Glucose-6-phosphatase was soon left behind when irregularities showed up with a control enzyme, acid phosphatase. After a gentle cell fractionation procedure, this activity was present at only one tenth of the level expected based on more violent extraction procedures. The activity then reappeared if extracts were left for several days in the refrigerator. As de Duve wrote, “…we could have rested satisfied with this result, dismissing the first series of assays as being due to one of those troublesome gremlins that so often infest laboratories, especially late at night. This would have been a pity, since chance had just contrived our first meeting with the lysosome.”
de Duve concluded that the acid phosphatase activity was latent because of a membrane-like barrier—initially believed to be the mitochondrial membrane. But analyzing the distribution of a single enzyme over many fractions, not just the contents of a single fraction as did many investigators, he found subtle differences in the distribution of acid phosphatase and mitochondrial enzymes. The differences were clinched when a centrifuge component broke, resulting in the use of a weaker table-top centrifuge that sedimented mitochondria but not the lighter lysosomes.
By 1955, five enzymes related to degradation had been localized to this fifth fraction, which had been added to the Claude's earlier quartet of nuclear, mitochondrial, microsomal, and supernatant fractions. The new entities were named lysosomes (de Duve et al., 1955). de Duve now had enough confidence in the biochemistry to enlist the EM expertise of Alex Novikoff, and together they tentatively identified a class of “dense bodies” as the probable structural correlate of biochemically defined lysosomes (Novikoff et al., 1956). Notwithstanding the presence of what Dorothy Bainton termed a few “excessively sad looking mitochondria” in these EM images (Bainton, 1981), the identification proved valid, and was supported by an independent study of “small droplets” by Straus (1956).
The de Duve approach was an excellent complement to that of George Palade, who started with EM pictures and then tried to ascribe functions to what he saw. de Duve, by contrast, started with the function (biochemistry) and studied it to prove the necessary existence of the structure—an approach that would also lead to the discovery of peroxisomes.