The membrane fraction of a rat liver post-mitochondria1 supernatant is largely derived from endoplasmic reticulum, but also contains plasma-membrane fragments and Golgi membranes (see, e.g., Depierre & Dallner, 1975). The subfractionation of microsomal preparations normally depends on the use of high centrifugal fields, and hence large hydrostatic pressures. We have attempted to avoid these harsh conditions. which may, for example, cause the detachment of previously membranebound ribosomes (Palmer et al.. 1978). by using a combination of gel filtration and afinity chromatography. Gel filtration, using Bio-Gel A-150 M, allows the separation of membrane vesicles, polyribosomes and soluble proteins from post-mitochondria1 supernatants (McCole et al., 1979). Further subfractionation could be based on differential lectin affinity for membrane glycoproteins. The carbohydrate moieties of these glycoproteins are generally considered to be on the ‘outside’ of plasmamembrane vesicles but those of the endoplasmic-membrane vesicles are unavailable for lectin binding and are considered to be ‘inside’ the vesicles (Winquist et a/., 1974). We therefore attempted to abstract plasma-membrane vesicles from postmitochondria1 supernatant by treatment with immobilized concanavalin A. The livers of male Sprague-Dawley rats (body wt. approx. 200g) were perfused with ice-cold saline (0.9% NaCI) to remove blood, before excision and homogenization with 2.5 vol. of 20 m ~ H e p e s 14(2hydroxyethy1)1 -piperazine-ethanesulphonic acidl, pH 7.5, containing 25 mM-KCI and 5 mMMgCI,. The homogenate was immediately centrifuged at 12000ga,, for 15 min. The resulting supernatant (post-mitochondrial supernatant) was incubated with agarose-bound concanavalin A (purchased from BDH Chemicals, Poole. Dorset. U.K.; 6mg of concanavalin A/ml of settled gel) for 30min at 7OC. The gel was then poured into a sintered tube and washed with homogenizing. medium until no more protein was eluted. This normally resulted in a dilution of the original post-mitochondrial-supernatant components by 3-4-fold. 5’Nucleotidase (AMPase) and glucose 6-phosphatase were used as the plasma-membrane and endoplasmic-membrane markers respectively (Emmelot et al., 1964). Table 1 shows the recovery of these two markers after incubation of post-mitochondria1 supernatant with various amounts of agarose-bound concanavalin A. The recovery of glucose 6-phosphatase is essentially complete (99 2%. n = 19). whereas the recovery of AMPase depends on the amount of lectin used. Filtration through agarose without lectin allows for almost complete recovery of both markers (glucose 6-phosphatase, 96 & 3%, n = 4; recovered AMPaseJglucose 6-phosphatase = 0.95 f 4, n = 4). Subsequent experiments were carried out using 12mg of lectin/ml of post-mitochondria1 supernatant. Under these conditions the recovery of AMPase was approx. 7% of that of glucose 6-phosphatase and that of another marker for plasma membrane. ouabain-sensitive ATPase (Emmelot et al., 1964), was 0-160/0. Initial attempts to recover these markers from the agarosebound lectin by washing with a-methyl mannoside (0.2-0.5 M) have yielded preparations of high specific activity/mg of phospholipid (40-90 times the activity of the post-mitochondrial supematant after lectin treatment) but disappointingly low recoveries ( <40% of original activity). Pre-incubation of post-mitochondria1 supernatant, before or after lectin treatment, with a-methyl mannoside does not affect measured enzyme activities. We believe that these data strongly suggest that concanavalin A will bind to and remove plasma-membrane fragments from a rat liver post-mitochondria1 supernatant and that immobilized lectins may be of general use in subcellular fractionation.
Read full abstract