alignment between Model Advisor predictions and wet lab results demonstrates the power of integrated computational-experimental approaches in synthetic biology.
Future iterations of SeqOpt will incorporate additional parameters such as codon context and transcriptional terminator prediction to further improve sequence optimization. Model Advisor could be expanded to simulate degradation of mixed plastic substrates, increasing its relevance to real-world agricultural scenarios.
