Chemotherapy 25: 204-213 (1979)
Nonspecific Aminoglycoside Cross-Resistance of Serratia marcescens1 Walter H. Traub and Paula I. Fukushima Department of Laboratory Medicine, University of California, and Clinical Microbiology Laboratory, VA Hospital, San Francisco, Calif.
Key Words. Amikacin • Gentamicin • Kanamycin • Netilmicin • Resistance, nonspecific • Resistance, plasmid-mediated • Serratia marcescens • Sisomicin
Introduction During the last 15 years, Serratia mar cescens, an opportunistic-pathogenic mi croorganism, gained increasing nosocomial significance throughout the Western Hemi sphere. This microorganism proved chromosomally resistant against various antimi crobial drugs (see Traub [11] for additional references); in addition, increasing numbers 1 This study was supported by the Medical Re search Service of the Veterans Administration (CMR-38T).
of clinical isolates of S. marcescens were shown to carry conjugative and nonconjugative resistance (R-)-plasmids [6, 15]. To compound matters, this microorganism re cently was demonstrated to yield phenotypic variants with combined resistance against chloramphenicol, nalidixic acid and trimeth oprim [14]. Two sets of late observations prompted us to examine phenotypic variants of 5. marcescens for nonspecific, combined re sistance against aminoglycoside antibiotics. First, a patient with biliary tree infection
Downloaded by: Univ. of California Santa Barbara 128.111.121.42 - 3/8/2018 1:42:06 PM
Abstract. 10 of 11 clinical isolates of Serratia marcescens yielded small numbers (10~7 to 10~8) of ‘gray’ colony-type variants after selection with either amikacin, gentamicin or kanamycin, of which most proved resistant against all of the following aminoglycoside anti biotics: amikacin, gentamicin, kanamycin, neomycin, netilmicin, sisomicin and strepto mycin. The level of resistance was not absolute, but rather low-level to intermediate. All except two of these ‘gray’, resistant phenotypic variants yielded ‘opaque’ revertants which were essentially indistinguishable from the parent (wild type) strains in terms of colonial morphology and antibiotic susceptibility. The ‘gray’ variants kinetically were more suscep tible to the bactericidal activity of fresh human senim than the ‘opaque’ variants. Prelimi nary evidence afforded the tentative conclusion that this nonspecific aminoglycosideresistance mechanism was not plasmid-mediated.
due to a multiple-drug-resistant strain of S. marcescens [11] yielded two phenotypic variants with respect to colonial morpholo gy: ‘gray-white’ (‘opaque’) variants proved susceptible to amikacin and carried a conjugative R-plasmid with 11 resistance (^-de terminants; ‘gray’ (small, transparent) co lony-type variants, on the other hand, were resistant against amikacin, and their resist ance against other aminoglycoside antibiot ics proved conjugally nontransferable. Sec ond, during the course of various in vitro conjugation experiments, several aminogly coside-sensitive recipient strains of S. mar cescens gave rise to tiny, transparent colo nies on selective kanamycin-containing MacConkey agar plates; these variants were resistant against amikacin and gentamicin as well. Therefore, a series of 5. marcescens isolates, all representative of nosocomially significant strains as determined with bacteriocin typing, along with isolates from the above referred to patient and a spontane ously ‘cured’ variant of S. marcescens [15], were examined for the presence of pheno typic variants with nonspecific cross-resist ance against aminoglycoside antibiotics. Materials and Methods Bacteria The following 11 clinical isolates of S. marces cens were employed: SF 15 (bacteriocin type 72); SF 44 (type 44); SF 67 (type 15); SF 77 (type 9); SF 85 (type 15); SF 93 (type 18); SF 119 (type 9); variant 65 (SE 154), spontaneously ‘cured’ [15], derived from isolate SE 154; SE 427, SE 431-agray, and SE 431-a-gray-white, all type 44 [11]. Escherichia coli strain ATCC 25922 served as a control for all antibiotic susceptibility tests. Media Tryptic Soy agar (TSA) and broth (TSB), Mueller-Hinton agar (MHA) and broth (MHB),
205
and Brain Heart Infusion broth (BHIB) were pur chased from Difco Laboratories, Detroit, Mich. MacConkey agar without added crystal violet (MAC) was obtained from BBL, Division of Becton, Dickinson and Company, Cockeysville, Md. The strains were maintained on TSA slants at 4 °C, from which they were transferred bimonth ly, and in a mixture of 0.5 ml of BHIB plus 0.5 ml of heat-inactivated bovine serum (Grand Island Biological Company, Grand Island, N. Y.) at -65 °C, respectively. Antimicrobial Drugs ‘High content’ antimicrobial disks were pro cured from Difco; disk contents in micrograms are stated in table III. Amikacin (AMIK; lot No. CD 76-329) and kanamycin (KM; lot No. 76F-1405) were received through the courtesy of Bristol Laboratories, East Syracuse, N. Y. Gentamicin (GM; lot No. GMC-5-M-6123), netilmicin (NETIL; lot No. 8344-53), and sisomicin (SISO; lot No. 9383-26) were a gift from Schering Corpora tion, Bloomfield, N. J. Aqueous stock solutions (2,560 ug/ml) were membrane-filter sterilized (0.2 fim filter units; Nalge Sybron Corporation, Rochester, N. Y.), dispensed as 1.5-mI aliquots into sterile 1 dram screw-capped vials, and frozen and kept stored at -65 °C until further use. Antimicrobial Susceptibility Tests Agar disk diffusion tests were carried out ac cording to the method of Bauer et al. [1]; large MHA plates measured 140 mm in diameter. Mi crotiter (Cooke Laboratory Products, Alexandria, Va.) broth dilution tests, utilizing MHB and bac terial inocula adjusted to yield approximately 1.5 X 106 bacteria/ml at 0 time, were performed as previously described [16]. U-shaped wells re ceived 0.05 ml of serially twofold diluted drug plus 0.05 ml of bacterial inoculum. The plates were incubated at 35 °C overnight (16-18 h). The minimal inhibitory concentrations (MICs) were in terpreted as the lowest concentration of drug that completely inhibited the growth of bacterial inoc ula. Selection of Nonspecifically Aminoglycoside-Resistant Variants o f S. marcescens With Selective MacConkey Agar. The 5. mar cescens isolates were inoculated into 2.5 ml of
Downloaded by: Univ. of California Santa Barbara 128.111.121.42 - 3/8/2018 1:42:06 PM
5. marcescens: Nonspecific Aminoglycoside Resistance
206
Traub/Fukushima
Bacteriocin Typing o f S. marcescens Isolates were typed as based on their suscepti bility to one or more of ten selected group A (phage tail) bacteriocins of S. marcescens in ac cordance with the technique of Traub et al. [17].
Results All isolates of S. marcescens examined with the aminoglycoside-selective MAC procedure yielded small numbers (10~7 to 10"8) of AMIK-, KM-, and variably GM-resistant variants (table I). Cell popula tions of two isolates (SF 67 and SF 85) did not contain GM-R variants. The AMIKand KM-R variants of isolate SF 85 still proved susceptible to GM (diameters of GM inhibition zones = 16 mm), but were
Table I. Selection of nonspecifically aminoglyco side-resistant variants from 8 isolates of S. marcescens S. marcescens isolate/ variant No.
SF 15 SF 44 SF 67 SF 77 SF 85 SF 93 SF 119 65 (SE 154)
Selective MacConkey agar with added : amikacin, 20 //g/ml
gentamicin, kanamycin, 10//g/ml 20//g/ml
13» 1 1 3 5 5 8 16
8 1 0 2 0 1 2 2
18 2 1 >200 1 7 12 32
1 Number of resistant variants isolated per selective MAC plate.
resistant against AMIK, KM, NM and SM; they were excluded from further examina tion. The aminoglycoside-resistant variants of the remaining S. marcescens isolates (ta ble II) invariably consisted of small, trans parent colonies on MAC plates, which ap peared ‘gray’ on chocolate agar plates. All ‘gray’, aminoglycoside-resistant variants ex cept variants No. 3 (SF 15-GM-R, colony No. 1) and No. 22 (65[SE 154]-GM-R, col ony No. 2; table II) proved phenotypically unstable; to varying degrees, the ‘gray’ var iants reverted to ‘opaque’ variants (opaque colonies on MAC plates; gray-white colo nies on chocolate agar plates) that were es sentially indistinguishable from the control parent (wild type) isolates with respect to colony morphology and antibiotic suscepti bility. It was noted that the TSB plus added AMIK-disk selective procedure (for S. mar cescens isolates SE 427 and SE 431-a-gray) was convenient and more economical than the selective MAC technique; however, this procedure did not yield quantitative data
Downloaded by: Univ. of California Santa Barbara 128.111.121.42 - 3/8/2018 1:42:06 PM
TSB and incubated at 35 °C overnight, after which 0.1-ml aliquots (approximately 3 X 108 cells) were inoculated and glass-spread onto selec tive MAC-pIates that contained 20 «g/ml AMIK, 10 Mg/ml GM, 20 ug/ml KM or no drug (control), respectively. The plates were incubated at 35 °C for 48 h, following which resistant variants (table I) were subcultured into 2.5 ml of TSB. These tubes were incubated at 35 °C overnight, and the growths obtained were examined preliminarily (disk method) for resistance against AMIK, GM, KM, neomycin and streptomycin on small MHA plates (100 mm diameter). Those aminoglycosideresistant variants (table II) that revealed uni form ‘opaque’ (medium-sized, gray-white) or ‘gray’ (small, transparent) colony morphology, were subcultured to chocolate agar (check for pu rity and phenotypic colony morphology), from which they were further tested for antibiotic sus ceptibility with the disk and the broth dilution methods. With Tryptic Soy Broth plus Added Amikacin Disks. Isolates SF 427 and SE 431-a-gray were in oculated into 2.5 ml of TSB and a total of five 10,«g AMIK disks were added aseptically to each tube. The tubes were incubated at 35 °C over night, after which the growths were subcultured to chocolate agar and further processed as above.
S. marcescens: Nonspecific Aminoglycoside Resistance
207
Table II. Sources of nonspecifically aminoglycoside-resistant variants of S. marcescens Code
S. marcescens isolate/variant No.
Derivation (selected with aminoglycoside antibiotic)
Colonial morphology
Phenotypic stability
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30
SF 15
amikacin-R, colony 2 amikacin-R, colony 2 gentamicin-R, colony 1 control amikacin-R, colony 1 gentamicin-R, colony 1 gentamicin-R, colony 1 control amikacin-R, colony 1 amikacin-R, colony 1 control amikacin-R, colony 2 amikacin-R, colony 2 gentamicin-R, colony 1 control gentamicin-R, colony 1 gentamicin-R, colony 1 control gentamicin-R, colony 1 gentamicin-R, colony 1 control gentamicin-R, colony 2 kanamycin-R, colony 1 control control amikacin-R (TSB2) control amikacin-R (TSB)2 control control
opaque gray gray opaque gray opaque gray opaque opaque gray opaque opaque gray gray opaque opaque gray opaque opaque gray opaque gray gray opaque opaque gray opaque gray opaque
stable not stable12 stable not stable stable not stable stable not stable stable not stable not stable stable not stable stable not stable stable not stable not stable not stable -
SF 44
SF 67
SF 77
SF 93
SF 119
65 (SE 154)
SE 427 SE 431-a-gray SE 431-a-gray-white E. coll ATCC 25922
with respect to the number of AMIK-R var iants present in the respective cell popula tions. The disk antibiograms of all isolates of S. marcescens tested (except isolate SF 85) and ‘gray’ (aminoglycoside-resistant) and ‘opaque’ (aminoglycoside-susceptible) pheno typic variants derived therefrom (table III)
indicated that regardless of the aminoglyco side antibiotic employed for selection, the majority of ‘gray’ variants proved resistant against AMIK, GM, KM, NM and SM. The Bauer-Kirby technique readily permitted detection of ‘opaque’ phenotypic revertants (double, concentric zones of inhibition) among ‘gray’ variants.
Downloaded by: Univ. of California Santa Barbara 128.111.121.42 - 3/8/2018 1:42:06 PM
1 Not stable = Opaque revertants observed. 2 These variants had been selected with amikacin disks added to tryptic soy broth (see ‘Materials and Methods’).
208
Traub/Fukushima
Code S. marcescens isolate/variant No.*1
AMIK 102
Amp 10
Carb 100
Ceph Chlor GM 30 30 10
KM 30
Na 30
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30
183 6/184 6 14 6 14/18 6/13 12 17 6 15 18 6/15 6/10 17 20 6 16 21 6/18 16 6 6/10 14 19 10/23 20 10/21 20 15
17 6/17 6 14 6 6 6 6 6 6 6 17 17 20 17 12/19 6 6 12 6 6 25 23 18 6 6 6 6 6 15
26 17/30 19 26 6 6 6 6 6 6 6 25 26 31 22 22/29 6 24 24 16/27 23 32 32 26 6 6 6 6 6 23
6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 19
23 12/24 8 21 6 19 12/19 23 24 10 18 13 6/14 6 15 22 12 21 22 14/23 22 9 11/17 20 6 6 6 6 6 20
24 27 20 25 28 25 28 29 27 33 23 25 27 28 25 26 28 25 26 28 25 36 24 28 23 25 26 23 24 24
SF 15-AMlK-R-opaque SF 15-AMIK-R-gray SF 15-GM-R-gray SF 15-control SF 44-AMIK-R-gray SF 44-AMlK-R-opaque SF 44-AMIK-R-gray SF 44-control SF 67-AMIK-R-opaque SF 67-AMlK-R-gray SF 67-control SF 77-AMI K-R-opaque SF 77-AMIK-R-gray SF 77-GM-R-gray SF 77-control SF 93-GM-R-opaque SF 93-GM-R-gray SF 93-control SF 119-GM-R-opaque SF 119-GM-R-gray SF 119-control 65 (SE 154)-GM-R-gray 65 (SE 154)-KM-R-gray 65 (SE 154)-control SE 427-control SE 427-AMIK-R-gray SE 431-a-gray-control SE 431-a-gray-AMIK-R-gray SE 431-a-gray-white-control E. coli ATCC 25922 (control)
22 25 23 21 18 20 19 23 22 18 19 23 22 23 22 21 17 21 20 23 20 27 28 23 6 6 6 6 6 21
22 14/24 13 20 6 17 13/17 21 19 11 17 21 13/21 9/13 20 21 12 20 22 16/23 22 13 13/18 19 11 6/17 14 6/14 13 21
AMIK = Amikacin; Amp = ampicillin; Carb = carbenicillin; Ceph = cephalothin; Chlor = chlorampheni col; GM = gentamicin; KM = kanamycin; Na = nalidixic acid; NM = neomycin; N f = nitrofurantoin; PB = polymyxin B; SM = streptomycin; Tc = tetracycline; SSS = triple sulfonamides; SXT = cotrimoxazole (tri methoprim + sulfamethoxazole). 1 Sec table II for sources of phenotypic variants. 2 Disk content of antimicrobial drug (//g). 3 Diameters of inhibition zones (in mm); disks measured 6 mm in diameter, therefore values of 6 indicated uninhibited bacterial growth. 4 Unstable phenotypic, nonspecifically aminoglycoside-resistant variants: left values refer to gray, resistant variants; values to the right of the crossbar indicate susceptible, opaque phenotypic revertants.
Downloaded by: Univ. of California Santa Barbara 128.111.121.42 - 3/8/2018 1:42:06 PM
Tabic III. Bauer-Kirby disk antibiograms of nonspecifically aminoglycoside-resistant variants of S. ntarcescens
209
S. marcescens: Nonspecific Aminoglycoside Resistance
Nf 300
PB 30
SM 10
Tc 30
SSS 300
SXT 1.25/ 23.75
19 12/20 11 19 6 17 14/16 18 18 12 17 20 17/22 12/14 20 18 11 17 19 14/21 20 11 11/17 19 8 6 11 6 10 17
11 11 16 11 6 10/19 9/14 12 11 10/15 10 10 14 17 13 12 15 11 11 11 11 24 24 15 8 6 10 6 11 24
6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 15
17 10/19 6 16 6 16/22 6/13 16 18 6 13 10 6 6 10 18/24 6 16 18 6/18 17 6 6/11 16 6 6 6 6 6 15
6 6/10 9 9 17 10 6 6 6 16 6 6 6 6 6 6/15 9 6 10 6 6 21 21 11 10 10 9 6/17 15 19
20 6/22 6 21 30 26 22 23 23 18 22 6 6 6 6 22 16 20 19 21 18 25 29 23 6 6 6 6 6 22
26 28 23 28 32 25 25 29 26 30 23 20 20 23 18 25 25 23 24 27 24 39 33 27 16 11 16 11 15 28
Discussion The data obtained in the present study permitted the following conclusions to be drawn. First, strains of S. marcescens con tained very small numbers (10~7 to 10~8) of aminoglycoside-resistant phenotypic var iants (‘gray’ colony type) that most likely would have remained undetected with con ventional methods of antibiotic susceptibili ty testing, i.e., the Bauer-Kirby disk tech nique (inoculum size ca. 7.5 X 107 bacteria, assuming that cotton swabs deliver 0.5 ml of inoculum adjusted to 1.5 X 108 bacteria/ ml) [1], and the microtiter, or for that
Downloaded by: Univ. of California Santa Barbara 128.111.121.42 - 3/8/2018 1:42:06 PM
NM 30
The MICs of AMIK, GM, KM, NETIL and SISO (table IV) essentially confirmed the data obtained with the disk agar diffu sion procedure. The ‘gray’ variants dis played low-level to intermediate, but not absolute resistance against the five amino glycoside antibiotics. The microtiter broth dilution technique, of course, failed to detect phenotypic reversion of ‘gray’ to ‘opaque’ variants, since we limited this procedure to determination of MICs only. In a related series of experiments, it was found that S. marcescens SF 15, SF 44, SF 67, SF 77, SF 93, SF 119 and variant 65 (SE 154) were ‘delayed serum-sensitive’ to the bactericidal activity of 80 vol% of fresh human serum, as did the ‘opaque’ variants (killing required from 1 to 4 h of exposure). Conversely, the ‘gray’ variants were killed more rapidly; in particular, the ‘gray’ AMIK-R variants derived from isolate SF 67 (code No. 10; table II) and from SF 77 (code No. 13; table II) were killed within 20 min [Traub, W. H. and Fukushima, P. unpubl. observations].
210
Traub/Fukushima
Table IV. MICs of amikacin, gentamicin, kanamycin, netilmicin and sisomicin against nonspecifically amino glycoside-resistant variants of S. marcescens Code
S. marcescens isolate/variant No.1
Amikacin
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30
SF 15-AMlK-R-opaque SF 15-AMIK-R-gray SF 15-GM-R-gray SF 15-control SF 44-AMIK-R-gray SF 44-AMlK-R-opaque SF 44-GM-R-gray SF 44-control SF 67-AMIK-R-opaquc SF 67-AMIK-R-gray SF 67-control SF 77-AMIK-R-opaque SF 77-AMlK-R-gray SF 77-GM-R-gray SF 77-control SF 93-GM-R-opaque SF 93-GM-R-gray SF 93-controi SF 119-GM-R-opaque SF 119-GM-R-gray SF 119-control 65 (SE 154)-GM-R-gray 65 (SE I54)-KM-R-gray 65 (SE 154)-control SE 427-control SE 427-AMIK-R-gray SE 431-a-gray-control SE 431-a-gray-AMIK-R-gray SE 431-a-gray-white-controI E. coli ATCC 25922 (control)
l2 32 64 2 128 4 16 2 0.5 32 2 1 16 32 2 2 32 2 0.5 16 2 32 16 8 2 16 2 16 2 2
Gentamicin
1 8 16 1 64 2 16 2 0.5 8 1 1 16 32 2 1 8 2 0.25 8 0.5 16 16 2 32 128 32 >128 32 2
Kanamycin
2 32 64 2 128 4 32 4 4 32 4 32 >128 >128 32 2 32 8 1 32 2 64 32 8 >128 >128 >128 >128 >128 4
Netilmicin
Sisomicin
2 32 64 4 128 1 8 1 0.5 16 2 1 16 32 2 0.25 16 2 i 16 4 16 16 1 0.5 4 I 4 0.25 0.25
0.5 8 32 2 64 0.5 4 0.5 0.5 8 1 1 8 64 2 0.125 8 0.5 0.25 8 0.5 8 8 0.5 4 32 4 32 2 1
matter, the tube broth dilution technique unrelated aminoglycosides as well; in other (recommended inoculum size: 10s to 106 words, these ‘gray’, selected phenotypic var bactcria/ml at 0 time) [4], Second, most, but iants usually were characterized by not all, of the ‘gray’ phenotypic variants se cross-resistance against all aminoglycosides lected with either AMIK, GM or KM re examined (table III, IV). The level of non vealed resistance against not only the partic specific aminoglycoside-resistance was not ular aminoglycoside utilized for selection, or absolute; rather, this nonspecific resistance against related aminoglycosides, but against proved low-level to intermediate (table IV).
Downloaded by: Univ. of California Santa Barbara 128.111.121.42 - 3/8/2018 1:42:06 PM
1 See table II for sources of phenotypic variants. 2 MICs = /rg/ml.
Third, this nonspecific aminoglycoside re sistance appeared to be not mediated by plasmids. Although numerous conjugative and nonconjugative R-plasmids had been previsouly detected among clinical isolates of S. marcescens [6, 15], the following evi dence mitigated against extrachromosomal resistance. S. marcescens variant 65 (SE 154) had spontaneously lost (‘curing’) a nonconjugative R-plasmid and proved gen uinely aminoglycoside-susceptible [15]; with two different experimental approaches, namely dye buoyancy density gradient cen trifugation [15] and agarose-gel electropho resis for bacterial plasmid and chromosomal deoxyribonucleic acid [12], this variant was shown to carry a cryptic miniplasmid (contour length: 0.7 /