ARCHIVES
OF BIOCHEMISTRY
Synthesis
AND
BIOPHYSICS
167,151-160
(1975)
of The Mannosyl-O-Serine
Glycoproteins
from Polyisoprenylphosphate (Hansen&a
of Chemistry,
Biochemistry
Mannose
Linkage
of
in Yeast
holstiil 1
R. K. BRETTHAUER2 Department
(Threonine)
AND
and Biophysics Program, Indiana 46556 Received September
S.
WV
University
of Notre
Dame, Notre Dame,
9, 1974
The mannolipid synthesized from GDP-mannose and lipid acceptors in a particulate enzyme preparation from the yeast Hansenula holstii (R. K. Bretthauer, S. Wu, and W. E. Irwin, (1973) Biochim. Biophys. Acta, 304, 736-747) has the properties of dolicholmonophosphate mannose. Transfer of [“Clmannose from exogenously supplied, purified mannolipid to endogenous protein acceptors of the particulate enzyme fraction has now been demonstrated. The synthesis of radioactive products which are insoluble in chloroformmethanol and water is dependent upon time and concentrations of substrate, particulate fraction protein, and detergent. Addition of MgCl, or MnCl, to incubation mixtures prepared in the absence of these ions had only small stimulatory effects (20-25%), suggesting that the reaction is not dependent upon metal ions. Relatively high concentrations (0.005 ~-0.05 M) of EDTA did partially inhibit the reaction, but this is considered to be due to secondary effects. Seventy percent of the radioactivity in the chloroform-methanol insoluble residue was solubilized with hot, neutral citrate buffer. The chromatographic properties of this material on Sephadex gels and on DEAE-Sephadex were very similar to the properties of glycoprotein products derived from GDP-[“Clmannose. The chloroform-methanol insoluble products were also solubilized with Pronase which subsequently resulted in the isolation of a radioactive glycopeptide that contained 25% of the radioactivity transferred from mannolipid. The radioactive component of this glycopeptide was shown by @-elmination experiments and by amino acid analyses to be [“Clmannose residues linked O-glycosiditally to serine and threonine residues. It was concluded, therefore, that one function of the mannolipid is to serve as mannosyl donor in the synthesis of the mannosyl-O-serine (threonine) linkage region of glycoproteins which may be part of the cell wall mannanprotein complex. Other mannose-containing products may also be synthesized from the mannolipid, as p-elimination of the chloroform-methanol insoluble fraction or of the Pronase soluble fraction did not result in recovery of all of the radioactivity as [“Clmannose.
The role in bacteria of polyisoprenol phosphates as glycosyl carriers from sugar nucleotides to polysaccharides has been established (1). Studies from several laboratories now suggest that lipids (dolichols)
may also play a role as intermediate glycosyl-carriers from sugar nucleotides in glycosylation reactions of eucaryotic cells (2-6). In particular, a mannosylated lipid characterized as dolichol phosphate mannose (7) has been implicated by several investigators as an intermediate glycosyl carrier in the synthesis of yeast cell wall mannan from GDP-mannose (8-11). The mannan component of the cell wall of many yeasts is generally considered to be a
* Supported by NSF grant GB 38356, and a grant from Miles Laboratories, Inc. Part of this work is from the doctoral thesis of S. W. 2 To whom correspondence should be addressed. 3 Present address: Department of Biochemistry, North Carolina State University, Raleigh, NC 27607. 151 Copyright 0 1975 by Academic Press, Inc. All rights of reproduction in any form reserved.
152
BRE’ITHAUER
polysaccharide-protein complex which has two kinds of carbohydrate units. One is a large, branched polysaccharide of mannose residues which is attached through Nacetylglucosamine to an asparagine residue of the protein moiety (12). The other is a short oligosaccharide of mannose residues which is linked to serine and/or threonine residues in the protein (12, 13). Therefore, several different kinds of glycosidic linkages involving mannose could possibly be formed with a mannolipid as glycosyl donor. We have recently demonstrated that, with particulate enzymes and acceptors from Hansen& holstii, GDP-mannose is a mannosyl donor to short oligosaccharides of mannose which are released from glycopeptides by alkali with a loss of serine and threonine residues (14). This characteristic p-elimination reaction for the glycosyl-Oserine (threonine) linkage can thus be used to determine which type of carbohydrate unit in mannan (or other glycoproteins) is being synthesized (or elongated) from radioactive substrates. This report describes experiments which demonstrate that the mannolipid previously characterized from H. ho&ii (11) is a mannosyl donor in the synthesis of the mannosyl-O-serine (threonone) linkage of glycoproteins endogenous to the particulate enzyme fraction.’ During the completion of this work, reports (15-17) have appeared from Tanner’s laboratory which also demonstrate that, in membrane preparations from Saccharomyces cereuisiue, mannosyl residues which are released from protein by b-elimination are derived from dolicholmonophosphate mannose. MATERIALS
AND
METHODS
Synthesis and purification of [V]mannolipid.” The particulate enzyme fraction from H. holstii NRRL-Y 2448 was prepared as previously described ’ Preliminary accounts of these studies have been presented at The Sixth Annual Miami Winter Symposium, January 14-18, 19’74, Miami, FL, and at the Biochemistry/Biophysics 1974 Meeting, June 2-7, 1974, Minneapolis, MN. ‘Throughout this report, the mannosylated lipid substrate which has the properties of dolicholmonophosphate [“Clmannose will be referred to as [“C]mannolipid.
AND WU (18). Phospholipids, extracted from acetone-dried H. hoktii cells as described previously (ll), were saponified in chloroform:methanol (2:l v/v) with 0.1 N NaOH for 5 min at 25°C before being using as a source of polyisoprenylphosphate for [“Clmannolipid synthesis. The synthesis of [L’C]mannolipid was carried out as follows: phospholipid (6 pmoles phosphorus) and 0.15 mmoles of Mg. EDTA were mixed and evaporated to dryness (19); then 0.75 mmole of Tris.HCl buffer, pH 7.5, 0.15 mmole of MgCI,, 90 mg of Triton X-100, 3.1 nmole of GDP-[“Clmannose (160 mCi/mmole) and 60 mg of particulate enzyme protein were added in the order indicated to make a total volume of 15 ml. After incubation at 25°C for 10 min, the reaction was terminated by addition of 90 ml of chloroform:methanol (2:l v/v). Four incubation mixtures (15 ml each) containing unlabeled GDP-mannose (0.2 rmolell5 ml) were prepared and the chloroform: methanol extracts combined with that from the mixture containing GDP[“Clmannose. The [“Clmannolipid was purified from the chloroform:methanol extracts after evaporation of the solvent by discarding any acetone-soluble material in the residue and discarding any material not soluble in ether:ethanol (1:l v/v). The ether:ethanol soluble material was treated in chloroform:methanol (2:l v/v) with 0.1 N NaOH for 5 min at 25°C. Further purification of the mannolipid was achieved by DEAE-cellulose and Sephadex LH-20 column chromatography as previously described (11). The final radioactive product gave a single radioactivity peak and a coincident spot with iodine vapor after thin layer chromatography with solvents D and E. The specific activity of the [“C]mannolipid, as calculated from the total GDP-mannose concentration, was 0.63 mCi/mmole. All of the radioactivity of the purified [“C]mannolipid was released by mild acid hydrolysis (0.1 N HCl at 100°C for 1 min in 50% methanol) and recovered as a mixture of [“Clmannose and methyl[“Clmannoside. The proton magnetic resonance spectrum (100 MHz) of a concentrated solution of the [“Clmannolipid in CDCl, revealed resonance signals at 4.93 r (protons of olefinic carbon atoms), 7.98 T (protons of methylene groups adjacent to olefinic carbon atoms), 8.34 and 8.42 T (protons of cis and bans methyl groups, respectively, on trisuhstituted olefinic groups), and 8.6-8.9 T and 9.1 r (protons of saturated isoprene units) (20, 21). The areas for the protons on unsaturated isoprene units were close to those calculated. Anomalies in the spectrum, as compared to pig liver dolichols (20). were the ratio of cis:trans isoprene units which was 12:7, and the high absorptions at 8.6-8.9 T and 9.1 T which were present in excess of those expected for dolichol to the extent of approximately eight saturated isoprene units. These anomalies most likely represent impurities, as the areas of the absorption peaks from 8.6-9.1 T relative to
SYNTHESIS
OF YEAST
those areas at lower field for the unsaturated isoprene units varied with different samples. Assay procedures. Incubation mixtures containing [“C]mannolipid as substrate were prepared by evaporating to dryness under nitrogen a mixture of the [“Clmannolipid and Mg.EDTA, followed by addition of Tris.HCl buffer, pH 7.5, MgCl, and Triton X-100. The particulate enzyme fraction was added last. Final concentrations of the components were 10 mM Mg.EDTA, 10 mM MgCl,, 50 mM Tris. HCl buffer, pH 7.5, 1 mg/ml Triton X-100, 4-5 mg/ml particulate fraction protein, and [L’C]mannolipid as indicated in the Figure and Table legends. After incubation at 25”C, 6 vol of chloroform:methanol (2:l v/v) were added and, after thorough mixing, the liquid phases were separated by low speed centrifugation. The liquid. phases were carefully removed from the interphase residue which was then washed twice by centrifugation from chloroform:methanol (2:l v/v) (the volume of each wash being equivalent to 6 times the volume of the original incubation mixture). The residue was then solubilized with a minimal amount of 6% NaOH at 100°C and the solution transferred to a paper disc for determination of radioactivity. The amount of radioactivity recovered from incubation mixtures containing heat denatured (3 min, 100°C) particulate fraction was independent of incubation time and was subtracted from the experimental values. Radioactivity measurements were routinely carried out by applying liquid samples to paper discs, immersing the dried discs in 8 ml of Beckman ReadySolv Solution V, and counting in a liquid scintillation counter. Amino acid analysis. Samples were hydrolyzed in 6 N HCl at 110°C for 24 h and analyzed on a Beckman Amino Acid Analyzer with a single column. Under these conditions, glucosamine eluted at the front of the third solvent between leucine and tyrosine. Chromatography. Paper chromatography was carried out in the descending direction on Schleicher and Schuell No. 589 Green Ribbon paper with the following solvents (in volume ratios): (A) butanol: pyridine: water (6:4:3); (B) methyl ethyl ketone: acetic acid: boric acid saturated-water (9:l:l); (Cl isobutyric acid: ammonium hydroxide: water (57:4:39). Sugars were located with the periodate-benzidine reagent. Thin layer chromatography was carried out on glass plates containing Adsorbosil-1 (Applied Science Laboratories) with the following solvents: (D) chloroform:methanol:acetic acid:water (30:15:4:2); (El diisobutyl ketone: acetic acid: water (20:15:2); F, chloroform: methanol: water (65:25:4). Lipids were visualized with iodine vapor. Materials. GDP- [“Clmannose was purchased from New England Nuclear, DEAE-cellulose from Bio-Rad Laboratories, Sephadexes from Pharmacia Fine Chemicals, and Pronase from Calbiochem.
153
GLYCOPROTEINS RESULTS
Properties of transfer reaction. Incubation of the purified [“Clmannolipid with the particulate enzyme fraction, Triton X-100, and MgCl, resulted in the timedependent incorporation of radioactivity into products which remained insoluble upon extraction of the incubation mixture with 6 vol of chloroform: methanol (2:l v/v) followed by washing of the insoluble residue with chloroform: methanol (2:l v/v). The distribution of radioactivity in the organic phase, the aqueous phase, and the insoluble residue from the initial chloroform: methanol extraction is shown in Fig. 1. The reaction proceeds for at least 2 h, at which time approximately 18% of the radioactive mannose from the substrate is recovered in the residue. Approximately 5% of the initial radioactivity was re-
2000
400
2 I500 e
300
-2
‘: s
P 0 d $ 2
-!
2 1000
200
:0 6 E
500
too
20
40
60 TIME
60
100
s y
120
(mm)
FIG. 1. Transfer of mannose from mannolipid substrate to endogenous protein acceptors. Standard incubation mixtures, each containing 2400 cpm of [“Clmannolipid, were incubated for the times indicated, denatured by addition of 6 co1 of chloroform: methanol (2:1), and processed as described in Methods. Radioactivity in: chloroform:methanol(2:1) (O-O); aqueous phase (A-A); residue (0-O). Radioactivity recovered from the residue of a heat denatured control incubation (78 cpm) was subtracted from the residue radioactivity for each time indicated.
154
BRETTHAUER
covered in the aqueous phase and was shown to be [“Clmannose by chromatography in solvents A and B. The radioactivity recovered in the organic phase was indistinguishable from the [“Clmannolipid used as substrate by chromatography on thin layer plates with solvents D and E, by elution from DEAE-cellulose (acetate) with dilute ammonium acetate in methanol, and by release of [14C]mannose upon mild acid hydrolysis. This fraction was therefore concluded to be unaltered substrate. The amount of [“Clmannose transferred from [“Clmannolipid to insoluble products was dependent on the presence of Triton X-100, 1 mg/ml being the optimal concentration (Fig. 2). The effects of other detergents have not been examined. The reaction was also dependent on the concentration of particulate enzyme fraction up to 9.6 mg/ml and concentration of substrate up to 2 x lo5 cpm/ml. An absolute requirement for metal ions in the reaction could not be demonstrated. Addition of MgCl, or MnCl, to standard incubation mixtures prepared without MgCl, (but still containing MgEDTA) resulted in only a small stimulatory effect. For either metal ion, 0.01 M concentration gave maximal stimulation (average of 25%). Omission of both MgEDTA and MgCl, from standard incubation mixtures had little effect on activity, and addition of either MgCl, alone or MgCl, and MnCl, again gave only a small (10-20s) stimulatory effect. These results, suggesting that metal ions (MnZ+ or Mg2+) are not required for the transfer reaction, are contradicted by the inhibitory effects observed upon addition of excess EDTA. Inhibition was proportional to EDTA concentration, 68% inhibition being observed with 0.05 M EDTA. This inhibition may be due either to chelation of endogenous metal ions which are essential for enzymatic activity, or to effects on the interaction of the components in this complex system which indirectly effects enzyme activity.
AND WU I
,
I TRITON
2 X-100 (mg/ml)
FIG. 2. Dependence of mannose transfer from mannolipid on Triton X-100 concentration. Standard incubation mixtures, each containing 4680 cpm of [“Clmannolipid and the indicated concentrations of Triton X-100, were incubated 60 min and then treated as described in Methods for determination of radioactivity in the residue. Radioactivity in the heat denatureu control incubation (105 cpm) was subtracted.
[“Clmannolipid). After incubating at 25” for 2 h, extraction with chloroform: methanol and water was carried out as previously described. The residue, containing 1.7 x 10’ cpm, was suspended in 10 ml of 0.2 M sodium citrate buffer, pH 7.0, and autoclaved at 120°C for 90 min. Seventy percent of the radioactivity (1.2 x 10’ cpm) was solubilized and was non-diffusable when dialyzed against water. Partial fractionation of the dialyzed products was achieved by filtration on a Sephadex G-100 column (1.5 x 100 cm, equilibrated with 0.05 M acetic acid) which resulted in 40 percent of the radioactivity (4.8 x lOa cpm) being excluded at the void volume and the remaining radioactivity (7.2 x lo3 cpm) being eluted near the total retention volume of the column. A similar elution profile was obtained for the radioactive products derived in the same manner from GDP- [“Clmannose with the exception that a larger percentage of the radioactive products was eluted at the void volume (14). Radioactive material that was solubilized by autoclaving heat denatured, control incubation mixtures (containing either Characterization of reaction products. GDP[“C]mannose or [“Clmannolipid as The standard incubation mixture de- substrate) was completely diffusable in the scribed in Methods was scaled up to 20 ml dialysis step. total volume (containing 1.8 x lo5 cpm of The Sephadex G-100 excluded fraction
SYNTHESIS
OF YEAST
derived from experiments with [“Clmannolipid as substrate was further fractionated into seven components by chromatography on a DEAE-Sephadex A-25 column. The elution profile (Fig. 3) shows that radioactive components are eluted with water and concentrations of KC1 up to 0.7 M. Protein, as monitored by absorption at 280 nm, elutes coincident with many of the radioactive components, suggesting that protein and carbohydrate are covalently attached in a macromolecular complex. This elution profile is very similar to that previously obtained for the radioactive glycoprotein products derived from GDP- [“Clmannose (14) and suggests that similar acceptors were mannosylated by both GDP-mannose and mannolipid. A second approach taken in characterizing the radioactive products was to treat the chloroform:methanol and water insoluble residue with Pronase, followed by isolation of soluble, radioactive glycopeptides. The procedures utilized and recoveries of radioactivity are listed in Table I. After extensive Pronase digestion, 89 percent of the radioactivity in the residue was solubilized. Of this radioactivity, 81% was retarded on Sephadex G-100. This retarded fraction was chromatographed on a Sephadex G-25 column from which 80% of the applied radioactivity eluted near the void volume in a single, symmetrical peak. Further chromatography of this fraction on a column of Dowex 50-X2 (pyridine)
w
0.1 Y KCL
0.2 Y KCL ELUTINQ
0.3Y
155
GLYCOPROTEINS
yielded a radioactive component which eluted at 0.14 M pyridinium formate and which contained 72% of the applied radioactivity (Fig. 4). A final gel filtration on Sephadex G-50 resulted in the recovery of 60% of the applied radioactivity in a peak which was TABLE
I
PURIFICATION OF [W]MANNOPE~E AFTER PRONASE DIGESTION OF’ENDOCENOUS PRODUCTS Fraction
Radioactivity
Chloroform:methanol insolublea Pronase soluble* Sephadex G-1W Sephadex G-25c Dowex 50 (Pyr’) Sephadex G-5@
cm
%
125,150
100
111,150 90,000 71,000 51,800 31,800
89 72 57 41 25
a The incubation mixture (20 ml), containing 1.2 x 10’ cpm of [Wlmannolipid substrate, was incubated at 30°C for 2 h before extracting with chloroform: methanol. bThe residue was suspended in 10 ml of 0.1 N Tris.HCl buffer, pH 7.5, containing 0.01 M CaCll. Pronase (1 mg) and a drop of toluene were added and after digestion at 37°C for 24 h, an additional 0.5 mg of Pronase was added. The pH was maintained at 7.5 by addition of 2 M Tris. After 88 h, the mixture was heated at 100°C for 3 min and the insoluble residue discarded. e Sephadex columns (1.5 x 100 cm) were equilibrated with 0.05 N acetic acid.
KCL
0.4Y
KCL
--0.7Y
KCL
AGENT
FIG. 3. Elution profile of solubilized radioactive products from DEAE-Sephadex A-25. Radioactive products solubilized by autoclaving at pH 7 and excluded from Sephadex G-100 were chromatographed on a column (2 x 58 cm) of DEAE-Sephadex A-25 (chloride) equilibrated with water. Elution was carried out stepwise with 150 ml each of water and KC1 solutions of concentrations indicated. Fractions of 3 ml were collected, 0.1 ml being used for determination of radioactivity.
156
BRETTHAUER
FRACTION
AND WU
NUMBER
FIG. 4. Elution profile of radioactive glycopeptides from Dowex 50-X2. The radioactive products, derived from Pronase digestion and gel filtration as described in the text, were chromatographed on a column (1 x 57 cm) of Dowex 50-X2 (pyridine) equilibrated with 1 mM pyridinium acetate, pH 3.0. Elution was carried out with a linear concentration gradient of pyridinium acetate (400 ml of 0.25 M pyridinium acetate, pH 3.0, into 400 ml of 1 mM pyridinium acetate, pH 3.0). Fractions of 3.9 ml were collected, 0.1 ml being used for measurement of radioactivity.
retarded by the gel and which had corresponding absorption at 230 nm, but no absorption at 280 nm (Fig. 5). Amino acid analysis of this fraction (tubes 46-50) revealed that aspartic acid, threonine, serine, glutamic acid, glycine, and alanine comprised 85% of the total amino acids (Table II). Serine and glycine were the two major amino acids, comprising 30% and 22%, respectively, of the total. The neutral sugar content, as measured by the phenolsulfuric acid method with mannose as standard, was 1.86 moles/mole aspartic acid. Hexosamines were not detectable on the amino acid analyzer. The type of linkage between the carbohydrate moiety and peptide moiety of the [“Clglycopeptide was elucidated by exposure to alkali and sodium borohydride. A radiochromatogram of the products from &elimination-reduction of the [“Clglycopeptide is shown in Fig. 6A. Seventy-five percent of the radioactivity in the [“Clglycopeptide was recovered as [“Clmannitol. No larger oligosaccharide alcohols were detected in this experiment or with use of solvent A. Chromatography of the [“Clglycopeptide which had not been exposed to alkali demonstrated that any possible contaminating [“Clmannose was absent (Fig. 6B). These experiments indicate that the [“Clmannose residues of
100
z z 50
FRACTION
f 0 P
NUMBER
FIG. 5. Elution profile of radioactive glycopeptide from Sephadex G-50. The major radioactive fraction (tubes 172-175, Fig. 4) from Dowex 50 was passed through a column (1.5 x 100 cm) of Sephadex G-50 equilibrated with 0.05 N acetic acid. Fractions of 2.6 ml were collected, 0.025 ml being used for measurement of radioactivity.
this glycopeptide are attached solely to serine or threonine residues. Supporting evidence for this conclusion was obtained from the amino acid composition after treatment of the [“Clglycopeptide with alkali. As seen in Table II, over 50% of the serine residues and 20% of the threonine residues are lost after B-elimination. The number of serine and threonine residues lost (1.90 moles/mole aspartate) is close to that expected from the neutral sugar content (1.86 moles/mole aspartate) if all the
SYNTHESIS
OF YEAST
mannose residues are 0-glycosidically linked to the peptide. The previous experiments, carried out on a purified glycopeptide which represented 25% of the total Pronase soluble radioactive products, do not prove that the only acceptor sites in the particulate enzyme fraction for mannose residues from mannolipid are serine and threonine residues. Experiments in which B-elimination was carried out on the chloroform-methanol insoluble residue or on the Pronase solubilized products invariably resulted in some radioactivity,, in addition to [“Clmannose, remaining at the origin of paper chromatograms developed with either solvents A or C. In some cases, this radioactivity accounted for as much as 30% of the radioacTABLE COMPOSITION
II
OF GLYCOPEPTIDE BEFORE AND AFTER EXPOSURE TO ALKALI
Component
Aspartate Threonine Serine Glutamate Glycine Alanine Valine Isoleucine Leucine Tyrosine Phenylalanine Histidine Lysine Neutral sugar”
Moles/mole
aspartate”
Control
After alkalib
1.00 0.73 3.18 0.98 2.41 0.99 0.34 Tr’ 0.30 Tr Tr Tr TI 1.86
1.00 0.58 1.43 1.21 2.20 1.02 0.41 Tr Tr Tr Tr Tr Tr ND’
a The amount. of sample applied to the amino acid analyzer contained 10 nmoles aapartate. All values have been adjusted to one mole aspartate. Acid hydrolysis was carried out in 6 N HCI at 110°C for 24 h. b Glycopeptide containing 10 nmoles aspartate was treated with 0.5 N NaOH for 30 min at 50°C in a total volume of 0.05 ml. The sample was neutralized with 0.05 ml of 0.5 N HCl, evaporated to dryness, and hydrolyzed in 0.4 ml of 6 N HCl at 110°C for 24 h. ‘Trace quantity, corresponding to less than 0.20 mole/mole aspartate. d Determined with phenol-sulfuric acid, using mannose as standard. ’ Not determined.
157
GLYCOPROTEINS
L--l 4
8
12
DISTANCE
16
20
(cm)
FIG. 6. Radiochromatogram of products obtained from [“Clglycopeptide after $-elimination-reduction. (A) The [“Clglycopeptide (5700 cpm) was dissolved in 0.5 ml of 0.5 M NaOH: 1.0 M NaBH,. After heating at 50°C for 30 min, 0.5 ml acetone was added and sufficient Dowex 50 (H+) to make the solution slightly acidic. The resin was centrifuged out and washed three times with 0.5 ml water. The combined supernatant fluids were evaporated to dryness five times from methanol:HCl (1OOO:l v/v). The residue was dissolved in water and chromatographed on paper with solvent B. Radioactivity was detected with a radiochromatogram scanner. (B) The [“Clglycopeptide (3000 cpml was chromatographed without exposure to alkaline NaBH,.
tivity in the chloroform-methanol insoluble fraction. The nature of this material has not been further investigated other than to demonstrate that acid hydrolysis (2 N HCl, lOO”C, 2 h) releases all of the radioactivity as [“Clmannose. DISCUSSION
The present report demonstrates that one function of the mannolipid derived from GDP-mannose in H. ho&ii is to mannosylate serine and threonine residues of glycoproteins. Because the extent of mannose transfer from mannolipid to endogenous acceptors is quite low, we have carried out experiments with purified mannolipid as substrate to characterize the reaction as being a typical enzyme-catalyzed reaction (i.e., the dependence on time, substrate concentration, and enzyme concentration, and the sensitivity to heat). The substrate
158
BRETTHAUER
utilized has been purified sufficiently so that characterization of the lipid moiety as a polyisoprenol was possible. Although a precise structure has not been arrived at, the chemical and physical properties of the mannolipid are entirely consistent with those of dolicholmonophosphate mannose as has been characterized from another yeast, S. cerevisiae, by Jung and Tanner (7). That the [“Clmannolipid could be converted in the presence of the particulate enzyme fraction to GDP- [“Clmannose which then serves as substrate is discounted on the following grounds: the reaction proceeds in the absence of added metal ions whereas the mannosylation of lipid with GDP-mannose (and the reverse reaction) has a strong requirement for either Mg2+ or Mn2+ ions (11); the particulate enzyme fraction is washed extensively during isolation to remove contaminating cellular metabolites such as GDP; the concentration of Triton X-100 used in the reaction mixture with mannolipid as substrate inhibits severely the transfer of mannose from GDP-mannose to endogenous acceptors (experiments not shown). Finally, as has recently been reported by Sharma et al. (17) with S. cerevisiae, and reported by us in preliminary form’ with H. holstii, GDP-mannose results in mannosylation of the mannose residue already attached to serine or threonine residues, thus allowing recovery of radioactive oligosaccharides after b-elimination. Such oligosaccharides have not been detected in our studies with only [“Clmannolipid as substrate. There does not appear to be a requirement for added metal ions (Mn2+ or Mg2+) in the transfer reaction from mannolipid to protein. Only a small stimulation (20-25%) was observed upon addition of 0.01 M MgCL, and 0.01 M MnCl, stimulated the reaction to even a smaller extent. The inhibition by EDTA is difficult to explain, but such inhibition has been observed by other investigators (22) in the metal-ion independent transfer of mannose from dolicholmonophosphate mannose to glycoprotein acceptors of mammalian membrane preparations (22, 23). It is not clear from
AND WU
the report of Sharma et al. (17) whether Mg2+ ion is required for the transfer reaction from mannolipid to protein acceptors in S. cerevisiae, although there does appear to be a requirement for metal ions for subsequent mannosylation reactions involving GDP-mannose. The nature of one of the products of mannose transfer from mannolipid has been convincingly shown to be a glycopeptide with 0-glycosidically bound mannose residues. The glycopeptide nature of the endogenous products is indicated by its susceptibility to Pronase digestion and cochromatography on a variety of supports of radioactivity and protein. Amino acid analysis of the purified radioactive glycopeptide revealed an abundance of serine and glycine with aspartate, threonine, glutamate and alanine comprising most of the remaining amino acids (Table II). A loss of serine and threonine, in amounts approximately equal to the neutral sugar content, and the recovery of only [14C]mannose upon treatment of the glycopeptide with alkali are properties expected for such glycopeptides. Whether this glycopeptide represents a fragment of cell wall mannanprotein complexes has not been firmly established. The similarity in elution patterns from DEAE-Sephadex of neutral citrate solubilized products from GDP-mannose (14) and mannolipid (Fig. 3) is consistent with the mannolipid products being a part of mannan, as glycopeptides from the GDP-mannose products have previously been shown to be characteristic of mannan (14). On the other hand, the composition of the purified glycopeptide described in the present report is quite different from those for glycopeptides derived from GDP-mannose (14). These latter glycopeptides contained larger amounts of mannose, and in addition contained some glucosamine. Also, the glycopeptides which underwent ,&elimination contained approximately 50% more threonine than serine. A comparison of compositions may not be significant, however, as the glycopeptide described in the present report represents only 25% of the radioactivity incorporated from [“Clmannolipid into chloroform-methanol insoluble products.
SYNTHESIS
OF YEAST
As other mannanlike products could have been discarded during purification, further structural studies are required. Our previous conclusions (11) from studies on mannosyl transfer reactions from GDP-mannose with the H. holstii particulate enzyme fraction, as well as reports of other investigators (8-10) using S. cereuisiae subcellular fractions, that carrier lipids were involved in the synthesis of cell wall mannan-protein complexes were based primarily on the kinetics of mannolipid synthesis and turnover. These kinetics tended to agree with the kinetics of synthesis of products which were precipitable with Fehlings solution and which yielded oligosaccharides of mannose upon acetolysis. More detailed analyses (14) of products indicated that, with GDP-mannose as substrate, only a small amount (less than one percent) of the mannose could be accounted for as either reducingterminal residues on oligosaccharides or as mannose itself upon p-elimination of purified glycopeptides which represented 50% of the radioactivity in the products solubilized by Pronase digestion. If these few mannose residues, linked 0-glycosidically, are the only products of mannose transfer from dolicholmonophosphate mannose, it is difficult to explain the kinetic correlations unless in our previous studies (ll), glycopeptides containing large numbers of 0-glycosidica.lly linked mannose residues were discarded in the purification procedure. It is also possible that dolicholmonophosphate mannose is involved in additional glycosylation reactions which results in alkali-stable products being formed. Indeed, as pointed out in the Results section, all of the radioactivity derived from [ “Clmannolipid in the chloroform-methanol insoluble fraction is not subject to p-elimination. Similar results were obtained in S. cerevisiae by Sharma et al. (17) who further demonstrated that the non-dialyzable radioactivity remaining after p-elimination became dialyzable after Pronase digestion. An interesting analogy exists between several mammalian glycoproteins and the polymannose segments of many yeast mannan-protein complexes in that mannose units are attached
159
GLYCOPROTEINS
to one or two N-acetylglucosamine residues in the core region which is directly attached to protein through a N-acylglycosylamine linkage (12, 24, 25). The recent reports (4, 6) on the possible assembly in mammalian tissues of this core region on dolicholpyrophosphate raises the question as to whether similar synthetic pathways may exist for yeast mannans, and whether the radioactive products from dolicholmonophosphate mannose which are not subject to P-elimination may represent core regions or partially synthesized core regions. Although we have not been able, up to this time, to detect any lipidbound radioactive components other than mannose, additional experiments are clearly needed to examine these possibilities. REFERENCES 1. ROTHFIELD, L., AND ROMEO, D. (1971) Bacterial. Reu. 35, 14-38. 2. LENNARZ, W. J., AND SCHER, M. G. (1972) Biochim. Biophys. Acta, 265, 417-441. 3. RICHARDS, J. B., AND HEMMING, F. W. (1972) Biochem. J. 130, 77-93. 4. BEHRENS, N. H., CARMINATTI, H., STANELONI, R. J., LELOIR, L. F., AND CANTARELLA, A. I. (1973) Proc. Nat. Acad. Sci. USA 70, 3390-3394. 5. WAECHTER, C. J., LUCAS, J. J., AND LENNAFZ, W. J. (1974) Biochem. Biophys. Res. Commun. 56, 343-350. 6. Hsu, A.-F., BAYNES, J. W., AND HEATH, E. C. (1974) Proc. Nat. Acad. Sci. USA 71, 2391-2395. 7. JUNG, P., AND TANNER, W. (1973) Eur. J. Biochem. 37, 1-6. 8. TANNER, W. (1969) Biochem. Biophys. Res. Commun. 35, 144-150. 9. SENTANDREU, R., AND LAMPEN, J. 0. (1971) Fed. Eur. Biochem. Sot. Lett. 14, 109-113. 10. SENTANDREU, R., AND LAMPEN, J. 0. (1972) Fed. EUF. Biochem. Sot. Lett. 27, 331-334. 11. BRETTHAUER, R. K., WV, S., AND IRWIN, W. E. (1973) Biochim. Biophys. Acta 304, 736-747. 12. SENTANDREU, R., AND NORTHCOTE, D. H. (1968) Biochem. J. 109, 419-432. 13. SENTANDREU, R., AND NORTHCOTE, D. H. (1969) Carbohyd. Res. 10, 584-585. 14. BRE~HAUER, R. K., AND CHEN, TSAY, G. (1974) Arch. Biochem. Biophys. 164, 118-126. 15. BABCZINSKI, P., AND TANNER, W. (1973) Biochem. Biophys. Res. Commun. 54, 1119-1124. 16. LEHLE, L., AND TANNER, W. (1974) Biochim. Biophys. Acta 350, 225-235.
160
BREmHAUER
17. SHARMA, C. B., BAEZCZINSKI,P., LEHLE, L., AND TANNER, W. (1974) Eur. J. Biochem. 46,35-41. 18. KOZAK, L. P., AND BREITHAUER, R. K. (1970) Biochemistry 9, 1115-1122. 19. BEHRENS, N. H., AND LELOIR, L. F. (1970) Proc. Nat. Acad. Sci. USA 66, 153-159. 20. BURGOS, J., HEMMING, F. W., PENNOCK, J. F., AND MORTON, R. A. (1963) Biochem. J. 88,470-482.
AND WU 21. FEENEY, J., AND HEMMING, F. W. (1967) Anal. Biochem. 20,1-15. 22. BAYNES, J. W., Hsu, A.-F., AND HEATH, E. C. (1973) J. Biol. Chem. 248, 5693-5704. 23. WAECHTER, C. J., LUCAS, J. J., AND LENNARZ, W. J. (1973) J. Biol. Chem. 248, 7570-7579. 24. TARENTINO, A. L., PLUMMER, T. H. JR., AND MALEY, F. (1974) J. Biol. Chem. 249, 818-824. 25. BALLOU, C. E. (1974) Adu. Enzymol. 40,239-270.
OF BIOCHEMISTRY
Synthesis
AND
BIOPHYSICS
167,151-160
(1975)
of The Mannosyl-O-Serine
Glycoproteins
from Polyisoprenylphosphate (Hansen&a
of Chemistry,
Biochemistry
Mannose
Linkage
of
in Yeast
holstiil 1
R. K. BRETTHAUER2 Department
(Threonine)
AND
and Biophysics Program, Indiana 46556 Received September
S.
WV
University
of Notre
Dame, Notre Dame,
9, 1974
The mannolipid synthesized from GDP-mannose and lipid acceptors in a particulate enzyme preparation from the yeast Hansenula holstii (R. K. Bretthauer, S. Wu, and W. E. Irwin, (1973) Biochim. Biophys. Acta, 304, 736-747) has the properties of dolicholmonophosphate mannose. Transfer of [“Clmannose from exogenously supplied, purified mannolipid to endogenous protein acceptors of the particulate enzyme fraction has now been demonstrated. The synthesis of radioactive products which are insoluble in chloroformmethanol and water is dependent upon time and concentrations of substrate, particulate fraction protein, and detergent. Addition of MgCl, or MnCl, to incubation mixtures prepared in the absence of these ions had only small stimulatory effects (20-25%), suggesting that the reaction is not dependent upon metal ions. Relatively high concentrations (0.005 ~-0.05 M) of EDTA did partially inhibit the reaction, but this is considered to be due to secondary effects. Seventy percent of the radioactivity in the chloroform-methanol insoluble residue was solubilized with hot, neutral citrate buffer. The chromatographic properties of this material on Sephadex gels and on DEAE-Sephadex were very similar to the properties of glycoprotein products derived from GDP-[“Clmannose. The chloroform-methanol insoluble products were also solubilized with Pronase which subsequently resulted in the isolation of a radioactive glycopeptide that contained 25% of the radioactivity transferred from mannolipid. The radioactive component of this glycopeptide was shown by @-elmination experiments and by amino acid analyses to be [“Clmannose residues linked O-glycosiditally to serine and threonine residues. It was concluded, therefore, that one function of the mannolipid is to serve as mannosyl donor in the synthesis of the mannosyl-O-serine (threonine) linkage region of glycoproteins which may be part of the cell wall mannanprotein complex. Other mannose-containing products may also be synthesized from the mannolipid, as p-elimination of the chloroform-methanol insoluble fraction or of the Pronase soluble fraction did not result in recovery of all of the radioactivity as [“Clmannose.
The role in bacteria of polyisoprenol phosphates as glycosyl carriers from sugar nucleotides to polysaccharides has been established (1). Studies from several laboratories now suggest that lipids (dolichols)
may also play a role as intermediate glycosyl-carriers from sugar nucleotides in glycosylation reactions of eucaryotic cells (2-6). In particular, a mannosylated lipid characterized as dolichol phosphate mannose (7) has been implicated by several investigators as an intermediate glycosyl carrier in the synthesis of yeast cell wall mannan from GDP-mannose (8-11). The mannan component of the cell wall of many yeasts is generally considered to be a
* Supported by NSF grant GB 38356, and a grant from Miles Laboratories, Inc. Part of this work is from the doctoral thesis of S. W. 2 To whom correspondence should be addressed. 3 Present address: Department of Biochemistry, North Carolina State University, Raleigh, NC 27607. 151 Copyright 0 1975 by Academic Press, Inc. All rights of reproduction in any form reserved.
152
BRE’ITHAUER
polysaccharide-protein complex which has two kinds of carbohydrate units. One is a large, branched polysaccharide of mannose residues which is attached through Nacetylglucosamine to an asparagine residue of the protein moiety (12). The other is a short oligosaccharide of mannose residues which is linked to serine and/or threonine residues in the protein (12, 13). Therefore, several different kinds of glycosidic linkages involving mannose could possibly be formed with a mannolipid as glycosyl donor. We have recently demonstrated that, with particulate enzymes and acceptors from Hansen& holstii, GDP-mannose is a mannosyl donor to short oligosaccharides of mannose which are released from glycopeptides by alkali with a loss of serine and threonine residues (14). This characteristic p-elimination reaction for the glycosyl-Oserine (threonine) linkage can thus be used to determine which type of carbohydrate unit in mannan (or other glycoproteins) is being synthesized (or elongated) from radioactive substrates. This report describes experiments which demonstrate that the mannolipid previously characterized from H. ho&ii (11) is a mannosyl donor in the synthesis of the mannosyl-O-serine (threonone) linkage of glycoproteins endogenous to the particulate enzyme fraction.’ During the completion of this work, reports (15-17) have appeared from Tanner’s laboratory which also demonstrate that, in membrane preparations from Saccharomyces cereuisiue, mannosyl residues which are released from protein by b-elimination are derived from dolicholmonophosphate mannose. MATERIALS
AND
METHODS
Synthesis and purification of [V]mannolipid.” The particulate enzyme fraction from H. holstii NRRL-Y 2448 was prepared as previously described ’ Preliminary accounts of these studies have been presented at The Sixth Annual Miami Winter Symposium, January 14-18, 19’74, Miami, FL, and at the Biochemistry/Biophysics 1974 Meeting, June 2-7, 1974, Minneapolis, MN. ‘Throughout this report, the mannosylated lipid substrate which has the properties of dolicholmonophosphate [“Clmannose will be referred to as [“C]mannolipid.
AND WU (18). Phospholipids, extracted from acetone-dried H. hoktii cells as described previously (ll), were saponified in chloroform:methanol (2:l v/v) with 0.1 N NaOH for 5 min at 25°C before being using as a source of polyisoprenylphosphate for [“Clmannolipid synthesis. The synthesis of [L’C]mannolipid was carried out as follows: phospholipid (6 pmoles phosphorus) and 0.15 mmoles of Mg. EDTA were mixed and evaporated to dryness (19); then 0.75 mmole of Tris.HCl buffer, pH 7.5, 0.15 mmole of MgCI,, 90 mg of Triton X-100, 3.1 nmole of GDP-[“Clmannose (160 mCi/mmole) and 60 mg of particulate enzyme protein were added in the order indicated to make a total volume of 15 ml. After incubation at 25°C for 10 min, the reaction was terminated by addition of 90 ml of chloroform:methanol (2:l v/v). Four incubation mixtures (15 ml each) containing unlabeled GDP-mannose (0.2 rmolell5 ml) were prepared and the chloroform: methanol extracts combined with that from the mixture containing GDP[“Clmannose. The [“Clmannolipid was purified from the chloroform:methanol extracts after evaporation of the solvent by discarding any acetone-soluble material in the residue and discarding any material not soluble in ether:ethanol (1:l v/v). The ether:ethanol soluble material was treated in chloroform:methanol (2:l v/v) with 0.1 N NaOH for 5 min at 25°C. Further purification of the mannolipid was achieved by DEAE-cellulose and Sephadex LH-20 column chromatography as previously described (11). The final radioactive product gave a single radioactivity peak and a coincident spot with iodine vapor after thin layer chromatography with solvents D and E. The specific activity of the [“C]mannolipid, as calculated from the total GDP-mannose concentration, was 0.63 mCi/mmole. All of the radioactivity of the purified [“C]mannolipid was released by mild acid hydrolysis (0.1 N HCl at 100°C for 1 min in 50% methanol) and recovered as a mixture of [“Clmannose and methyl[“Clmannoside. The proton magnetic resonance spectrum (100 MHz) of a concentrated solution of the [“Clmannolipid in CDCl, revealed resonance signals at 4.93 r (protons of olefinic carbon atoms), 7.98 T (protons of methylene groups adjacent to olefinic carbon atoms), 8.34 and 8.42 T (protons of cis and bans methyl groups, respectively, on trisuhstituted olefinic groups), and 8.6-8.9 T and 9.1 r (protons of saturated isoprene units) (20, 21). The areas for the protons on unsaturated isoprene units were close to those calculated. Anomalies in the spectrum, as compared to pig liver dolichols (20). were the ratio of cis:trans isoprene units which was 12:7, and the high absorptions at 8.6-8.9 T and 9.1 T which were present in excess of those expected for dolichol to the extent of approximately eight saturated isoprene units. These anomalies most likely represent impurities, as the areas of the absorption peaks from 8.6-9.1 T relative to
SYNTHESIS
OF YEAST
those areas at lower field for the unsaturated isoprene units varied with different samples. Assay procedures. Incubation mixtures containing [“C]mannolipid as substrate were prepared by evaporating to dryness under nitrogen a mixture of the [“Clmannolipid and Mg.EDTA, followed by addition of Tris.HCl buffer, pH 7.5, MgCl, and Triton X-100. The particulate enzyme fraction was added last. Final concentrations of the components were 10 mM Mg.EDTA, 10 mM MgCl,, 50 mM Tris. HCl buffer, pH 7.5, 1 mg/ml Triton X-100, 4-5 mg/ml particulate fraction protein, and [L’C]mannolipid as indicated in the Figure and Table legends. After incubation at 25”C, 6 vol of chloroform:methanol (2:l v/v) were added and, after thorough mixing, the liquid phases were separated by low speed centrifugation. The liquid. phases were carefully removed from the interphase residue which was then washed twice by centrifugation from chloroform:methanol (2:l v/v) (the volume of each wash being equivalent to 6 times the volume of the original incubation mixture). The residue was then solubilized with a minimal amount of 6% NaOH at 100°C and the solution transferred to a paper disc for determination of radioactivity. The amount of radioactivity recovered from incubation mixtures containing heat denatured (3 min, 100°C) particulate fraction was independent of incubation time and was subtracted from the experimental values. Radioactivity measurements were routinely carried out by applying liquid samples to paper discs, immersing the dried discs in 8 ml of Beckman ReadySolv Solution V, and counting in a liquid scintillation counter. Amino acid analysis. Samples were hydrolyzed in 6 N HCl at 110°C for 24 h and analyzed on a Beckman Amino Acid Analyzer with a single column. Under these conditions, glucosamine eluted at the front of the third solvent between leucine and tyrosine. Chromatography. Paper chromatography was carried out in the descending direction on Schleicher and Schuell No. 589 Green Ribbon paper with the following solvents (in volume ratios): (A) butanol: pyridine: water (6:4:3); (B) methyl ethyl ketone: acetic acid: boric acid saturated-water (9:l:l); (Cl isobutyric acid: ammonium hydroxide: water (57:4:39). Sugars were located with the periodate-benzidine reagent. Thin layer chromatography was carried out on glass plates containing Adsorbosil-1 (Applied Science Laboratories) with the following solvents: (D) chloroform:methanol:acetic acid:water (30:15:4:2); (El diisobutyl ketone: acetic acid: water (20:15:2); F, chloroform: methanol: water (65:25:4). Lipids were visualized with iodine vapor. Materials. GDP- [“Clmannose was purchased from New England Nuclear, DEAE-cellulose from Bio-Rad Laboratories, Sephadexes from Pharmacia Fine Chemicals, and Pronase from Calbiochem.
153
GLYCOPROTEINS RESULTS
Properties of transfer reaction. Incubation of the purified [“Clmannolipid with the particulate enzyme fraction, Triton X-100, and MgCl, resulted in the timedependent incorporation of radioactivity into products which remained insoluble upon extraction of the incubation mixture with 6 vol of chloroform: methanol (2:l v/v) followed by washing of the insoluble residue with chloroform: methanol (2:l v/v). The distribution of radioactivity in the organic phase, the aqueous phase, and the insoluble residue from the initial chloroform: methanol extraction is shown in Fig. 1. The reaction proceeds for at least 2 h, at which time approximately 18% of the radioactive mannose from the substrate is recovered in the residue. Approximately 5% of the initial radioactivity was re-
2000
400
2 I500 e
300
-2
‘: s
P 0 d $ 2
-!
2 1000
200
:0 6 E
500
too
20
40
60 TIME
60
100
s y
120
(mm)
FIG. 1. Transfer of mannose from mannolipid substrate to endogenous protein acceptors. Standard incubation mixtures, each containing 2400 cpm of [“Clmannolipid, were incubated for the times indicated, denatured by addition of 6 co1 of chloroform: methanol (2:1), and processed as described in Methods. Radioactivity in: chloroform:methanol(2:1) (O-O); aqueous phase (A-A); residue (0-O). Radioactivity recovered from the residue of a heat denatured control incubation (78 cpm) was subtracted from the residue radioactivity for each time indicated.
154
BRETTHAUER
covered in the aqueous phase and was shown to be [“Clmannose by chromatography in solvents A and B. The radioactivity recovered in the organic phase was indistinguishable from the [“Clmannolipid used as substrate by chromatography on thin layer plates with solvents D and E, by elution from DEAE-cellulose (acetate) with dilute ammonium acetate in methanol, and by release of [14C]mannose upon mild acid hydrolysis. This fraction was therefore concluded to be unaltered substrate. The amount of [“Clmannose transferred from [“Clmannolipid to insoluble products was dependent on the presence of Triton X-100, 1 mg/ml being the optimal concentration (Fig. 2). The effects of other detergents have not been examined. The reaction was also dependent on the concentration of particulate enzyme fraction up to 9.6 mg/ml and concentration of substrate up to 2 x lo5 cpm/ml. An absolute requirement for metal ions in the reaction could not be demonstrated. Addition of MgCl, or MnCl, to standard incubation mixtures prepared without MgCl, (but still containing MgEDTA) resulted in only a small stimulatory effect. For either metal ion, 0.01 M concentration gave maximal stimulation (average of 25%). Omission of both MgEDTA and MgCl, from standard incubation mixtures had little effect on activity, and addition of either MgCl, alone or MgCl, and MnCl, again gave only a small (10-20s) stimulatory effect. These results, suggesting that metal ions (MnZ+ or Mg2+) are not required for the transfer reaction, are contradicted by the inhibitory effects observed upon addition of excess EDTA. Inhibition was proportional to EDTA concentration, 68% inhibition being observed with 0.05 M EDTA. This inhibition may be due either to chelation of endogenous metal ions which are essential for enzymatic activity, or to effects on the interaction of the components in this complex system which indirectly effects enzyme activity.
AND WU I
,
I TRITON
2 X-100 (mg/ml)
FIG. 2. Dependence of mannose transfer from mannolipid on Triton X-100 concentration. Standard incubation mixtures, each containing 4680 cpm of [“Clmannolipid and the indicated concentrations of Triton X-100, were incubated 60 min and then treated as described in Methods for determination of radioactivity in the residue. Radioactivity in the heat denatureu control incubation (105 cpm) was subtracted.
[“Clmannolipid). After incubating at 25” for 2 h, extraction with chloroform: methanol and water was carried out as previously described. The residue, containing 1.7 x 10’ cpm, was suspended in 10 ml of 0.2 M sodium citrate buffer, pH 7.0, and autoclaved at 120°C for 90 min. Seventy percent of the radioactivity (1.2 x 10’ cpm) was solubilized and was non-diffusable when dialyzed against water. Partial fractionation of the dialyzed products was achieved by filtration on a Sephadex G-100 column (1.5 x 100 cm, equilibrated with 0.05 M acetic acid) which resulted in 40 percent of the radioactivity (4.8 x lOa cpm) being excluded at the void volume and the remaining radioactivity (7.2 x lo3 cpm) being eluted near the total retention volume of the column. A similar elution profile was obtained for the radioactive products derived in the same manner from GDP- [“Clmannose with the exception that a larger percentage of the radioactive products was eluted at the void volume (14). Radioactive material that was solubilized by autoclaving heat denatured, control incubation mixtures (containing either Characterization of reaction products. GDP[“C]mannose or [“Clmannolipid as The standard incubation mixture de- substrate) was completely diffusable in the scribed in Methods was scaled up to 20 ml dialysis step. total volume (containing 1.8 x lo5 cpm of The Sephadex G-100 excluded fraction
SYNTHESIS
OF YEAST
derived from experiments with [“Clmannolipid as substrate was further fractionated into seven components by chromatography on a DEAE-Sephadex A-25 column. The elution profile (Fig. 3) shows that radioactive components are eluted with water and concentrations of KC1 up to 0.7 M. Protein, as monitored by absorption at 280 nm, elutes coincident with many of the radioactive components, suggesting that protein and carbohydrate are covalently attached in a macromolecular complex. This elution profile is very similar to that previously obtained for the radioactive glycoprotein products derived from GDP- [“Clmannose (14) and suggests that similar acceptors were mannosylated by both GDP-mannose and mannolipid. A second approach taken in characterizing the radioactive products was to treat the chloroform:methanol and water insoluble residue with Pronase, followed by isolation of soluble, radioactive glycopeptides. The procedures utilized and recoveries of radioactivity are listed in Table I. After extensive Pronase digestion, 89 percent of the radioactivity in the residue was solubilized. Of this radioactivity, 81% was retarded on Sephadex G-100. This retarded fraction was chromatographed on a Sephadex G-25 column from which 80% of the applied radioactivity eluted near the void volume in a single, symmetrical peak. Further chromatography of this fraction on a column of Dowex 50-X2 (pyridine)
w
0.1 Y KCL
0.2 Y KCL ELUTINQ
0.3Y
155
GLYCOPROTEINS
yielded a radioactive component which eluted at 0.14 M pyridinium formate and which contained 72% of the applied radioactivity (Fig. 4). A final gel filtration on Sephadex G-50 resulted in the recovery of 60% of the applied radioactivity in a peak which was TABLE
I
PURIFICATION OF [W]MANNOPE~E AFTER PRONASE DIGESTION OF’ENDOCENOUS PRODUCTS Fraction
Radioactivity
Chloroform:methanol insolublea Pronase soluble* Sephadex G-1W Sephadex G-25c Dowex 50 (Pyr’) Sephadex G-5@
cm
%
125,150
100
111,150 90,000 71,000 51,800 31,800
89 72 57 41 25
a The incubation mixture (20 ml), containing 1.2 x 10’ cpm of [Wlmannolipid substrate, was incubated at 30°C for 2 h before extracting with chloroform: methanol. bThe residue was suspended in 10 ml of 0.1 N Tris.HCl buffer, pH 7.5, containing 0.01 M CaCll. Pronase (1 mg) and a drop of toluene were added and after digestion at 37°C for 24 h, an additional 0.5 mg of Pronase was added. The pH was maintained at 7.5 by addition of 2 M Tris. After 88 h, the mixture was heated at 100°C for 3 min and the insoluble residue discarded. e Sephadex columns (1.5 x 100 cm) were equilibrated with 0.05 N acetic acid.
KCL
0.4Y
KCL
--0.7Y
KCL
AGENT
FIG. 3. Elution profile of solubilized radioactive products from DEAE-Sephadex A-25. Radioactive products solubilized by autoclaving at pH 7 and excluded from Sephadex G-100 were chromatographed on a column (2 x 58 cm) of DEAE-Sephadex A-25 (chloride) equilibrated with water. Elution was carried out stepwise with 150 ml each of water and KC1 solutions of concentrations indicated. Fractions of 3 ml were collected, 0.1 ml being used for determination of radioactivity.
156
BRETTHAUER
FRACTION
AND WU
NUMBER
FIG. 4. Elution profile of radioactive glycopeptides from Dowex 50-X2. The radioactive products, derived from Pronase digestion and gel filtration as described in the text, were chromatographed on a column (1 x 57 cm) of Dowex 50-X2 (pyridine) equilibrated with 1 mM pyridinium acetate, pH 3.0. Elution was carried out with a linear concentration gradient of pyridinium acetate (400 ml of 0.25 M pyridinium acetate, pH 3.0, into 400 ml of 1 mM pyridinium acetate, pH 3.0). Fractions of 3.9 ml were collected, 0.1 ml being used for measurement of radioactivity.
retarded by the gel and which had corresponding absorption at 230 nm, but no absorption at 280 nm (Fig. 5). Amino acid analysis of this fraction (tubes 46-50) revealed that aspartic acid, threonine, serine, glutamic acid, glycine, and alanine comprised 85% of the total amino acids (Table II). Serine and glycine were the two major amino acids, comprising 30% and 22%, respectively, of the total. The neutral sugar content, as measured by the phenolsulfuric acid method with mannose as standard, was 1.86 moles/mole aspartic acid. Hexosamines were not detectable on the amino acid analyzer. The type of linkage between the carbohydrate moiety and peptide moiety of the [“Clglycopeptide was elucidated by exposure to alkali and sodium borohydride. A radiochromatogram of the products from &elimination-reduction of the [“Clglycopeptide is shown in Fig. 6A. Seventy-five percent of the radioactivity in the [“Clglycopeptide was recovered as [“Clmannitol. No larger oligosaccharide alcohols were detected in this experiment or with use of solvent A. Chromatography of the [“Clglycopeptide which had not been exposed to alkali demonstrated that any possible contaminating [“Clmannose was absent (Fig. 6B). These experiments indicate that the [“Clmannose residues of
100
z z 50
FRACTION
f 0 P
NUMBER
FIG. 5. Elution profile of radioactive glycopeptide from Sephadex G-50. The major radioactive fraction (tubes 172-175, Fig. 4) from Dowex 50 was passed through a column (1.5 x 100 cm) of Sephadex G-50 equilibrated with 0.05 N acetic acid. Fractions of 2.6 ml were collected, 0.025 ml being used for measurement of radioactivity.
this glycopeptide are attached solely to serine or threonine residues. Supporting evidence for this conclusion was obtained from the amino acid composition after treatment of the [“Clglycopeptide with alkali. As seen in Table II, over 50% of the serine residues and 20% of the threonine residues are lost after B-elimination. The number of serine and threonine residues lost (1.90 moles/mole aspartate) is close to that expected from the neutral sugar content (1.86 moles/mole aspartate) if all the
SYNTHESIS
OF YEAST
mannose residues are 0-glycosidically linked to the peptide. The previous experiments, carried out on a purified glycopeptide which represented 25% of the total Pronase soluble radioactive products, do not prove that the only acceptor sites in the particulate enzyme fraction for mannose residues from mannolipid are serine and threonine residues. Experiments in which B-elimination was carried out on the chloroform-methanol insoluble residue or on the Pronase solubilized products invariably resulted in some radioactivity,, in addition to [“Clmannose, remaining at the origin of paper chromatograms developed with either solvents A or C. In some cases, this radioactivity accounted for as much as 30% of the radioacTABLE COMPOSITION
II
OF GLYCOPEPTIDE BEFORE AND AFTER EXPOSURE TO ALKALI
Component
Aspartate Threonine Serine Glutamate Glycine Alanine Valine Isoleucine Leucine Tyrosine Phenylalanine Histidine Lysine Neutral sugar”
Moles/mole
aspartate”
Control
After alkalib
1.00 0.73 3.18 0.98 2.41 0.99 0.34 Tr’ 0.30 Tr Tr Tr TI 1.86
1.00 0.58 1.43 1.21 2.20 1.02 0.41 Tr Tr Tr Tr Tr Tr ND’
a The amount. of sample applied to the amino acid analyzer contained 10 nmoles aapartate. All values have been adjusted to one mole aspartate. Acid hydrolysis was carried out in 6 N HCI at 110°C for 24 h. b Glycopeptide containing 10 nmoles aspartate was treated with 0.5 N NaOH for 30 min at 50°C in a total volume of 0.05 ml. The sample was neutralized with 0.05 ml of 0.5 N HCl, evaporated to dryness, and hydrolyzed in 0.4 ml of 6 N HCl at 110°C for 24 h. ‘Trace quantity, corresponding to less than 0.20 mole/mole aspartate. d Determined with phenol-sulfuric acid, using mannose as standard. ’ Not determined.
157
GLYCOPROTEINS
L--l 4
8
12
DISTANCE
16
20
(cm)
FIG. 6. Radiochromatogram of products obtained from [“Clglycopeptide after $-elimination-reduction. (A) The [“Clglycopeptide (5700 cpm) was dissolved in 0.5 ml of 0.5 M NaOH: 1.0 M NaBH,. After heating at 50°C for 30 min, 0.5 ml acetone was added and sufficient Dowex 50 (H+) to make the solution slightly acidic. The resin was centrifuged out and washed three times with 0.5 ml water. The combined supernatant fluids were evaporated to dryness five times from methanol:HCl (1OOO:l v/v). The residue was dissolved in water and chromatographed on paper with solvent B. Radioactivity was detected with a radiochromatogram scanner. (B) The [“Clglycopeptide (3000 cpml was chromatographed without exposure to alkaline NaBH,.
tivity in the chloroform-methanol insoluble fraction. The nature of this material has not been further investigated other than to demonstrate that acid hydrolysis (2 N HCl, lOO”C, 2 h) releases all of the radioactivity as [“Clmannose. DISCUSSION
The present report demonstrates that one function of the mannolipid derived from GDP-mannose in H. ho&ii is to mannosylate serine and threonine residues of glycoproteins. Because the extent of mannose transfer from mannolipid to endogenous acceptors is quite low, we have carried out experiments with purified mannolipid as substrate to characterize the reaction as being a typical enzyme-catalyzed reaction (i.e., the dependence on time, substrate concentration, and enzyme concentration, and the sensitivity to heat). The substrate
158
BRETTHAUER
utilized has been purified sufficiently so that characterization of the lipid moiety as a polyisoprenol was possible. Although a precise structure has not been arrived at, the chemical and physical properties of the mannolipid are entirely consistent with those of dolicholmonophosphate mannose as has been characterized from another yeast, S. cerevisiae, by Jung and Tanner (7). That the [“Clmannolipid could be converted in the presence of the particulate enzyme fraction to GDP- [“Clmannose which then serves as substrate is discounted on the following grounds: the reaction proceeds in the absence of added metal ions whereas the mannosylation of lipid with GDP-mannose (and the reverse reaction) has a strong requirement for either Mg2+ or Mn2+ ions (11); the particulate enzyme fraction is washed extensively during isolation to remove contaminating cellular metabolites such as GDP; the concentration of Triton X-100 used in the reaction mixture with mannolipid as substrate inhibits severely the transfer of mannose from GDP-mannose to endogenous acceptors (experiments not shown). Finally, as has recently been reported by Sharma et al. (17) with S. cerevisiae, and reported by us in preliminary form’ with H. holstii, GDP-mannose results in mannosylation of the mannose residue already attached to serine or threonine residues, thus allowing recovery of radioactive oligosaccharides after b-elimination. Such oligosaccharides have not been detected in our studies with only [“Clmannolipid as substrate. There does not appear to be a requirement for added metal ions (Mn2+ or Mg2+) in the transfer reaction from mannolipid to protein. Only a small stimulation (20-25%) was observed upon addition of 0.01 M MgCL, and 0.01 M MnCl, stimulated the reaction to even a smaller extent. The inhibition by EDTA is difficult to explain, but such inhibition has been observed by other investigators (22) in the metal-ion independent transfer of mannose from dolicholmonophosphate mannose to glycoprotein acceptors of mammalian membrane preparations (22, 23). It is not clear from
AND WU
the report of Sharma et al. (17) whether Mg2+ ion is required for the transfer reaction from mannolipid to protein acceptors in S. cerevisiae, although there does appear to be a requirement for metal ions for subsequent mannosylation reactions involving GDP-mannose. The nature of one of the products of mannose transfer from mannolipid has been convincingly shown to be a glycopeptide with 0-glycosidically bound mannose residues. The glycopeptide nature of the endogenous products is indicated by its susceptibility to Pronase digestion and cochromatography on a variety of supports of radioactivity and protein. Amino acid analysis of the purified radioactive glycopeptide revealed an abundance of serine and glycine with aspartate, threonine, glutamate and alanine comprising most of the remaining amino acids (Table II). A loss of serine and threonine, in amounts approximately equal to the neutral sugar content, and the recovery of only [14C]mannose upon treatment of the glycopeptide with alkali are properties expected for such glycopeptides. Whether this glycopeptide represents a fragment of cell wall mannanprotein complexes has not been firmly established. The similarity in elution patterns from DEAE-Sephadex of neutral citrate solubilized products from GDP-mannose (14) and mannolipid (Fig. 3) is consistent with the mannolipid products being a part of mannan, as glycopeptides from the GDP-mannose products have previously been shown to be characteristic of mannan (14). On the other hand, the composition of the purified glycopeptide described in the present report is quite different from those for glycopeptides derived from GDP-mannose (14). These latter glycopeptides contained larger amounts of mannose, and in addition contained some glucosamine. Also, the glycopeptides which underwent ,&elimination contained approximately 50% more threonine than serine. A comparison of compositions may not be significant, however, as the glycopeptide described in the present report represents only 25% of the radioactivity incorporated from [“Clmannolipid into chloroform-methanol insoluble products.
SYNTHESIS
OF YEAST
As other mannanlike products could have been discarded during purification, further structural studies are required. Our previous conclusions (11) from studies on mannosyl transfer reactions from GDP-mannose with the H. holstii particulate enzyme fraction, as well as reports of other investigators (8-10) using S. cereuisiae subcellular fractions, that carrier lipids were involved in the synthesis of cell wall mannan-protein complexes were based primarily on the kinetics of mannolipid synthesis and turnover. These kinetics tended to agree with the kinetics of synthesis of products which were precipitable with Fehlings solution and which yielded oligosaccharides of mannose upon acetolysis. More detailed analyses (14) of products indicated that, with GDP-mannose as substrate, only a small amount (less than one percent) of the mannose could be accounted for as either reducingterminal residues on oligosaccharides or as mannose itself upon p-elimination of purified glycopeptides which represented 50% of the radioactivity in the products solubilized by Pronase digestion. If these few mannose residues, linked 0-glycosidically, are the only products of mannose transfer from dolicholmonophosphate mannose, it is difficult to explain the kinetic correlations unless in our previous studies (ll), glycopeptides containing large numbers of 0-glycosidica.lly linked mannose residues were discarded in the purification procedure. It is also possible that dolicholmonophosphate mannose is involved in additional glycosylation reactions which results in alkali-stable products being formed. Indeed, as pointed out in the Results section, all of the radioactivity derived from [ “Clmannolipid in the chloroform-methanol insoluble fraction is not subject to p-elimination. Similar results were obtained in S. cerevisiae by Sharma et al. (17) who further demonstrated that the non-dialyzable radioactivity remaining after p-elimination became dialyzable after Pronase digestion. An interesting analogy exists between several mammalian glycoproteins and the polymannose segments of many yeast mannan-protein complexes in that mannose units are attached
159
GLYCOPROTEINS
to one or two N-acetylglucosamine residues in the core region which is directly attached to protein through a N-acylglycosylamine linkage (12, 24, 25). The recent reports (4, 6) on the possible assembly in mammalian tissues of this core region on dolicholpyrophosphate raises the question as to whether similar synthetic pathways may exist for yeast mannans, and whether the radioactive products from dolicholmonophosphate mannose which are not subject to P-elimination may represent core regions or partially synthesized core regions. Although we have not been able, up to this time, to detect any lipidbound radioactive components other than mannose, additional experiments are clearly needed to examine these possibilities. REFERENCES 1. ROTHFIELD, L., AND ROMEO, D. (1971) Bacterial. Reu. 35, 14-38. 2. LENNARZ, W. J., AND SCHER, M. G. (1972) Biochim. Biophys. Acta, 265, 417-441. 3. RICHARDS, J. B., AND HEMMING, F. W. (1972) Biochem. J. 130, 77-93. 4. BEHRENS, N. H., CARMINATTI, H., STANELONI, R. J., LELOIR, L. F., AND CANTARELLA, A. I. (1973) Proc. Nat. Acad. Sci. USA 70, 3390-3394. 5. WAECHTER, C. J., LUCAS, J. J., AND LENNAFZ, W. J. (1974) Biochem. Biophys. Res. Commun. 56, 343-350. 6. Hsu, A.-F., BAYNES, J. W., AND HEATH, E. C. (1974) Proc. Nat. Acad. Sci. USA 71, 2391-2395. 7. JUNG, P., AND TANNER, W. (1973) Eur. J. Biochem. 37, 1-6. 8. TANNER, W. (1969) Biochem. Biophys. Res. Commun. 35, 144-150. 9. SENTANDREU, R., AND LAMPEN, J. 0. (1971) Fed. Eur. Biochem. Sot. Lett. 14, 109-113. 10. SENTANDREU, R., AND LAMPEN, J. 0. (1972) Fed. EUF. Biochem. Sot. Lett. 27, 331-334. 11. BRETTHAUER, R. K., WV, S., AND IRWIN, W. E. (1973) Biochim. Biophys. Acta 304, 736-747. 12. SENTANDREU, R., AND NORTHCOTE, D. H. (1968) Biochem. J. 109, 419-432. 13. SENTANDREU, R., AND NORTHCOTE, D. H. (1969) Carbohyd. Res. 10, 584-585. 14. BRE~HAUER, R. K., AND CHEN, TSAY, G. (1974) Arch. Biochem. Biophys. 164, 118-126. 15. BABCZINSKI, P., AND TANNER, W. (1973) Biochem. Biophys. Res. Commun. 54, 1119-1124. 16. LEHLE, L., AND TANNER, W. (1974) Biochim. Biophys. Acta 350, 225-235.
160
BREmHAUER
17. SHARMA, C. B., BAEZCZINSKI,P., LEHLE, L., AND TANNER, W. (1974) Eur. J. Biochem. 46,35-41. 18. KOZAK, L. P., AND BREITHAUER, R. K. (1970) Biochemistry 9, 1115-1122. 19. BEHRENS, N. H., AND LELOIR, L. F. (1970) Proc. Nat. Acad. Sci. USA 66, 153-159. 20. BURGOS, J., HEMMING, F. W., PENNOCK, J. F., AND MORTON, R. A. (1963) Biochem. J. 88,470-482.
AND WU 21. FEENEY, J., AND HEMMING, F. W. (1967) Anal. Biochem. 20,1-15. 22. BAYNES, J. W., Hsu, A.-F., AND HEATH, E. C. (1973) J. Biol. Chem. 248, 5693-5704. 23. WAECHTER, C. J., LUCAS, J. J., AND LENNARZ, W. J. (1973) J. Biol. Chem. 248, 7570-7579. 24. TARENTINO, A. L., PLUMMER, T. H. JR., AND MALEY, F. (1974) J. Biol. Chem. 249, 818-824. 25. BALLOU, C. E. (1974) Adu. Enzymol. 40,239-270.
