Observed by Scanning Electron Microscopy

The strain of P. aeruginosa used was a Fisher serotype 3,7, originally isolated from a patient with severe burns. The susceptibility of the strain to antibiotics used singly and in combination was determined by the tube dilution method (12). Minimal inhibitory concentrations (MICs) were determined by observing the lowest dilution of antibiotic that prevented visible growth of an inoculum of 105 organisms incubate,d for 18 h at 37°C. When pairs of antibiotics were tested in combination, 0.5 ml of each concentration of antibiotic in saline (total volume, 1 ml) was added to each tube, and 1 ml of Trypticase soy broth containing 105 organisms per ml was added to the antibiotic mixture. When three antibiotics were tested, 1.5 ml of the bacterial suspension in Trypticase soy broth was added to the 1.5-ml total volume of mixed antibiotic solution. The MICs were determined in the same manner as when single antibiotics were used. The Barenbaum criteria were used to determine synergism (2). The antibiotic concentrations in the specimens observed in the SEM were one dilution lower than the MICs as determined by tube dilution susceptibility test, i.e., if the MIC was 12 tig/ml, the concentration of antibiotic in the specimen prepared for observation in the SEM was 6 Ag/ml. In control observations, chloramphenicol and tetracycline were used at 200 and 25 ,ug/ml, respectively, both singly and in combination. Subcultures on thioglycolate medium were made from the tubes examined in the SEM to be sure no contaminating organisms were present that would give spurious morphological findings. For this purpose, 0.1 ml of the antibiotic-containing specimens was added to 10 ml of thioglycolate medium and observed for growth at 37°C after 48 h. The SEM preparations were made as follows. At 18 h, the time at which the results of the tube dilution susceptibility tests were read, the bacteria in the tubes were fixed by the addition of 2 ml of 2.6% glutaraldehyde in phosphate buffer (pH 7.3) to each tube. The fixed bacteria were allowed to settle for 24 h at 40C. Supernatants were decanted, and the undisturbed pellet was suspended in 2 to 4 ml of deionized water that had been filtered through a 0.20 im membrane filter (Nalge Co., Rochester, N.Y.). A 0.4-ml sample of this suspension was transferred to a Gelman 0.45-,um TCM cellulose nitrate filter (Gelman Instrument Co., Ann Arbor, Mich.) in a funnel-type 25-m holder (Millipore Corp., Bedford, Mass.). The bacteria were washed with 125 ml of filtered deionized water, dehydrated in a graded ethanol series (50, 70, 85, 95, 100, 100, and 100%), and critical-point dried in a Samdri PVT-3 (Tousimis Research Corp., Rockville, Md.) (1). The filters were coated with gold-palladium in a high-vacuum evaporator while resting on an omnidirectional orbital rotator (Denton Vacuum, Cherry Hill, N.J.). The specimens were examined in a JSM-35C SEM (JEOL USA, Medford, Mass.). The microscopic fields to be photographed were selected by random scanning, and at least 25 fields were photographed (Polaroid type 55 P/N film) for each concentration of antibiotic or antibiotics. Morphological characteristics seen in the photographs were classified and quantitated without the observer's knowing the antibiotics to which the bacteria had been exposed. Morphological changes in the bacteria were designated as "frequent" if they were seen in more than 10% of the bacteria and as "less frequent" if they were seen in 2 to 10% of the bacteria. The entire experiment was performed using prepara-