Whether Escherichia coli K-12 strain W3110 can enter the "viable but nonculturable" state was studied with sterile and nonsterile water and soil at various temperatures. In nonsterile river water, the plate counts of added E. coli cells dropped to less than 10 CFU/ml in less than 10 days. Acridine orange direct counts, direct viable counts, most-probable-number estimates, and PCR analyses indicated that the added E. coli cells were disappearing from the water in parallel with the number of CFU. Similar results were obtained with nonsterile soil, although the decline of the added E. coli was slower. In sterile water or soil, the added E. coli persisted for much longer, often without any decline in the plate counts even after 50 days. In sterile river water at 37؇C and sterile artificial seawater at 20 and 37؇C, the plate counts declined by 3 to 5 orders of magnitude, while the acridine orange direct counts remained unchanged. However, direct viable counts and various resuscitation studies all indicated that the nonculturable cells were nonviable. Thus, in either sterile or nonsterile water and soil, the decline in plate counts of E. coli K-12 strain W3110 is not due to the cells entering the viable but nonculturable state, but is simply due to their death. MATERIALS AND METHODS Bacterial strains and preparation of inocula. The E. coli K-12 strain used was the standard prototrophic wild-type strain W3110 (1). For the studies requiring a plasmid-bearing strain, strain W3110 was transformed with the plasmid pBR322 (2). This particular plasmid was used to permit the inclusion of ampicillin and tetracycline in the direct viable count (DVC) method and to facilitate the PCR measurements employed. Fresh cultures of the strains W3110 and W3110(pBR322) that had been grown for 14 h in Luria-Bertani (LB) medium [plus ampicillin and tetracycline in the case of W3110(pBR322)] were washed with sterile 0.9% saline, adjusted to the desired cell concentration, and added to the water microcosms in 10 ml of inocula or to the soil microcosms in 3 ml of inocula. Media and chemicals. Levine eosin methylene blue (EMB) agar, tryptone, yeast extract, brain heart infusion medium, and Bacto-agar were obtained from Difco Laboratories (Detroit, Mich.). Cycloheximide, nalidixic acid, acridine orange, 2-(p-iodophenyl)-3-(p-nitrophenyl)-5-phenyltetrazolium chloride (INT), and glycine betaine were obtained from Sigma Chemical Co. (St. Louis, Mo.). LB medium (35) was used to grow the strains; LB agar was LB medium with 15 g of Bacto-agar per liter. Plating of nonsterile soil suspensions on LB agar gave low total counts, mainly because the colonies were often obscured by mats of fungal growth and because of the tendency of many of the soil bacteria to swarm. Use of soil extract (SEC) agar gave higher total counts, and inclusion of cycloheximide prevented both fungal growth and bacterial swarming. Soil extract was prepared by suspending 1 kg of soil and 0.5 g of calcium carbonate in 1 liter of distilled water. The suspension was autoclaved for 60 min...