Postantibiotic Effect and Delay of Regrowth in Strains Carrying Mutations That Save Proteins or RNA

Abstract

The postantibiotic effect (PAE) values found for proteinase-defective (Lon⁻) Escherichia coli and RNase-defective E. coli exposed to antibiotics were reduced (31 to 60% and 35 to 50%, respectively) in comparison with the control (AB1157), and in the recA13 mutant these values were about 0.4 h with all drugs. Nalidixic acid, under anaerobic conditions, induced no PAE (0 to 0.1 h) in AB1157. A delay in regrowth (0.2 to 0.26 h) was noted with dnaA46(Ts), gyrA43(Ts), and gyrB41(Ts) mutants cultured for 2 h at 43°C. These findings suggest that when proteins and RNA are saved, the cell rapidly resumes the original growth rate.

Persistent suppression of bacterial growth following short antibiotic exposure has been well documented with a variety of microorganisms; this phenomenon is known as postantibiotic effect (PAE) (2, 3,22, 29). The length of the in vitro PAE directly correlated with the antibiotic concentration employed, the exposure time, and the species studied (3, 10, 21).

It is reasonable to assume that delayed regrowth may depend upon the time the cell needs to restore physiological functions after nonlethal damage due either to the antibiotic or to the persistence of a drug at a bacterial binding site (3, 6, 7, 13, 27).

Turnover of bacterial components represents a physiological process in which proteolysis and enzymatic degradation of mRNA play a key role in the regulatory circuits of bacteria (11, 25).

The purpose of this work was to evaluate a PAE in a cytoplasmic background where proteins or mRNA cannot be degraded during antibiotic treatment by virtue of a specific mutation. The role of DNA or protein synthesis inhibition in the absence of antibiotics during the period of time required by bacteria to resume their physiological growth rate was also studied.

(A preliminary account of this work was presented at the 101st General Meeting of the American Society for Microbiology, Orlando, Fla., 2001.)

The bacterial strains were derivatives of Escherichia coli K-12 obtained from Mary K. B. Berlyn (1) of the E. coli Genetic Stock Center and are detailed elsewhere (4, 5). In some experiments, E. coli B, a laboratory strain, together with its proteinase-negative mutant BL21 (31) (purchased from Pharmacia Biotech, Milan, Italy) were also studied. AB1157 was used as a control strain. Microorganisms carrying characteristics relevant for this study belonged to the proteinase-negative PAM161 strain with mutations lon-22 and sulB25 (16). Briefly, any DNA damage activates the SOS DNA repair response mediated by the recA gene product, which induces the expression of several genes including sulB, whose product inhibits formation of septa (32). When DNA is repaired, the latter protein is degraded by the Lon protease (12); in Lon⁻ mutants, septum formation does not resume and this leads to cell death. Mutation in the sulB gene enables Lon-defective bacteria to divide. N464 is an RNase-negative (rna-19 and rnb-464) strain where mRNA degradation is impaired (33). AB2463 (recA13) cannot induce an SOS response (15) and produces nonnucleate cells due to continued protein synthesis, but spontaneous errors introduced during DNA duplication cannot be repaired and chromosome synthesis stops (16). CTR4610 [dna46(Ts)] (13), KNK453 [gyrA43(Ts)] (18), and N4177 [gyrB41(Ts)] (24) cultured at 43°C were defective for functions required for DNA metabolism, but they continued to divide and form nonnucleated cells because the remaining cellular functions were not affected. Strains were cultured in Mueller-Hinton (MH) medium supplemented with thymine (25 mg/liter) when required.

Antimicrobial agents (amikacin, chloramphenicol, imipenem, and nalidixic acid) were obtained from Sigma Chemical Co. (Milan, Italy). MICs were determined using NCCLS methods (28). The MICs of nalidixic acid and amikacin for the recA13 mutant and in the experiments carried out under anaerobic conditions appeared in some cases to vary in an unpredictable way. To circumvent this problem, MIC determinations were repeated simultaneously with the PAE evaluation. Only the experiments where the expected MICs matched those used for PAE determination are reported in Table 1.

TABLE 1.

PAE induced by various antibiotics in E. coli K-12 and in E. coli B and their derivatives carrying the indicated mutations

Strain or phenotypea	PAE range (h)b	MIC (mg/liter)
Chloramphenicol
AB1157	1.7 ± 0.3	4
Lon−	1.0 ± 0.25c	8
RNase−	1.0 ± 0.1c	4
Rec−	0.4 ± 0.1c	4
B	1.6 ± 0.15	2
BL21	0.85 ± 0.1c	2
Nalidixic acid
AB1157	1.6 ± 0.3	4
Lon−	1.1 ± 0.1c	4
RNase−	0.8 ± 0.15c	4
Rec−	0.4 ± 0.15c	2
B	1.4 ± 0.2	4
BL21	0.9 ± 0.3d	4
Imipenem
AB1157	1.0 ± 0.25	0.12
Lon−	0.4 ± 0.1c	0.12
RNase−	0.5 ± 0.1c	0.12
Rec−	0.4 ± 0.1c	0.06
B	1.15 ± 0.2	0.12
BL21	0.50 ± 0.15c	0.12
Amikacin
AB1157	1.5 ± 0.3	4
Lon−	1.0 ± 0.4c	4
RNase−	1.0 ± 0.2c	2
Rec−	0.4 ± 0.1c	2
B	1.3 ± 0.15	2
BL21	0.45 ± 0.3c	2
Nalidixic acid (AN)
AB1157	0.0 ± 0.1	8

^a(AN), experiments carried out under anaerobic conditions.

^bValues are means ± standard deviations of the means for at least five separate experiments.

^cP ≤ 0.01.

^dP = 0.05, using the Student t test to compare experimental results with those obtained with the respective control strain.

The PAE was estimated in accordance with the method of Craig and Gudmundsson (3). Log-phase bacteria were adjusted to 10⁶ to 10⁷ CFU/ml and divided into two portions. The antibiotic (4× MIC) was added to one sample. After 1 h in a shaking water bath (37°C), the drug was inactivated by a 1:1,000 dilution of the cultures in a prewarmed antibiotic-free medium and reincubated. Bacterial counts were determined at time zero, immediately after drug dilution, and at each hour after removal for 6 to 7 h by a pour-plate technique. The PAE was defined as the difference in time required by treated and untreated cultures of the same microorganism to increase by 1 log in CFU number.

Bacterial starvation was obtained by filtration on a 45-μm-pore-size filter of an appropriate concentration of microorganisms. Bacteria were washed with cold minimal salt buffer (MSB) and finally resuspended in the original volume of salt solution. This suspension was then diluted to a final concentration of about 10⁷ CFU/ml in three flasks containing prewarmed (37°C) MSB, MH broth, and MH broth with chloramphenicol at 4× MIC. After 1 h of incubation all samples were diluted 1:1,000 in prewarmed broth. The PAE was determined as described above.

The anaerobic environment was established by culturing the test bacteria for 24 h in broth covered with 2 cm of liquid paraffin. The broth used, antibiotic dilutions for the MIC determination, and flasks employed for anaerobic experiments were incubated in an anaerobic glove box (model 1024; Forma Scientific, Marietta, Ohio) for 24 h. The standard inoculum of bacteria was dispersed in the containers, and then they were covered with 2 cm of liquid paraffin and transferred to an aerobic environment. The PAE was evaluated as described above; bacterial samples were collected using a Hamilton syringe.

When thermosensitive mutants were studied, the test was modified as follows. Bacterial cultures incubated at 32°C were adjusted to the usual cell density and divided into two portions. One sample was incubated at 43°C for 2 h (treated culture), and the other was maintained at 32°C (untreated culture). Both cultures were then diluted 1/100 to facilitate bacterial counts in prewarmed broth (32°C), and the number of viable microorganisms was determined as described above. The difference in time required by treated and untreated cultures of the same microorganism in order to increase by 1 log in CFU number is the delay in regrowth.

First, the period of time of growth suppression was tested by suspending a bacterial culture of AB1157 in an MSB solution. As reported in Fig. 1, no significant difference was found in the period of time of regrowth suppression between the bacteria resuspended in MSB solution and those exposed to chloramphenicol.

FIG. 1.

Comparison between the PAE induced in E. coli AB1157 exposed for 1 h to chloramphenicol (4× MIC) and the delay in regrowth when bacteria were resuspended in MSB for the same period of time. Filled symbols, before dilution; open symbols, after 1/1,000 dilution; ⧫ and ◊, control; ▴ and ▵, chloramphenicol; • and ○, MSB.

The PAE was then determined for a strain which is defective in proteolysis activity. As shown in Table 1, the presence of a defective proteinase function enables cells to make a rapid recovery of the normal growth rate; this was found with all the antibiotics tested irrespective of their mode of action. The differences in the PAE values between proteinase-deficient strains and control strains ranged from 0.5 h (nalidixic acid and amikacin) to 0.7 h (chloramphenicol). The same experiments were repeated using an E. coli B strain and its proteinase-negative derivative (BL21), which is used for DNA cloning purposes (31). The results obtained were similar to those registered with E. coli K-12.

The third step involved the induction of a PAE in a strain lacking RNase activity. The periods of regrowth suppression registered with this strain were shorter than those observed with the control. The difference was about 0.5 h for amikacin and imipenem, 0.7 for chloramphenicol, and 0.8 for nalidixic acid (Table 1).

The role of the recA13 mutation in the duration of the PAE was also tested. The results reported in Table 1 indicate that the persistent suppression of bacterial growth registered with the recA13 mutant was of shorter duration than that obtained with the other mutants.

The role of an anaerobic environment, where 4-quinolones exhibit bacteriostatic activity (19), in the induction of a PAE was then tested. As displayed in Table 1, no PAE (0 to 0.1 h) was induced by nalidixic acid in the organism tested under these experimental conditions.

The following experiments were performed to verify the delay in regrowth when DNA synthesis is arrested by high temperature while protein synthesis is not affected. In a comparison between regrowth of E. coli gyrA43(Ts), gyrB41(Ts), and dna46(Ts) mutants at 43°C and at a permissive temperature, regrowth was delayed by 0.25 ± 0.15 (mean ± standard deviation), 0.26 ± 0.15, and 0.2 ± 0.15 h at 43°C. All of these conditional lethal mutants demonstrated a rapid recovery (0.2 to 0.26 h) of the normal growth rate after 2 h of incubation at the permissive temperature.

The present findings indicate that protein synthesis is one of the factors involved in the establishment of a PAE.

The delay in regrowth obtained with bacteria resuspended in MSB supplies a first indication. This environment, in fact, is reminiscent of that obtained with chloramphenicol-treated bacteria; in the first condition, however, protein synthesis is affected without the use of a specific antimicrobial agent. All bacteria need time to reconstruct their protein pool; this is known as a PAE if it occurs following antibiotic treatment, and it is known as bacterial adaptation if the delay in regrowth is the result of starvation.

An opposite situation was found when determining a PAE for a proteinase-negative strain. In this organism, proteins cannot readily be destroyed, so the bacterial cell maintains its enzymatic pool for a longer period of time than is the case with a proteinase-positive strain (12, 17); therefore, irrespective of the drug employed, the PAE values are lower for the mutant than for the control.

Other interesting data were obtained with the RNase-negative strain, where mRNA is not readily degraded after the accomplishment of its biological role (33); again, the period of regrowth suppression was shortened with this organism. Thus, mRNA appears to be another factor that influences the length of a PAE.

Another indication of the role of protein synthesis in the duration of a PAE comes from experiments carried out under anaerobic conditions using nalidixic acid. In this case, the activity of the quinolone is no longer bactericidal; rather, a bacteriostatic effect is noted (19). This is probably due to the lack of SOS induction under anaerobic conditions (9); thus, cell division is not blocked (8, 20, 32). Bacteria continue to divide, producing nonnucleate cells. DNA synthesis is stopped, while the other physiological activities are not influenced. When the drug is reduced to an inactive level, the cell has all the proteins needed to restart DNA synthesis and other functions.

Finally, the recA13 mutant does not exhibit a marked PAE after exposure to various compounds. The product of the recA13 gene drives a great variety of physiological functions. It is possible that in a RecA-defective strain, the lon gene cannot be fully expressed, showing a Lon⁻ phenotype (8, 15, 16, 20, 23, 32). When the antibiotic is eliminated, the rare survivors have only to recover the drug-damaged intracellular compounds using the complete and unaltered enzymatic pool. This requires a shorter period of time than is required by a wild-type microorganism, which could explain the low values of PAE found with this mutant.

Considering the experiments carried out with thermosensitive mutants, where DNA synthesis but not cellular division is halted (14, 18, 26, 30), the results obtained indicate that under these conditions proteins are saved. When the cultures are transferred at the permissive temperature, the intact enzymatic pool restarts DNA synthesis and other synthesis without a significant time lag.

In conclusion, the results registered here demonstrate for the first time a correlation between the length of the period of growth suppression and protein and mRNA synthesis.