Representative Experiments
We will use our first test protein, beta-lactamase, as an example to demonstrate the process of experimental design and execution.
Class B beta-lactamase which confers resistance to the beta-lactam antibiotics, including penicillins, cephalosporins and carbapenems. Acts via hydrolysis of the beta-lactam ring. Has penicillin-, cephalosporin- and carbapenem-hydrolyzing activities. (beta-lactamase - Pseudomonas aeruginosa | UniProtKB | UniProt)
Figure 1: 3D structure of natural beta-lactamase (queried from UniProtKB database)
Experimental Procedure
First, based on the results after running the model, we select the protein with the highest score among the proteins that the model considers successfully optimized and obtain its amino acid sequence (Figure 3). We used the neurosnap online platform for structural prediction to initially determine whether the protein maintained its original basic structure and function after model optimization. By comparing it with the natural structure, we can determine that the optimized beta-lactamase retains its basic structure and function (Figure 2).
Figure 3: Amino acid sequence of the optimized protein
Figure 2: Predicted protein structure of optimized beta-lactamase (neurosnap-Alphafold2 online tool)
We used the online platform NovoPro to perform reverse translation to obtain the first version of the DNA gene sequence (Figure 4), and then continued to use the NovoPro platform to perform codon optimization to obtain the final gene sequence for the experiment (Figure 5).
Figure 4: Initial version of the beta-lactamase gene sequence (model-optimized version)
Figure 5: Codon-optimized beta-lactamase gene sequence (model-optimized version)
We selected the plasmid vector pht01-ATAAAA, which is suitable for expressing the inserted gene. This plasmid vector has an ampicillin resistance gene (Figure 6).
Figure 6: Gene map of plasmid vector pht01-ATAAA (SnapGene software)
The presence of beta-lactamase brings ampicillin resistance to bacteria, and the optimization direction of our model is to enhance this function of beta-lactamase. Therefore, we replaced the original Amp resistance gene on the plasmid vector with our target gene sequence (knocking out the original resistance gene, creating a gap in the plasmid vector, and ligating the target beta-lactamase gene at the gap), so that the resistance to ampicillin-class antibiotics shown by the bacteria after introducing the plasmid vector comes entirely from beta-lactamase (Figure 7).
Figure 7: Plasmid map showing the replacement of the original Amp(E) gene with the beta-lactamase gene
In the SnapGene software, we designed the primers used to synthesize the gene, using forward and reverse primers for PCR to synthesize the target gene. We designed the primers required for ligation and used PCR to amplify the plasmid backbone from the plasmid vector (the fragment excluding the original Amp resistance gene) (Figures 8, 9).
Figure 8: Plasmid map showing the details of the primers for the target gene (V53G is the name for the model-optimized version, indicating the optimized sites and method, i.e., the beta-lactamase gene)
Figure 9: Primer distribution for the synthesis of the target beta-lactamase gene
Primer design is a key design step. For the natural beta-lactamase gene, we used the same process. After summarizing all the designed primers, we sent them to a biological company that has a long-term cooperative relationship with our laboratory for primer synthesis (Figure 10).
Figure 10: List of related primers (primer name + primer sequence (5'→3') + primer function)
We used the synthesized primers to perform PCR to synthesize the optimized beta-lactamase gene and the natural beta-lactamase gene, and to amplify the plasmid backbone fragment from the pht01-ATAAAA plasmid (length 7kb). We used agarose gel electrophoresis to verify the size of the PCR products and recovered the target gene fragments of the corresponding size. We used ligase to connect the plasmid backbone and the gene fragments to construct a new plasmid. We introduced the plasmid into competent E. coli cells and enriched them in a solid medium containing chloramphenicol (the plasmid vector contains a chloramphenicol resistance gene) to amplify the plasmid, and verified whether the plasmid ligation was successful (colony PCR) (Figures 12, 13).
Figure 12: Electrophoresis results (a, the upper part is the PCR synthesized gene, the upper left part is the optimized beta-lactamase gene, the upper right part is the natural beta-lactamase gene, the lower part of a is the electrophoresis for recovering the target gene fragment, the two on the left are the optimized beta-lactamase gene, the two on the right are the natural beta-lactamase gene; b, the electrophoresis result of the product after PCR amplification of the plasmid backbone, both are 7kb DNA fragments; c, the electrophoresis result of the product after colony PCR, a total of nine colonies were verified)
Figure 13: Chloramphenicol medium culture results (a, growth of E. coli transformed with the plasmid containing the optimized beta-lactamase gene; b, growth of E. coli transformed with the plasmid containing the natural beta-lactamase gene)
We extracted the plasmids from the successfully transformed E. coli and introduced them into new E. coli for protein expression. Based on the resistance to ampicillin antibiotics shown by the E. coli (growth on a medium containing ampicillin-class antibiotics), we determined whether the optimized protein would make the E. coli more resistant. We designed the same process for the natural beta-lactamase protein to form a single control condition with the optimized protein. We used wild-type E. coli without any relevant plasmid as a blank control.
In the bacterial culture verification stage of the experiment, we obtained three different types of E. coli (transformed with the plasmid containing the optimized beta-lactamase gene (Class A E. coli), transformed with the plasmid containing the natural beta-lactamase gene (Class B E. coli), and not transformed with any foreign plasmid (Class C E. coli)), forming three control groups. Each control group of E. coli was inoculated onto three culture plates, making three parallel controls. This resulted in a total of nine plates, which were cultured simultaneously under the same conditions to observe and compare the growth of E. coli (Figure 14).
Figure 14: Culture status on ampicillin-containing medium (a, growth of Class A E. coli; b, growth of Class B E. coli; c, growth of Class C E. coli)
Analysis and Discussion
According to the electrophoresis results, the gene fragment obtained by PCR is around 800bp (Figure 12a), and the obtained plasmid backbone fragment is around 7kb (Figure 12b), which are both the target lengths during design, so the PCR was successful. According to the results of the first round of culture (Figure 13), E. coli can grow on the medium containing chloramphenicol, which indicates that the plasmid was successfully transformed into E. coli and expressed the chloramphenicol resistance protein. According to the results of colony PCR (Figure 12c), the band size is consistent with the design length, which indicates that the target gene was successfully ligated with the plasmid backbone. In the second round of culture (Figure 14), with three groups of controls, the growth of Class A E. coli was the best, with 201, 252, and 246 single colonies on the three plates respectively; the growth of Class B E. coli was slightly worse than Class A, with 98, 147, and 105 single colonies on the three plates; the growth of Class C E. coli was the least, with 8, 19, and 3 single colonies on the three plates. This shows that the optimized beta-lactamase can indeed make E. coli show stronger resistance. From this point of view, the optimization of the PROTEUS model is successful.
References: (For reference only, not cited)
[1] Johnny Zakhour,L’Emir Wassim El Ayoubi,Souha S. KanjCA. Metallo-beta-lactamases: mechanisms, treatment challenges, and future prospects[J]. Expert Review of Anti-infective Therapy,2024,Vol.22(4): 189-201.
[2] John Z Chen, Douglas M Fowler, Nobuhiko Tokuriki. Comprehensive Exploration of the Translocation, Stability and Substrate Recognition Requirements in VIM-2 Lactamase[J]. eLife,2020,Vol.9: e56707.
[3] Ana Paula Barbosa do NascimentoCA,Fernando Medeiros Filho,Heidi Pauer,et al. Characterization of a SPM-1 metallo-beta-lactamase-producing Pseudomonas aeruginosa by comparative genomics and phenotypic analysis[J]. Scientific Reports,2020,Vol.10(1): 13192.
[4] Yen Hai Le,Hoa Thi Thanh Hoang,Diep Thi Khong,et al. Contamination of retail market meat with extended-spectrum beta-lactamase genes in Vietnam[J]. International Journal of Food Microbiology,2025,Vol.430: 111061.
[5] Jacqueline Findlay, Otavio Hallal Ferreira Raro, Laurent Poirel,et al. Molecular analysis of metallo-beta-lactamase-producing Pseudomonas aeruginosa in Switzerland 2022-2023[J]. European journal of clinical microbiology & infectious diseases : official publication of the European Society of Clinical Microbiology,2024,Vol.43(3): 551-557.