Studies of the mechanism of tunicamycin in hibition of IgA and IgE secretion by plasma cells.

Tunicamycin, an antibiotic which blocks the formation of N-acetylglucosamine-lipid intermediates, thereby preventing glycosylation of glycoproteins, inhibits the secretion of IgA and IgE by MOPC 315 mouse plasma cells and IR162 rat plasma cells, respectively. At 0.5 microng of tunicamycin per ml, D-[14C]glucosamine incorporation into newly synthesized immunoglobulin was inhibited greater than 90% while the overall rate of protein synthesized was much less inhibited (40% in the case of MOPC 315 cells and 13% in the case of IR162 cells). This dose of tunicamycin produced an 85% inhibition of IgA secretion by the MOPC 315 cells and a complete inhibition of intact IgE secretion by the IR162 plasma cells. In contrast, tunicamycin had little effect on the secretion of normally nonglycosylated lambda light chains or on cell-free protein synthesis, demonstrating that tunicamycin is not a general inhibitor of protein synthesis or a non-specific inhibitor of protein secretion. No enhancement of intracellular degradation of nonglycosylated immunoglobulin could be demonstrated. Electron microscopy of tunicamycin-treated MOPC 315 cells revealed marked dilatations of the rough endoplasmic reticulum, and direct immunofluorescence indicated that the dilated rought endoplasmic reticulum contained IgA. These data indicate that glycosylation of newly synthesized IgA and IgE may be necessary for normal secretion to occur.     

Effect of tunicamycin on IgM, IgA, and IgG secretion by mouse plasmacytoma cells

Tunicamycin, an antibiotic that prevents glycosylation of glycoproteins by blocking the formation of N-acetylglucosamine-lipid intermediates, was used to study the importance of glycosylation for the secretion of immunoglobulins by mouse plasmacytoma lines that produce immunoglobulins of different classes. Biosynthetically labeled secreted and intracellular immunoglobulins were measured by immunoprecipitation assays. Tunicamycin, at a concentration of 0.5 mug/ml produced an 81% inhibition of IgM secretion by MOPC 104E plasma cells without significantly affecting the initial rate of synthesis of intracellular IgM. No increase in the intracellular degradation of nonglycosylated IgM could be demonstrated. Tunicamycin also produced a 64% average inhibition of IgA secretion by several mouse IgA-secreting plasmacytoma lines. In contrast, despite inhibiting the incorporation of D-[14C] glucosamine into newly synthesized IgG, tunicamycin only produced a 28% average inhibition of IgG secretion, which was only slightly more than the nonspecific inhibition of secretion of the normally nonglycosylated lambda2 light chains by variant MOPC 315 plasmacytomas. These data indicate that the extent of inhibition of immunoglobulin secretion produced by tunicamycin depends on the immunoglobulin class produced by the plasma cell.     

Post-translational fate of variant MOPC 315 lambda chains in Xenopus oocytes and mouse myeloma cells

The post-translational fates of three immunoglobulin lambda chain variants of MOPC 315 were investigated in mouse plasmacytoma cell lines and in mRNA-microinjected Xenopus oocytes. Quite unexpectedly we found that one non-secretory variant chain (lambda-43) underwent extensive post-translational N-glycosylation: however the presence of the oligosaccharide moiety did not account for the nonsecretory phenotype nor did it affect the rate of degradation of this lambda chain. Another variant chain (lambda-47) at first believed to be non-secretory, was found to be secreted from oocytes at a very low level, but mostly as a lambda-lambda dimer. In myeloma cells a low level of lambda-47 chain was secreted and again lambda-lambda dimers were the favoured secretory form. The secretory lambda-48 chain also formed lambda-lambda dimers, whereas lambda-43, which was never secreted, was only found as a monomeric lambda chain in both oocytes and myeloma cells. A similar relationship between assembly and secretion was found when oocytes were coinjected with MOPC 21 heavy (gamma 1) chain mRNA and MOPC 315 lambda chain mRNAs. The wild type lambda chain (lambda-48) was able to assemble with the gamma chain in a covalently bound tetramer (gamma gamma lambda lambda). The variant lambda-47 chain was also able to form gamma gamma lambda lambda tetramers, whereas the lambda-43 was not, even when glycosylation was prevented by tunicamycin. Both types of tetramer were secreted. These data reinforce the idea that conformational changes play a major role in the routing of secretory proteins and that the cellular mechanisms by which these changes are recognized are not cell-type specific.     

Antisera to the secretory component recognize the murine Fc receptor for IgA

Studies were undertaken to determine a possible structural relationship between the secretory component (SC) and the receptor for IgA (Fc alpha R). An IgA-mediated rosetting technique was used to assess the presence of Fc alpha R+ cells in various lymphoid tissues from normal BALB/c mice and mice bearing an IgA plasmacytoma (MOPC 315). Tissues from the MOPC 315-bearing BALB/c mice were found to have a significantly higher percentage of Fc alpha R+ cells; thus, nonadherent spleen cells from MOPC 315-bearing mice were used as a source of Fc alpha R+ cells in these studies. The cells were preincubated with anti-SC and then assayed for the ability of IgA to bind to the Fc alpha R. Antisera to SC from various species inhibited the formation of IgA-mediated rosettes, although preincubation of the Fc alpha R+ cells with antisera directed against other cell surface molecules (e.g., Thy1.2, Lyt1, Lyt2, Fc gamma R, MHC class I and II) or preimmune sera had no significant effect on IgA-mediated rosette formation. Preabsorption of the anti-SC with secretory IgA or with free SC removed the inhibitory effect; preabsorption with myeloma IgA had no effect. These data suggest that SC and Fc alpha R are related serologically and may be structurally related, possible in the IgA-binding region.     

The murine IgA-secreting plasmacytoma MOPC-315 expresses somatostatin receptors

We have previously shown that some neuropeptides had a profound effect on in vitro Ig synthesis (especially IgA) and mitogen-driven murine lymphocyte proliferation. MOPC-315, an IgA-secreting plasmacytoma line, has been extensively used in studies of the regulation of IgA synthesis. In this report we show that the neuropeptide somatostatin (SOM) inhibits proliferation ([3H]thymidine uptake) of MOPC-315 and also inhibits IgA synthesis in vitro. MOPC-315 cells bind both fluorescent SOM and [125I]SOM specifically. On cytofluorimetric analysis, 68 +/- 6.8% (mean +/- SE, n = 7) of MOPC 315 cells labeled with fluorescent SOM and this staining was compatible by incubation with an excess of unlabeled peptide. Specific [125I]SOM binding increased linearly with cell concentration, was rapid and achieved equilibrium after 20 min at 4 degrees C. It was temperature-dependent, readily reversible, and under equilibrium conditions demonstrated a dissociation constant of 1.6 +/- 0.7 nM (mean +/- SE, n = 5). Scatchard analysis showed that MOPC-315 cells had 40,733 +/- 16,050 (mean +/- SE) binding sites for SOM per cell. The characteristics of the interactions of SOM with MOPC-315 cells suggest a specific receptor-mediated mechanism whereby this neuropeptide may modulate lymphocyte function.     

Processing of MOPC 315 immunoglobulin A oligosaccharides: evidence for endoplasmic reticulum and trans Golgi alpha 1,2-mannosidase activity

The processing of asparagine-linked oligosaccharides on the alpha- chains of an immunoglobulin A (IgA) has been investigated using MOPC 315 murine plasmacytoma cells. These cells secrete IgA containing complex-type oligosaccharides that were not sensitive to endo-beta-N- acetylglucosaminidase H. In contrast, oligosaccharides present on the intracellular alpha-chain precursor were of the high mannose-type, remaining sensitive to endo-beta-N-acetylglucosaminidase H despite a long intracellular half-life of 2-3 h. The major [3H]mannose-labeled alpha-chain oligosaccharides identified after a 20-min pulse were Man8GlcNAc2 and Man9GlcNAc2. Following chase incubations, the major oligosaccharide accumulating intracellularly was Man6GlcNAc2, which was shown to contain a single alpha 1,2-linked mannose residue. Conversion of Man6GlcNAc2 to complex-type oligosaccharides occurred at the time of secretion since appreciable amounts of Man5GlcNAc2 or further processed structures could not be detected intracellularly. The subcellular locations of the alpha 1,2-mannosidase activities were studied using carbonyl cyanide m-chlorophenylhydrazone and monensin. Despite inhibiting the secretion of IgA, these inhibitors of protein migration did not effect the initial processing of Man9GlcNAc2 to Man6GlcNAc2. Furthermore, no large accumulation of Man5GlcNAc2 occurred, indicating the presence of two subcellular locations of alpha 1,2-mannosidase activity involved in oligosaccharide processing in MOPC 315 cells. Thus, the first three alpha 1,2-linked mannose residues were removed shortly after the alpha-chain was glycosylated, most likely in rough endoplasmic reticulum, since this processing occurred in the presence of carbonyl cyanide m-chlorophenylhydrazone. However, the removal of the final alpha 1,2-linked mannose residue as well as subsequent carbohydrate processing occurred just before IgA secretion, most likely in the trans Golgi complex since processing of Man6GlcNAc2 to Man5GlcNAc2 was greatly inhibited in the presence of monensin.     

Biosynthesis of immunoglobulin A (IgA). Secretion and addition of carbohydrate to monomer and polymer forms of a mouse myeloma protein

Cell suspensions of mouse plasma-cell tumour MOPC 315 secreting predominantly IgA (immunoglobulin A) monomer and dimer were incubated with radioactive leucine, mannose, galactose and fucose for various periods of time. The amounts of secreted and intracellular immunoglobulins were measured by co-precipitation with specific antibody, and the molecular species present were assessed by electrophoresis in polyacrylamide gels. Analysis of the secreted myeloma protein demonstrated that monomer and dimer IgA molecules are identical with respect to carbohydrate composition and rate of secretion. Within the cell, the myeloma protein is almost entirely accounted for by monomer units which either leave the cell as such or are polymerized with the addition of J chain close to the time of secretion. The results support the concept of a stepwise addition of carbohydrate residues to IgA immunoglobulin during the process of secretion. Similar patterns of carbohydrate assembly were found for the monomer or dimer molecules. Mannose residues are added at an early stage, whereas fucose is added close to the time of secretion. Galactose is also added early, but some may also be incorporated at a later stage. Control of IgA polymerization is considered unlikely to reflect regulation at the level of carbohydrate addition, and it is suggested that the critical controlling factor is the J chain.     

Formation of lipid-linked oligosaccharides by MOPC 315 plasmacytoma cells. Decreased synthesis by a nonsecretory variant

MOPC 315 is a BALB/c plasmacytoma which secretes a trinitrophenol-binding IgA lambda 2 paraprotein. We have investigated the incorporation of [3H]mannose into lipid-linked oligosaccharide precursors in wild-type MOPC 315/J and variant nonsecretory 315/P cells. In pulse labeling experiments, no differences could be detected in the ability of the two cell types to incorporate [3H]mannose into lipid-linked oligosaccharides containing 5 or less mannose residues. In contrast, quantitation of the incorporation of [3H]mannose into larger lipid-linked oligosaccharides and proteins revealed a 49 and 40% decrease, respectively, in the 315/P cells compared to wild-type cells. Further characterization of the lipid-linked structures documented a marked decrease in glucosylated oligosaccharides isolated from 315/P cells. When membranes from the two cell lines were analyzed for their ability to transfer [3H]glucose from UDP-[3H]glucose to [3H]glucosylphosphoryldolichol, an apparent deficiency was noted in the 315/P preparations. However, if assay conditions were adjusted to include AMP in the reaction mixtures, no differences in the in vitro synthesis of [3H]glucosylphosphoryldolichol or [3H]glucose-labeled oligosaccharide-lipid could be detected. In these reactions AMP was found to prevent hydrolysis of UDP-[3H]glucose by inhibiting nucleotide pyrophosphatase (EC 3.6.1.9), the specific activity of which was determined to be more than 100 times greater in variant 315/P compared to wild-type MOPC 315/J cells. This large difference in specific activity was not accompanied by similar differences in the activity of several other enzymes analyzed. A decrease in whole cell UDP-glucose pool size was not detected in 315/P cells. Therefore, if nucleotide pyrophosphatase is important for the control of substrates for glycosylation, it must regulate nucleotide sugar levels at a site other than the cytoplasm of cells, perhaps at the location of synthesis of the larger lipid-linked oligosaccharides.     

Cyclic AMP in plasmacytoma cells

Summary In Balb/c plasmacytoma cells (MOPC 315) the synthesis and secretion of IgA in vitro was suppressed by prior incubation of the tumor cells with anti-Balb/c alloantibody. In tumor cells so treated, the levels of cyclic AMP were found to be increased. The alloantibody, either as alloantiserum or ascitic fluid globulins from allogeneic strains of mice immunized with spleen cells, had been differentially absorbed with Balb/c spleen cells. Such absorption reduces the cytotoxic antibody effect to below detectable levels, as measured by Trypan Blue exclusion, but permits retention of substantial levels of the suppressive antibody effect. The range of dilutions of alloantibody causing the increase in cAMP corresponded with those that caused suppression of IgA secretion. When tumor cells were exposed to agents that increased the level of cAMP either endogenously or exogenously, suppression of secretion of IgA was found, similar to that caused by the alloantibody. When submaximal concentrations of suppressive antibody were combined with submaximal levels of the cAMP-increasing agents greater degrees of suppression were found, suggesting an additive effect. No cytotoxic effect on these tumor cells, as indicated by Trypan Blue exclusion, was caused by these cAMP-increasing agents at the concentrations shown, either alone or in combination with the suppressive antibody.     

Resonance Raman-spectroscopic studies of the hapten features involved in the binding of 2,4-dinitrophenyl haptens by the mouse myeloma proteins MOPC 315 and MOPC 460.

The binding of four dinitrophenyl haptens to the mouse myeloma proteins MOPC 315 IgA (immunoglobulin A) and MOPC 460IgA was studied by resonance Raman spectroscopy. Isotopic substitution with 15N and 2H was used to assign features in the resonance Raman spectra of the free haptens. Changes in each of these features on binding to the proteins could then be attributed to interactions of the proteins' binding sites with either the p-NO2 or the o-NO2/amine regions of the haptens. The interactions between a given hapten and MOPC 315 IgA are often quite distinct from those between the same hapten and MOPC 460 IgA. Moreover, for both antibodies the nature of the R side chain in a Dnp-NHR (Dnp, 2,4-dinitrophenyl) compound appears to modify the interactions between the Dnp chromophore and the protein. Thus, with the haptens studied, there is no unique set of contacts between the Dnp group and the binding site. The contacts expected between epsilon-2,4-dinitrophenyl-L-lysine and the site on MOPC 315 IgA, on the basis of a recent model for this site [Dwek, Wain-Hobson, Dower, Gettins, Sutton, Perkins & Givol (1977) Nature (London) 266, 31--37] were not detected. However, the contacts between this hapten and the site on MOPC 460 IgA were closer to those predicted by the model for MOPC 315 IgA.     

Glycosylation causes an apparent block in translation of immunoglobulin heavy chain

Analysis of nascent heavy chains isolated from MPC11 (gamma 2b heavy chains) and MOPC 21 (gamma 1 heavy chains) mouse myeloma cells demonstrates an accumulation of nascent heavy chains which are slightly smaller in mass (approximately 35,000 daltons) than nascent heavy chains which have just been glycosylated (approximately 38,000 daltons). The accumulation of 35,000-dalton nascent heavy chain appears to be a consequence of the glycosylation process since tunicamycin, an inhibitor of glycosylation, abolishes the apparent translational block manifested by the accumulation of 35,000-dalton nascent chains. Tunicamycin also causes a 15 to 25% increase n the relative rate of synthesis of heavy chain compared to the corresponding rate of synthesis of the nonglycosylated light chain synthesized by the same cell. These results suggest that the translation block, caused by the glycosylation process, of heavy chain synthesis contributes to the imbalance of heavy chain and light chain biosynthesis observed in malignant and normal lymphoid cells.     

Selective removal of alpha heavy-chain glycosylation sites causes immunoglobulin A degradation and reduced secretion.

The importance of carbohydrate in the secretion of immunoglobulin A (IgA) has previously been suggested by results of studies with tunicamycin, which prevents N-linked glycosylation of all cell glycoproteins. To directly evaluate the role of individual oligosaccharides in the secretion of IgA, we have used site-directed mutagenesis to selectively eliminate the two N-linked attachment sites reported to be glycosylated in alpha heavy chains. Transfected wild-type and mutant alpha genes were expressed in kappa light-chain-producing MPC-11 variant myeloma cells, and secretion kinetics of the IgAs were compared. Removal of either or both glycosylation sites led to intracellular alpha heavy-chain degradation and a 90 to 95% inhibition of IgA secretion. These results reveal that both N-linked oligosaccharides of the alpha heavy chain are essential for intracellular stability and normal secretion of IgA. This suggests that the key function of carbohydrate here is to maintain proper conformation of the glycoprotein. We also found that when expressed in the MPC-11 variant cells, alpha heavy chains were glycosylated at a third, normally unused site.     

Binding of antigen-bearing fluorescent liposomes to the murine myeloma tumor MOPC 315.

Small unilamellar lipid vesicles bearing the DNP-hapten on their surfaces and containing the water-soluble fluorescent dye carboxyfluorescein were formed by sonication. These vesicles were incubated with cells from the murine myeloma tumor MOPC 315, which secrete and also bear on the cell surface an immunoglobulin with affinity for the nitrophenyl hapten. At 0 degrees C the cells bound an average of several thousand vesicles at saturation. This binding was specific for the nitrophenyl hapten on the vesicle since it was abolished by an excess of soluble nitrophenyl derivative, by omission of the hapten from the vesicle, or by substitution for MOPC 315 of a tumor lacking receptors for the nitrophenyl hapten. Specific binding of vesicles was greater when cells were incubated at 37 degrees C. The study suggests that ligand-bearing vesicles can be a useful marker for cell surface immunoglobulin. However, in spite of the ability to "target" vesicles to cell surface determinants, binding did not result in increased delivery of vesicle contents to the cytoplasm.     

Studies of the biosynthesis of the mannose 6-phosphate receptor in receptor-positive and -deficient cell lines

Phosphomannosyl residues present on lysosomal enzymes are specifically recognized by the mannose 6-phosphate receptor protein. This interaction results in the selective targeting of lysosomal enzymes to lysosomes. While this pathway is operative in many cell types, we have found four cultured cell lines that are deficient in the ability to bind lysosomal enzymes containing phosphomannosyl residues to their intracellular or surface membranes (Gabel, C., D. Goldberg, and S. Kornfeld, 1983, Proc. Natl. Acad. Sci. USA, 80:775-779). These cells appear to segregate lysosomal enzymes by an alternate intracellular pathway. To determine the basis for the lack of mannose 6-phosphate receptor activity in these cell lines, we studied the biosynthesis of the receptor in receptor-positive (BW5147) and receptor-deficient (P388D1 and MOPC 315) cells. The cells were labeled with [2-3H]mannose or [35S]methionine and the receptor was immunoprecipitated with an antireceptor antiserum. BW5147 cells synthesize a receptor protein whose size increases after translation/glycosylation. MOPC 315 cells produce an apparently normal receptor and degrade it rapidly. P388D1 cells fail to synthesize any detectable receptor. The receptor from BW5147 and MOPC 315 cells is a glycoprotein with both high mannose and complex asparagine-linked oligosaccharides. The complex-type units become fully sialylated and remain so during long periods of chase.     

Glucose-dependent glycosylation of secretory glycoprotein in mouse myeloma cells.

The mouse myeloma tumor, MOPC-46, produces a kappa-type immunoglobulin light chain that may be isolated from the urine of tumor-bearing animals. This protein possesses a single carbohydrate side chain, attached by glycosylamine linkage to asparagine residue 28. When viable single cell suspensions of the tumor are incubated in vitro in minimum essential medium containing sodium pyruvate as a source of carbon and energy, the major protein synthesized and secreted corresponds to a nonglycosylated form of the kappa light chain. However, when glucose or mannose are substituted for sodium pyruvate as a source of carbon, the immunoglobulin light chain is synthesized and secreted in the fully glycosylated, native form. The dependence of normal glycosylation of the protein on the presence of either glucose or mannose in the medium is relatively specific for these compounds since substitution with either fructose, galactose, glycerol, ribose, or N-acetylglucosamine was ineffective. The nonglycosylated protein produced in the presence of sodium pyruvate was characterized as nonglycosylated MOPC-46 light chain by immunoprecipitation and gel electrophoresis. An identical nonglycosylated protein was produced by tumor cells in the presence of glucose when the incubation mixtures contained tunicamycin.     

Influence of the N-linked oligosaccharides on the biosynthesis, intracellular routing, and function of the human asialoglycoprotein receptor

The human asialoglycoprotein receptor (ASGP-R) is a membrane glycoprotein of 46,000 Da which possesses two N-linked oligosaccharide chains (Schwartz, A. L., and Rup, D. (1983) J. Biol. Chem. 258, 11249-11255). In order to examine the role of N-linked oligosaccharides in the biosynthesis, intracellular routing, and function of the ASGP-R, we have used Hep G2 cells, which have a large number of ASGP-R, and two inhibitors of glycosylation, swainsonine and tunicamycin. In the presence of swainsonine, newly synthesized ASGP-R is a 43,000-Da species which is endoglycosidase H-sensitive, appears on the Hep G2 cell surface, and specifically binds 125I-asialoorosomucoid (ASOR). In the presence of tunicamycin newly synthesized ASGP-R is a 34,000-Da nonglycosylated species which appears on the Hep G2 cell surface where it specifically binds 125I-ASOR. There is no major effect on subsequent uptake and degradation of 125I-ASOR in cells whose ASGP-R was synthesized in the presence of tunicamycin. The turnover of ASGP-R synthesized in the presence of either swainsonine or tunicamycin is not significantly altered from that found for the normal 46,000-Da species. Thus, it appears that the two N-linked oligosaccharide chains of the human ASGP-R do not play a major role in the intracellular routing, turnover, or function of ASGP-R.     

The allogeneic effect on tumor growth. II. Suppression of both ascitic and solid MOPC 315 plasmacytoma by the graft-vs-host reaction, with pathologic correlation.

The growth of an ascitic murine plasmacytoma, MOPC 315, can be retarded in CAF1 hybrid host mice by the i.p. injection of donor lymphoid cells. The graft-vs-host reaction can be established by a variety of donor cells, including parental BALB/c and A/J and congenic inbred B10.D2 which share the major histocompatibility locus with BALB/c(H-2d). Optimal results are consistently obtained when parental BALB/c spleen cells are injected before tumor inoculation, and a second dose of donor spleen cells injected 1 week later. This aloogeneic effect on tumor growth is manifested by delayed appearance of the tumor and prolonged host survival. Pathologic studies on the ascites tumor indicated that the allogeneic effect suppresses the initial appearance and early growth of the plasmacytoma. However, once established, MOPC 315 grows rapidly and fatally in both control mice and recipients of donor lymphoid cells. Further, a subcutaneous implant of MOPC 315 is suppressed by an allogeneic effect established either i.v. with BALB/c spleen cells before tumor inoculation or by BALB/c spleen cells administered subcutaneously at the time of MOPC 315 implant. Thirty percent of mice treated by i.v. or subcutaneous donor lymphoid cells were tumor free at 150 days after tumor inoculation.     

Gene transfer of immunoglobulin light chain restores heavy chain secretion

Several lines of evidence suggest that immunoglobulin (Ig) light (L) chain plays a role in the secretion of heavy (H) chain. For example, myeloma variant lines, which synthesize the Ig H chain but not the L chain, fail to secrete H chain protein. Here we have tested directly the role of chain assembly in the control of Ig secretion by the transfer of functional L chain genes into two such L chain-defective myeloma mutants. A lambda 2 or kappa L chain gene was introduced into variant lines of the mouse myelomas MOPC 315 (IgA, lambda 2) or PC7 (IgM, kappa), respectively. Although the two mutant lines are unable to secrete the H chain they produce, rescue of secretion of complete Ig protein molecules (IgA or IgM) was observed after transfection. These results imply that the secretory apparatus of these cells is intact and that the failure to secrete free H chain reflects a structural feature intrinsic to that protein. The implications of these results with respect to control of secretion of multi-subunit proteins are discussed.     

Secretion of an active nonglycosylated form of the repressible acid phosphatase of Saccharomyces cerevisiae in the presence of tunicamycin at low temperatures

The role of mannan chains in the formation and secretion of active acid phosphatase of yeast (Saccharomyces cerevisiae), a repressible cell surface mannoprotein, was studied in yeast protoplast systems by using tunicamycin at various temperatures. At 30 degrees C, tunicamycin-treated protoplasts did not produce active acid phosphatase; however, at 25 or 20 degrees C they formed and secreted active enzyme. This form of acid phosphatase gave 59-, 57-, and 55-kDa bands on SDS-PAGE which neither bound to concanavalin A Sepharose, nor changed in molecular weight upon treatment with endoglycosidase H, indicating that the peptides are nonglycosylated. The nonglycosylated form, like its glycosylated counterpart, is a dimer on the basis of gel permeation chromatography. The Km for para-nitrophenyl-phosphate and Ki for inorganic phosphate of both glycosylated and nonglycosylated acid phosphatases were almost the same. These results suggested that 1) the conformation of the nonglycosylated acid phosphatase secreted at low temperatures is probably identical with that of the glycosylated one, and 2) the conformation of acid phosphatase is very important for its secretion. The rate of intracellular transport of nonglycosylated acid phosphatase is about one-fourth that of the glycosylated enzyme, indicating that glycosylation facilitates the transport of acid phosphatase proteins.