Linked Pools of Processed a-Mannosidase in Dictyostelium discoideum*

We are studying the fate of a-mannosidase, a lyso- somal enzyme, in Dictyostelium discoideum. a-Man-nosidase is synthesized as a 150,000-dalton precursor which becomes proteolytically cleaved to mature (56,000-62,000 dalton) forms. When cells are shifted into starvation buffer (5 mM phosphate, pH 6.5), the enzyme is secreted. We compared the kinetics of secre- tion of newly processed a-mannosidase with that of bulk forms. After 2 h in phosphate buffer less than 20% of the newly processed forms but as much as 50% of the bulk enzyme activity was secreted. During the course of the experiments, the total a-mannosidase activity remains constant, cells remain viable, and there is no evidence of degradation of enzyme. Fur- thermore, a 4-h chase period prior to starvation is required before the newly processed forms are secreted to the same extent as the bulk forms. On the basis of these results we propose that the enzyme is present in at least two pools and that it is transferred from a newly processed to an efficiently secreted pool. Cells secrete several

We are studying the fate of a-mannosidase, a lysosomal enzyme, in Dictyostelium discoideum. a-Mannosidase is synthesized as a 150,000-dalton precursor which becomes proteolytically cleaved to mature (56,000-62,000 dalton) forms. When cells are shifted into starvation buffer (5 m M phosphate, pH 6.5), the enzyme is secreted. We compared the kinetics of secretion of newly processed a-mannosidase with that of bulk forms. After 2 h in phosphate buffer less than 20% of the newly processed forms but as much as 50% of the bulk enzyme activity was secreted. During the course of the experiments, the total a-mannosidase activity remains constant, cells remain viable, and there is no evidence of degradation of enzyme. Furthermore, a 4-h chase period prior to starvation is required before the newly processed forms are secreted to the same extent as the bulk forms. On the basis of these results we propose that the enzyme is present in at least two pools and that it is transferred from a newly processed to an efficiently secreted pool.
Cells secrete either uncleaved precursor or mature forms of several lysosomal enzymes. Most precursor molecules enter functional lysosomes, but a minor fraction has been shown to be transported directly to the media by fibroblasts (I, 2), Chinese hamster ovary cells (3, 4), and porcine kidney cells (5,6). The extent of secretion of precursor forms is much greater in cells which are genetically defective in elements of the phosphomannosyl recognition system (2,4 ) or in cells which have been poisoned with amines (1, 7).
In normal cells most precursor molecules enter functional lysosomes where they are soon converted by proteases to mature products. For example, pulse-labeled a-iduronidase precursors were found in an endoplasmic reticulum fraction of human fibroblasts and could be chased into a lysosomal fraction in which they were cleaved (8). In vitro experiments of Frisch and Neufeld (9) lend support to these ideas.
The contents of functional lysosomes are secreted under certain conditions. For example, kidney cells have been shown to exocytose lysosomal enzymes in a coordinated manner (lo), particularly in response to hormonal stimulation (11). Macrophages and polymorphonuclear leukocytes secrete enzymes * This research was funded in part by United States Public Health Service Research Grant AM21424. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "aduertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
$ after stimulation by zymosan granules (In), and mast cells secrete histamine and acid hydrolases in response to a variety of secretagogues (13). Although it often has gone undemonstrated, it is reasonable to assume that these secreted hydrolases are mature cleaved forms. We have reported the secretion of both unprocessed precursors and mature forms of amannosidase by D. discoideum (14). Analysis of the latter is the focus of this paper.
D. discoideum synthesizes acid hydrolases which either accumulate intracellularly or in the medium. Measurements of changes in net enzyme activity levels indicated that the extent and rate of secretion of these enzymes is controlled by genetic (15), developmental (16), and nutritional factors (17). For example a-mannosidase, @-hexosaminidase, and p-glucosidase are extensively secreted when cells are shifted from growth medium to starvation buffer. We previously reported on the biosynthesis and processing of a-mannosidase from this organism (14, 18). Many of these findings have since been confirmed by Mierendorf et al. (19). We purified the enzyme, and prepared a specific antibody against it. The antibody was utilized in pulse-chase studies which indicated that 150,000-and 80-000-dalton precursors are processed to a group of mature forms (56,000-62,000 daltons). We noticed that, under conditions which would be expected to produce marked secretion of a-mannosidase, little of the newly processed forms were secreted.
Here we document the finding that newly processed CYmannosidase polypeptides are secreted to a much lesser extent than the bulk enzyme activity, when growing cells are subjected to starvation. We present further evidence that the radiolabeled polypeptides are indeed derived from a-mannosidase, and we show results which establish a kinetic relationship between newly processed and efficiently secreted forms. The possibility that mature a-mannosidase exists in two linked pools and the relation of the results to an understanding of the generation of lysosome heterogeneity are discussed.

MATERIALS AND METHODS
D. discoideum (strain Ax2) was generously donated by Dr. Claudette Klein (St. Louis University School of Medicine) and was maintained in log phase by daily dilution with HL-5 medium (20). The purification of a-mannosidase and preparation of antibody directed against this enzyme were performed as previously described (14). D. discoideum mutants were gifts from Dr. Randall Dimond (University of Wisconsin-Madison).
[%]Methionine was purchased from Amersham Corp. Pansorbin, a commercial preparation of formaldehydetreated Staphylococcus aureus was purchased from Calbiochem. Protosol (a tissue and gel solubilizer) was purchased from New England Nuclear.
Secretion of a-Mannosidase Forms during Staruation-Cells (2-6 x IO6 cells/ml) were centrifuged at 200 X g for 8 min. The medium was aspirated and cells were resuspended to 7.5 X lo6 cells/ml in HL- Processing and secretion of newly synthesized amannosidase during starvation. Cells (7.5 X lofi cells/ml) were incubated 3.5 h in 5 ml of HL-5 medium. ["SSjMethionine (600 pCi) was added and the incuhation was continued for 1.5 h. The suspension was made 50 mM in methionine and centrifuged. Cells were resuspended to their original volume in starvation buffer. Aliquots (1 ml) were removed at the indicated times, cells and medium were separated, and samples were assayed for enzyme activities. Subsequent treatment of cells and medium, which included immunoprecipitation, SDS-polyacrylamide gel electrophoresis and autofluorography, is described under "Materials and Methods." A is a reproduction of the exposed autofluorograph. The gel concentration was 7.5%. Film exposure time was 10 days. C and M refer to cells and medium, respectively. tu-Man, is the 150,000-dalton precursor form of tu-mannosidase. tu-Man, is the group of mature forms (56,000-62,000 daltons). H depicts the relationship between starvation time and secretion of tu-Man,, bulk tu-mannosidase forms, and acid phosphatase. The direct counting method was used to quantitate the distribution of labeled tu-mannosidase forms: X, n-mannosidase activity; 0, tr-Man,; 0, acid phosphatase activity. cells were resuspended in 5 mM phosphate, pH 6.5 (starvation buffer). At the indicated times, 1-ml aliquots were removed, cells were separated from medium, and cells were lysed (14). The laheled polypeptides in each fraction were immunoprecipitated and subjected to SDS'-polyacrylamide gel electrophoresis. Then autofluorography or autoradiography was performed as previously described (14).

(growth medium). Resuspended cells were incubated in the presence
Two methods were used to quantitate the extent of secretion of laheled tu-mannosidase forms. 1) Radioactive bands corresponding to trMan, were cut from the dried gel, solubilized in Protosol and counted in a liquid scintillation counter. 2) Densitometric tracings of the exposed film were obtained and peak heights of the nMan, were measured. In each case background noise from a portion of the gel  Comparison of extents of secretion of newly labeled and bulk n-mannosidase. Cells (7.5 X lofi cells/ml) were incubated 2 h in 5 ml of HL-5 medium. The suspension was centrifuged and cells were resuspended to their original volume in a defined medium (22) minus methionine. ["S]Methionine (850 pCi) was added and the incubation was continued for 20 min. The suspension was centrifuged, cells were resuspended in 5 ml of HL-5 medium and incubated 1 h to allow for precursor processing. Cells were centrifuged and resuspended in 5 ml of starvation buffer. One-ml aliquots were removed at the indicated times. Gel patterns and secretion percentages were obtained as described in the legend to Fig. 1. A is a reproduction of the exposed autoradiograph. The gel concentration was 7.5%. Film exposure time was 8 days. R depicts the extent of secretion of newly synthesized and hulk tu-mannosidase from starved cells. Densitometric tracings were used to quantitate the distribution of labeled (Ymannosidase: X, n-mannosidase activity; O, labeled o-mannosidase. which did not contain samples was subtracted. The data were reported as per cent secretion: 9; secretion = Media Media and cells x 100 In some figures these results are compared to the per cent secretion of enzyme activity. For these measurements, aliquots of lysed cells or medium were removed before immunoprecipitation. a-Mannosidase was assayed as previously described (14). Acid phosphatase was assayed using 4-methylumhelliferyl phosphate made 1 mM in 50 mM sodium acetate buffer, pH 4.5 (21). Per cent secretion was calculated as described for n-Man,.

RESULTS
Secretion of Newly Processed and Bulk Forms of a-Mannosidase-Newly processed n-mannosidase is secreted to a much lesser extent than bulk forms of the enzyme when D. discoideum are starved. Evidence for this is shown in Fig. 1 (23) were incubated at a density of 7.5 X lofi cells/ml for 4 h in HL-5; 170 pCi of ["'SSjmethionine were added for each milliliter of suspension and the incubation continued for 4 h. The suspension was made 10 mM in methionine and 1 ml aliquots were lysed as described under "Materials and Methods." Subsequent treatment of lysed suspensions which included immunoprecipitation SDS-polyacrylamide gel electrophoresis, and autoradiography is described under "Materials and Methods." Film exposure time was 7 days and the gel concentration was 7.5%; hnes a and d, wild type; lane b, M-1; lane c, "4. tions for 1.5 h, chased with unlabeled methionine, then shifted to starvation conditions and incubated for the indicated periods. For each sample, cells and media were separated and cells were lysed. Enzyme activity in both fractions was measured. Labeled polypeptides were immunoprecipitated with anti-a-mannosidase and the immunoprecipitates were subjected to SDS-polyacrylamide gel electrophoresis. The pattern shown in Fig. 1A was obtained.2 The fraction of the newly processed forms secreted was compared to the fraction of bulk enzyme activity secreted as shown in Fig. 1B. A lesser fraction of the newly processed forms than of the bulk enzyme was secreted. The minimal secretion of acid phosphatase confirms the findings of Dimond et al. (21) and indicates that little if any cell damage occurred during starvation.
Quantitation of the fraction of newly processed n-mannosidase secreted in this experiment was complicated by the transfer of radioactivity from intracellular precursor to product during the starvation period. This might lead to miscalculation of the extent of secretion. To avoid this complication, kinetic experiments were modified as described in Fig. 2. Cells were briefly labeled with [:%]methionine in a defined growth medium lacking methionine to obtain sufficient incorporation of label. The label was then chased with unlabeled methionine in growth medium for 1 h to allow complete processing of precursor previous to starvation. The remainder of the analysis was performed as described above. As shown in Fig. 2, by comparison to bulk enzyme activity, little newly processed a-This experiment is similar to, but not identical with Fig. 4R in Ref. 14. The latter employed a 20-min pulse and was designed to detect secretion of precursor. Precursor is barely detectable in the 2h chase time of the experiment presented here in Fig. 1. The 90-min pulse period employed was designed to increase the sensitivity of detection of secretion of mature forms. This condition allows for considerable processing of precursor and build-up of mature forms before the chase. As a consequence relatively less precursor is seen. mannosidase was secreted. We also considered the possibility that differential degradation of intracellular and extracellular newly processed forms could lead to this result. This has been eliminated since the total amounts of both a-mannosidase activity and newly processed a-mannosidase remain constant during the experiment. Furthermore, in experiments not shown, labeled n-mannosidase in secretions was found to be undegraded for as long as 24 h.
Confirmation That the Labeled Polypeptides Are Derived from a-Mannosidase-We had previously shown that the labeled polypeptides were immunoprecipitated by an antiserum specific for a-mannosidase and that they co-migrated on SDS gels with purified a-mannosidase (14). Still it was formally possible that the labeled bands might not be the enzyme. This could explain the discrepancies in secretion of the bands and the enzyme activity. Two results yield strong evidence against this possibility. As shown in Fig. 3 Cells were resuspended to their original volume in starvation buffer. Aliquots (1 ml) were removed at the indicated times, and cells and medium were separated and assayed for n-mannosidase activity. Further treatment of cells and medium is described in the legend to Fig. 1. The gel concentration was 7.5%. Film exposure time was 10 days. Gel slices were excised and counted to determine the extent of secretion of continuously labeled n-mannosidase: 0, mmannosidase activity; X, labeled tr-mannosidase.
by guest on March 23, 2020 http://www.jbc.org/ Downloaded from discoideum is starved. This idea is in consonance with the differential extents of secretion of newly processed and bulk a-mannosidase (Figs. 1 and 2). Trivial interpretations such as miscalculations of secretion rates due to replenishment of intracellular newly processed enzyme by precursor, or differential degradation of intracellular and extracellular enzyme, have been eliminated by the results of control experiments. The possibility that the labeled polypeptides are not derived from a-mannosidase is also unlikely since they were immu-+ ACTIUITY efficiently secreted pool. Cells (7.5 X IO6 cells/ml) were incubated FIG. 5. Transit of newly processed a-mannosidase to an 2 h in 5 ml of HL-5 medium.
[35S]Methionine (850 pCi) was added and the incubation continued for 1 h. The suspension was made 10 mM in methionine and at 0, 1, 2, 3, and 4 h, 1-ml aliquots were removed. These aliquots were centrifuged and cells were resuspended in 1 ml of starvation buffer. After 2 h, cells and medium were separated, assayed for a-mannosidase activity, and processed as described under "Materials and Methods." The gel concentration was 8.5%. Film exposure time was 14 days. Cells which were continuously labeled with [35S]methionine as described in Fig. 4 were included as a control. Bars in the graph represent the per cent secretion of a-Man,,, as quantitated by densitometry. The line indicates the per cent secretion of a-mannosidase activity.
thionine to the same extent as two wild type control cultures (this also confirms the results of Mierendorf et al. (19)). Furthermore, after labeling cells for three generations with [:"SS]methionine it was possible to demonstrate the co-secretion of steady state labeled polypeptides and enzyme activity (Fig. 4). The results in Fig. 4 also obviate another potential source of miscalculation. We assume the enzyme activity is derived predominantly from mature enzyme. Mierendorf et al. (19) have presented evidence to support the idea that the precursor is enzymatically active and accounts for 20% of the activity secreted during growth. We did not determine the fraction of activity secreted during starvation which is due to precursor. However, the close correlation between per cent secretion of enzyme activity and of mature polypeptides following starvation shown here indicates that correction for secretion of precursor activity is unnecessary.
Transit of Newly Processed a-Mannosidase to an Efficiently Secreted Pool-Could the reason for the discrepancy in the extents of secretion of a-Man, and enzyme activity be that the a-Man, was in a different pool from the bulk enzyme? If this were the case, then a cohort of pulse-labeled a-Man, molecules should in time enter the efficiently secreted pool and be secreted to the same extent as enzyme activity or steady state labeled a-Man,.
This result was obtained in experiments of the type described in Fig. 5. Cells were labeled for 1 h in growth medium then chased with methionine for the indicated periods previous to starvation. The distribution of enzyme activity and labeled polypeptides was assessed after the cells were starved for 2 h. In the absence of a chase period the usual low secretion of newly processed forms was observed. As the chase period was lengthened, i.e. as the age of the cohort of enzyme molecules increased, the extent of secretion increased until at 4 h it equaled the extent of secretion of enzyme activity and of long term labeled polypeptides. Furthermore, the results indicate that the half-life for transit between the two pools is approximately 2 h.

DISCUSSION
One way to explain these results is to assume that amannosidase resides in a presecretory pool before it enters a pool of enzyme which can be efficiently secreted when D. noprecipitated by antibody specific for a-mannosidase and have the same mobility as the purified enzyme on SDS gels (14), are absent in extracts from a-mannosidase-negative mutants (Fig. 3), and are co-secreted with a-mannosidase when steady state pools are labeled (Fig. 4). The idea of separate linked pools is further supported by pulse-chase experiments (Fig. 5 ) which indicate the transfer of the enzyme from the presecretory to the efficiently secreted pool.
The physical nature of these pools is unclear. If the fate of acid hydrolases in D. discoideum is analogous to that in higher organisms (8,9), it is converted from proenzyme to mature enzyme in functional lysosomes. Since the enzyme in the presecretory pool has been proteolytically processed, it seems reasonable to assume that this pool is in a population of functional lysosomes. However, it is also possible that cleavage of acid hydrolase precursors in D. discoideum may occur in prelysosomal organelles (Golgi, Gerl, transport vesicles) in a manner similar to proinsulin (24) or prolactin (25). If cleavage takes place in functional lysosomes, it is possible that the presecretory pool is in a different population of lysosomes than the efficiently secreted pool. This view was suggested as one of the explanations of the bimodal distribution of administered asialoceruloplasmin in rat liver (26). If correct it would extend to lysosomes the well documented pool separation already demonstrated for the secretory vesicle-condensing vacuole system by electron microscopy (27).
An alternate explanation of our results is that young cleaved enzyme is chemically different from old cleaved enzyme. The chemical difference may allow young enzyme to be sequestered within a lysosome in a manner which supports its retention by the cell. Indeed, differential sequestration of 0glucuronidase forms based on their ability to interact with the phosphomannosyl receptor has been reported (28, 29) in higher organisms. This particular recognition system is an unlikely candidate for such differential sequestration in D. discoideum because essentially all a-mannosidase molecules secreted during starvation bear the phosphomannosyl recognition marker (30). Also, numerous attempts to detect the receptor in this organism have been unsuccessful. However, the general mechanism is a reasonable one especially since indirect evidence for the involvement of other recognition systems in lysosome biogenesis has been reported (31-33).
Experiments are underway to determine whether the existence of the two pools is the result of heterogeneity of lysosomal vesicles or of a-mannosidase structure. These studies may highlight mechanisms of sorting which occur after proteins leave the Golgi region. It should be stressed that the secretion of acid hydrolases induced by starvation may be different in mechanism from secretion during growth. Experiments are planned to determine if the pools of mature enzyme described here also are involved in other secretion modes. The reported findings may alert others studying lysosome biogenesis to the possibility of linked pools of acid hydrolases in higher organisms.