We have developed a quantitative assay to monitor the oxidative burst (H2O2 production) of polymorphonuclear leukocytes (PMNL) using single cell analysis by flow cytometry, and have examined whether PMNL respond to membrane stimulation with an all-or-none oxidative burst. During incubation with normal neutrophils, dichlorofluorescin diacetate diffused into the cells, was hydrolyzed to 2',7'-dichlorofluorescin (DCFH) and was thereby trapped within the cells. The intracellular DCFH, a nonfluorescent fluorescein analogue, was oxidized to highly fluorescent 2',7'-dichlorofluorescein (DCF) by PMNL stimulated by phorbol myristate acetate (PMA). That the oxidative product was DCF was shown by excitation/emission spectra and by mass spectrometry of the product from PMA-stimulated PMNL. Normal resting and PMA-stimulated PMNL oxidized 6.9 +/- 0.7 and 160 +/- 13 attomoles DCF per cell, respectively, in 15 min. Absence of calcium and magnesium ions and/or addition of 2 mM EDTA did not inhibit DCF formation by PMNL stimulated by 100 ng/ml PMA. Since EDTA prevented aggregation of PMNL (even when stimulated by 100 ng/ml PMA), which would prevent accurate flow cytometric analysis, further experiments were performed with EDTA in the medium. A close correlation between average DCFH oxidation and hexose monophosphate shunt stimulation was demonstrated using cells from patients whose PMNL had oxidative metabolic defects of varying severity. Intracellular DCFH was also oxidized by reagent H2O2 or oxygen derivatives generated by glucose oxidase + glucose or by xanthine oxidase + acetaldehyde; DCFH oxidation by these systems was inhibited by catalase but unchanged by superoxide dismutase. The data indicate that the DCFH oxidation assay is quantitatively related to the oxidative metabolic burst of PMNL, and they strongly suggest that the reaction is mediated by H2O2 generated by the PMNL. Incubation of PMNL with varying concentrations of PMA caused graded responses by all PMNL present; i.e., 1 ng/ml PMA caused a mean response of 34% maximal with a single population of responding PMNL (rather than 66% resting and 34% fully stimulated as predicted by the all-or-none hypothesis). Thus, with these assay conditions, oxidative product formation by PMNL occurs as a graded response to membrane stimulation by PMA.