Carbohydrate chains on yeast carboxypeptidase Y are phosphorylated


Carbohydrate chains on yeast carboxypeptidase Y are phosphorylated

(glycoprotein high mannose oligosaccharides/nuclear magnetic resonance/a-mannosidase/phosphodiester)


Department of Biochemistry, University of California, Berkeley, California 94720

Contributed by Clinton Ballou, January 12, 1981


Carboxypeptidase Y, a vacuolar enzyme from Saccharomyces cerevisiae, was digested with endo-fi-N-acetyl-Dglucosaminidase H to release the four oligosaccharide chains that are linked to asparagine in the glycoprotein. The oligosaccharides were fractionated into a neutral and acidic component, and the latter proved to phosphorylated. From its gel filtration pattern, the neutral fraction was shown to be a mixture of at least four homologs, the smallest of which had a proton NMR spectrum almost identical to that given by an IgM oligosaccharide with eight mannoses and one N-acetylglucosamine [Cohen, R. E. & Ballou, C. E. (1980) Biochemistry 19, 4345-4358]. The yeast oligosaccharide has one additional mannose unit in an al-+3 or al-*6 linkage, whereas the larger homologs appear to have two, three, and four more mannose units. One phosphorylated oligosaccharide with a mannose/phosphate ratio of 12.5 was reduced with NaB3H4 and then subjected to mild acid hydrolysis. This released mannose and mannobiose that were glycosidically linked to the phosphate group, whereas complete acid hydrolysis yielded Dmannose 6-phosphate. The recovered oligosaccharide phosphomonoester, which contained 11 or 12 mannose units, was digested exhaustively with a-mannosidase, and the product of this reaction was treated with alkaline phosphatase, which yielded radioactive Man3GlcNAcH2. These results suggest that the mannosidase-resistant phosphorylated oligosaccharide has the structure Man-- P--6aMan--aMan--6fBMan--s4GlcNAcH2, in which some of the phosphate groups are substituted with mannobiose instead ofmannose. A second phosphorylated oligosaccharide with a mannose/ phosphate ratio of 6.5 probably contains two phosphodiester groups, but its structure has not been investigated in detail.

Proc. Natl. Acad. Sci. USA Vol. 78, No. 4, pp. 2244-2248, April 1981.


The yeast Saccharomyces cerevisiae has extracellular periplasmic glycoproteins (mannoproteins), such as invertase, that contain =50% mannose (1) and intracellular glycoproteins, such as carboxypeptidase Y, that contain 15% mannose (2) and are similar in some respects to mammalian glycoproteins (3). The asparagine-linked carbohydrate chains of the external invertase contain 100-150 mannose units that are differentiated into a core unit of 11 mannoses and an outer chain of 90-140 mannoses (4). From the composition and number of the carbohydrate chains in carboxypeptidase Y, the average chain size appears to be -13 mannoses (5)-i.e., the size ofa core unit. From these facts, one might postulate that the externalization of a mannoprotein is associated with or dependent on the addition of the outer chain and that intracellular mannoproteins lack a signal that specifies outer-chain addition (6). Although it is an attractive idea, a strong argument against this hypothesis is our recent isolation of S. cerevisiae mutants that make and secrete mannoproteins, including invertase, that appear to possess only the core oligosaccharide units (7).

To determine whether the carbohydrate chains have any role in specifying the compartmentation of glycoproteins in yeast, we have undertaken a detailed comparison of the core oligosaccharide units from intracellular and extracellular mannoproteins. This preliminary report on the carbohydrate chains from carboxypeptidase Y shows that they are similar to oligosaccharides from mammalian glycoproteins and that some ofthe chains are phosphorylated. Because carboxypeptidase Y is found in the yeast vacuole (8), an organelle somewhat analogous to the lysosome of higher organisms, the phosphorylation of the carbohydrate chains may have a role in determining the localization of the enzyme, as it appears to have with some lysosomal enzymes (9).

Materials and methods

Materials. Ion exchange resins and gel filtration materials came from Bio-Rad and phenylmethanesulfonyl fluoride (toluenesulfonyl fluoride) was froim Sigma. Bacterial alkaline phosphatase was from Worthington, jack bean a-mannosidase was prepared according to Li (10), and endo-p-N-acetyl-D-glucosaminidase H was purified from Streptomyces plicatus (American Type Culture Collection 27800) according to Tarentino et al. (11). Carboxypeptidase Y was isolated from compressed bakers' yeast (Universal Foods, San Francisco) by the procedure of Kuhn et al. (12). Reference oligosaccharides were from an earlier study (13).

Methods. Alkaline phosphatase digestion was done for 24 hr in 1.0 M Tris-HCl (pH 8.0); a-mannosidase digestion was for 24 hr in 50mM NaOAc, pH 4.5/0.1 mM ZnCl2; and endoglucosaminidase H digestion was for 96 hr in 50 mM Na citrate (pH 5.5). All reactions were carried out at 37°C. Carboxypeptidase Y was treated with toluenesulfonyl fluoride according to Kuhn et al. (14). Mannose 6-phosphate was determined by a coupled enzyme assay (15) as modified by Hashimoto et al. (16). A correction was made for the amount of NADPH produced by an acid hydrolysate of bovine serum albumin. Partial acetolysis of oligosaccharides was carried out in acetic acid/acetic anhydride/ sulfuric acid, 10:10:1 (vol/vol) at 40°C for 7 hr (17), and the recovered acetylated products were deacetylated in methanol containing Na methoxide. Mild acid hydrolysis to cleave glycosylphosphate bonds was done in 0.01 M HC1 at 100°C for 30 min. Carboxypeptidase Y was methylated by a standard procedure (13).

Oligosaccharide samples (50-500 ,ug) were reduced in 500 *,l of 0.1 M NaHCO3 by adding 200 uCi (1 Ci = 3.7 X 10"° becquerels) of NaB3H4 as a dimethyl sulfoxide solution (4 mCi/ ml). After 12 hr at 23°C, 5 mg of NaBH4 was added and the reaction was allowed to proceed for another 12 hr. Dowex 50 (H+) was added until effervescence ceased, the tube contents were filtered through a Pasteur pipette plugged with glass wool, and the filtrate was evaporated to dryness under reduced pressure. One drop of glacial acetic acid and 1 ml of methanol were added to the residue, and the mixture was refluxed for 1 min and then evaporated to dryness under a stream ofN2. This treatment was repeated four times. For neutral samples, an acidic noncarbohydrate radioactive contaminant was removed by passing a water solution of the reduced product through a small column of Dowex 1 (acetate). Acidic reduced oligosaccharides were purified on a Bio-Gel P-4 (>400 mesh) column (1 X 25 cm) in 0.1 M acetic acid. Finally, the reduced compounds were desalted by passage through a Sephadex G-10 column (0.5 x 25 cm) in water.

Oligosaccharide sizes were estimated by gel filtration on a calibrated Bio-Gel P-4 column (>400 mesh) (0.6 x 115 cm) in 0.1 M NH4OAc containing 0.02% NaN3 to prevent microbial growth. Fractions of 3 drops were collected, and the elution positions were determined relative to the reference compounds ,/Man-*4GlcNAcH2, aMan--3(3Man-*4GlcNAcH2, aMan--> 2aMan--2aMan--3/3Man--*4GlcNAcH2, Man6GlcNAcH2 (from ovalbumin) (13), Man8GlcNAcH2 (from IgM) (13), and Man,,GlcNAcH2 (from yeast mannoprotein) (7). On this same column, bovine serum albumin was eluted at fraction 60 and mannose was eluted at fraction 176.

Polyacrylamide gel (10%)/0. 1% NaDodSO4 electrophoresis was done by the discontinuous system of Laemmli (18), with a running buffer ofTris HCl (pH 7.2). Samples were reduced with 1% mercaptoethanol/1% NaDodSO4 at 100°C for 3 min. Gels were stained for protein with Coomassie brilliant blue R-250. 'H NMR and 31P NMR were carried out as described (13, 16).


Isolation of Carboxypeptidase Y and Digestion with Endoglucosaminidase H. From 100 lb ofyeast cake, =500 mg ofpure carboxypeptidase Y was obtained. As noted by Kuhn et al. (12), two peaks of enzyme activity were found at the DEAE-Sephadex A-50 isolation step, here designated CPY-I and CPY-II. These two peaks were separated, and each gave a single peak when it was rechromatographed. The carbohydrate content of CPY-I was 14.4% and that of CPY-II was 14.7%. Hasilik and Tanner (5) have reported 14.1% carbohydrate for the enzyme. Both carboxypeptidase preparations contained organic phosphate, the mannose/phosphate molar ratios being 12.5 for CPYI and 10. 8 for CPY-II. Both fractions ofcarboxypeptidase Y were inactivated with toluenesulfonyl fluoride (14) to prevent autodigestion and digestion of the endoglucosaminidase H during treatment to release the carbohydrate chains. The yield of inactivated CPY-I was 260 mg and that of inactivated CPY-II was 230 mg. Both preparations gave single major bands that were indistinguishable on acrylamide/NaDodSO4 gel electrophoresis (Fig. 1).

Methylation of intact CPY-I, followed by hydrolysis, reduction, and acetylation of the methylated derivative, yielded the alditol acetates of 2,3,4,6-tetra-O-methylmannose (5 mol), 2,4,6-tri-0-methylmannose (3 mol), 3,4,6-tri-0-methylmannose (2 mol), 2,4-di-0-methylmannose (1 mol), and 3,4,-di-0- methylmannose (2 mol). The two di-O-methyl derivatives are those expected for the branching at positions 2 and 3 ofthe 1-*6- linked backbone mannose that is characteristic of yeast glycoprotein core oligosaccharides (13). Also typical are the two tri- 0-methyl ethers that reflect the presence of mannose in 1-*2 and 1--3 linkage. Samples of inactivated CPY-I and CPY-II (150 mg each) were digested with 0.75 unit ofendoglucosaminidase H for 96 hr (19). As shown in Fig. 1, both carboxypeptidase preparations were converted quantitatively to single faster-migrating protein bands that were again indistinguishable. The enzyme digests were then fractionated on a Bio-Gel P-30 (100-200 mesh) column (2 x 90 cm) by elution with water (Fig. 2). The separated protein peak (fractions 23-31) contained very little mannose; 95% of the mannose was distributed in a neutral peak (fractions 78-93) and a broader peak (fractions 32-71) associated with most of the phosphate. The latter was divided into three fractions: fraction a with a mannose/phosphate ratio of6.5, fraction b with a ratio of 8.0, and fraction c with a ratio of 12.5. These results suggest that fraction a has two phosphate groups per oligosaccharide chain, that fraction c has one, and that fraction b is an overlapping mixture of fractions a and c.

Nature of the Neutral Oligosaccharide. The material in fractions 78-93 (fraction CPY-IId; see Fig. 2) was combined, and its neutral nature was confirmed by the absence of phosphate and the failure ofthe carbohydrate to bind when passed through a Dowex 1 (acetate) column. The sample, chromatographed on a Bio-Gel P4 (>400 mesh) column (2 X 190 cm) by elution with water, had at least four components that appeared to differ from each other by increments of about one mannose unit (Fig. 3A). The following characterization was limited to the smallest oligosaccharide of the mixture.

Fractions 114-121 (see Fig. 3A) were combined, and the material was reduced with NaB3H4 (see Materials and Methods). This reduced sample was then chromatographed on the calibrated Bio-Gel P-4 column, from which it eluted in the position ofMan9GIcNAcH2 (Fig. 3B). From this, we conclude that the larger oligosaccharides in Fig. 3A range in size from Man11GlcNAc to Man13GlcNAc.

The MangGlcNAcH2 sample gave an anomeric proton NMR spectrum (Fig. 4) that was almost identical to that of an oligosaccharide Man8GlcNAcH2 obtained from IgM and designated rGP-563-I (13). This latter oligosaccharide has the structure shown in bold type in Fig. 4. The carboxypeptidase oligosaccharide has one more mannose unit, however, and, from the shoulder (g) on peak b, we conclude that about halfofthe chains have an unsubstituted al-46-linked mannose, whereas the excess area under peak a,e,e' would be consistent with about half of the chains having an additional terminal al-*3-linked mannose. The postulated modifications are shown in italics on the structure in Fig. 4.

Nature of the Phosphorylated Component of Carboxypeptidase Y. CPY-I and CPY-II were hydrolyzed in 1 M trifluoroacetic acid at 120°C for 2 hr. The acid was evaporated under a stream of N2, and the residue was assayed enzymically for Dmannose 6-phosphate (15, 16). The results (Table 1) show that the phosphate was almost quantitatively accounted for as mannose 6-phosphate.

Phosphorylated oligosaccharide CPY-IIc [fractions 56-71 (see Fig. 2)] was passed through a Bio-Gel P-4 (>400 mesh) column (2 X 190 cm) and eluted with 0.1 M NH4OAc in 3-ml fractions. The majority of the carbohydrate appeared in a symmetrical peak between fractions 97 and 113. The combined material was analyzed by proton-decoupled 31P NMR, and it gave a single signal at -1.7 ppm (pH 7.0 in 0.1 M Na EDTA), a chemical shift characteristic of phosphate diesters (20). S. cerevisiae mannoproteins often contain diesterified phosphate in which one component of the diester is mannose or mannobiose in glycosyl linkage that is subject to release by mild acid hydrolysis (21). Phosphorylated CPY-IIc oligosaccharide was heated at 100TC for 30 min in 0.01 M HCI, and the products were separated on a Bio-Gel P-2 (>400 mesh) column (0.2 x 114 cm) by elution with 0.1 M NH4OAc (Fig. 5). Peaks corresponding to mannose and mannobiose were obtained, and their identities were confirmed by paper chromatography (not shown). The weight ratio oligosaccharide residue/mannose plus mannobiose was 21, whereas a ratio of about 12 would be expected if all of the phosphate groups were esterified with mannose. Incomplete hydrolysis of the diester bond is the most likely explanation for this result because subsequent phosphatase digestion gave only a 70% yield of neutral oligosaccharide. In confirmation ofthis conclusion, a second mild acid hydrolysis of the recovered oligosaccharide gave sufficient additional mannose to bring the weight ratio oligosaccharide/mannose plus mannobiose to 10.6.

A sample of the acidic CPY-IIc oligosaccharide was reduced with NaB3H4, and the product was subjected to mild acid hydrolysis and treatment with alkaline phosphatase to dephosphorylate the monoester oligosaccharide component. This gave a neutral radioactive oligosaccharide that was separated into two peaks on the calibrated Bio-Gel P-4 column. These two appear to differ by one mannose unit, and their sizes correspond to Man11GlcNAcH2 and Man12GlcNAcH2 (not shown).

Location of the Mannose 6-Phosphate in the Acid-Stable Oligosaccharide. Strong acid hydrolysis of intact carboxypeptidase Y gives mannose 6-phosphate (see above), so this structure must be present in the acidic oligosaccharide recovered after mild acid hydrolysis. The location of the mannose phosphate was determined by subjecting the NaB3H4-reduced oligosaccharide to exhaustive digestion with a-mannosidase. The recovered acidic radioactive product was dephosphorylated with alkaline phosphatase and then fractionated on the calibrated Bio-Gel P-4 column (Fig. 6). It eluted in the position of Man3GlcNAcH2. Assuming the radioactive product was derived from an oligosaccharide such as that in Fig. 4, this suggested three possible structures: aMan-*6aMan--36,Man-> 4GlcNAcH2, aMan--2(or 3)aMan--6,8Man--)4GlcNAcH2, and aMan--2aMan--3f3Man--4GlcNAcH2. To discriminate between these, we used partial acetolysis (17); only the oligosaccharides with 1--6 linkages would be degraded readily in this reaction. As shown in Fig. 6, the acetolysis ofMan3GlcNAcH2 labeled only in the glucosaminitol residue yields a radioactive product having the elution position of ManGlcNAcH2. This is consistent with either of the first two structures above. Because the original phosphorylated oligosaccharide (CPY-TIc) contains 11 or 12 mannoses, however, the backbone probably has four 1-+6-linked mannose units, which would preclude attachment of phosphate to position 6 of the third mannose in the mannosidase- resistant tetrasaccharide. From this consideration, we favor the structure aMan--2 (or 3) aMan- 6,3Man-*4GlcNAcH2 for this oligosaccharide.


Yeast carboxypeptidase Y contains phosphate in diester form that is linked as mannosylphosphate and mannobiosylphosphate units to position 6 of mannose in some of the oligosaccharide units of the glycoprotein. These observations show a striking parallel between this yeast vacuolar hydrolytic enzyme and the lysosomal enzymes of higher organisms, some of which are also phosphorylated (9). Whether the phosphorylated oligosaccharide chains ofcarboxypeptidase Y specify the localization of this enzyme in the vacuole is unknown, but it is certain that the mere presence of phosphate in a mannoprotein does not determine location; some secreted yeast mannoproteins also carry mannosylphosphate and mannobiosylphosphate groups in the outer chain (22). The novelty in the present report is the discovery ofphosphate in the core oligosaccharide ofan intracellular yeast mannoprotein.

It has been reported (12), and we confirm, that partially purified carboxypeptidase Y is resolved on DEAE-Sephadex into two peaks of enzyme activity. It is possible that this separation is based on different phosphate contents. The molar ratios of mannose/phosphate were 12.5 in CPY-I and 10.8 in CPY-II. As carboxypeptidase contains =50 mannose units per molecule (5), this means that there may be four phosphates in CPY-I and five in CPY-LI. The additional phosphate group in CPY-II would be consistent with its later elution from DEAE-Sephadex. In our isolation, the ratio CPY-I/CPY-II was -0. 8, whereas Kuhn et al. (12) found a ratio of about 0.5. The finding of four or five phosphate groups per enzyme molecule and the fact that one-fourth ofthe four oligosaccharide chains are not phosphorylated suggests that some oligosaccharide chains must carry more than one phosphate group. The elution pattern of the oligosaccharides from the Bio-Gel P-30 column does reveal heterogeneity in the extent of phosphorylation, and the early fraction (CPY-IIa) appears to have about twice as much phosphate as the later one (CPY-IIc). By 31P NMR we have shown that all of the phosphate in carboxypeptidase Y is diesterified and, because strong acid hydrolysis yields mannose 6-phosphate quantitatively, it is clear that the phosphate is esterified to position 6 of mannose units in some of the oligosaccharide chains. In most or all of the phosphodiester units, the other attached group is mannose or mannobiose in an acid-labile glycosylphosphate linkage. The phosphorylated mannose unit ofthe carboxypeptidase Y oligosaccharide CPY-IIc is probably attached as a side chain to the first al-+6-linked mannose unit of the backbone by an al-*2 or al-*3 linkage (Fig. 7). Ifthe phosphorylated oligosaccharides are related structurally to those of the outer chain (22), the linkage is probably al-42. This structure has obvious similarity to that reported by Tabas and Kornfeld (23) for a mouse lymphoma P-glucuronidase in which N-acetylglucosamine phosphate is esterified to the core oligosaccharide.

All of the carbohydrate chains on carboxypeptidase Y are not phosphorylated, and one of the neutral oligosaccharides was shown to have a structure almost identical to that of an IgM chain with eight mannoses (13). The unphosphorylated carboxypeptidase Y oligosaccharide fraction (CPY-IId) is composed of at least four homologs that appear to have 9, 11, 12, and 13 mannoses, respectively. This neutral fraction makes up about one-fourth of the total oligosaccharide component and could all be localized at a particular site in the polypeptide chain, but we have not investigated this feature. Methylation of intact carboxypeptidase Y showed the presence of 2- and 3-0-substituted, as well as 2,6- and 3,6- doubly 0-substituted mannoses. Because the carbohydrate chains are heterogeneous, the observed ratios of these units do not fit any single structure well, but all of the linkages are expected in the yeast core oligosaccharides. The somewhat high proportion of 1-3-linked mannose is consistent with the presence of this linkage in the mannobiose unit previously found attached to phosphate in the mannoprotein outer chain (22).

That the observed heterogeneity of the carboxypeptidase-Y oligosaccharides results from phosphatase and mannosidase digestion during isolation of the enzyme seems unlikely for several reasons. Some of the neutral oligosaccharides are larger than the corresponding part ofthe acidic ones; none ofthe oligosaccharides contains monoesterified phosphate, which would presumably be an intermediate in any degradative pathway; and the smallest neutral oligosaccharide detected contained nine mannose units, whereas smaller oligosaccharides are readily formed by digestion with jack bean a-mannosidase. Yeast mannoproteins show a high species polymorphism (24), and the carbohydrate chains of the external mannoprotein, invertase, are known to be heterogeneous in size (25) and in the structure of the core oligosaccharides (4). As we show in the present study, carboxypeptidase Y, an intracellular mannoprotein that possesses only core oligosaccharides, has a very heterogeneous assortment ofcarbohydrate chains that differ in both size and state of phosphorylation. An important aim in extending this study will be to elucidate the origin and function of this diversity.


1. Gascon, S., Neumann, N. P. & Lampen, J. 0. (1968) J. Biol. Chem. 243, 1573-1577.

2. Aibara, S., Hayashi, R. & Hata, T. (1971) Agric. Biol. Chem. 35, 658-666.

3. Kornfeld, R. & Kornfeld, S. (1976) Annu. Rev. Biochem. 45, 217-237.

4. Lehle, L., Cohen, R. E. & Ballou, C. E. (1979)J. Biol. Chem. 254, 12209-12218.

5. Hasilik, A. & Tanner, W. (1978) Eur. J. Biochem. 91, 567-575.

6. Eylar, E. H. (1965)J. Theor. Biol. 10, 89-113.

7. Ballou, L., Cohen, R. E. & Ballou, C. E. (1980) J. Biol. Chem. 255, 5986-5991.

8. Matile, P. & Wiemken, A. (1967) Arch. Mikrobiol. 56, 148-155.

9. Kaplan, A., Achord, D. T. & Sly, W. S. (1977) Proc. Nati. Acad. Sci. USA 74, 2026-2030.

10. Li, Y.-T. (1967) J. Biol. Chem. 242, 5474-5480.

11. Tarentino, A. L., Trimble, R. B. & Maley, F. (1978) Methods Enzymol. 50, 574-580.

12. Kuhn, R. W., Walsh, K. A. & Neurath, H. (1974) Biochemistry 13, 3871-3877.

13. Cohen, R. E. & Ballou, C. E. (1980) Biochemistry 19, 4345-4358.

14. Kuhn, R. W., Walsh, K. A. & Neurath, H. (1976) Biochemistry 15, 48814885.

15. Gawehn, K. (1974) Methods Enzymatic Anal.. 3, 1263-1267.

16. Hashimoto, C., Cohen, R. E. & Ballou, C. E. (1980) Biochemistry 19, 5932-5938.

17. Rosenfeld, L. & Ballou, C. E. (1974) Carbohydr. Res. 32, 287-298.

18. Laemmli, U. K. (1970) Nature (London) 27, 680-685.

19. Trimble, R. B. & Maley, F. (1977) Biochem. Biophys. Res. Commun. .78, 935-944.

20. Van 'Wazer, J. R., Callis, C. F., Shoolery, J. N. & Jones, R. C. (1956) 1. Am. Chem. Soc. 78, 5715-5726.

21. Rosenfeld, L. & Ballou, C. E. (1974) J. Biol. Chem. 249, 2319-2321.

22. Ballou, C. E. (1976) Adv. Microb. Physiol. 14, 93-158.

23. Tabas, I. & Kornfeld, S. (1980) J. Biol. Chem. 255, 6633-6639.

24. Ballou, C. E ..& Raschke, W. C. (1974) Science 184, 127-134.

25. Tarentino, A. L., Plummer, T. H., Jr. & Maley, F. (1974)J. Biol. Chem. 249, 818-824.