<!doctype html> Results

Results

Overview

Build Responsory Parts

We totally submitted more than 200 parts. After testing the circuits with our engineering expression system, we finally confirmed 4 of them that fit the system well and showed us satisfactory results. They are NDM-23 with GST tag(BBa_K2933161), SPG-1 with SUMO tag(BBa_K2933183), ElBla II with SUMO tag(BBa_K2933184) and AFM-1 with GST tag(BBa_K2933164).

Screen Out Broad-spectrum Inhibitors

With our specific fluorescent probe CDC-1, we firstly explored the most suitable high-throughput screening system, then measured related enzyme kinetics parameters. Then we screened out effective inhibitors for these four metallo-beta-lactamases from FDA drug libraries and traditional Chinese medicine libraries. Each beta-lactamase has at least 10 inhibitors.

Ensure Inhibitors’ Influence

After initial screening effective inhibitors, further than that, we tested the inhibitory mechanism type of each inhibitors to corresponding beta-lactamase and measured their inhibitory constants, IC₅₀ and Ki. Then we tested them in living bacteria to evaluate.

Data and Results

Performance of Genetic Circuits

We successfully constructed our parts, expressed and purified four kinds of metallo-beta-lactamase using parts constructed which are NDM-23, SPG-1, AFM-1 and ElBla II.Below shows our purification results.

Fig. 1 The picture of PCR, SDS page and gel filtration A. NDM-23 B. SPG-1 C. AFM-1 D. ElblaII

Establishment of Fluorescent HTS System

Following the principles mentioned in Design part, we completed establishment of HTS system of 4 MBLs. Each kind of enzyme owns its unique characteristics, and performs activity highly under different conditions.

For NDM-23

Fig.2 Determination of enzyme concentration.

We set gradients then decided the proper concentration in system of NDM-23 is 1.51nM, through comparing R² and EC₈₀ value. And this number of concentration is less than usual in HTS system, which shows our circuit works well to provide beta-lactamase with high activity.

Fig.3 Determination of buffer condition.

The first picture shows the effect of NaCl concentration to protein activity, and the second is about Zn²⁺ concentration. These results indicate that this kind of metallo-beta-lactamase depends on low concentration of zinc ion, which shows the same as reports.^[1] And under the circumstance with lower pH, the enzyme can be inactivated. Its most suitable pH is around 8.5.
To explore the potential influence of substrate’s and compounds’ solvent, DMSO, we change the percentage of DMSO in reaction system when diluting protein, without changing the final concentrations for each components. Data shows that at the range of 2% to 10%, the effect of DMSO can be ignored, and RSD value is less than 5%.

Chart.1 HTS system of NDM-23

Components	Volume	Concentration
NDM-23	94μL	1.51nM
NaCl		140mM
ZnCl₂		25μM
pH		8.0
CDC-1	5μL	8.5μM
Compound	1μL	20μM
Total	100μL

For SPG-1

Fig.4 Determination of enzyme concentration.

The proper concentration in system of SPG-1 is 24.72nM, which shows a bit lower activity compared with NDM-23.

Fig.5 Determination of buffer condition.

Experimental data shows that this kind of MBL gets higher activity under higher pH of 9.5.

Chart.2 HTS system of SPG-1

Components	Volume	Concentration
SPG-1	94μL	24.72nM
NaCl		35mM
ZnCl₂		25μM
pH		9.5
CDC-1	5μL	8.5μM
Compound	1μL	20μM
Total	100μL

For AFM-1

Fig.6 Determination of enzyme concentration.

The proper concentration in system of AFM-1 is 0.85nM.

Fig.7 Determination of buffer condition.

Compared with other 3 MBLs, AFM-1 fits higher salt concentration and higher zinc concentration.

Chart.3 HTS system of AFM-1

Components	Volume	Concentration
AFM-1	94μL	0.85nM
NaCl		280mM
ZnCl₂		100μM
pH		8.5
CDC-1	5μL	8.5μM
Compound	1μL	20μM
Total	100μL

For beta-lactamase II [Erythrobacter litoralis HTCC2594] (ElBla II)

Fig.8 Determination of enzyme concentration.

The proper concentration in system of ElBla II is 1.887nM.

Fig.9 Determination of buffer condition.

These results indicate that this kind of metallo-beta-lactamase depends on low concentration of NaCl.

Chart.4 HTS system of beta-lactamase II [Erythrobacter litoralis HTCC2594]

Components	Volume	Concentration
ElBla II	94μL	1.887nM
NaCl		35mM
ZnCl₂		25μM
pH		8.5
CDC-1	5μL	8.5μM
Compound	1μL	20μM
Total	100μL

Determination of Enzyme Activity

As a step of system establishment, K_m and k_cat of MBLs are required to determine in order to decide the concentration of substrate. These kinetic parameters are our brilliant achievements as well. We use Michaelis-Menten and Lineweaver-Burk plotting the same data then choose the best fit one, comparing their R² to improve the accuracy of values of constants.

Fig.10 Michaelis-Menten plot and Lineweaver-Burk plot for NDM-23

Chart.5 Kinetic parameters of NDM-23. As reports^[2]NDM-1’s K_m with the same substrates is 17.6μM, which is very close to our results of NDM-1’s variant NDM-23.

Constant	Value
K_m	1.50×10^-5 mol·L^-1
V_max	18.9 Fluorescent unit·s^-1
k_cat	7.97 s^-1
k_cat/K_m	5.30×10⁵ mol^-1·L·s^-1

Fig.11 Michaelis-Menten plot and Lineweaver-Burk plot for SPG-1

Chart.6 Kinetic parameters of SPG-1.

Constant	Value
K_m	2.243×10^-5 mol·L^-1
V_max	167.64 Fluorescent unit·s^-1
k_cat	4.31 s^-1
k_cat/K_m	1.92×10⁵mol^-1·L·s^-1

Fig.12 Michaelis-Menten plot and Lineweaver-Burk plot for AFM-1

Chart.7 Kinetic parameters of AFM-1.

Constant	Value
K_m	9.21×10^-6 mol·L^-1
V_max	57.55 Fluorescent unit·s^-1
k_cat	43.06 s^-1
k_cat/K_m	4.68×10⁵ mol^-1·L·s^-1

Fig.13 Michaelis-Menten plot and Lineweaver-Burk plot for ElBla II

Chart.8 Kinetic parameters of ElBla II.

Constant	Value
K_m	1.00×10^-5 mol·L^-1
V_max	86.79 Fluorescent unit·s^-1
k_cat	29.26 s^-1
k_cat/K_m	2.93×10⁶ mol^-1·L·s^-1

Effective Inhibitors We Screened Out

Through high-throughput screening with fluorescent probe CDC-1, we successfully found several possible inhibitor molecules, from thousands of compounds. Among these results, the broad-spectrum inhibitor towards all four MBLs has emerged and several compounds are found to be inhibitors of a MBL. However, this is not our destination. Some kinetic parameters should be measured to validate their effectiveness, half inhibitory concentration (IC₅₀) and Ki.

Chart.9 Results of screened inhibitors to NDM-23, SPG-1, AFM-1, ElBla II

Tannic Acid: a broad-spectrum arrow

Fig.14 IC₅₀ of Tannic acid for NDM-23, SPG-1, AFM-1 and ElBla II

Tannic Acid is a kind of ingredient of Chinese medical and is widespread in rhubarb. And it belongs to polyphenols, a class of Phenolic compounds (PCs). PCs is one of the most diverse groups of plant secondary metabolite which contain aromatic rings with one or more hydroxyl or methoxyl groups. The antimicrobial activity of PCs is dependent on their chemical structure, in particular, the substitution position in the benzene ring and the saturated chain length. It has been reported that the antimicrobial action of PCs is related to the inactivation of cellular enzymes, by changing membrane permeability^[3].

With the emergence of multidrug resistant bacteria, there is a need to develop not only one kind of MBL inhibitors, but compounds with dual MBLs inhibition properties for use in combination therapies^[5]. Therefore we highlight broad-spectrum inhibitors. What we have found is that tannic acid have inhibitory effect on all of four MBLs we obtained. It is a special arrow targeting at different metallo-beta-lactamases.

Fig.15 Inhibitory mechanism of tannic acid to NDM-23

With different concentrations of inhibitors, the fit lines passing the origin, which shows that its mechanism should be reversible inhibition. Then with different concentrations of substrate, fit lines gather at Y axis, which shows that its mechanism is competitive inhibition.

In our measurement of kinetic constants and mechanisms, tannic acid has a IC₅₀ value in microlevel of four MBLs, which is distinctly lower than other researchers’ results^[3]. After ensuring its inhibitory mechanism of reversible and competitive, we can calculate the Ki value of tannic acid to NDM-23 is 0.1075μM according to the equation between IC₅₀ with Ki^[4].

Binding to the active site of NDM-23, this result is in harmony with modeling prediction, more explanation can be seen in their docking discussion.

Arrows from Chinese Medicine Library

From the 24 effective inhibitors of NDM-23 screened out before, we selected 8 of them to test. Four are from Chinese medicine library, tannic acid, corilagin, ginkolic acid C15:1, and (-)-Epigallocatechin gallate(EGCG). There is a similarity of them that they all belongs to phenolic compounds mentioned above. However, with distinctions on structure, their IC₅₀ and inhibitory mechanisms are not same.

Fig.16 IC₅₀ and inhibitory mechanism of Corilagin for NDM-23

Corilagin also shows a good inhibitory effect via IC₅₀ value which is at the level of micrometer. When the concentration of corilagin changes, its V_max value changes while K_m value is same, which shows that its mechanism is uncompetitive type.

Fig.17 IC₅₀ and inhibitory mechanism of (-)-Epigallocatechin gallate(EGCG) for NDM-23.

Having highly similar structure with tannic acid, EGCG was tested to have same inhibitory mechanism with it, reversible and competitive inhibition. It can be seen as a useful inhibitor as well. And its Ki value is 8.36μM.

Fig.18 IC₅₀ and inhibitory mechanism of (-)-Epigallocatechin gallate(EGCG) for NDM-23

Arrows from FDA Approved Libraries

From FDA approved drug libraries, we have screened out more potent inhibitor compounds.

Fig.19 IC₅₀ and inhibitory mechanism of Adapalene for NDM-23, and IC₅₀ of Adapalene for AFM-1. The curve shows that it is irreversible inhibition.

Fig.20 IC₅₀ and inhibitory mechanism of Silver sulfadiazine for NDM-23.

Fig.21 IC₅₀ and inhibitory mechanism of Fingolimod (FTY720) HCl for NDM-23.

Fig.22 IC₅₀ of other specific inhibitors we screened, including Etravirine(TMC125) for NDM-23, Punicalagin for SPG-1 and Isoconazole nitrate for ElBla II.

Chart.10 Kinetic constants of inhibitors

Protein	Inhibitor	IC₅₀(μM)	Inhibitory mechanism	Ki(μM)
NDM-23	Tannic acid	0.1687	Reversible, competitive	0.1075
	Corilagin	0.8069	Reversible, uncompetitive	0.2903
	Adapalene	8.307	Irreversible	-
	Ginkgolic acid C15:1	4.969	Irreversible	-
	(-)-Epigallocatechin gallate	13.07	Reversible, competitive	8.36
	Etravirine (TMC125)	3.352	-	-
	Silver sulfadiazine	0.6774	Reversible, uncompetitive	0.2437
	Fingolimod (FTY720) HCl	7.721	Reversible, competitive	4.94
SPG-1	Tannic acid	0.1674	Reversible, competitive	0.1124
SPG-1	Punicalagin	0.7235	-	-
AFM-1	Tannic acid	1.175	Reversible, competitive	0.6108
AFM-1	Adapalene	25.84	Irreversible	-
ElBla II	Tannic acid	2.561	Reversible, competitive	1.3838
ElBla II	Isoconazole nitrate	8.618	-	-

Monitoring in living bacterial cells with antibiotics

Experiments proving the accuracy of the system

Firstly, NDM-23 protein solution at 3.02nM was mixed with cefazolin at 250 μM, and the characteristic absorption peak of cefazolin at 273nm obtained by our experiment was continuously measured.

Fig.23 The curve of specific absorption peak of antibiotics mixed with NDM-23 was revealed.

According to the figure, we can clearly see that the specific absorption peak value of antibiotics decreases with time, which proves that we can choose the specific absorption peak of antibiotics as the criterion to evaluate whether antibiotics are decomposed or not.

Secondly, after phosphate buffer treatment, E. coli strain transferred to NDM-23 beta-lactamase gene was diluted to OD₆₀₀=0.15, and the strain was mixed with cefazolin with a final concentration of 250 μM, and the peak value of cefazolin at 273nm was continuously measured. The results are shown below.

Fig.24 The curve of specific absorption peak of antibiotics mixed with E. coli with the NDM23 beta-lactamase gene was revealed

This figure shows that the parts constructed by us can operate in E.coli and play its original functions, which proves the feasibility of our next step——to evaluate the therapeutic effect of inhibitors

At the same time, the phosphate buffer used in the preparation of bacterial solution and the E.coli without the transfer of NDM-23 beta-lactamase gene were mixed with cefazolin with a final concentration of 250 μM respectively as blank control and negative control, and the specific absorption peak value of cefazolin at 273nm was continuously determined.

Fig.25 The curve of specific absorption peak of antibiotics mixed with the phosphate buffer and the E.coli without the transfer of NDM23 beta-lactamase gene was revealed. The one up is antibiotic combined with phosphate buffer and the one down is antibiotic combined with the E coli without the transfer of NDM-23 beta-lactamase gene.

Establishment of measurement system

Since NDM-23 beta-lactamase is metallo-beta-lactamase, it should be able to hydrolyze most kinds of beta-lactam antibiotics. We put the cefazolin, faropenem and e.faecalis respectively in the form of a final concentration of 250 microns and mixed them with E. coli with the NDM-23 beta-lactamase gene. Through continuous determination of experiments we get the specific absorption peak and peak value of the three kinds of antibiotics, in order to find the optimum substrate. We also measured the characteristic peak absorption of the nonbeta-lactam antibiotic tetracycline over time as proof of experimental accuracy. The results are as follows:

Fig.26 The curve of specific absorption peak of antibiotics mixed with E. coli with the NDM-23 beta-lactamase gene was revealed. The peaks of faropenem, meropenem and tetracycline were measured at 307nm, 300nm and 360nm, respectively

According to the figure, tetracycline was not broken down by E.coli that transferred the NDM-23 beta-lactamase gene, proving the accuracy of the experiment. Faropenem and meropenem are well decomposed by E. coli too.

After determining the substrate of the method，we establish a system for the determination of viable bacteria. We have choose the best condition of E.coil for the measurement for different beta-lactamase. For NDM-23, the OD of E.coil is 0.02 and the induction time is 2h. For SPG-1, the OD is 0.4 and the induction time is 2h. For Elbla II, the OD is 0.1 and the induction time is 3h.For AFM-1, the OD is 0.1 and the induction time is 3h.

We used EDTA as a positive control for therapeutic efficacy evaluation. As a metal ion chelator, EDTA can chelate the zinc ion of metal beta-lactamase and inactivate the lactamase. The result are as follows.

Fig.27 EDTA inhibition curve of NDM23. The top curve is the decomposition curve of cefazolin at different concentration of EDTA. The below curve shows the inhibition curve of EDTA according to the inhibition rate.

As a metal ion chelator, EDTA can inhibit metal beta-lactamase well. The inhibitory effect of EDTA measured by our evaluation system is close to the literature value (IC₅₀ values from the literature is 1.63 μM), which also proves that our therapeutic effect evaluation system is feasible

Evaluation of therapeutic effect of NDM-23

After determining the reaction system, we diluted the concentration gradient of several inhibitors and incubated them with the prepared E.coli which has been transferred into NDM-23 beta-lactamase gene in dark for 1h, then added cefazolin with a final concentration of 250 μM and continuously measured the light absorption value at 273nm. The results are as follows.

Fig.28 Corilagin inhibition curve of NDM23. The top curve is the decomposition curve of cefazolin at different concentration of Corilagin. The below curve shows the inhibition curve of Corilagin according to the inhibition rate.

Corilagin is a water-soluble tannin. It can be extracted from longan, and the content of underleaf pearl and Tibetan sweet root corilla jing is also relatively high. We screened it from the Chinese medicine library and showed a good inhibitory effect on the protein level of NDM-23 beta-lactamase. Corilagin maintained a stable inhibitory effect in the evaluation of therapeutic effect.

Fig.29 EGCG inhibition curve of NDM23. The top curve is the decomposition curve of cefazolin at different concentration of EGCG. The below curve shows the inhibition curve of EGCG according to the inhibition rate.

Epigallocatechin gallate is the main component of green tea polyphenols which is a catechin monomer isolated from tea leaves. It stands out from our modeling work. Our modeling group screened potential inhibitors by homologous modeling followed by molecular docking. EGCG scored highly in our modeling work, indicating that it binds well with NDM-23.EGCG also showed good inhibitory effect in our evaluation of therapeutic effect, and its IC₅₀ was relatively low, which could be used in combination with cefazolin to achieve good bacteriostatic effect.

Evaluation the therapeutic effect of screened broad-spectrum beta-lactamase inhibitors

After high-throughput screening, tannic acid was screened as the inhibitor of four beta-lactamases which are NDM-23, SPG-1, ELBLA2-1 and AFM-1. We have used the UV visible method to assess the effectiveness of the treatment. The results are as follows .

Fig.30 Tannic acid inhibition curve of NDM-23. The top curve is the decomposition curve of cefazolin at different concentration of Tannic acid. The below curve shows the inhibition curve of Tannic acid according to the inhibition rate.

Fig.31 Tannic acid inhibition curve of SPG. The top curve is the decomposition curve of cefazolin at different concentration of SPG. The below curve shows the inhibition curve of SPG according to the inhibition rate.

Fig.32 Tannic acid inhibition curve of ElBla II. The top curve is the decomposition curve of cefazolin at different concentration of ElBla II. The below curve shows the inhibition curve of ElBla II according to the inhibition rate.

Fig.33 Tannic acid inhibition curve of AFM-1. The top curve is the decomposition curve of cefazolin at different concentration of AFM-1. The below curve shows the inhibition curve of AFM-1 according to the inhibition rate.

Tannic acid is a tannic acid derived from gallnut. According to the outcome of Evaluation the therapeutic effect, we can see that Tannic acid has a broad spectrum of inhibitory effects on beta-lactamase. Tannic acid has different abilities to inhibit different beta-lactamase in living bacteria.

Brief summary of the project

After our high-throughput screening of FDA approved drug libraries, we have screened out good inhibitors and evaluated the therapeutic effect of inhibitors with better inhibition rate. We have identified several representative inhibitors. We also found a broad spectrum of inhibition of metal beta-lactamase inhibitors - tannic acid. It is expected that tannic acid can be used as an inhibitor of metal beta-lactamase in combination with beta-lactam antibiotics to treat drug-resistant bacteria.

Chart.11 Our Excellent inhibitors

Protein	Inhibitor	IC₅₀(μM)	Living bacterium IC₅₀(μM)	Inhibitory mechanism
NDM-23	Corilagin	0.8069	38.02	Reversible, uncompetitive
	(-)-Epigallocatechin gallate	13.07	15.73	Reversible, competitive
	Tannic acid	0.1687	31.62	Reversible, competitive
SPG-1	Tannic acid	0.1674	42.66	Reversible, competitive
AFM-1	Tannic acid	1.175	25.12	Reversible, competitive
ElBla II	Tannic acid	2.561	28.18	Reversible, competitive

Winding Roads

Future Plans

Co-crystallization of beta-lactamase and inhibitor complex

Clinical antibiotic-resistant bacteria test

Minimum Inhibitory Concentration

References

[1] Lin-Cheng Ju, Zishuo Cheng, Walter Fast, Robert A. Bonomo, Michael W. Crowder, The Continuing Challenge of Metallo-beta-Lactamase Inhibition: Mechanism Matters, Trends in Pharmacological Sciences, Volume 39, Issue 7, 2018, 635-647.
[2] van Berkel SS, Brem J, Rydzik AM, et al. Assay platform for clinically relevant metallo-beta-lactamases. J Med Chem. 2013;56(17):6945–6953.
[3]Mandal S M , Dias R O , Franco O L . Phenolic Compounds in Antimicrobial Therapy[J]. Journal of Medicinal Food, 2017:jmf.2017.0017.
[4]Cer RZ, Mudunuri U, Stephens R, Lebeda FJ. IC₅₀-to-Ki: a web-based tool for converting IC₅₀ to Ki values for inhibitors of enzyme activity and ligand binding. Nucleic Acids Res. 2009;37(Web Server issue):W441-W445.
[5] Brem, Jürgen, Van Berkel S S , Aik W S , et al. Rhodanine hydrolysis leads to potent thioenolate mediated metallo-β-lactamase inhibition[J]. Nature Chemistry, 2014, 6(12):1084-1090.

↑

Team:TJUSLS China/Results