The systems

We based our biosensor on two endogenous two-component systems, which naturally grant bacteria resistance against β-lactam antibiotics. These systems are VbrK/VbrR, which are found in the gram-negative bacterium Vibrio parahemolyticus and BlaR1/BlaI, present in the gram-positive bacterium Staphylococcus aureus. Both systems initiate the transcription of a β-lactamase gene as a response to β-lactam antibiotics binding to the membrane-bound receptor.

VbrK is a sensor histidine kinase that plays a pivotal role in the response to β-lactam antibiotics. Upon binding a β-lactam, VbrK undergoes auto-phosphorylation, leading to the activation of the downstream response regulator VbrR by transferring the phosphate group. The activated VbrR, in turn, acts as a transcriptional activator for the β-lactamse gene (Li et al 2016).

The sensor membrane protein BlaR1 on the other hand responds to β-lactam antibiotics by activating its intracellular metalloprotease domain. This activation initiates a cascade of events where the repressor BlaI, responsible for suppressing the transcription of the β-lactamase gene, is cleaved. This cleavage now prevents BlaI from dimerisation, which renders it incapable of hindering transcription. Consequently, transcription is no longer inhibited in this state and expression of the β-lactamase is initiated (Alexander et al 2023).

To utilize these systems for our biosensor, we will exchange the expression cassette for the endogenous β-lactamase gene. Using molecular cloning, we will implement reporter genes behind the endogenous promotor region where the response regulators induce gene expression. This way, binding of β-lactam antibiotics will lead to the gene expression of our reporter proteins, enabling the reliable detection of β-lactam antibiotics in different samples.

Workflow overview of two-component biosensor.

Figure: Molecular mechanism of our biosensor, examplementary for the VbrK/VbrR two-component system. Binding of β-lactam antibiotics (here penicillin) leads to the activation and auto-phosphorylation of VbrK, the membrane-bound histidine kinase. Upon activation, a phospho-transfer to the response regulator protein VbrR occurs. In its phosphorylated state, VbrR acts as an transcriptional activator and leads to the expression of the desired reporter gene. Created with BioRender.com.

Reliable Reporter Genes for Robust Application

To go beyond simply desinging a biosensor, we want to provide a system that makes biosensors accessible and user-friendly for every consumer.

Our system for around-the-clock monitoring of wastewater is based on the encapsulation of our whole-cell biosensor within alginate (Moya-Ramírez et al 2022). Alginate may ring a bell - it is the hydrogel bobas that bubble tea are made of. These boba-biosensors are introduced into water flows. While the water passes through the hydrogel, the cells remain entrapped in it. The readout is achieved through the subsequent fluorescence emitted by the GFP reporter gene that is expressed upon β-lactam binding.

An alternative strategy is to offer the biosensor in the form of a paper test strip. In this context, the β-galactosidase acts as the reporter enzyme, facilitating the conversion of the substrate X-Gal from a colorless state to a blue dye. These cells can be preserved in a refrigerator for several months after treatment with a drying protectant solution, followed by lyophilization, until they are reactivated by exposure to the test solution (Stocker et al 2003). In this case, the read-out does not require a specific light or spectral filters. It is visible by the naked eye and presents an affordable and easy solution for smaller wastewater treatment plants, clinics or even households.

The Biosensor

Since both, Vibrio parahaemolyticus and Staphylococcus aureus are biosafety level 2 organisms, we decided to use Escherichia coli for our microbial whole-cell biosensor. We designed the biosensor as a level 2 Golden Gate construct. For the level 1 constructs, we selected the following components:

  • pOdd1

    Houses the membrane-bound receptor (VbrK or BlaR1) with a His-tag

  • pOdd2

    Contains the response regulator (VbrR or BlaI) also tagged with a His-tag

  • pOdd3

    Encompasses the specific promoter sequence followed by a reporter gene, which could be either GFP or β-galactosidase.

Together with an empty pOdd4, they assemble forming our complete biosensor. The level 1 plasmids can be used for protein expression in BL21 cells followed by Ni-NTA purification and as a template for mutations as a result of the dry lab team.

Schematic for plasmid design of the Golden Gate assembly.

Figure: Plasmid design for the Golden Gate assembly of our biosensor, exemplemantary for VbrK/VbrR. The upper panel shows the level 1 plasmids, the pOdd1 containing the transcriptional unit (TU) for VbrK, pOdd2 containing the TU for VbrR and pOdd3 containing the promotor region of VbrR with the β-galactosidase gene. The pOdd4-spacer plasmid also needed for assembly is not shown here.

In case of the VbrK/VbrR two-component system, the binding site of VbrR in the promoter region of the β-lactamase gene is unkown. To that end, we utilize promoter prediction tools such as SAPPHIRE (Coppens and Lavigne 2020) and integrate estimated binding sites of other studies using local alignments.

Increasing Receptor Specificity

The used receptors in our biosensor are able to detect antibiotics of the β-lactam class without further specificity (Li et al. 2016, Alexander et al. 2023). Therefore, those receptors will not be able to determine the abundance of a specific β-lactam antibiotic, e.g. penicillin. If the specificity of our receptors is increased towards a specific β-lactam antibiotic, our biosensor would achieve a higher level of detail. To that end, we aimed to construct a repertoire of receptor mutations that are more specific to:

Structure of penicillin G.
Penicillin G
Structure of ampicillin.
Ampicillin

For that purpose, the dry lab team makes use of docking, homology modeling, and protein interface design to simulate many different mutations and evaluate their binding affinity to the specific β-lactam antibiotics.

In our research, we utilize a diverse set of tools and software to perform covalent dockings, assess mutants, and visualize proteins effectively. Key components of our toolkit include MOE (MOE 2022) for docking, PyMOL (PyMOL) and OpenEye (Hawkins et al. 2007) for visualization, ROSETTA (Moretti et al. 2016) for modeling and interface design, and PLIP (Adasme et al. 2021) for detailed protein-ligand interaction analysis. This comprehensive approach empowers us to gain profound insights into complex molecular interactions.