Membrane association may limit the use of acid phosphatase as a lysosomal marker [proceedings].

Although acid B-glycerophosphatase (acid phosphatase, EC 3.1.3.2) has long been the chosen marker for lysosomes in cell fractionation (de Duve et al., 1955; Reid, 1972), there are differences in distribution between acid phosphatase and other hydrolases when rat liver is fractionated. In particular, acid-phosphatase-rich lysosomes sediment more slowly (Sellinger et a/., 1960; Rahman & Cerny, 1969; Burge & Hinton, 1971) and are less dense (Beaufay et al., 1964) than other lysosomes. Furthermore, unlike most lysosomal enzymes, one isoenzyme of acid phosphatase seems to be associated with the lysosomal membrane (Sloat &Allen, 1969; Dobrota & Hinton, 1974). The present study gives further evidence of a membrane-located acid phosphatase, and we conclude that acid phosphatase should not be used as the sole marker for lysosomes in high-resolution cell-fractionation experiments. Livers of hooded rats of the University of Surrey strain were homogenized in 0 . 2 5 ~ sucrose in 5 m~-Tris/HCl/buffer, pH 7.4, with a Potter-Elvehjem homogenizer. From the homogenate, successive pellets were obtained : a crude nuclear fraction (4000g-min), a crude mitochondrial+ lysosomal fraction (1 72 500g-min) and a microsomal fraction (6.5 x 106g-min). The crude nuclear fraction was further fractionated with an A-XI1 zonal rotor (Hinton et al., 1970) and the mitochondrialflysosomal fraction was fractionated by rate sedimentation in an HS zonal rotor (Burge & Hinton, 1971) and subsequent isopycnic banding of the lysosome-rich region in a B-XIV zonal rotor (Dobrota & Hinton, 1974). In some experiments the lysosomes were disrupted before isopycnic banding by treatment with a Polytron FT35 OD homogenizer (Kinematica, Lucerne, Switzerland) for 1 min at 13000rev./min at 4°C. Microsomal fractions were subfractionated by isopycnic ffotation (Norris et al., 1974). Assays for enzymeactivities and protein concentration have been described previously (Hinton & Norris, 1972; Prosper0 et al., 1973). In some experiments latent activities were released by Triton X-100 (0.1 %) or digitonin (0.0016%) (Jacques et al., 1975) rather than by freeze-thawing. As in previous work (Burge & Hinton, 1971 ; Dobrota & Hinton, 1974), we found that lysosomes rich in acid phosphatase sedimented more slowly and banded at a slightly lower density than did lysosomes enriched in other acid hydrolases. Slight differences were observed between the distributions of the other hydrolases (ribonuclease 11, acid phosphodiesterase, 8-galactosidase, 8-glucuronidase and cathepsin D), but these were small compared with the differences between each of these enzymes and acid phosphatase. Isopycnic banding of disrupted lysosomes yielded a zone of smooth membranes, of density 1 . I 6g/ml, that contained acid phosphatase but only small amounts of other enzymes (Dobrota & Hinton, 1974). Of the total amount loaded on the gradient, only 54 % of the acid phosphatase was released as soluble protein, compared with 73 %of the B-galactosidase, 74% of the ribonuclease I1 and 61 % of the acid phosphodiesterase, whereas 18 % of the acid phosphatase was recovered in the smooth-membrane region compared with 5 %, 4 % and 9%, respectively, of the other enzymes. The distribution of membrane-bound acid phosphatase is quite different from that of membrane protein (mostly due to endoplasmic-reticulum fragments). It therefore seems unlikely that the membrane-associated acid phosphatase is due to non-specific absorption, but attempts to confirm the presence of lysosome membranes by assaying Bglucosidase (Tappel, 1969) failed as this enzyme was absent from the livers of our rats. Electron microscopy of the smooth-membrane fraction separated from disrupted lysosomes showed several types of structure, including small strips of membrane with * Permanent address: Cancer Institute, University of Cairo, Cairo, Egypt.