Pitfalls in ecto-5'-nucleotidase enzyme cytochemistry as demonstrated by the immunogold-labelling technique on macrophages.

We have used both the enzyme cytochemical method with lead nitrate as a capture agent and an immunological method at the electron microscope level to localize plasma membrane 5'-nucleotidase in rat peritoneal resident macrophages during the initial interactions of latex beads or heat-killed Escherichia coli with the cell during phagocytosis. In macrophages at rest, cytochemical reaction product was evenly distributed along the external surface of the plasma membrane. However, when the cells were phagocytosing latex beads or bacteria, reaction product covered the entire surface of the adhering particles. To determine whether the apparent redistribution of 5'-nucleotidase onto the adhering particle was fact or artifact, we localized 5'-nucleotidase using a monoclonal antibody and an immunogold labelling technique. In macrophages binding or beginning to ingest bacteria, gold particles were distributed along the plasma membrane, except at the sites of cell-bacterium internalization. More significantly, the adhering bacteria were free of gold particles and therefore had no 5'-nucleotidase on their surfaces. Latex beads proved to be unsuitable as a test particle because the gold particles stuck to them non-specifically. We conclude that the artifactual redistribution of lead-phosphate reaction product is a major drawback of enzyme cytochemical methods when used on cell surfaces and that the immunogold labelling technique is more reliable.