- Email: [email protected]
Potential of macrolide antibiotics to inhibit protein synthesis of Pseudomonas aeruginosa: suppression of virulence factors and stress response
J Infect Chemother (2000) 6:1–7
© Japan Society of Chemotherapy 2000
SPECIAL ARTICLE Kazuhiro Tateda · Yoshikazu Ishii · Tetsuya Matsumoto Takao Kobayashi · Shuichi Miyazaki · Keizo Yamaguchi
Potential of macrolide antibiotics to inhibit protein synthesis of Pseudomonas aeruginosa: suppression of virulence factors and stress response
Received: September 13, 1999 / Accepted: December 7, 1999
Abstract Recently we have reported that sub-minimum inhibitory concentrations (MICs) of macrolide antibiotics, such as erythromycin, clarithromycin, and azithromycin, induce loss of viability of Pseudomonas aeruginosa with longer incubation periods. In the present study we examined the effects of sub-MICs of macrolide antibiotics on protein synthesis and the expression of heat shock proteins (Gro-EL) in P. aeruginosa and the association of these factors with the viability of P. aeruginosa. In seven strains of P. aeruginosa clinical isolates, inhibition of protein synthesis was generally observed in bacteria grown on agar with sub-MIC azithromycin (8 µg/ml) at 24 h, and this was followed by loss of viability after an additional 24-h incubation. The inhibition of protein synthesis was shown in bacteria treated with sub-MICs of erythromycin and clarithromycin, but not with sub-MICs of other antibiotics examined (josamycin, tobramycin, ofloxacin, clindamycin, and ceftazidime) even at relatively high sub-MICs. In the heat shock condition (45°C), strong expression of the heat shock protein Gro-EL was induced in bacteria grown on antibiotic-free medium, whereas there was a delay of such a response in bacteria exposed to 4 µg/ml of azithromycin. Reflecting these results, an abrupt reduction of viability in azithromycin-treated bacteria was observed within 3 h in the heat shock condition. Western blot analysis, using specific antibody for GroEL, demonstrated that erythromycin, clarithromycin, and azithromycin, at concentrations of 0.5–2 µg/ml, inhibited the
This study was partly supported by the 7th Yasushi Ueda Prize, which was awarded to Dr. K. Tateda in 1996. The Yasushi Ueda Prize is awarded annually to researchers who make major contributions to the field of infection and chemotherapy. K. Tateda (*) · Y. Ishii · T. Matsumoto · T. Kobayashi · S. Miyazaki · K. Yamaguchi Department of Microbiology, Toho University School of Medicine, 5-21-16 Ohmori-nishi, Ohta-ku, Tokyo 143-8540, Japan Tel. 181-3-3762-4151 (ext. 2397); Fax 181-3-5493-5415 e-mail: [email protected]
expression of lower-molecular weight Gro-EL bands in the constitutive state. These results indicated that macrolides, at concentrations far below the MICs, suppressed protein synthesis in P. aeruginosa, an effect which may be associated with the inhibition of P. aeruginosa virulence and its loss of viability with longer incubation. Moreover, it is likely that the macrolides may sensitize bacteria to stresses, as these antibiotics induced alterations in a major stress protein, Gro-EL, in constitutive and inducible states. Key words P. aeruginosa · Macrolides · Protein synthesis · Virulence · Stress responses
Introduction Pseudomonas aeruginosa is one of the most important bacterial pathogens in patients with chronic pulmonary diseases, such as cystic fibrosis1,2 and diffuse panbronchiolitis.3 Colonization of the organism in the lungs of these patients ultimately causes death because of respiratory failure or the consequences of respiratory infection, de spite of patients receiving treatment with various potent antibiotics. It is most important to inhibit P. aeruginosa infection or to eradicate this organism from such patients. It is well known that treatment with certain macrolide antibiotics, such as erythromycin, clarithromycin, and azithromycin, alleviates the clinical symptoms and improves the prognosis of patients with chronic pulmonary infection, including that with P. aeruginosa.4,5 As these antibiotics at therapeutic doses are neither bactericidal nor bacteriostatic to P. aeruginosa, it has been speculated that the macrolides may affect the virulence factors of this organism,6,7 or the host defense mechanisms,8–10 or both.11 Recently we have reported that certain macrolide antibiotics, such as erythromycin, clarithromycin, and azithromycin, sensitize bacteria to serum-bactericidal activity.12 These effects were well associated with alterations in bacterial hydrophobicity and cell-surface structures, such as lipopolysaccharide and outer
2
membrane proteins.13 Moreover, these antibiotics induced decreased viability of P. aeruginosa, possibly through the inhibition of protein synthesis, when bacteria were exposed to the antibiotics for prolonged times.14 Although these effects may explain, at least in part, the clinical efficacy of these macrolide antibiotics, the precise mechanisms remain unclear. In the present study, we examined the effects of macrolides and other classes of antibiotics at sub-minimum inhibitory concentrations (MICs) on the viability and protein synthesis of bacteria, and the expression of a heat shock protein (Gro-EL) in constitutive and inducible states.
Materials and methods
lowest concentration of antibiotic that inhibited the visible growth of bacteria.15 Determination of viability of bacteria grown on agar with or without antibiotics After growth on Mueller-Hinton agar in the presence or absence of antibiotics, bacteria were harvested at various intervals and suspended in physiological saline. After being washed by centrifugation, each bacterial suspension was adjusted to an optical density at 660 nm of 0.100. The viable bacterial counts in these suspensions were determined using serially diluted samples spread on Mueller-Hinton agar after overnight incubation at 35°C. Bacterial viability was expressed as the percentage of viable bacteria compared with that of bacteria grown on agar without antibiotics.
Bacterial strains P. aeruginosa PAO-1 was kindly provided by B.H. Iglewski (University of Rochester School of Medicine and Dentistry, Rochester, NY). Mucoid phenotype bacteria (strains CF-3, CF-31, and CF-54), isolated from patients with cystic fibrosis, were kindly provided by P.H. Edelstein (Department of Pathology and Laboratory Medicine and Department of Medicine, University of Pennsylvania School of Medicine). Nonmucoid phenotype (strains 2231, 2232, TMS478, and S6) was isolated from patients with diffuse panbronchiolitis (Toho University hospital, Tokyo, Japan). These strains were stored at 280°C until used. Antibiotics The antibiotics used were kindly provided by the comparies listed here. As 14-membered macrolide antibiotics: we used erythromycin (Dainippon Pharmaceutical, Osaka, Japan), clarithromycin (Taisho Pharmaceutical, Tokyo, Japan), and oleandomycin (Pfizer Laboratories, Groton, CT, USA). As a 16-membered macrolide antibiotic we used: josamycin (Yamanouchi Pharmaceutical, Tokyo, Japan), and the 15membered macrolide antibiotic we used was azithromycin (Pfizer Laboratories). As other classes of antibiotics, we used ceftazidime (Tanabe Pharmaceutical, Osaka, Japan), ofloxacin (Daiichi Pharmaceutical, Tokyo, Japan), minocycline (Lederle Japan, Tokyo, Japan), tobramycin (Shionogi Pharmaceutical, Osaka, Japan), and clindamycin (Japan Upjohn, Tokyo, Japan).
Pulse-labeling experiments for assessing protein synthesis in P. aeruginosa Bacteria were grown on agar with or without antibiotics, and suspended in M9 minimal medium. Pulse-labeling was initiated by adding 10 µCi 35S-methionine to 5 ml of culture, as reported previously.14 Portions (0.3 ml) were taken at intervals (5, 10, 15, 30, and 60 min) and total proteins were precipitated with trichloroacetic acid (final concentration, 5%) by incubation for 15 min on ice. Precipitates were collected (10 000 g for 3 min) and then washed with acetone and finally dissolved in a solution containing 1% sodium dodecyl sulfate (SDS), 50 mM Tris-HCl (pH 8.0), and 1 mM ethylenediaminetetraacetic acid (EDTA). These samples were analyzed electrophoretically using SDS polyacrylamide gel (10%), and this was followed by fluorography. Preparation of antiserum to Gro-EL Antiserum to Gro-EL was prepared in male Japanese white rabbits by the intramuscular injection of purified Escherichia coli Gro-EL emulsified with Freund’s adjuvant. An aliquot of 0.5 ml of antigen was injected once a week for 2 weeks, followed, after 2 weeks’ rest, by a 0.5-ml injection once a week for 2 weeks. Sera were obtained from the animal 1 week after the last injection and were stored at 280°C until used.
Antimicrobial susceptibility test
Immunoblotting procedure
MICs of the antibiotics against P. aeruginosa were determined by the agar dilution method, using serial twofold dilutions of each individual antibiotic in Mueller-Hinton agar (Difco Laboratories, Detroit, MI, USA). On the agar that was to contain the antibiotics, an organism quantity of approximately 105 log-phase was inoculated, and the MICs were determined, after 24 h of incubation at 35°C, as the
After electrophoresis, bacterial proteins were transferred to a nitrocellulose membrane with a semi-dry electroblotting apparatus. The membrane was immunoblotted first with the rabbit serum against Gro-EL, and then with peroxidaseconjugated goat anti-rabbit immunoglobulin, and was developed with a horseradish peroxidase (HRP) kit (Bio-Rad Laboratories, Hercules, CA, USA).
3
Results MICs of macrolides and other classes of antibiotics to P. aeruginosa The MICs of macrolide antibiotics to P. aeruginosa PAO-1 were, respectively, 128, .128, 128, .128, and .128 µg/ml for azithromycin, erythromycin, clarithromycin, josamycin, and oleandomycin. The MICs of other classes of antibiotics to strain PAO-1 were, respectively, 16, 0.5, 32, 0.5, and .128 µg/ml for ceftazidime, tobramycin, minocycline, ofloxacin, and clindamycin. The MICs of azithromycin to the seven strains of clinical isolates of P. aeruginosa were between 32 and .128 µg/ml. Effects of sub-MIC azithromycin on protein synthesis and viability of P. aeruginosa Seven strains of P. aeruginosa were grown on agar with or without 8 µg/ml of azithromycin for 24 h, and then bacterial protein synthesis was evaluated by 35S-methionine pulse labeling method in dot-blotting (Fig. 1). In all control bacteria, protein synthesis was detected within 15 min after addition of the radioisotope. In contrast, delay of protein synthesis was clearly demonstrated in all azithromycintreated cells. In four strains in particular (2231, 2232, CF-3, CF-31), only trace amounts of radioisotope incorporation were observed even after 60-min incubation. These data indicated that bacteria exposed to azithromycin at levels far below the MIC showed inhibitior of protein synthesis although loss of viability of the bacteria was not achieved in this condition. However, additional 24-h incubation on azithromycin-containing agar (total, 48-h incubation) induced a clear decrease in viability in all seven strains examined (Table 1).
Fig. 1. Effects of sub-minimum inhibitory concentration (MIC) azithromycin on protein synthesis of Pseudomonas aeruginosa
Effects of several antibiotics at sub-MICs on protein synthesis of P. aeruginosa PAO-1 We showed that a heat shock protein, Gro-EL, was the major fraction in the protein synthesis of P. aeruginosa PAO-1, by 35S-methionine pulse labeling and immunoprecipitation, using specific antibody for Gro-EL (Fig. 2). Next, we examined the effects of several antibiotics at sub-MICs on Gro-EL expression, as an indicator of protein synthesis of P. aeruginosa PAO-1. Figure 3 shows expression of the Gro-EL band 30 min after addition of the radioisotope in bacteria grown on agar for 24 h with various concentrations of antibiotics. With azithromycin, erythromycin, and clarithromycin, inhibition of protein synthesis was demonstrated at concentrations of 2, 16, and 16 µg/ml, respectively. In contrast, no alterations in protein synthesis were observed with the other antibiotics examined (josamycin, tobramycin, ofloxacin, clindamycin, and ceftazidime), even at relatively high sub-MICs.
Table 1. Effects of sub-MIC azithromycin on viability of P. aeruginosa PAO-1 Strains
2231 2232 TMS-478 CF-3 CF-31 CF-54 S-6
Azithromycin MICs (µg/ml) 128 128 32 .128 .128 128 128
% Viabilitya 24 h
48 h
130.8 62.3 140.2 61.3 130.9 45.8 55.9
4.5 0.4 ,0.1 0.6 0.4 1.6 ,0.1
MIC, Minimum inhibitory concentration a Bacteria were grown on agar with azithromycin (8 µg/ml) for 24 and 48 h, and then bacterial viability was examined
4
Effects of sub-MIC azithromycin on protein synthesis, Gro-EL expression, and viability of bacteria in heat-shock condition We compared protein synthesis in bacteria treated with or without sub-MIC azithromycin (4 µg/ml) at normal temperature (35°C) or under heat-shock conditions (45°C) (Fig. 4). Different patterns of protein synthesis profiles were demonstrated, at normal temperature and under heatshock conditions, particularly in the timing and extent of Gro-EL induction. In the heat-shock condition, strong ex-
pression of Gro-EL was induced from 5 min after addition of the radioisotope in bacteria grown on antibiotic-free medium. A normal response was observed in bacteria grown on agar with azithromycin (4 µg/ml) for 12 h. In contrast, delay of protein synthesis was dramatically demonstrated in bacteria exposed to azithromycin for more than 24 h, although Gro-EL was resistant to such effects of azithromycin under these conditions. Well associated with these results, an abrupt reduction of viability in azithromycin-treated bacteria was observed within 3 h in azithromycin-free medium in the heat-shock condition (Fig. 5). Effects of macrolide antibiotics on expression of Gro-EL in the constitutive state Western blot analysis, using specific polyclonal antibody for Gro-EL, demonstrated that erythromycin and clarithromycin induced inhibition in Gro-EL bands at concentrations of 1–2 µg/ml, whereas no changes were observed in bacteria treated with josamycin and oleandomycin, even at 16 µg/ml (Fig. 6). Azithromycin (0.5 µg/ml) also inhibited the expression of lower-molecular weight Gro-EL bands in the constitutive state (data not shown). These data indicated that erythromycin, clarithromycin, and azithromycin at sub-MICs affected the expression of Gro-EL of P. aeruginosa in the constitutive state, although these conditions were neither bactericidal nor bacteriostatic to the bacteria.
5
10 15 30 60 5 10 15 30 60
Time after addition of radioisotope (min) Fig. 2. Protein synthesis and heat shock protein (GroEL) expression in P. aeruginosa PAO-1
Discussion The present data demonstrated that some macrolide antibiotics, such as erythromycin, clarithromycin, and azithro-
Fig. 3. Effects of several antibiotics at sub-MICs on protein synthesis of P. aeruginosa PAO-1. P. aeruginosa PAO1 was grown on agar with several antibiotics for 24 h, and then Gro-EL expression 30 min after the addition of radioisotope was compared
5 Fig. 4. Effects of sub-MIC azithromycn on protein synthesis of P. aeruginosa PAO-1 in heat-shock condition
Control
Azithromycin 4 µg/ml
Fig. 5. Changes of viability of P. aeruginosa PAO-1 grown on agar with or without sub-MIC azithromycin in heat-shock condition. Bacteria were grown on agar with axithomycin (4 µg/ml) for 24 h, and then changes of viability were examined at 45°C
Control
Azithromycin 4 µg/ml
mycin, specifically suppressed protein synthesis of P. aeruginosa at concentrations far below their MICs, i.e., at concentrations which may be achievable at infection sites in the lungs. Our data also showed that these macrolides sensitized bacteria to heat shock stress, which was well associated with alterations in the expression of a major stress protein, Gro-EL, in the constitutive and inducible states. Macrolide antibiotics bind to the 50S subunit of bacterial ribosomes and increase the dissociation of peptidyl-tRNA from the ribosomes, thus inhibiting protein synthesis.16 As the ribosome target is intracellular and macrolides must cross two membrane barriers in gram-negative bacteria, the mechanism of uptake into bacteria is of central importance to macrolide action on P. aeruginosa. Several investigators have demonstrated that higher incubation temperatures and alkaline conditions tended to increase macrolide uptake and lower the MIC.17,18 We have previously reported that P. aeruginosa accumulated azithromycin during longer incubation periods,14 and this was associated with alterations in bacterial hydrophobicity and cell surface structures, such as outer membrane proteins and lipopolysaccharide.13 Haemophilus influenzae, which are gramnegative bacteria, are known to be more susceptible to macrolides, probably because of the lack of an O-antigenic side chain,19 which would increase the permeability of the outer membrane to hydrophobic agents.20 Further investigations of the association between macrolide-induced alterations in cell surface structure and the uptake of macrolides into bacteria are requited for the better understanding of macrolide efficacy in the clinical setting.
6 Fig. 6. Effects of macrolide antibiotics on expression of GroEL of P. aeruginosa PAO-1. P. aeruginosa PAO-1 was grown on agar with antibiotics for 24 h, and then expression of GroEL was examined by western blot analysis
Erythromycin
Clarithromycin
Oleandomycin
In-vitro determinations of bacterial susceptibility to antimicrobial agents are done by measuring the growthinhibiting concentration of antibiotics in uniformly defined conditions, such as with a bacterial inoculum size of 105–106 CFU, and 18- to 24-h incubation at 35°C. Although the results of these in-vitro tests are usually consistent with clinical responses, they may not be concordant under certain circumstances. An increasing number of reports indicates that subinhibitory levels of antibiotics can affect bacteria in ways other than the expected antimicrobial action.21,22 As the amount of antibiotic at a bacterial infection site during therapy may fall below inhibitory levels, it is tempting to speculate that some of these subinhibitory effects may contribute to the clinical efficacy of certain antibiotics, especially in patients in whom antibiotics are administered for long-term periods. Although it is generally accepted that P. aeruginosa is completely resistant to macrolide antibiotics, the present data clearly showed inhibition of protein synthesis in P. aeruginosa when bacteria were exposed to sub-MICs of erythromycin, clarithromycin, and azithromycin. Moreover, loss of viability was shown with 48-h, but not with 24-h incubation. These data indicate that the inhibition of protein synthesis by macrolides is subtle with the usual incubation time, but may be lethal with longer exposure. On the other hand, josamycin, a 16-membered macrolide, did not inhibit protein synthesis of P. aeruginosa PAO-1, even at 32 µg/ml. Erythromycin and clarithromycin are 14-membered, and share similar structures, except for a difference in one side chain residue of the macrolide aglycone ring. In contrast, azithromycin is a 15-membered macrolide, and it differs structually from erythromycin by the presence of a nitrogen at position 9a in the macrolide ring. These results suggest that the action of azithromycin, erythromycin, and clarithromycin on P. aeruginosa may be related to fine structures, such as the substitutions on the lactone ring and/or the sugar composition of these antibiotics, in addition to their simple classification as being 14-, 15- or 16-membered.
In the other classes of antibiotics tested, no effects were observed on P. aeruginosa PAO-1 with tobramycin, ofloxacin, ceftazidime, and clindamycin. Tobramycin and clindamycin are protein synthesis inhibitors, as are the macrolide antibiotics, although the anti-bacterial mechanisms in these two antibiotics and macrolides are slightly different. It is possible that the protein synthesis-inhibitory activity that occurs with longer incubation may be specific for some protein synthesis inhibitors, such as azithromycin, erythromycin, and clarithromycin, with varying degrees of activity. It has been reported that macrolides inhibit the production of several exoproducts, such as exoenzymes,7,11,23,24 exopolysaccharide,4,6 and pigment.24 Our results, of exposure-dependent protein synthesis inhibition, support these reports, because protein synthesis systems may be involved in several steps in the expression of these factors. For example, the stress protein GroEL, which is known to play a critical role in the assembly, transport, and secretion of several exoproducts,25,26 as well as a wide range of other proteins, was strongly affected by the macrolides erythromycin, clarithromycin, and azithromycin. The present data clearly demonstrated that macrolide antibiotics affected Gro-EL expression in the constitutive and inducible states, and that this effect was well associated with loss of viability in the heat shock condition. Macrolide effects on stress responses in bacteria may be a promising subject for study, as infecting bacteria in the body are undoubtedly subjected to various stresses, such as phagocytosis by leukocytes and complement-mediated killing. In conclusion, we have reported specific protein synthesis-inhibitory activity in macrolides at sub-MICs. The present data may explain, at least in part, the clinical efficacy of macrolide antibiotics against persistent P. aeruginosa infection in lungs, in conditions in which bacteria are continuously exposed to these antibiotics, although the concentration is far below MIC levels. Acknowledgments We thank Shogo Kuwahara for critical reading of the manuscript and helpful suggestions.
7
References 14. 1. Neu HC. The role of Pseudomonas aeruginosa in infections. J Antimicrob Chemother 1983;11(Suppl B):1–13. 2. Wood RE, Boat TF, Doershuk CF. Cystic fibrosis. Am Rev Respir Dis 1976;113:833–77. 3. Homma H, Yamanaka A, Tanimoto S, Tamura M, Chijimatsu Y, Kira S, et al. Diffuse panbronchiolitis: a disease of the transitional zone of the lung. Chest 1983;83:63–9. 4. Kobayashi H. Airway biofilm disease: clinical manifestations and therapeutic possibilities using macrolides. J Infect Chemother 1995;1:1–15. 5. Kudoh S, Uetake K, Hagiwara K, Hirayama M, Hus LH, Kimura H, et al. Clinical effect of low-dose long-term erythromycin chemotherapy on diffuse panbronchiolitis (in Japanese with English abstract). Jpn J Thorac Dis 1987;25:632–42. 6. Ichimiya T, Yamazaki T, Nasu M. In-vitro effects of antimicrobial agents on Pseudomonas aeruginosa biofilm formation. J Antimicrob Chemother 1994;34:331–41. 7. Kita E, Sawaki M, Oku D, Hamuro A, Mikasa K, Konishi M, et al. Suppression of virulence factors of Pseudomonas aeruginosa by erythromycin. J Antimicrob Chemother 1991;27:273– 84. 8. Fernandes AC, Anderson R, Theron AJ, Joone G, Van Rensburg CEJ. Enhancement of human polymorphonuclear leukocyte motility by erythromycin in vitro and in vivo. S Afr Med J 1984;66: 173–7. 9. Fraschini F, Scaglione F, Ferrara F, Marelli O, Braga PC, Teodori F. Evaluation of the immunostimulating activity of erythromycin in man. Chemotherapy 1986;32:286–90. 10. Kita E, Sawaki M, Nishikawa F, Mikasa K, Yagyu Y, Takeuchi S, et al. Enhanced interleukin production after long-term administration of erythromycin stearate. Pharmacology 1990;41:177– 83. 11. Hirakata Y, Kaku M, Mizukane R, Ishida K, Furuya N, Matsumoto T, et al. Potential effects of erythromycin on host defense systems and virulence of Pseudomonas aeruginosa. Antimicrob. Agents Chemother 1992;36:1922–7. 12. Tateda K, Hirakata Y, Furuya N, Ohno A, Yamaguchi K. Effects of sub-MICs of erythromycin and other macrolide antibiotics on serum sensitivity of Pseudomonas aeruginosa. Antimicrob. Agents Chemother 1993;37:675–80. 13. Tateda K, Ishii Y, Hirakata Y, Matsumoto T, Ohno A, Yamaguchi K. Profiles of outer membrane proteins and lipopolysaccharide of Pseudomonas aeruginosa grown in the presence of sub-MICs of
15.
16.
17.
18.
19. 20. 21. 22.
23.
24.
25. 26.
macrolide antibiotics and their relation to enhanced serum sensitivity. J Antimicrob Chemother 1994;34:931–42. Tateda K, Ishii Y, Matsumoto T, Furuya N, Nagashima M, Matsunaga T, et al. Direct evidence for antipseudomonal activity of macrolides: exposure-dependent bactericidal activity and inhibition of protein synthesis by erythromycin, clarithromycin, and azithromycin. Antimicrob Agents Chemother 1996;40:2271–5. National Committee for Clinical Laboratory Standards. Methods for dilution antimicrobial susceptibility tests for bacteria that grow aerobically. Approved standard M7-A2. Villanova, PA.: National Committee for Clinical Laboratory Standards; 1990. Menninger JR, Otto DP. Erythromycin, carbomycin, and spiramycin inhibit protein synthesis by stimulating the dissociation of peptidyl-tRNA from ribosomes. Antimicrob. Agents Chemother 1982;21:811–8. Dette GA, Knothe H, Kaula S. Modifying effects of pH and temperature on (14°C) erythromycin uptake into Staphylococcus aureus – relation to antimicrobial activity. Zentralbl Bakteriol Hyg A 1987;265:393–403. Sabath LD, Lorian V, Gerstein D, Loder PB, Finland M. Enhancing effect of alkalinization of the medium on the activity of erythromycin against gram-negative bacteria. Appl Microbiol 1968;16: 1288–92. Inzana TJ. Electrophoretic heterogeneity and interstrain variation of the lipopolysaccharide of Haemophilus influenzae. J Infect Dis 1983;148:492–9. Nikaido H, Vaara M. Molecular basis of bacterial outer membrane permeability. Microbiol Rev 1985;49:1–32. Darveau RP, Cunningham MD. Influence of subinhibitory concentrations of cephalosporins on the serum sensitivity of Pseudomonas aeruginosa. J Infect Dis 1990;162:914–21. Veringa E, Box A, Rozenberg-Arska M, Verhoef J. Monobactam antibiotics in subinhibitory concentrations enhance opsonophagocytosis and serum bacteriolysis in certain Escherichia coli strains. Drugs Exp Clin Res 1998;14:1–8. Mizukane R, Hirakata Y, Kaku M, Ishii Y, Furuya N, Ishida K, et al. Comparative in vitro exoenzyme-suppressing activities of azithromycin and other macrolide antibiotics against Pseudomonas aeruginosa. Antimicrob. Agents Chemother 1994;38:528–33. Molinari G, Guzman CA, Pesce A, Schito GC. Inhibition of Pseudomonas aeruginosa virulence factors by subinhibitory concentrations of azithromycin and other macrolide antibiotics. J Antimicrob Chemother 1993;31:681–8. Georgopoulos C, Welch WJ. Role of the major heat shock proteins as molecular chaperones. Ann Rev Cell Biol 1993;9:601–34. Hendrick JP, Hartl FU. Molecular chaperone functions of heat-shock proteins. Ann Rev Biochem 1993;62:349–84.
© Japan Society of Chemotherapy 2000
SPECIAL ARTICLE Kazuhiro Tateda · Yoshikazu Ishii · Tetsuya Matsumoto Takao Kobayashi · Shuichi Miyazaki · Keizo Yamaguchi
Potential of macrolide antibiotics to inhibit protein synthesis of Pseudomonas aeruginosa: suppression of virulence factors and stress response
Received: September 13, 1999 / Accepted: December 7, 1999
Abstract Recently we have reported that sub-minimum inhibitory concentrations (MICs) of macrolide antibiotics, such as erythromycin, clarithromycin, and azithromycin, induce loss of viability of Pseudomonas aeruginosa with longer incubation periods. In the present study we examined the effects of sub-MICs of macrolide antibiotics on protein synthesis and the expression of heat shock proteins (Gro-EL) in P. aeruginosa and the association of these factors with the viability of P. aeruginosa. In seven strains of P. aeruginosa clinical isolates, inhibition of protein synthesis was generally observed in bacteria grown on agar with sub-MIC azithromycin (8 µg/ml) at 24 h, and this was followed by loss of viability after an additional 24-h incubation. The inhibition of protein synthesis was shown in bacteria treated with sub-MICs of erythromycin and clarithromycin, but not with sub-MICs of other antibiotics examined (josamycin, tobramycin, ofloxacin, clindamycin, and ceftazidime) even at relatively high sub-MICs. In the heat shock condition (45°C), strong expression of the heat shock protein Gro-EL was induced in bacteria grown on antibiotic-free medium, whereas there was a delay of such a response in bacteria exposed to 4 µg/ml of azithromycin. Reflecting these results, an abrupt reduction of viability in azithromycin-treated bacteria was observed within 3 h in the heat shock condition. Western blot analysis, using specific antibody for GroEL, demonstrated that erythromycin, clarithromycin, and azithromycin, at concentrations of 0.5–2 µg/ml, inhibited the
This study was partly supported by the 7th Yasushi Ueda Prize, which was awarded to Dr. K. Tateda in 1996. The Yasushi Ueda Prize is awarded annually to researchers who make major contributions to the field of infection and chemotherapy. K. Tateda (*) · Y. Ishii · T. Matsumoto · T. Kobayashi · S. Miyazaki · K. Yamaguchi Department of Microbiology, Toho University School of Medicine, 5-21-16 Ohmori-nishi, Ohta-ku, Tokyo 143-8540, Japan Tel. 181-3-3762-4151 (ext. 2397); Fax 181-3-5493-5415 e-mail: [email protected]
expression of lower-molecular weight Gro-EL bands in the constitutive state. These results indicated that macrolides, at concentrations far below the MICs, suppressed protein synthesis in P. aeruginosa, an effect which may be associated with the inhibition of P. aeruginosa virulence and its loss of viability with longer incubation. Moreover, it is likely that the macrolides may sensitize bacteria to stresses, as these antibiotics induced alterations in a major stress protein, Gro-EL, in constitutive and inducible states. Key words P. aeruginosa · Macrolides · Protein synthesis · Virulence · Stress responses
Introduction Pseudomonas aeruginosa is one of the most important bacterial pathogens in patients with chronic pulmonary diseases, such as cystic fibrosis1,2 and diffuse panbronchiolitis.3 Colonization of the organism in the lungs of these patients ultimately causes death because of respiratory failure or the consequences of respiratory infection, de spite of patients receiving treatment with various potent antibiotics. It is most important to inhibit P. aeruginosa infection or to eradicate this organism from such patients. It is well known that treatment with certain macrolide antibiotics, such as erythromycin, clarithromycin, and azithromycin, alleviates the clinical symptoms and improves the prognosis of patients with chronic pulmonary infection, including that with P. aeruginosa.4,5 As these antibiotics at therapeutic doses are neither bactericidal nor bacteriostatic to P. aeruginosa, it has been speculated that the macrolides may affect the virulence factors of this organism,6,7 or the host defense mechanisms,8–10 or both.11 Recently we have reported that certain macrolide antibiotics, such as erythromycin, clarithromycin, and azithromycin, sensitize bacteria to serum-bactericidal activity.12 These effects were well associated with alterations in bacterial hydrophobicity and cell-surface structures, such as lipopolysaccharide and outer
2
membrane proteins.13 Moreover, these antibiotics induced decreased viability of P. aeruginosa, possibly through the inhibition of protein synthesis, when bacteria were exposed to the antibiotics for prolonged times.14 Although these effects may explain, at least in part, the clinical efficacy of these macrolide antibiotics, the precise mechanisms remain unclear. In the present study, we examined the effects of macrolides and other classes of antibiotics at sub-minimum inhibitory concentrations (MICs) on the viability and protein synthesis of bacteria, and the expression of a heat shock protein (Gro-EL) in constitutive and inducible states.
Materials and methods
lowest concentration of antibiotic that inhibited the visible growth of bacteria.15 Determination of viability of bacteria grown on agar with or without antibiotics After growth on Mueller-Hinton agar in the presence or absence of antibiotics, bacteria were harvested at various intervals and suspended in physiological saline. After being washed by centrifugation, each bacterial suspension was adjusted to an optical density at 660 nm of 0.100. The viable bacterial counts in these suspensions were determined using serially diluted samples spread on Mueller-Hinton agar after overnight incubation at 35°C. Bacterial viability was expressed as the percentage of viable bacteria compared with that of bacteria grown on agar without antibiotics.
Bacterial strains P. aeruginosa PAO-1 was kindly provided by B.H. Iglewski (University of Rochester School of Medicine and Dentistry, Rochester, NY). Mucoid phenotype bacteria (strains CF-3, CF-31, and CF-54), isolated from patients with cystic fibrosis, were kindly provided by P.H. Edelstein (Department of Pathology and Laboratory Medicine and Department of Medicine, University of Pennsylvania School of Medicine). Nonmucoid phenotype (strains 2231, 2232, TMS478, and S6) was isolated from patients with diffuse panbronchiolitis (Toho University hospital, Tokyo, Japan). These strains were stored at 280°C until used. Antibiotics The antibiotics used were kindly provided by the comparies listed here. As 14-membered macrolide antibiotics: we used erythromycin (Dainippon Pharmaceutical, Osaka, Japan), clarithromycin (Taisho Pharmaceutical, Tokyo, Japan), and oleandomycin (Pfizer Laboratories, Groton, CT, USA). As a 16-membered macrolide antibiotic we used: josamycin (Yamanouchi Pharmaceutical, Tokyo, Japan), and the 15membered macrolide antibiotic we used was azithromycin (Pfizer Laboratories). As other classes of antibiotics, we used ceftazidime (Tanabe Pharmaceutical, Osaka, Japan), ofloxacin (Daiichi Pharmaceutical, Tokyo, Japan), minocycline (Lederle Japan, Tokyo, Japan), tobramycin (Shionogi Pharmaceutical, Osaka, Japan), and clindamycin (Japan Upjohn, Tokyo, Japan).
Pulse-labeling experiments for assessing protein synthesis in P. aeruginosa Bacteria were grown on agar with or without antibiotics, and suspended in M9 minimal medium. Pulse-labeling was initiated by adding 10 µCi 35S-methionine to 5 ml of culture, as reported previously.14 Portions (0.3 ml) were taken at intervals (5, 10, 15, 30, and 60 min) and total proteins were precipitated with trichloroacetic acid (final concentration, 5%) by incubation for 15 min on ice. Precipitates were collected (10 000 g for 3 min) and then washed with acetone and finally dissolved in a solution containing 1% sodium dodecyl sulfate (SDS), 50 mM Tris-HCl (pH 8.0), and 1 mM ethylenediaminetetraacetic acid (EDTA). These samples were analyzed electrophoretically using SDS polyacrylamide gel (10%), and this was followed by fluorography. Preparation of antiserum to Gro-EL Antiserum to Gro-EL was prepared in male Japanese white rabbits by the intramuscular injection of purified Escherichia coli Gro-EL emulsified with Freund’s adjuvant. An aliquot of 0.5 ml of antigen was injected once a week for 2 weeks, followed, after 2 weeks’ rest, by a 0.5-ml injection once a week for 2 weeks. Sera were obtained from the animal 1 week after the last injection and were stored at 280°C until used.
Antimicrobial susceptibility test
Immunoblotting procedure
MICs of the antibiotics against P. aeruginosa were determined by the agar dilution method, using serial twofold dilutions of each individual antibiotic in Mueller-Hinton agar (Difco Laboratories, Detroit, MI, USA). On the agar that was to contain the antibiotics, an organism quantity of approximately 105 log-phase was inoculated, and the MICs were determined, after 24 h of incubation at 35°C, as the
After electrophoresis, bacterial proteins were transferred to a nitrocellulose membrane with a semi-dry electroblotting apparatus. The membrane was immunoblotted first with the rabbit serum against Gro-EL, and then with peroxidaseconjugated goat anti-rabbit immunoglobulin, and was developed with a horseradish peroxidase (HRP) kit (Bio-Rad Laboratories, Hercules, CA, USA).
3
Results MICs of macrolides and other classes of antibiotics to P. aeruginosa The MICs of macrolide antibiotics to P. aeruginosa PAO-1 were, respectively, 128, .128, 128, .128, and .128 µg/ml for azithromycin, erythromycin, clarithromycin, josamycin, and oleandomycin. The MICs of other classes of antibiotics to strain PAO-1 were, respectively, 16, 0.5, 32, 0.5, and .128 µg/ml for ceftazidime, tobramycin, minocycline, ofloxacin, and clindamycin. The MICs of azithromycin to the seven strains of clinical isolates of P. aeruginosa were between 32 and .128 µg/ml. Effects of sub-MIC azithromycin on protein synthesis and viability of P. aeruginosa Seven strains of P. aeruginosa were grown on agar with or without 8 µg/ml of azithromycin for 24 h, and then bacterial protein synthesis was evaluated by 35S-methionine pulse labeling method in dot-blotting (Fig. 1). In all control bacteria, protein synthesis was detected within 15 min after addition of the radioisotope. In contrast, delay of protein synthesis was clearly demonstrated in all azithromycintreated cells. In four strains in particular (2231, 2232, CF-3, CF-31), only trace amounts of radioisotope incorporation were observed even after 60-min incubation. These data indicated that bacteria exposed to azithromycin at levels far below the MIC showed inhibitior of protein synthesis although loss of viability of the bacteria was not achieved in this condition. However, additional 24-h incubation on azithromycin-containing agar (total, 48-h incubation) induced a clear decrease in viability in all seven strains examined (Table 1).
Fig. 1. Effects of sub-minimum inhibitory concentration (MIC) azithromycin on protein synthesis of Pseudomonas aeruginosa
Effects of several antibiotics at sub-MICs on protein synthesis of P. aeruginosa PAO-1 We showed that a heat shock protein, Gro-EL, was the major fraction in the protein synthesis of P. aeruginosa PAO-1, by 35S-methionine pulse labeling and immunoprecipitation, using specific antibody for Gro-EL (Fig. 2). Next, we examined the effects of several antibiotics at sub-MICs on Gro-EL expression, as an indicator of protein synthesis of P. aeruginosa PAO-1. Figure 3 shows expression of the Gro-EL band 30 min after addition of the radioisotope in bacteria grown on agar for 24 h with various concentrations of antibiotics. With azithromycin, erythromycin, and clarithromycin, inhibition of protein synthesis was demonstrated at concentrations of 2, 16, and 16 µg/ml, respectively. In contrast, no alterations in protein synthesis were observed with the other antibiotics examined (josamycin, tobramycin, ofloxacin, clindamycin, and ceftazidime), even at relatively high sub-MICs.
Table 1. Effects of sub-MIC azithromycin on viability of P. aeruginosa PAO-1 Strains
2231 2232 TMS-478 CF-3 CF-31 CF-54 S-6
Azithromycin MICs (µg/ml) 128 128 32 .128 .128 128 128
% Viabilitya 24 h
48 h
130.8 62.3 140.2 61.3 130.9 45.8 55.9
4.5 0.4 ,0.1 0.6 0.4 1.6 ,0.1
MIC, Minimum inhibitory concentration a Bacteria were grown on agar with azithromycin (8 µg/ml) for 24 and 48 h, and then bacterial viability was examined
4
Effects of sub-MIC azithromycin on protein synthesis, Gro-EL expression, and viability of bacteria in heat-shock condition We compared protein synthesis in bacteria treated with or without sub-MIC azithromycin (4 µg/ml) at normal temperature (35°C) or under heat-shock conditions (45°C) (Fig. 4). Different patterns of protein synthesis profiles were demonstrated, at normal temperature and under heatshock conditions, particularly in the timing and extent of Gro-EL induction. In the heat-shock condition, strong ex-
pression of Gro-EL was induced from 5 min after addition of the radioisotope in bacteria grown on antibiotic-free medium. A normal response was observed in bacteria grown on agar with azithromycin (4 µg/ml) for 12 h. In contrast, delay of protein synthesis was dramatically demonstrated in bacteria exposed to azithromycin for more than 24 h, although Gro-EL was resistant to such effects of azithromycin under these conditions. Well associated with these results, an abrupt reduction of viability in azithromycin-treated bacteria was observed within 3 h in azithromycin-free medium in the heat-shock condition (Fig. 5). Effects of macrolide antibiotics on expression of Gro-EL in the constitutive state Western blot analysis, using specific polyclonal antibody for Gro-EL, demonstrated that erythromycin and clarithromycin induced inhibition in Gro-EL bands at concentrations of 1–2 µg/ml, whereas no changes were observed in bacteria treated with josamycin and oleandomycin, even at 16 µg/ml (Fig. 6). Azithromycin (0.5 µg/ml) also inhibited the expression of lower-molecular weight Gro-EL bands in the constitutive state (data not shown). These data indicated that erythromycin, clarithromycin, and azithromycin at sub-MICs affected the expression of Gro-EL of P. aeruginosa in the constitutive state, although these conditions were neither bactericidal nor bacteriostatic to the bacteria.
5
10 15 30 60 5 10 15 30 60
Time after addition of radioisotope (min) Fig. 2. Protein synthesis and heat shock protein (GroEL) expression in P. aeruginosa PAO-1
Discussion The present data demonstrated that some macrolide antibiotics, such as erythromycin, clarithromycin, and azithro-
Fig. 3. Effects of several antibiotics at sub-MICs on protein synthesis of P. aeruginosa PAO-1. P. aeruginosa PAO1 was grown on agar with several antibiotics for 24 h, and then Gro-EL expression 30 min after the addition of radioisotope was compared
5 Fig. 4. Effects of sub-MIC azithromycn on protein synthesis of P. aeruginosa PAO-1 in heat-shock condition
Control
Azithromycin 4 µg/ml
Fig. 5. Changes of viability of P. aeruginosa PAO-1 grown on agar with or without sub-MIC azithromycin in heat-shock condition. Bacteria were grown on agar with axithomycin (4 µg/ml) for 24 h, and then changes of viability were examined at 45°C
Control
Azithromycin 4 µg/ml
mycin, specifically suppressed protein synthesis of P. aeruginosa at concentrations far below their MICs, i.e., at concentrations which may be achievable at infection sites in the lungs. Our data also showed that these macrolides sensitized bacteria to heat shock stress, which was well associated with alterations in the expression of a major stress protein, Gro-EL, in the constitutive and inducible states. Macrolide antibiotics bind to the 50S subunit of bacterial ribosomes and increase the dissociation of peptidyl-tRNA from the ribosomes, thus inhibiting protein synthesis.16 As the ribosome target is intracellular and macrolides must cross two membrane barriers in gram-negative bacteria, the mechanism of uptake into bacteria is of central importance to macrolide action on P. aeruginosa. Several investigators have demonstrated that higher incubation temperatures and alkaline conditions tended to increase macrolide uptake and lower the MIC.17,18 We have previously reported that P. aeruginosa accumulated azithromycin during longer incubation periods,14 and this was associated with alterations in bacterial hydrophobicity and cell surface structures, such as outer membrane proteins and lipopolysaccharide.13 Haemophilus influenzae, which are gramnegative bacteria, are known to be more susceptible to macrolides, probably because of the lack of an O-antigenic side chain,19 which would increase the permeability of the outer membrane to hydrophobic agents.20 Further investigations of the association between macrolide-induced alterations in cell surface structure and the uptake of macrolides into bacteria are requited for the better understanding of macrolide efficacy in the clinical setting.
6 Fig. 6. Effects of macrolide antibiotics on expression of GroEL of P. aeruginosa PAO-1. P. aeruginosa PAO-1 was grown on agar with antibiotics for 24 h, and then expression of GroEL was examined by western blot analysis
Erythromycin
Clarithromycin
Oleandomycin
In-vitro determinations of bacterial susceptibility to antimicrobial agents are done by measuring the growthinhibiting concentration of antibiotics in uniformly defined conditions, such as with a bacterial inoculum size of 105–106 CFU, and 18- to 24-h incubation at 35°C. Although the results of these in-vitro tests are usually consistent with clinical responses, they may not be concordant under certain circumstances. An increasing number of reports indicates that subinhibitory levels of antibiotics can affect bacteria in ways other than the expected antimicrobial action.21,22 As the amount of antibiotic at a bacterial infection site during therapy may fall below inhibitory levels, it is tempting to speculate that some of these subinhibitory effects may contribute to the clinical efficacy of certain antibiotics, especially in patients in whom antibiotics are administered for long-term periods. Although it is generally accepted that P. aeruginosa is completely resistant to macrolide antibiotics, the present data clearly showed inhibition of protein synthesis in P. aeruginosa when bacteria were exposed to sub-MICs of erythromycin, clarithromycin, and azithromycin. Moreover, loss of viability was shown with 48-h, but not with 24-h incubation. These data indicate that the inhibition of protein synthesis by macrolides is subtle with the usual incubation time, but may be lethal with longer exposure. On the other hand, josamycin, a 16-membered macrolide, did not inhibit protein synthesis of P. aeruginosa PAO-1, even at 32 µg/ml. Erythromycin and clarithromycin are 14-membered, and share similar structures, except for a difference in one side chain residue of the macrolide aglycone ring. In contrast, azithromycin is a 15-membered macrolide, and it differs structually from erythromycin by the presence of a nitrogen at position 9a in the macrolide ring. These results suggest that the action of azithromycin, erythromycin, and clarithromycin on P. aeruginosa may be related to fine structures, such as the substitutions on the lactone ring and/or the sugar composition of these antibiotics, in addition to their simple classification as being 14-, 15- or 16-membered.
In the other classes of antibiotics tested, no effects were observed on P. aeruginosa PAO-1 with tobramycin, ofloxacin, ceftazidime, and clindamycin. Tobramycin and clindamycin are protein synthesis inhibitors, as are the macrolide antibiotics, although the anti-bacterial mechanisms in these two antibiotics and macrolides are slightly different. It is possible that the protein synthesis-inhibitory activity that occurs with longer incubation may be specific for some protein synthesis inhibitors, such as azithromycin, erythromycin, and clarithromycin, with varying degrees of activity. It has been reported that macrolides inhibit the production of several exoproducts, such as exoenzymes,7,11,23,24 exopolysaccharide,4,6 and pigment.24 Our results, of exposure-dependent protein synthesis inhibition, support these reports, because protein synthesis systems may be involved in several steps in the expression of these factors. For example, the stress protein GroEL, which is known to play a critical role in the assembly, transport, and secretion of several exoproducts,25,26 as well as a wide range of other proteins, was strongly affected by the macrolides erythromycin, clarithromycin, and azithromycin. The present data clearly demonstrated that macrolide antibiotics affected Gro-EL expression in the constitutive and inducible states, and that this effect was well associated with loss of viability in the heat shock condition. Macrolide effects on stress responses in bacteria may be a promising subject for study, as infecting bacteria in the body are undoubtedly subjected to various stresses, such as phagocytosis by leukocytes and complement-mediated killing. In conclusion, we have reported specific protein synthesis-inhibitory activity in macrolides at sub-MICs. The present data may explain, at least in part, the clinical efficacy of macrolide antibiotics against persistent P. aeruginosa infection in lungs, in conditions in which bacteria are continuously exposed to these antibiotics, although the concentration is far below MIC levels. Acknowledgments We thank Shogo Kuwahara for critical reading of the manuscript and helpful suggestions.
7
References 14. 1. Neu HC. The role of Pseudomonas aeruginosa in infections. J Antimicrob Chemother 1983;11(Suppl B):1–13. 2. Wood RE, Boat TF, Doershuk CF. Cystic fibrosis. Am Rev Respir Dis 1976;113:833–77. 3. Homma H, Yamanaka A, Tanimoto S, Tamura M, Chijimatsu Y, Kira S, et al. Diffuse panbronchiolitis: a disease of the transitional zone of the lung. Chest 1983;83:63–9. 4. Kobayashi H. Airway biofilm disease: clinical manifestations and therapeutic possibilities using macrolides. J Infect Chemother 1995;1:1–15. 5. Kudoh S, Uetake K, Hagiwara K, Hirayama M, Hus LH, Kimura H, et al. Clinical effect of low-dose long-term erythromycin chemotherapy on diffuse panbronchiolitis (in Japanese with English abstract). Jpn J Thorac Dis 1987;25:632–42. 6. Ichimiya T, Yamazaki T, Nasu M. In-vitro effects of antimicrobial agents on Pseudomonas aeruginosa biofilm formation. J Antimicrob Chemother 1994;34:331–41. 7. Kita E, Sawaki M, Oku D, Hamuro A, Mikasa K, Konishi M, et al. Suppression of virulence factors of Pseudomonas aeruginosa by erythromycin. J Antimicrob Chemother 1991;27:273– 84. 8. Fernandes AC, Anderson R, Theron AJ, Joone G, Van Rensburg CEJ. Enhancement of human polymorphonuclear leukocyte motility by erythromycin in vitro and in vivo. S Afr Med J 1984;66: 173–7. 9. Fraschini F, Scaglione F, Ferrara F, Marelli O, Braga PC, Teodori F. Evaluation of the immunostimulating activity of erythromycin in man. Chemotherapy 1986;32:286–90. 10. Kita E, Sawaki M, Nishikawa F, Mikasa K, Yagyu Y, Takeuchi S, et al. Enhanced interleukin production after long-term administration of erythromycin stearate. Pharmacology 1990;41:177– 83. 11. Hirakata Y, Kaku M, Mizukane R, Ishida K, Furuya N, Matsumoto T, et al. Potential effects of erythromycin on host defense systems and virulence of Pseudomonas aeruginosa. Antimicrob. Agents Chemother 1992;36:1922–7. 12. Tateda K, Hirakata Y, Furuya N, Ohno A, Yamaguchi K. Effects of sub-MICs of erythromycin and other macrolide antibiotics on serum sensitivity of Pseudomonas aeruginosa. Antimicrob. Agents Chemother 1993;37:675–80. 13. Tateda K, Ishii Y, Hirakata Y, Matsumoto T, Ohno A, Yamaguchi K. Profiles of outer membrane proteins and lipopolysaccharide of Pseudomonas aeruginosa grown in the presence of sub-MICs of
15.
16.
17.
18.
19. 20. 21. 22.
23.
24.
25. 26.
macrolide antibiotics and their relation to enhanced serum sensitivity. J Antimicrob Chemother 1994;34:931–42. Tateda K, Ishii Y, Matsumoto T, Furuya N, Nagashima M, Matsunaga T, et al. Direct evidence for antipseudomonal activity of macrolides: exposure-dependent bactericidal activity and inhibition of protein synthesis by erythromycin, clarithromycin, and azithromycin. Antimicrob Agents Chemother 1996;40:2271–5. National Committee for Clinical Laboratory Standards. Methods for dilution antimicrobial susceptibility tests for bacteria that grow aerobically. Approved standard M7-A2. Villanova, PA.: National Committee for Clinical Laboratory Standards; 1990. Menninger JR, Otto DP. Erythromycin, carbomycin, and spiramycin inhibit protein synthesis by stimulating the dissociation of peptidyl-tRNA from ribosomes. Antimicrob. Agents Chemother 1982;21:811–8. Dette GA, Knothe H, Kaula S. Modifying effects of pH and temperature on (14°C) erythromycin uptake into Staphylococcus aureus – relation to antimicrobial activity. Zentralbl Bakteriol Hyg A 1987;265:393–403. Sabath LD, Lorian V, Gerstein D, Loder PB, Finland M. Enhancing effect of alkalinization of the medium on the activity of erythromycin against gram-negative bacteria. Appl Microbiol 1968;16: 1288–92. Inzana TJ. Electrophoretic heterogeneity and interstrain variation of the lipopolysaccharide of Haemophilus influenzae. J Infect Dis 1983;148:492–9. Nikaido H, Vaara M. Molecular basis of bacterial outer membrane permeability. Microbiol Rev 1985;49:1–32. Darveau RP, Cunningham MD. Influence of subinhibitory concentrations of cephalosporins on the serum sensitivity of Pseudomonas aeruginosa. J Infect Dis 1990;162:914–21. Veringa E, Box A, Rozenberg-Arska M, Verhoef J. Monobactam antibiotics in subinhibitory concentrations enhance opsonophagocytosis and serum bacteriolysis in certain Escherichia coli strains. Drugs Exp Clin Res 1998;14:1–8. Mizukane R, Hirakata Y, Kaku M, Ishii Y, Furuya N, Ishida K, et al. Comparative in vitro exoenzyme-suppressing activities of azithromycin and other macrolide antibiotics against Pseudomonas aeruginosa. Antimicrob. Agents Chemother 1994;38:528–33. Molinari G, Guzman CA, Pesce A, Schito GC. Inhibition of Pseudomonas aeruginosa virulence factors by subinhibitory concentrations of azithromycin and other macrolide antibiotics. J Antimicrob Chemother 1993;31:681–8. Georgopoulos C, Welch WJ. Role of the major heat shock proteins as molecular chaperones. Ann Rev Cell Biol 1993;9:601–34. Hendrick JP, Hartl FU. Molecular chaperone functions of heat-shock proteins. Ann Rev Biochem 1993;62:349–84.