STUDY OF THE EFFECT OF PROTEIN SYNTHESIS INHIBITORS ON GROWING ESCH ERICH IA CO LI BACTERIA USING ELECTROCHEMICAL SENSORS

Background. The study of the mechanisms of action of antibiotics requires the inte­ grated use of traditional microbiological and physicochemical methods. The aim. To study the response of Escherichia coli bacteria to the action of four antibiotics, inhibitors of protein synthesis, using combined approach. Methods. Bacteria were grown under aerobic conditions on minimal M9 medium with glucose. Tetracycline, kanamycin, streptomycin and chloramphenicol have been tested. The effect of antibiotics on survival (CFU) and growth rate was determined. Respiratory activity, sulfide production, extracellular potassium, as well as pH and Eh of the medium were measured using electrochemical sensors directly in a growing culture in the "real time", membrane potential was measured using a DiBAC dye and a Leica DM2000 fluorescent microscope. Results. The tested antibiotics were divided into two groups according to their prop­ erties. Tetracycline and chloramphenicol showed a pronounced bacteriostatic effect, growth inhibition began immediately after the addition of antibiotics and occurred at a high rate. Both antibiotics inhibited respiration, stimulated sulfide production and an Eh jump. Bacteria treated with tetracycline and chloramphenicol retained the ability to maintain membrane potential and intracellular potassium better. Inhibition of respiration led to a decrease in glucose catabolism, as evidenced by a lower rate of acidification of the medium compared to the control. Growth inhibition with streptomycin and kanamycin was initiated with a 30-minute delay. Both antibiotics showed a bactericidal effect, did not stimulate sulfide production and Eh jump, did not inhibit respiration, but caused a drop in membrane potential and intracellular potassium. High respiratory activity promoted glucose catabolism, as evidenced by the rapid acidification of the medium. Of interest is the detection of kanamycin-induced sulfide production during E. coli growth on MOPS medium. Conclusion. An analysis of the data obtained indicates that the use of electro­ chemical sensors in combination with traditional methods is a promising approach to studying the mechanisms of action of antibiotics.

The abrupt changes in the oxidation-reduction poten tial to negative values (Eh jumps) measured by the platinum electrode in aerobic cultures of Escherichia coli and other bacteria, growing in a minimal medium were previously detected when growth ceased due various stresses [1]. Later it was shown that these jum ps are associated with a reversible increase in the level of sulfide or cysteine in the medium [2]. We also observed the accumulation of sulfide and sulfide-mediated Eh jumps during exposure of E. coli to certain antibiotics, including chloram pheni col, tetracycline and high doses of ciprofloxacin, while the other antibiotics, such as ampicillin, cefotaxime, kanamycin and streptomycin did not stimulate the formation of sulfide [1][2][3]. It was of interest to investigate the rea son for the different response of bacteria to the action of various antibiotics. It was suggested that differenc es in the response of bacteria to treatment with differ ent antibiotics may be related to the traits of their effect on growth and metabolic activity. To test this hypothe sis, we examined the response of growing E. coli to three antibiotics, including tetracycline, kanamycin and strep tomycin, known to inhibit protein synthesis in bacteria. For comparison, the list of antibiotics under study includ ed the well-studied chloramphenicol, which has the abil ity to induce sulfide release [3].
To solve the problems, a combined approach was ap plied, including the measurement of growth parameters (growth rate and the ability to form colonies, CFU), and con tinuous in vivo real-time registration of a number of im portant physiological parameters using electrochemical sensors.

M A T E R IA LS A N D M ETH O D S O F R ESEA R C H
The object of study was the gram -negative bacte ria Escherichia coli BW25113 (wt) and the knockout mu tant JW2663 (AgshA), deficient in synthesis of glutathione (GSH), both from the Keio international collection. Bacte ria were cultured on M9 minimal medium [4] with glucose as a source of carbon and energy under aerobic condi tions at 37 °C in a thermostatically controlled orbital shak er at 150 rpm. To determine the levels of extracellular po tassium, bacteria were grown on M9 medium with a low content of potassium. Where indicated, MOPS (3-[N-morpholine]propanesulfonic acid) medium was used. Colony formation capacity (CFU) and the specific growth rate (p) were determined in a conventional manner.
Measurements using electrochemical ion-selective sensors were carried out in flasks with aerobically grow ing E. coli cultures. All parameters were continuously and synchronously processed in real time using a hard ware-software complex, including various registration blocks. The partial pressure of oxygen (dO2) was measured using an InPro 6800 Clark electrode (Mettler Toledo, USA) on a modified BioFlo 110 controller (New Brunswick Sci entific Co., USA). The concentration of extracellular sulfide ion was determined using a sulfide-specific ion-selective XC-S2-001 electrode (Sensor Systems Company, Russian Federation), the redox potential (Eh, ORP) was recorded with an ERP-105 platinum electrode (Measuring Technol ogies, Russian Federation), the level of extracellular potas sium ions (K+) was recorded with a K+-selective electrode ELIS-121K (Measuring Technologies, RF). Results were ex pressed in mV (1 mV = 15 pM K+). Potentiometric data from these electrodes were processed with cpX-2 digital pX me ters (IBP, Pushchino, Russian Federation). Synchronization of the primary data received from these sensors was car ried out using the RS-232 and Modbus protocols. Graphical visualization of real time monitoring of dO2, Eh, S2-and K+ ions is presented in the form of typical curves from a se ries of experiments.
Each result is indicated as the mean value of at least three independent experim ents ± the standard error of the mean (SEM). Significant difference was analyzed by Student's t-test. A p-value of 0.05 was used as the cut off for statistical significance. The results were analyzed by means of Statistica 6 (StatSoft Inc., USA).

Determination of growth parameters under the action of antibiotics
All studies were performed using cultures in the mid dle of the logarithmic phase of growth, in the range of opti cal density of the culture (OD600) 0.4-0.8. During the growth of bacteria cultured in the absence of any treatments (con trol), the specific growth rate (p) gradually decreased from the maximum value (0.71 ± 0.01) to 0.43 ± 0.02 h-1 ( Fig. 1). In preliminary experiments, the concentration of the antibi otic was selected so that its effect was manifested in the re gion of the optimal bactericidal concentration, with the ex ception of tetracycline. The effect of this antibiotic was stud ied at different concentrations.
In the response of bacteria to the action of aminoglyco sides kanamycin (30 pg/ml) and streptomycin (30 pg/ml), three phases were distinguished. The first phase, dur ing which there was no statistically significant growth inhibition, lasted for the first 30 minutes after the addi tion of the antibiotic. In the second phase, which last ed for 15 min, a rapid drop in p to about 0.2 h-1 was ob served. Then the growth rate slowly dropped to zero. Un der the action of tetracycline (30 pg/ml) and chloramphen icol (25 pg/ml), the phase of a rapid decrease in the growth rate was observed, during which, 15 min after the addi tion of antibiotics, p decreased to 0.3 h-1, and the phase of a slow decrease in the growth rate continued until to a complete stop (Fig. 1).
In e xp erim en ts w ith te tra cyclin e , the d e p e n d ence of the effect on the concentration of the antibi otic was studied. During the first 45 min after the addi tion of 2 and 30 pg/ml of tetracycline, growth was inhib ited in the same way, but upon treatment with 30 pg/ml, the grow th stopped by the end of cultivation, while at 2 pg/ml, growth continued at a rate of 0.2 h-1. Inter estingly, under the action of a low dose of tetracycline (0.1 pg/ml) for 105 min of cultivation, the growth rate did not differ from that observed in the control; however, at the end of cultivation, the p value was the same as after treatment with 2 mg/ml. F IG . 1.

Effect o f kan am ycin (Kan) (30 pg/m l), streptom ycin (Str) (30 pg/m l), tetracycline (Tet) (30 pg/m l) a n d ch lo ra m p h e n ico l (Cam ) (25 pg/m l) on the specific grow th rate (p) o f E. co li (wt) grow ing on M9 m e dium with glucose. The arrow indicates the time w hen an tibio tic was added
Data characterizing the colony-form ing capacity un der the action of antibiotics are presented as the differ ence between the logarithms of CFU in samples taken dur ing cultivation and the values at the beginning of cultiva tion (AlogCFU) (Fig. 2). In the absence of any treatments, an increase in the number of CFU was observed as bio mass accumulated. Chloram phenicol, exhibiting bacte riostatic properties, did not cause significant cell death. Tetracycline (30 pg/ml) had a significant bacteriostatic effect and a weak bactericidal effect. The value of CFU by the end of incubation in the presence of tetracycline decreased by about two times. Kanamycin and strep tom ycin at a concentration of 30 pg/m l had a pro nounced bactericidal effect, reducing the number of CFU by 5.9 x 103 and 0.9 x 103 times, respectively. An inverse relationship was found between the rate of growth in hibition (the rate of decrease in p) and the bactericidal action of antibiotics. Thus, kanamycin and streptomycin had the most pronounced bactericidal effect, stopping growth 45 minutes after the addition of the antibiotic. On the contrary, tetracycline and, in particular, chloram phenicol, which sharply inhibited the growth of bacte ria as early as 15 min after their introduction, had rather a bacteriostatic effect. Fig. 1

. The arrow in dica tes the time w hen a n tib io tic w as ad ded
Respiratory activity of bacteria and pH changes Under our conditions, despite intensive agitation of the culture medium, as the biomass accumulated, the res piratory activity of growing cells led to a gradual decrease in the oxygen concentration in the medium to complete and irreversible exhaustion (Fig. 3). The addition of tetra cycline or chloramphenicol to a growing culture of E. coli caused a rapid inhibition of respiration, most pronounced in the culture treated with 30 pg/ml tetracycline. After 120 min of incubation with the antibiotic, an increase in dO2 in the medium was observed almost to the baseline level. Treatment with 0.1 pg/ml tetracycline caused a gradual de crease in O2 consumption, the effect of 2 pg/ml was similar to 30 pg/ml, but less pronounced (Fig. 3).
Unlike chloramphenicol and tetracycline, cells treated with aminoglycosides continued to show high respirato ry activity during the first 60-80 minutes. Notably, the rate of oxygen uptake by kanamycin was even higher than in cul ture untreated with the antibiotic, and in the case of strep tomycin, the same as in untreated cells. In the second phase of the response to the action of aminoglycosides, respiration inhibition was observed, which was absent in cells not treat ed with antibiotics (Fig. 3).
During experiments, we found that the effect of kana mycin on the respiratory activity of growing E. coli is high ly dependent on the com position of culture medium. In this series of experiments, E. coli BW25113 was cultured on minimal synthetic MOPS medium with glucose as a car bon source. Under these conditions, treatment with kanamycin caused irreversible inhibition of respiration as ear ly as 10 min after its addition; the bacteria in this case be haved in the same way as when treated with a high con centration of tetracycline. Fig. 1

. The arrow indicates the time when antibiotic was added
During aerobic growth of E. coli on the M9 minimal medium containing glucose as the only source of car bon and energy, a gradual decrease in pH is observed due to the accumulation of products of incomplete glucose oxidation [6]. The degree of change in the pH of the cul ture medium during continuous monitoring can indirect ly indicate the rate of carbohydrate metabolism in bacte ria and serve as an additional integral indicator of the state of the culture. The choice of the pH sensor and the oper ating mode of the controller made it possible to achieve high sensitivity and stability of indications. The measure ment results were expressed as the difference between the values at the time of antibiotic addition and in the fi nal (ApH). Under our conditions, during normal growth, a gradual decrease in the pH of the medium from 6.86 to 6.66 (ApH = 0.2) was recorded with a noticeable accel eration after the 90th minute of incubation. This coincides with the time of complete depletion of dissolved oxygen in the medium and the transition from aerobic to anaero bic catabolism.
Under the action of tetracycline and chloram phen icol, only a slight decrease in pH was noted (ApH = 0.04 and ApH = 0.07, respectively). On the contrary, treatment with am inoglycosides was accompanied by a decrease in pH, comparable to that observed in culture without treat ment, and even greater in the case of kanamycin. In gen eral, in aerobic cultures, observed changes in pH correlate well with changes in respiratory activity. Higher respira tory activity promoted a higher rate of glucose utilization and the accumulation of acidifying metabolites, which is re corded as a greater decrease in pH.

Redox status of the medium (Eh)
and sulfide production In the absence of treatments, Eh gradually changed from +190 mV (at the point corresponding to the time of ad dition of antibiotics) to -150 mV at the end of the experi ment. The transition of Eh to negative values is associated with the gradual depletion of oxygen in the medium, which confirms the role of O2 as one of the main redox-active com ponents in aerobic bacterial cultures. In the antibiotic-free medium, sulfide production was maintained at a constant low level (Fig. 4a).
Tetracycline and chloramphenicol, when added to E. coli culture, induced a jump in Eh to negative values with ampli tudes of 101.5 ± 6 and 133.8 ± 7.5 mV, respectively. The ad dition of tetracycline in an amount of 0.1 pg/ml did not sig nificantly affect the redox potential of the medium, while the addition of 2 pg/ml caused the Eh jum p with ampli tude of 82 ± 4 mV. Simultaneously, tetracycline and chlo ramphenicol caused a sharp, short-term increase in endog enous sulfide production with amplitudes of 15 ± 2.5 mV and 36.6 ± 2.7 mV, respectively (Fig. 4a). Chloramphenicol stimulated the largest drop in the potential of the sulfide electrode in terms of amplitude and duration among all test ed antibiotics. A dose-dependent pattern of action of tetra cycline was revealed (Fig. 4a). A close relationship was not ed between the degree of inhibition of the specific growth rate and respiratory activity, and the production of sulfide under the action of the tested antibiotics.
Treatment of growing bacteria with aminoglycosides (kanamycin and streptomycin) did not cause an abrupt change in the potential of the Eh electrode, however, it sig nificantly accelerated its decrease (compared to the control), during the first 60 minutes for kanamycin and 80 minutes for streptomycin, after which gradual return of the poten tial to the base value was recorded. The observed chang es in Eh under these conditions can be associated with changes in dO2. There were also no jumps in the potential of the sulfide electrode under the action of both antibiotics on the cells of the parental strain E. coli BW25113.
We have previously shown that stress-induced pro duction of extracellular sulfide in some cases can be signif icantly increased in E. coli mutants deficient in the synthe sis of glutathione (GSH) [2]. To test this effect, in the case of kanamycin, we used the knockout mutant E. coli JW2663 AgshA, deficient in glutathione synthesis. When grown on M9 medium, this mutant, like the parental strain, did not produce sulfide under the action of 30 pg/ml kanamycin. In the mutant (but not in the parent) grow ing on MOPS medium with glucose, kanamycin-induced jum ps in the potential of the sulfide electrode were re corded (Fig. 4b). Their am plitude was about 100 mV, which was significantly higher than in the other exam ined above cases. Fig. 1

Changes in membrane potential and extracellular K+ levels
In the absence of antibiotics, the growth of E. coli was accompanied by an uptake of potassium from the me dium in an amount proportional to the number of cells, which was recorded as a gradual decrease in the poten tial of the potassium-specific electrode. Under these con ditions, bacteria m aintained the membrane potential at a constant level, and the number of stained cells did not exceed 1 %. Treatment with chloramphenicol or tetracy cline caused a decrease in membrane potential in no more than 6 % of cells.
After treatment with chloramphenicol, the bacteria continued to uptake potassium at a rate close to that ob served in the absence of the antibiotic. 30 min after the ad dition of chloramphenicol, a slowdown in potassium up take was observed which corresponded in time to the end of the phase of rapid growth inhibition and a decrease in respiratory activity. Tetracycline affected potassium up take in a dose-dependent manner. At a concentration of 0.1 pg/ml, there was no change in the rate of potassi um uptake within 60 minutes after treatment (compared to the control), while at 2 pg/ml of antibiotic a significant slowdown in potassium absorption was noted already af ter 30 minutes. The addition of 30 pg/ml tetracycline com pletely stopped the entry of K+ into the cells.
When aminoglycosides were added, a two-phase pat tern of changes in the concentration of extracellular po tassium was observed. In the first phase, the cells contin ued to uptake potassium at a rate similar to that observed in the absence of treatment. 45 min after the addition of streptomycin and 60 min after the addition of kanamy cin, the second phase began, in which the concentration of potassium in the medium began to increase, indicating the loss of the ability of bacterial cells to retain K+. An in crease in the level of extracellular potassium may be asso ciated with cell lysis or a drop in membrane potential. In deed, after the treatment of bacteria with streptomycin, already after 30 minutes, there was a significant loss of MP in 6.3 ± 0.6 % of cells. By the end of the experiment, MP lost 19.2 ± 0.04 % of cells. The action of kanamycin caused a less pronounced effect: after 60 minutes of incubation with the antibiotic, the number of stained cells began to in crease (4.65 ± 1.05 %), by the 120th minute 19.6 ± 1.6 % of the cells lost MP. In general, under the action of all tested antibiotics, the observed changes in the membrane poten tial were closely associated with the ability of bacterial cells to retain potassium. In addition, a correlation was found be tween growth inhibition and the ability of cells to maintain the membrane potential.

d i s c u s s i o n
In most cases, the monitoring of changes in the para meters of growing bacterial cultures is carried out by mi crob iological and physicochem ical m ethods in sam ples taken by centrifugation or filtration. In the pro cess of sampling, cells are exposed to various influenc es that are stressful for them. When a sample taken from a growing aerobic culture, bacteria, being in a centrifuge tube, consume oxygen for several seconds and go into a state characteristic of anaerobic or microaerobic growth. Conversely, under membrane filtration, cells are exposed to increased oxygen content. A similar situation is ob served when cells are placed on agar surface in Petri plates. As a consequence, the obtained values of the meas ured parameters may not correspond to those observed in situ. This is especially true for parameters reflecting the respiratory activity and transmembrane ion fluxes, which are characterized by a rapid response to changes in environmental conditions.
A feature of the above research was the use of electro chemical ion-selective sensors installed directly in flasks with aerobically growing cultures, which allowed to simul taneously and continuously measuring several important parameters in real time. Using this approach, a compre hensive real-time monitoring of such important parameters of a growing E. coliculture as respiratory activity, redox po tential, sulfide production, and changes in the levels of ex tracellular K+ ions under the action of antibiotics that block protein synthesis was carried out, in conjunction with classi cal microbiological methods (determination specific growth rate and colony-forming ability, as well as membrane poten tial). The results obtained indicate that these parameters, although to varying degrees, are closely related (Table 1). Growth inhibition of E. coli by tetracycline and chloram phenicol began immediately after the addition of the antibi otic and occurred very quickly. Both antibiotics had a weak er bactericidal effect and a pronounced bacteriostatic effect. It is important that it was tetracycline and chlorampheni col that stimulated the production of sulfide and the jump in Eh. As shown earlier, the extracellular reductant sulfide, interacting with a platinum electrode, shifts its potential to negative values, which is recorded as a jum p in Eh [2]. Inhibition of respiration led to a significant decrease in glu cose catabolism, as evidenced by a lower rate of acidifica tion of the medium compared to the control. At the same time, cells treated with tetracycline and chloramphenicol retained better the ability to maintain the membrane po tential and intracellular potassium.
An essential characteristic of the action of streptomycin and kanamycin is the 30-minute delay in growth inhibition after they have been added to a growing culture. Both an tibiotics had a bactericidal effect, did not stimulate the pro duction of sulfide (and, accordingly, a jump in Eh), did not in hibit respiration, but stimulated a drop in the membrane potential and, accordingly, in intracellular potassium. Con tinued respiration in the presence of both antibiotics pro moted glucose catabolism, as evidenced by acidification of the medium.
Collins et al. showed that treatment of E. coli with a range of bacteriostatic antibiotics that inhibit protein synthesis (including chloramphenicol and tetracycline) re sulted in rapid suppression of cellular respiration. In con trast, three bactericidal antibiotics (ampicillin, gentamicin, and norfloxacin) accelerated respiration. Growth inhibition is associated with suppression of central metabolism, while killing is linked to activation of metabolism [7].
Our data largely confirm the data discussed above. How ever, there are differences. Of the two bactericidal antibi otics, kanamycin stimulated respiratory activity and, ju d g ing by the pH change, showed a higher metabolic activi ty, while streptomycin did not have these effects. Notably, when E. coli was cultured on the MOPS medium kanamycin caused irreversible inhibition of respiration.
We assume that, in our case, the key event determining the differences in the action of bacteriostatic and bacteri cidal antibiotics may be the presence of the delay between the addition of the antibiotic and the onset of growth inhi bition in the first case and its absence in the second case. It is possible that, in turn, this is a consequence of the dif ferent rates of penetration of antibiotics into cells. ppGpp, the signaling molecule of the stringent response, which is the master regulators of the bacterial response to various stresses, may play an important role. It is known that the ac tivities associated with the action of ppGpp are largely de termined by the growth rate of bacteria at the time of stress exposure [8], including the action of antibiotics [9]. How ever, it should be noted that chloram phenicol, which does not have a bactericidal effect, inhibits the synthesis of ppGpp [10].
The scientific literature concerning studies of the ac tion of antibiotics contains a number of conflicting data. One of the reasons for this is the use of different culture conditions by different authors. In this regard, it is of in terest to detect the production of sulfide under the ac tion of kanamycin on the E. coli knockout mutant, when the growth medium is changed from M9 to MOPS medi um. These results indicate that the observed effects can largely depend not only on the concentration of the ac tive substance, but also on the composition of the me- Note. Growth inhibition -growth inhibition starts immediately after antibiotic addition (+ ); BS -bacteriostatic action; BC -bactericidal action; RA -respiratory activity in the first phase of response to antibiotic action, inhibition (4 ), stimulation (T ), no effect (-) ; H2S -presence ( ! ) or absence (-) of production; MP -decreased membrane potential (4 ), no effect (-); pH -decrease of by the end of cultivation.

Antibiotics
dium, which is important to consider when interpreting the results. In E. coli, the release of sulfide into the medium is part of the process aimed at maintaining cysteine homeostasis during stress-induced arrest of protein synthesis. An ex cess of cysteine in this situation can lead to stimulation of the Fenton reaction, during which hydrogen perox ide oxidizes the free divalent iron ion with the formation of a highly toxic hydroxyl radical that can damage DNA, proteins and lipids. Cysteine in this situation can act as a reductant that reduces oxidized iron [11]. To keep cysteine at a safe level, bacteria incorporate some of the cysteine into glutathione and some of it is excreted into the envi ronment. In addition, part of cysteine is cleaved to form sulfide, which is released into the environment and under our conditions can be detected by a sulfide-specific elec trode [2]. Based on the presented scheme, the production of sulfide by kanamycin, observed by us in the MOPS me dium, can be associated with the stimulation of the Fenton reaction due to the presence in this medium (in contrast to the M9 medium) of an increased amount of iron. At the same time, these data confirm the important role of glu tathione in maintaining cysteine homeostasis in E. coli un der stress. It should be noted that endogenously produced H2S renders multiple bacterial species highly resistant to ox idative stress and various classes of antibiotics [12,13].

C O N C LU SIO N
The results obtained in this work not only confirm our previous data, but also provide a deeper understand ing of the mechanisms of bacterial response to the action of antibiotics, inhibitors of protein synthesis. Although the studied antibiotics inhibit the same function in E. coli, two of them are bacteriostatic and two are bactericidal. This is accompanied by significant differences not only in such general physiological parameters as growth rate and survival, but also in respiratory activity, the rate of utiliza tion of the carbon and energy source, stress-induced sulfide production, and the ability to maintain membrane poten tial and intracellular potassium. The increase in the num ber of pathogenic microorganisms resistant to antibiotics, observed in recent years, requires the improvement of ap proaches and methods for the development of new anti microbial drugs and increasing the effectiveness of widely used ones. The data obtained in this work indicate the prom ise of using electrochemical sensors for screening chemical compounds for antimicrobial activity.

Funding
This work has been supported by a grant of the Russian Science Foundation (RSF № 22-14-00093).