ph modulation of aminoglycoside resistance in Staphylococcus epidermidis harbouring 6'-/V-aminoglycoside acetyltransferase

Journal of Antimicrobial Chemotherapy (1996) 37, 881-889 ph modulation of aminoglycoside resistance in Staphylococcus epidermidis harbouring 6'-/V-aminoglycoside acetyltransferase E. Culebras**, J. L. Martinez**, F. Baquero* and J. C. P6rez-Diaz* "Centro Nactonal de Biotecnologia (CSIC), Campus (JAM, Cantoblanco, 28049-Madrid; h Servicio de Microbiologia, Hospital Ramon y Cajal, Carretera de Colmenar Km 9,100, 28034-Madrid, Spain The kinetic constants of the aminoglycoside-modifying enzyme 6'-/V-aminoglycoside acetyltransferase (AAC(6')IV) from the clinical strain Staphylococcus epidermidis RYC 13036 differed depending on whether tobramycin and amikacin (glucosamine group) or gentamicin and netilmicin (garosamine group) were used as substrates. Acetylation of the glucosamine antibiotics was highly susceptible to substrate inhibition which increased with ph whereas the garosamine group compounds showed limited substrate inhibition over a wide ph range. These differences in activity correlated with MIC values of S. epidermidis RYC 13036 for different aminoglycosides. Aminosugars moiety and ph markedly influenced the AAC(6')FVaminoglycoside interactions. Introduction Resistance to gentamicin and the related aminoglycosides, tobramycin and kanamycin, which is primarily due to the production of antibiotic-inactivating enzymes, is common in nosocomial strains of staphylococci isolated in hospitals around the world (Graham, 1981; Wiedemann & Kresken, 1984; Guimaraes, Sage & Noone, 1985; Davies, 1994). The recent spread of multiresistant methicillin-resistant staphylococcus (Mulligan &, 1993; Lencastre et al., 1994; Swartz, 1994) may increase this antibiotic resistance problem. In the case of antibiotic-inactivating enzymes, enzymatic modification of an aminoglycoside does not necessarily imply a clinical resistance to this drug. Resistance to an aminoglycoside appears to be a function of the ratio between accumulation and modification velocities of the drug. This ratio is determined by the amount of antibiotic at the site of infection, the permeability characteristics of bacteria, and the activity of the antibiotic-inactivating enzyme. Hence the knowledge of kinetic constant values is essential for the understanding of this process. Several studies with the enzymes aminoglycoside nucleotidyl transferase (2") (Bongaerts & Molendijik, 1984), gentamicin acetyltransferase I (Williams & Northrop, 1978) and kanamycin acetyltransferase (Radika & Northrop, 1984), have shown that aminoglycoside resistance mediated by inactivating enzymes is a function of the hydrolytic constant V^K^. 'Corresponding author. 881 0305-7453/96/050881 + 09 $12.00/0 1996 The British Society for Antimicrobial Chemotherapy

882 E. Culebras et at. Although environmental ph is known to affect the activities of many antibiotics, including aminoglycosides (Bryan, 1984; Thrupp, 1980), little is known about the effect of this factor on the kinetic constants of antibiotic-inactivating enzymes. Values of ph can change between 5.5 and 7.5 in different tissues of healthy individuals (Livermore & Corkill, 1992), and the range may be even wider in patients suffering infectious diseases (Hamill & Maki, 1986; Gudmundson et al., 1991). It has been shown that changes in both extracellular and intracellular ph affect the expression of several genes (Olson, 1993). In general, bacteria are able to maintain the internal ph constant even when external ph changes by up to 2 units. However, it is known that internal ph cannot be maintained when bacteria are growing under stress conditions (Booth, 1985) and changes in external ph have been found to trigger transient changes of internal ph, which may produce changes in both the expression of certain genes such as the Escherichia coli SOS response (Dri & Moreau, 1994), and the activity of certain enzymes. In the present work, we have determined the kinetic constants of the aminoglycoside-inactivating enzyme 6'-JV-aminoglycoside acetyltransferase (AAC(6')IV) from Staphylococcus epidermidis using antibiotics of clinical interest. We have also evaluated the effect of 6'-aminosugar functional groups and of ph on the kinetic constants. Chemicals Materials and methods The antibiotics were obtained from the following suppliers: gentamicin and netilmicin from Shering Corp., Bloomfield USA, tobramycin from Eli Lilly & Co., Indianapolis USA, and amikacin from Bristol-Squibb Laboratories, Syracuse, USA. The radioactive compounds were obtained from Amersham UK. Lysozyme and lysostaphin were obtained from Sigma Chemical Co., St. Louis, USA. Antibiotic susceptibility tests Antibiotic susceptibility was evaluated by the Mueller-Hinton broth dilution technique, according to the protocols of the National Committee for Clinical Laboratory Standards (NCCLS, 1990). When needed, the medium was buffered at the desired ph with 0.1 M phosphate buffer. Aminoglycoside-modifying enzyme assays Bacterial extracts were prepared and the enzymes were assayed using the phosphocellulose paper-binding technique described previously (Haas & Dowding, 1975). The reaction was allowed to proceed for 30 min at 37 C. In the case of kinetic studies, assay mixtures contained the appropriate substrates, 0.1 M Tris-maleate buffer ph 5.5, 6.5 or 7.5, 10 mm MgCl 2, 2.5 mm dithiothreitol and 1 mm 1- M C-ACCOA (1500 cpm/nmol) in a total volume of 250 nl. Each reaction was started by the addition of 50 /il of the crude bacterial extracts. All kinetic studies were performed at 37 C, and determination of the kinetic constants was made as described (Culebras et al., 1994).

ph regulation of AAC(60 activity 883 Kinetic studies of acetyllransferase activity Results The clinical isolate 5. epidermidis RYC13036 (Culebras et al., 1994) harbouring the bifunctional enzyme AAC(6')IV-APH(2") (Le Goffic, Moreau & Masson, 1977) was used in our work. Since acetyltransferase activity is responsible for most of the aminoglycoside resistance in staphylococci, we analyzed the kinetic constants of this enzyme. As Table I shows, the acetyltransferase enzyme had a very high K m value for netilmicin and a very low value for tobramycin both at ph 5.5 and 6.5. At the same ph values, amikacin and gentamicin exhibited intermediate K m values. All these were lower at ph 7.5. At all ph values tested, the enzyme was most active against tobramycin and least active against netilmicin. The value of the V^. constant was also clearly dependent on ph, showing highest values at ph 6.5 using tobramycin, amikacin or gentamicin as substrates. At ph 7.5 these values were slightly lower and they were lower still at ph 5.5. Results with netilmicin were different, with K.^, values rising with increasing ph. As Table I shows, the Knu at ph 7.5 was three times higher than at ph 6.5, whereas at ph 5.5 the activity of the enzyme was nearly undetectable. Substrate inhibition ph influence As substrate inhibition might be relevant for the susceptibility to aminoglycosides of bacteria producing antibiotic-inactivating enzymes, we evaluated the K, of AAC(6')IV to tobramycin, gentamicin, amikacin and netilmicin. The analysis was performed at the same three ph values (5.5, 6.5 and 7.5) previously used for determining the K m and V mmi values. As Figure 1 shows, the type of 6'-aminosugar of the molecule (Figure 2) appeared to have a dramatic effect on the substrate inhibition. The members of the kanamycin family (with a glucosamine-like 6'-aminosugar) showed an inhibition that increased with increasing ph. The opposite was the case for the members of the gentamicin family (groups of aminoglycosides with garosamine) which presented the highest K t values at ph 7.5 (Table II). Also for all ph values tested, the K, values for gentamicin and netilmicin were large, which indicates that substrate inhibition is not significant for members of this family of aminoglycosides. Discussion It has been shown that the hydrolytic constant (V<mJKm) has a close relationship with MIC values for aminoglycoside-inactivating enzymes were compared (Radika & Northrop, 1984). With our strain, this relationship existed when MICs for tobramycin, amikacin and netilmicin, but not gentamicin. The strain used in this study has APH(2") activity, and since this is more active against gentamicin (Bongaerts & Vliegenthart, 1988), the phenotypic behaviour of S. epidermidis RYC13036 towards this antibiotic results from activities of both AAC(6'), and APH(2"), the second being especially active against gentamicin. Substrate inhibition was shown to be relevant for the kanamycins but not for the gentamicins. It has been suggested (Umezawa, Yagisawa & Sama, 1975) that the enzyme AACX60 has two substrate recognition regions, one being the catalytic region, the

Table 1. Apparent kinetic constants of AAC(6')IV and MICs of S. epidermidis RYC13036 at different phs Aminoglycoside Tobramycin Gentamicin Netilmicin ph5.5 216 571 653 m K m (JIM) ph6.5 77 364 297 2440 ph7.5 26 36 77 259 (pmol/mg/min) ph5.5 ph6.5 ph7.5 3384 1458 206 26 5845 2278 2331 1067 4554 2101 2237 3140 ph5.5 16.67 2.05 0.31 b 'mmxf **-m ph6.5 75.95 6.26 7.85 0.43 ph7.5 179.24 57.72 29.04 4.77 MIC (mg/l) ph5. 5 ph 6.5 ph7.5 The value was too high to be measured accurately *Since the value of /f» was not obtainable for this antibiotic at ph 5.5. V^,/K m was not determined, but is assumed to be very low because the value for AC was very high. 64 32 512 8 128 16 128 4 512 8 128 4 w E. re 1 % s.

ph regulation of AAC(60 activity 885 ph5.5 ph6.5 ph7.5 0.5 1.0 1.5 0.5 1.0 1.5 0.5 1.0 1.5 20 30 40 10 20 30 40 10 20 30 40 i itain 1C1 80 60 40 - - Gei 20 n * t 1 i 1 i 1/v (mg.ml/pmol) Fignre 1. Kinetics of excess substrate inhibition: The K, values of AAC(6') from S. epidermidis at different phs were calculated using amikacin, tobramycin, netilmicin and gentamicin substrates.

886 E. Culebru et at. AMINOHEXOSE(') OH 2, 6 DIAMINE Gentamicin C Tbbramycin Netilmicin NH 2 3AMIN0HEXOSET) 0 (R): NHCCHCHJCHNHJ OH (R): NHCjHg Netilmicin 6AMINE Gentamicin B CHJNHJ GLUCOSAMINE Tbbramycin OH 2AMINE Gentamicin A GAROSAMINE Gentamicin Figure 2. Aminoglycoside structures, relevant features of amikacin, gentamicin, tobramycin and netilmicin are shown. other being involved in substrate inhibition. According to our results, both the \-N substituents and the aminohexose (3") are important for the inhibition. Hence, the glucosamine sugar increases substrate inhibition, since the K, values were much lower for tobramycin and amikacin than for the garosamine-containing aminoglycosides gentamicin and netilmicin. On the other hand, the fact that antibiotics with l-n substitutions (amikacin and netilmicin) had higher K t than their non-substituted counterparts (tobramycin and gentamicin), suggests that the \-N side chain reduces inhibition. The values of ph vary between different tissues, and also depend on the site of infection and underlying patient condition. Metabolic acidosis is a well-known complication of septic shock (Hamill & Maki, 1986). Although acidic ph has been Table II. Effect of ph on K { of different aminoglycosides K(HM) Aminoglycoside ph 5.5 ph 6.5 ph 7.5 Tobramycin Gentamicin Netilmicin 6.47 22.06 22.56 64.46 2.3 8.7 48.2 75.0 0.59 4.99 79.00 168.00

ph regulation of AAC(60 activity 887 shown to inhibit the cellular uptake of aminoglycosides (Damper & Epstein, 1981; Eisenberg et al., 1984), there is no available information on the effect of this on the activity of aminoglycoside-inactivating enzymes. In the case of AAC(6'), we have shown that the different functional groups in the antibiotic moiety influence the kinetic constants of the aminoglycoside-modifying-enzymes and hence help to determine the resistance to one specific aminoglycoside. Since ph affects the ionization state not only of these groups, but also of the groups of the amino-acids at the active site of the enzyme, this parameter may influence the antibiotic resistance mediated by AAQ60. We have shown that, in all cases, the values of V^/Km increased with ph, whereas this was not true for the values of K- At acidic ph, substrate inhibition did not occur. When ph increased, the values of K decreased for kanamycins, whereas they increased for gentamicins (Table II). Therefore, at neutral ph, substrate inhibition could be relevant for the kanamycins, whereas it is not important for the gentamicins. In fact, this inhibition counteracted the high value shown by V^/KM at high ph in the case of amikacin but not of tobramycin. Thus, at ph 7.5, amikacin was more effective than tobramycin and gentamicin against staphylococci harbouring AAC(6'). Despite the ten-fold lower value of Vm^/K m for netilmicin than for amikacin, substrate inhibition produced by amikacin resulted in there being less difference between MICs than might be anticipated. A similar effect was also detectable at ph 6.5, but differences in favour of amikacin were less marked because of the lower inhibition observed at this ph. At ph 5.5, the effect of substrate inhibition was overcome by the decrease in aminoglycoside permeability at low ph. The effect of ph on substrate inhibition may be important in clinical practice, and should also be considered in the design and development of new aminoglycosides. Our data indicate not only that aminoglycosides with glucosamine moeities interact with enzymes in a different way from those with garosamine, but also that kinetic constants for both groups of antibiotics vary in opposite directions in response to ph changes. The effect of ph on the kinetic constants of the /Mactamase TEM-1, as well as on the activity of different /J-lactam antibiotics against TEM-1-producing bacteria have been described previously (Livermore & Corkill, 1992). Detailed studies on the effect of ph on the activity of aminoglycoside-modifying enzymes have been published (Gates & Northrop, 1988), but we believe that this is the first report of the relationship between this effect and the activity of aminoglycosides against AAC(6') producing bacteria. These data point to a role of ph in modulating antibiotic resistance mediated by inactivating enzymes. The effect is seen not only on enzymatic activity {V^K^ and K,) but on the permeability of the organism for the antibiotic as well. Since antibiotics and enzymes (even those belonging to the same family) respond differently to ph changes, much more work is needed to clarify the role of this physiological effector on the mechanisms of antibiotic resistance. The understanding of this modulation might contribute to the development of new therapeutic protocols and to the design of new antibiotics. References Bongaerts, G. P. & Molendijik, L. (1984). Relation between aminoglycoside 2'-O-nucleotidyltransferasc activity and aminoglycoside resistance. Antimicrobial Agents and Chemotherapy 25, 234-7.

888 E. Culebras el al. Bongaerts, G. P. & Vliegenthart, J. S. (1988). Effect of aminoglycoside concentration on reaction rates of aminoglycoside-modifying enzymes. Antimicrobial Agents and Chemotherapy 32, 740-6. Booth, I. R. (1985). Regulation of cytoplasmic ph in bacteria. Microbiological Reviews 49, 359-78. Bryan, L. E. (1984). Mechanisms of action of aminoglycoside antibiotics. In Contemporary Issues in Infectious Diseases, Vol. 1. New Dimensions in Antimicrobial Therapy (Root, R. K. & Sande, M. A., Eds), pp. 17-36. Churchill Livingstone, Edinburgh. Culebras, E., Perez-Diaz, J. C, Blazquez, J., Alonso, R. & Baquero, F. (1994). In-vitro synergy between aminogjycosides deployed against Staphylococcus spp. harbouring an 6'-aminoglycoside acetyltransferase, 2"-aminoglycoside phosphotransferase enzyme. Journal of Antimicrobial Chemotherapy 33, 747-55. Damper, P. D. & Epstein, W. (1981). Role of the membrane potential in bacterial resistance to aminogjycoside antibiotics. Antimicrobial Agents and Chemotherapy 20, 803-8. Davies, J. (1994). Inactivation of antibiotics and the dissemination of resistance genes. Science 264, 375-82. de Lencastre, H., de Jonge, B. L., Matthews, P. R. & Tomasz, A. (1994). Molecular aspects of methicillin resistance in Staphvlococcus aureus. Journal of Antimicrobial Chemotherapy 33, 7-24. Dri, A. & Moreau, V. (1994). Control of the LexA regulon by ph: evidence for a reversible inactivation of the LexA repressor during the growth cycle of Escherichia coli. Molecular Microbiology 12, 621-9. Eisenberg, E. S., Mandel, L. J., Kaback, H. R. & Miller, M. H. (1984). Quantitative association between electrical potential across the cytoplasmic membrane and early gentamicin uptake and killing in Staphylococcus aureus. Journal of Bacteriology 157, 863 7. Gates, C. A. & Northrop, D. B. (1988). Determination of the rate-limiting segment of aminoglycoside nucleotidyltransferase 2"-I by ph- and viscosity-dependent kinetics. Biochemistry 27, 3834-42. Graham, D. R. (1981). Gentamicin-resistant staphylococci. Lancet ii, 698. Gudmundsson, A., Erlendsdottir, H., Gottfredsson, M. & Gudmundsson, S. (1991). Impact of ph and cationic supplementation on in vitro postantibiotic effect. Antimicrobial Agents and Chemotherapy 35, 2617-24. Guimaraes, M. H., Sage, R. & Noone, P. (1985). The comparative activity of eleven aminocyclitol antibiotics against 773 aerobic Gram-negative rods and staphylococci isolated from infected hospitalized patients. Journal of Antimicrobial Chemotherapy 16, 555 61. Haas, M. J. & Dowding, J. E. (1975). Aminoglycoside-modifying enzymes. Methods in Enzymology 43, 611-8. Hamill, R. J. & Maki, D. K. (1986). Endotoxin shock in man caused by gram-negative bacilli. In Handbook of Endotoxin, Vol. 4. Clinical Aspects of Endotoxin Shock (Proctor, R. A., Ed.), pp. 55-126. Elsevier Science Publishers B. V. Amsterdam. Le Goffic, F., Moreau, N. & Masson, M. (1977). Are some aminoglycoside antibiotics inactivating enzymes polyfunctional? Annales de Microbwlogie 128B, 465-9. Livermore, D. M. & Corkill, J. E. (1992). Effects of CO, and ph on inhibition of TEM-1 and other /Mactamases by penicillanic acid sulfones. Antimicrobial Agents and Chemotherapy 36, 1870-6. Mulligan, M. E., Murray-Leisure, K. A., Ribner, B. S., Standiford, H. C, John, J. F., Korvick, J. A. et al. (1993). Methicillin-resistant Staphylococcus aureus: a consensus review of the microbiology, pathogenesis, and epidemiology with implications for prevention and management. American Journal of Medicine 94, 313-28. National Committee for Clinical Laboratory Standards. (1990). Methods for Dilution Antimicrobial Susceptibility Tests for Bacteria that Grow Aerobically Second Edition: Approved Standard M7-A2. NCCLS. Villanova, PA. Olson, E. R. (1993). Influence of ph on bacterial gene expression. Molecular Microbiology 8, 5-14. Radika, K. & Northrop, D. B. (1984). Correlation of antibiotic resistance with V^K, ratio of enzymatic modification of aminoglycosides by kanamycin acetyltransferase. Antimicrobial Agents and Chemotherapy 25, 479-82.

ph regulation of AAC(60 activity 889 Swartz, M. N. (1994). Hospital-acquired infections: diseases with increasingly limited therapies. Proceedings of the National Academy of Sciences USA 91, 2420-7. Thrupp, L. D. (1980). Susceptibility testing of antibiotics in liquid media. In Antibiotics in Laboratory Medicine (Lorian, V., Ed.), pp. 73-113. Williams & Wilkins Co., Baltimore. Umezawa, Y., Yagisawa, M., Sawa, T., Takeuchi, T. & Umezawa, H. (1975). Aminoglycoside 3'-phosphotransferase III, a new phosphotransferase resistance mechanism. Journal of Antibiotics 28, 845-53. Wiedemann, B. & Krcsken, M. (1984). The incidence and development of resistance in Staphylococcus aureus from three European countries. Journal of Antimicrobial Chemotherapy 14, Suppl. D, 27-34. Williams, J. W. & Northrop, D. B. (1978). Substrate specificity and structure-activity relationships of gentamicin acetyltransferase I. Journal of Biological Chemistry 253, 5908-14. {Received 25 July 1995; returned 22 August 1995; revised 15 September 1995; accepted 28 December 1995)