We employed the Engineering Design Cycle as a systematic method to enhance our CADABRA enzyme cocktail's efficacy in eliminating antibiotics from wastewater during our iGEM odyssey in 2023. This approach encompassed a sequence of stages that provided us with valuable guidance throughout our project.
1) Identify the Problem:
Analyzing the results from last year’s project, we realized that the VIM-2 metallo-β-lactamase (MBL)-based version of the CADABRA cocktail suffers from some limitations in its efficiency. As visualized in our urine decontamination test (https://2022.igem.wiki/bulgaria/proof-of-concept), the meropenem levels in this experiment were reduced from 750 ug/ml to approximately 0.4 ug/ml in 12 hours. Our main goal for this year was to accomplish complete removal of the antibiotic.
Furthermore, while investigating the problem of antibiotic pollution in Bulgaria, we came across data indicating a concerning prevalence of antibiotic-resistant bacteria in some significant rivers within our country. Notably, the most pronounced levels of resistance were observed concerning the β-lactam antibiotic ampicillin, which is commonly employed in animal farming (Tsvetanova, Z.; Najdenski, H. Antimicrobial Resistance of Heterotrophic Bacteria and Enterobacteriaceae Inhabiting an Anthropogenic-Affected River Stretch in Bulgaria. Processes 2023, 11, 2792. https://doi.org/10.3390/pr11092792)
Ampicillin
2) Research and Gather Information
In our quest for the requisite information, we conducted extensive literature research and arranged several meetings with medical professionals, including specialists in medical microbiology, hematology, and clinical pharmacology. To gain deeper insights into the issue of antibiotic contamination in clinical settings, we also engaged in discussions with the head of the government agency responsible for hazardous hospital waste management.
3) Brainstorm, Generate Ideas and Select the Best Solution
As we contemplated methods to enhance the efficiency of carbapenem degradation, two potential opportunities emerged: the elevation of MBL enzyme production to yield a more concentrated cocktail or the augmentation of MBL enzymatic activity. Our choice was to concentrate our efforts on the latter option. In pursuit of improving MBL activity, our initial step was to assess whether any naturally occurring MBL enzyme displayed greater activity than VIM-2. Subsequently, our plan involved implementing directed evolution techniques to further boost its enzymatic efficiency.
In our endeavor to combat ampicillin contamination in the rivers, our initial strategy was to determine the prevalent form of this antibiotic's usage in animal farming. Subsequently, our plan assessing both the original CADABRA cocktail and the enhanced version to determine their efficacy in degrading this β-lactam antibiotic in aquatic environments.
4) Design and Prototype
We designed two novel MBL coding sequences, one based on IMP-1 (Q79MP6), and the other derived from a modified version of NDM, incorporating the mutations found in variants 2 and 5. These sequences were synthesized as codon-optimized gBlock fragments by IDT. To simplify the cloning process, we included VF2 and VR primer annealing sites at both ends of the fragment, along with the BioBrick Prefix and Suffix sequences. Subsequently, we introduced these constructs into a vector featuring a strong promotor and medium RBS (BBa_K608003) following the rules of the BioBrick RFC 10 assembly protocol. Positive transformants were selected on growth media with chloramphenicol and ampicillin and further verified by colony PCR with primers VF2 and VR.
NDM MBL (doi.org/10.3390/ijms232416083)
5) Test and Evaluate
To test the newly generated constructs, we applied the following methods
• Antibiograms – antibiograms were prepared following the EUCAST criteria, utilizing discs loaded with imipenem, meropenem, and imipenem + EDTA. The distinction in zone diameters between imipenem and imipenem + EDTA indicated a MBL production in our transformants.
• Meropenem degradation by cell free extract – we prepared cell free extracts from IMP and NDM producing strains using a combination of centrifugation (2700 g for 10 min) and filtration (0.22 um filter). The complete elimination of cells was confirmed by inoculating portions of the cell-free extract onto agar plates with a suitable growth medium. Subsequently, we examined the extract's ability to degrade carbapenems by treating MH medium with 50 µg/ml of meropenem. The residual antibiotic levels were determined using the agar well dilution method, and notably, complete meropenem degradation was observed.
Agar well dilution method
• Carbapenem hydrolysis monitoring in real time – we applied ultraviolet-visible (UV-Vis) spectroscopy to monitor the degradation of meropenem. A sample containing the cell-free extract and meropenem was prepared in a quartz cuvette, and we measured the absorbance at 298 nm for 10 minutes using a JENWAY 6715 UV-Vis spectrophotometer. The slope of the resulting line was used as an indicator for the enzymatic activity of our MBL extract.
6) Iterate, Refine and Improve
Based on the findings from our initial tests, we concluded that both of the newly constructed variants resulted in the production of functional MBL enzymes. Notably, the hybrid NDM-2&5 variant exhibited the most significant MBL activity, as indicated by the antibiograms and kinetic curves. To assess its performance relative to the original VIM-2 utilized in 2022, we conducted a broth dilution assay to determine the effective MIC (Minimum Inhibitory Concentration) of meropenem for strains producing VIM-2 and NDM-2&5. The results revealed that E. coli DH10B blaVIM-2 exhibited no growth inhibition at 25 µg/ml of meropenem, while the E. coli DH10B blaNDM-2&5 clone displayed inhibition at concentrations exceeding 200 µg/ml, representing a substantial improvement of more than eightfold.
Broth dilution assay to evaluate the meropenem’s MIC of our MBL producers
However, we believed that we had not fully harnessed the potential of our chosen MBLs. To maximize their capabilities, our first step was to employ a directed evolution strategy to enhance our NDM-2&5 enzyme. We opted for a dual gene-specific mutator system capable of introducing transition mutations at consistent frequencies within a target sequence in vivo. (Seo D, Koh B, Eom GE, Kim HW, Kim S. A dual gene-specific mutator system installs all transition mutations at similar frequencies in vivo. Nucleic Acids Res. 2023 Jun 9;51(10):e59. doi: 10.1093/nar/gkad266. PMID: 37070179; PMCID: PMC10250238.). To utilize its mutagenic potential, first we cloned our blaNDM-2&5 construct in a plasmid with T7 promotor and RBS (BBa_K525998). While this approach would have effectively introduced random mutations into our sequence, we also needed a straightforward method for screening our clones to identify MBL production. Our attempts to induce expression using the T7 promoter in our construct within E. coli KRX cells did not yield MBL production after L-rhamnose induction, as indicated by test antibiograms. To address this challenge, we opted to subclone our fragment into a pSB1A2 vector (BBa_J23101) that includes the J23101 Anderson constitutive promoter. We anticipated that this promoter would facilitate protein expression, while the T7 promoter could be reserved for mutagenesis purposes. To test this hypothesis, we created a construct comprising J23101-T7-RBS followed by the coding sequence of amilGFP (BBa_K592010). Upon observation of the transformants under blue light, clear signs of amilGFP expression became evident.
Construct J23101-T7-RBS-amilGFP CDS in pSB1A2 (E. coli DH10B cells, no gene for T7 RNA polymerase)
Once we validated the functionality of our approach, we proceeded to subclone the T7-RBS-blaNDM-2&5 fragment into the BBa_J23101 plasmid. Positive transformants were chosen based on their meropenem resistance and confirmed through colony PCR verification. In this configuration, MBL production occurred consistently from the J23101 promoter, facilitating straightforward functional assessments.
Regrettably, the constructs based on pSB1A2 proved unsuitable for mutagenesis. Firstly, this vector is characterized by a high copy number, whereas the system necessitated a low copy number target. Additionally, it contains a standard β-lactamase gene, conferring resistance to ampicillin, which could have interfered with our MBL when assessing our extract for ampicillin degradation. To address these challenges, we carried out subcloning of our T7-RBS-amilGFP and T7-RBS-blaNDM-2&5 fragments into the low copy number SC101 ori-based vector pSB4K5 (kanamycin resistance).
Construct J23101-T7-RBS-amilGFP CDS in pSB4K5 (E. coli DH10B cells, no gene for T7 RNA polymerase)
The resulting constructs were employed for mutagenesis purposes. In this process, plasmid DNA from our test construct was extracted from positive transformants and subsequently reintroduced into competent cells that harbored plasmids pDae029 (eMutaT7[TadA-8e] – Addgene 187620), pDae069 (eMutaT7[PmCDA1] with an optimized RBS – Addgene 187621), or pDae079 (eMutaT7[transition] – Addgene 187622). Following this, amilGFP-expressing clones were cultivated on media containing both chloramphenicol and kanamycin, and mutagenesis was initiated by the addition of L-arabinose. This procedure underwent three rounds of mutagenesis, with colony monitoring conducted after each round. Multiple clones with altered colour, brightness, and/or size were readily detectable after the second round. Their number was highest in the cells containing the pDae079 mutagenesis plasmid.
Initial transformants and the resulting colonies with altered colour, brightness, and/or size due to the random mutagenesis
Next, we re-transformed the pSB4K5-J23101-T7-RBS-blaNDM-2&5 construct into competent E. coli DH10B cells with pDae079. Three positive transformants were selected on media with ampicillin and 10 rounds of mutagenesis were performed with each of them. Aliquots of each clone from every round were combined and incubated in liquid LB medium enriched with elevated meropenem concentrations. Remarkably, we witnessed bacterial growth in cultures exposed to 600 µg/ml of meropenem, representing a remarkable improvement of more than 24-fold compared to the original VIM-2 construct.
The improved activity of the mutagenized NDM-2&5 MBL
To validate the effectiveness of the newly developed CADABRA extract based on NDM-2&5 in eliminating carbapenems from hospital wastewater, we replicated last year's experiment using artificial urine contaminated with 750 µg/ml of meropenem. We followed the same procedure, and this time we observed the complete removal of antibiotics from the sample. Our first goal was accomplished!
NDM-CADABRA vs artificial urine with meropenem
Concerning the challenge of ampicillin degradation, we evaluated the performance of the original VIM-2 CADABRA and the newly prepared, non-mutagenized NDM-2&5 CADABRA against elevated ampicillin concentrations (1 mg/ml) within a short treatment duration (45 minutes). The outcomes demonstrated that our new version surpassed the original by at least fourfold.
NDM-2&5- and VIM-2-based CADABRA against high levels of ampicillin
With this we demonstrated our engineering success.