We have built over 200 hundred parts to load different kinds of MBLs. For the sake of characterizing these parts, it is an innovation for us to combine two usual methods, detecting Fluorescent Intensity and Absorbance for each, to do the screening and evaluation of inhibitors.
With fluorescent probe
Using CDC-1 as a fluorescent probe hydrolyzed by MBLs , we can simply measure the value of fluorescent intensity at same intervals of seconds. Then the reaction velocity was calculated via linear fitting to demonstrate.
Firstly, the system of HTS was established by a series of experiments in different conditions. At the same time of determining the concentration of substrate, enzymatic parameters including Km, Kcat were measured with fluorescent probe.
Determination of enzyme(NDM-23) concentration. Three curves show the reaction velocity of different enzyme concentrations.
Determination of enzyme(NDM-23) concentration. The time starts after 5 mins incubation to model HTS system adding. Original protein concentration is 1.4g/ml and its final concentration in system after diluting 1000 times is 48.22nM.
Determination of enzyme concentrations in 32000-time dilution and more
R² fit values (0-300s) and rE for different enzyme concentrations
Secondly, in the process of HTS with thousands of small molecular compounds, changing of the values of fluorescent intensity shows their inhibitory effects to some extent. Some values are much smaller than those of negative controls, thus we paid more attention to them doing repeated screening. And conditions in some wells are opposite, having higher fluorescent values may indicate that there is a fluorophore in the compound’ s original structure. In the repeated screening, data is used for the calculation of quenching rate and inhibitory rate.
Real-time reaction curve in HTS. Line 1 are negative controls. Curves show that most compounds in this plate may not as well play a role on beta-lactamase, except for F4.
Inhibitory curve of Tannic acid of NDM-23.
Columns of fluorescent quenching, inhibitor Tannic acid to NDM-23.
Thirdly, IC50 in vitro was tested with difference concentrations of inhibitors. Initial velocity was tested the same way. Also, in order to characterize the inhibitory mechanism further, different concentrations of inhibitors connected to different concentrations of substrate were added then tested.
IC50 curve of Tannic acid to 1.51nM NDM-23 in vitro.
Inhibitory mechanism of Tannic Acid of NDM-23. With different concentrations of inhibitors, the fit lines passing the origin, which shows that its mechanism should be reversible inhibition. And the Lineweaver-Burk plot of different concentrations of inhibitors shows that it’s competitive inhibition type.
By UV-Vis spectroscopy
Absorbance of specific peak related to antibiotics shows the effect of inhibitors in living E.coli cells which has been transformed with MBLs’ gene circuits. To ensure the accuracy of UV-vis test, we dedicatedly picked 96-well transparent plate and the living condition of bacterial liquid, as well as set several controls. UV-vis spectroscopy is a convenient way to show how fast inhibitors play its role on MBLs and the therapy influence with antibiotics.
Fig 9. UV- vis spectra of cefazolin hydrolysis by 3 nM purified NDM-23 (A), in the presence of E. coli cells (OD600 = 0.15) expressing NDM23 (B), E. coli cells (OD600 = 0.15) without NDM23 plasmid (C), and in the supernatant of sample A (D). All samples were prepared with 50 mM phosphate buﬀer, pH 7.0.
We used GraphPad Prism software to analyze amount of data.
Use "nonlinear fit" – "straight line" to calculate the initial velocity of each reaction a.k.a. its slope value. Relative plots were drawn automatically. For absorbance, the initial velocity and IC50 was measured the same way. And the initial value and final value of absorbance displayed by antibiotics were subtracted to show the change of absorbance.
As for other kinds of nonlinear fit, we calculated:
EC 80 value. Set log(concentration of protein) as X, the rate of emission as Y. Use "nonlinear fit" – "log(agonist) vs. response—Find ECanything", input 80 as the value of F parameter.
Km Input [S] as X, initial velocity as Y. Use "nonlinear fit" – "Michaelis-Menten" to fit Michaelis plot of MBL. At the same time the software will calculate kinetic constants Km, Vmax automatically.
Kcat Kcat=Vmax/[E]*fluorescent calibration value.
IC50 Import log[I] as X, inhibitory rates as Y. Use "nonlinear fit" – "log(inhibitor) vs. normalized response – Variable slope" to fit IC50 curve, and its value would be calculated automatically.
In a nutshell, we set up a measurement system combining of testing fluorescent intensity as well as absorbance. It yields benefit to our inhibitor screening and assessment. Specific fluorescent substrate is used through all the procedure of HTS, and difference absorbance of antibiotics is monitored in living bacterial cells by UV-vis.
 Assay Platform for Clinically Relevant Metallo-β-lactamases. Sander S. van Berkel, Jürgen Brem, Anna M. Rydzik, Ramya Salimraj, Ricky Cain, Anil Verma, Raymond J. Owens, Colin W. G. Fishwick, James Spencer, and Christopher J. Schofield. Journal of Medicinal Chemistry 2013 56 (17), 6945-6953
 TIAN Cong-hui, TANG Yan-ting, WANG Quan, ZHOU Hong-gang. Establishment and Application of a Model for Drug Screening Targeting Neprilysin Proteinase. China Biotechnology, 35(2), 52-58.
 ZHU Yun-peng, WANG Peng, XIA Bo-ran, TANG Yan-ting. Screening and Inhibition Kinetics of SARS Coronavirus Main Protease Inhibitors. China Biotechnology, 36(4), 35-42.
 Ying Ge, Ya-Jun Zhou, Ke-Wu Yang, Yi-Lin Zhang, Yang Xiang and Yue-Juan Zhang. Real-time activity assays of b-lactamases in living bacterial cells: application to the inhibition of antibiotic-resistant E. coli strains. Royal society of Chemistry. 2017