characterization oligosaccharides lipid-linked ... · biogel p-4 (200-400 mesh) was from bio-rad....

Plant Physiol. (1982) 70, 12-200032-0889/82/70/0012/09/$00.50/0

Characterization of the Oligosaccharides from Lipid-LinkedOligosaccharides of Mung Bean Seedlings'

Received for publication December 14, 1981 and in revised form February 23, 1982

HIDETAKA HORI AND ALAN D. ELBEINDepartment of Biochemistry, The University of Texas Health Science Center, San Antonio, Texas 78284

ABSTRACT

Lipid-linked oligosaccharides were synthesized with the particulate en-zyme preparation from mung bean (Phaseols aureus) seedlngs in thepresence of GDP-1J4C1 mannose. The oligosaccharides were released fromthe lipids by mild acid hydrolysis and purified by several passages on BiogelP4 columns. Five different oUgosacchardes were purified in this way.Based on their relative elution constants (Kd) compared to a variety ofstandard oligosaccharides, they were sized as (mannose-acetylglucosa-mine) Man7GlcNAc2, Man5GIcNAc2, Man3GlcNAc2, Man2GlcNAc2, andManGlcNAc2. These oUgosaccharides were treated with endoglucosamin-idase H and a- and 8i-mannosidase, and the products were examined onBiogel P4 columns. They also were subjected to a number of chemicaltreatments including analysis of the reducing sugar by NaB3H4 reduction,methylation analysis, and in some cases acetolysis. From these data, thelikely structures of these oligosaccharides are as foUows: E, Man,B-GlcNAc-GlcNAc; D, Mana1-*3Manfi-GlcNAc-GlcNAc; C, Manal-+2Mana-1 3Man#-GlcNAc-GlcNAc; B, Manal-2Mana1-*2Mana1--*3(ManaI-.6)Man,8-GlcNAc-GlcNAc; and A, Manal-2Mana1--+2Manal- 3(Manal-3lManal- 6I Manal--.6)Man8-GkcNAc-GlcNAc.The synthesis of the Man7GlcNAc2 was greatly diminished when tunica-mycin (10 pg/ml) was added to the incubation mixtures.

Many plant and animal glycoproteins have an oligosaccharidecomposed of mannose, N-acetylglucosamine (GlcNAc) and pos-sibly other sugars, that is attached to the protein in a GlcNAc-asparagine bond (10, 17, 32). The biosynthetic pathway for theformation of this oligosaccharide involves a series of membrane-bound glycosyl transferases that utilize dolichol-linked saccharideintermediates (10, 32). In the animal systems, it is now well-established that these enzymes catalyze the transfer of GlcNAc,mannose, and glucose to dolichyl-phosphate to form a dolichyl-pyrophosphoryl-oligosaccharide having the composition,Glc3Man9GlcNAc2 (5, 20, 22). This oligosaccharide is then trans-ferred to the protein (6, 7, 24, 28). Several of the intermediatelipid-linked oligosaccharides have been isolated from animal tis-sues and the oligosaccharides have been characterized (4, 15, 32,33). Plant cells also contain glycoproteins having oligosaccharidesattached in GlcNAc--asparagine bonds (10), and the structure ofthe oligosaccharide of soybean lectin has been shown to be a high-mannose type (21).Although the general pathway of assembly of the lipid-linked

saccharides of plant cells is known (1-3, 8, 13, 19), the details offormation of the various lipid intermediates remain to be estab-lished (10). Thus, few of the oligosaccharides have been isolatedor characterized, nor are the steps in the synthesis known. Re-

' Supported by grants from the National Institutes of Health (AM21800) and the Robert A. Welch Foundation.

cently, however, an oligosaccharide-lipid with the compositionGlc3Man9GlcNAc2 was identified in several plant tissues (18, 27).

This report describes the isolation and characterization of fiveof the major oligosaccharides from the lipid-linked oligosaccha-rides synthesized by mung bean particulate enzyme preparations.The oligosaccharides were purified by chromatography on BiogelP-4 and were characterized by enzymic and chemical procedures.

MATERIALS AND METHODS

Materials. GDP-[U-14C]mannose (269 mCi/mmol) and UDP-[6-3H]GlcNAc (6.6 Ci/mmol) were purchased from New EnglandNuclear. Unlabeled sugar nucleotides and type III a-mannosidasewere from Sigma. Pronase was obtained from Calbiochem andBiogel P-4 (200-400 mesh) was from Bio-Rad. Oligosaccharidestandards, Glc3Man9GlcNAc2 and Man5GlcNAc2 were kindlysupplied by Drs. C. Hubbard and P. W. Robbins, MassachusettsInstitute of Technology. Purified yeast mannan was a generousgift from Dr. C. Ballou, University of California at Berkeley. Allother chemicals were obtained commercially and were of the bestgrade available.

Preparation and Assay of Particulate Enzyme. Mung beanseedlings were grown on moist vermiculate in the dark at 25°Cfor 3 d. The sprouts were homogenized in 50 mm Tris buffer, pH7.5 containing 2 mm fl-mercaptoethanol, 1 mm EDTA, 0.1 mMMgCl2, 0.5% PVP, and 8% sucrose. One hundred g of mung beansprouts were placed in 50 ml of the above medium and blendedtwice for 10 s each in a Waring Blendor. The homogenate waspassed through eight layers of cheese cloth, and the filtrate wascentrifuged at 3000g for 10 min to remove whole cells and otherlarge particles. The supernatant liquid was then centrifuged at105,000g for 30 min to obtain the particulate enzyme. The pelletwas resuspended in 50 mm Tris buffer containing 10 mM MgC12,10 mm MnCl2, and 2 mm ,8-mercaptoethanol (0.5 ml medium/10g sprouts) with a Dounce homogenizer and used as the enzymesource (25 ,ug protein/,ld).

Reaction mixtures for assay of the particulate enzyme containedthe following components in a final volume of 0.5 ml: GDP-['4C]mannose (0.04 ,Ci) or UDP-[3H]GlcNAc (0.1 ,IuCi), 0.02 mmunlabeled UDP-GlcNac, 50 mm Tris buffer, pH 7.5, and varyingamounts of enzyme. When UDP-[3HJGlcNAc was used as thesubstrate, the unlabeled UDP-GlcNAc was omitted. Reactionmixtures were preincubated for 5 min without enzyme or withoutlabeled sugar nucleotides, and the reaction was initiated by theaddition of the missing component. Incubations were done atroom temperature for varying periods of time.The reactions were terminated by the addition of 2 ml of

CHC13:CH30H (1:1) and 0.5 ml of H20. The mixture was stirredvigorously, and the layers were separated by centrifugation. Thelower CHC13 layer was removed and saved, and the aqueous phaseand interface were extracted with another 1 ml of CHC13. Thecombined CHC13 phases, which contained lipid-linked monosac-charides and lower mol wt lipid-linked saccharides, were extractedwith CHC13:CH30H:H20 (3:48:47) before being placed in scintil-

12https://plantphysiol.orgDownloaded on February 1, 2021. - Published by

Copyright (c) 2020 American Society of Plant Biologists. All rights reserved.

https://plantphysiol.org

CHARACTERIZATION OF LIPID-LINKED OLIGOSACCHARIDES

lation vials for counting. The particulate material that remainedat the interface during these extractions was isolated by centrifu-gation after adding 1 ml of CH30H to dissolve any remainingCHC13. The pellets were washed two times with 50%o CH30H andonce with 1O00o CH30H before being suspended inCHCl3:CH30H:H20 (10:10:3) to extract the LLO.2 Aliquots ofthis extraction were placed in scintillation vials for counting (13).

Chemical Methods. LLO were hydrolyzed in 0.02 N HC1 in 20%methanol at 100°C for 20 min in screw capped tubes. Aftercooling, the hydrolysate was extracted with chloroform to removelipids, and the aqueous phase was saved for the isolation ofoligosaccharides.

Oligosaccharides released by mild acid hydrolysis were reducedwith NaB3H4. Samples were dissolved in 50 ,ul of H20 and I dropof I N NH40H was added, followed by 250 ,uCi of NaB3H4. Themixture was allowed to stand for 24 h at 4°C, and the reactionwas terminated by the addition of a few drops of acetic acid. Thesamples were concentrated to dryness on a rotary evaporator, andborate was removed by repeated evaporation with methanol. Thesample was passed through Biogel P-4 columns to isolate theradioactive oligosaccharides. The reduced material was then hy-drolyzed in 3 N HCI at 100°C for 4 h, concentrated to dryness,and subjected to paper chromatography.

Methylation Analysis. Oligosaccharides were purified on col-umns of Biogel P4 and were subjected to methylation accordingto the method of Hakamori (14) as modified by Sanford andConrad (25). The lyophilized samples were dissolved in 2 ml ofdimethylsulfoxide under N2 and sonicated with methylsulfmyl-carbanion at 50°C for S h. The mixtures were chilled and 2 ml ofCH3I were added. The mixture was sonicated for 2 h at 4°C witha further addition of 2 ml of CH3I after 1 h. The samples wereallowed to stand at room temperature overnight and then werepassed through columns of Sephadex LH-20 equilibrated with80% CH30H to remove dimethyl sulfoxide, methylsulfmyl carb-anion, and other salts. Columns were run with 80%1o methanol. Theeluates from this column were hydrolyzed in 2 N H2S04 at 100°Cfor 4 h in screw capped tubes. The hydrolysate was neutralizedwith Ba(OH)2, and the supernatants were concentrated to dryness,dissolved in chloroform, and analyzed by TLC.

Enzymic Digestions. Jack bean a-mannosidase was from Sigma.,B-Mannosidase was purified from Aspergillus fumigatus as de-scribed previously (11). Endoglucosaminidase H was obtainedfrom Miles Laboratories.

Oligosaccharides were treated with the various glycosidases asdescribed below. Purified oligosaccharides from Biogel P4 weretreated with 10 munits of endoglucosaminidase H in 0.1 ml of 0.05M Na citrate buffer, pH 6.0, at 30°C for 48 h under a tolueneatmosphere. Products were analyzed by Biogel P4.Samples were dissolved in 0.1 ml of 0.1 M acetate buffer, pH

5.0, containing 0.4 mm ZnCl2 and digested with 0.5 to 1.0 units ofa-mannosidase for 48 h at 30°C under a toluene atmosphere.Oligosaccharides were also dissolved in 0.1 ml of 0.1 M acetatebuffer, pH 4.0, and digested with 0.1 to 0.2 units offi-mannosidaseat 30°C for 48 h under a toluene atmosphere. At 24 h, another 0.2units of enzyme was added. Samples were analyzed by chroma-tography on Biogel P-4. Reactions were terminated by the additionof 100% ethanol.The residues remaining after extraction of the various lipids

were washed three times with H20 and once with 20 mm Trisbuffer, pH 8.5, containing 5 mm CaCl2. The residue was suspendedin 1 ml of 20 mm Tris buffer, pH 8.5, containing 5 mm CaCl2, and1 mg of pronase was added. Samples were incubated at 37°C for72 h under a toluene atmosphere. An additional 1 mg of pronasewas added at 24 and 48 h. Reactions were terminated by theaddition of TCA to a final concentration of 5%. The precipitatewas removed by centrifugation, and the supernatant liquids wereextracted with diethyl ether to remove TCA.

2 Abbreviation: LLO, lipid-linked oligosaccharides.

Chromatographic Methods. Paper chromatography of sugarswas done on Whatman 3 MM paper in 1-butanol:pyridine:0. 1 NHC1 (5:3:2). TLC of methylated sugars was done on Silica gelF-254 plates (0.5 mM thickness) in benzene:acetone:water:ammonium hydroxide (50:200:3:1.5). Radioactivity was de-tected on paper or on thin-layer plates by cutting papers into stripsor by scraping plates and counting in the liquid scintillationcounter. Standard sugars on paper chromatograms were detectedwith the silver nitrate dip (30) and on thin-layer plates by charringafter spraying with a mixture of H2SO4 and ethanol.

Oligosaccharides were separated and purified on columns ofBiogel P-4 (200-400 mesh) equilibrated and run in 0.1 M aceticacid. Biogel P-4 columns were I x 100 or 1.5 x 150 cm. Aliquotsof each fraction were removed for the determination of radioac-tivity.

RESULTS

Biosynthesis of LLO. The particulate enzyme from mung beanseedlings catalyzes the incorporation of mannose from GDP-[ 14C]mannose and GlcNAc from UDP-[3H]GlcNAc into lipid-linkedmonosaccharides (i.e. soluble in CHC13:CH30H:H20 [1:1:1]) andinto a series of LLO (mostly soluble in CHCl3:CH30H:H20110:10:3]). The incorporation of mannose into LLO was linearwith time of incubation for 5 to 10 min, and then leveled off withlonger incubations (data not shown). When unlabeled UDP-GlcNAc (0.01 ,umol) was added to the incubation mixtures, therewas a 30 to 40o increase in the amount of mannose incorporatedinto LLO probably due to an increase in the amount of N,N'-diacetylchitobiosyl-pyrophosphoryl-polyisoprenol which canserve as an acceptor of mannose. On the other hand, the additionof unlabeled UDP-glucose (0.01 ,umol) to the incubation mixturescaused a 30 to 40%o decrease in mannose incorporation suggestingthat mannose and glucose compete for the same acceptors.The incorporation of GIcNAc from UDP-[3HJGlcNAc into the

CHCI3:CH3OH: H20 (1:1:1) soluble lipids (i.e. GlcNAc- and N,N'-diacetylchitobiosyl-pyrophosphoryl-polyrenol) was linear withtime for 20 to 30 min, whereas incorporation of this sugar intoLLO was at least 10-fold lower and leveled off in I to 2 min (datanot shown). In this case, the addition of unlabeled GDP-mannose(0.005 ,umol) significantly inhibited the incorporation of GlcNAcinto the mono- and diGlcNAc-lipids by 50%o or greater, whileGlcNAc incorporation into LLO was stimulated about 2- to 3-fold.

Identification of LLO. The LLO formed from GDP-[ "C]man-nose at various times of incubation were isolated by solventextraction and the oligosaccharides were released by mild acidhydrolysis. The nature of the oligosaccharides was then examinedon Biogel P-4 columns. Figure I shows the profiles of [14C]mannose-oligosaccharides obtained during incubations of 1 to 60min. After 1 min, several oligosaccharides were observed thatmigrated slower than the heptasaccharide standard (M5N2), aswell as a radioactive peak emerging in the maltose area and alarge radioactive peak in the mannose area. With increasing timesof incubation, the amount of mannose incorporated into thesepeaks increased, and several new and faster migrating peaks weredetected. At least five major oligosaccharides were separated onthis column, and the characterization of these species is detailedin the remaining portion of this paper. It should be pointed outthat the addition of unlabeled UDP-glucose, even at concentra-tions as high as 10 mm, did not alter the labeling pattern seen inFigure 1. Thus, in contrast to studies in animal systems, UDP-glucose did not give rise to any detectable peaks in theGlc3Man9N2, Glc2Man9N2, or Glc,Man9N2 areas of the column.The lipid-linked saccharides labeled with UDP-[3H]GlcNAc

were also isolated, and the saccharides were examined on BiogelP-4 (Fig. 2). In this case, three major peaks were observed, one ofwhich eluted near the maltose area while the other two eluted nearthe stachyose standard. The fastest migrating peak is apparently

13

https://plantphysiol.orgDownloaded on February 1, 2021. - Published by Copyright (c) 2020 American Society of Plant Biologists. All rights reserved.


HORI AND ELBEIN

0

x

I

a.-

U

00

4r

FRACTION NUMBERFIG. 1. Biogel P4 column chromatography of the ['4CJmannose-la-

beled oligosaccharides from LLO isolated at various times of incubation.Incubations were as described in the text, and oligosaccharides wereobtained by mild acid hydrolysis and chromatographed on a I x 100 cmcolumn of Biogel P4. Standards are Glc3Man9GlcNAc2 (G3M9N2);Man5GlcNAc2 (M5N2), and maltose (Mal).

the trisaccharide, Man-fl-GIcNAc-GlcNAc, while the next peak isN,N'-diacetylchitobiose, and the slowest peak is GlcNAc.

Effect of Tunicamycin on the Formation of LLO. The antibiotictunicamycin has been shown to inhibit the formation of GlcNAc-pyrophosphoryl-polyisoprenol in both animal (33) and plant (12)tissues. Thus, it was of interest to examine the effect of thisantibiotic on the various plant lipid-linked saccharides. Incubationmixtures were prepared with GDP-[ "C]mannose and unlabeledUDP-GlcNAc as described in Figure 1, except that 10 or 50 ,ug oftunicamycin was added. After incubation, the LLO were isolatedand the oligosaccharides were chromatographed on Biogel P4(Fig. 3). In the control incubations, two major radioactive peakswere observed near the M5N2 standard as well as a radioactivepeak in the maltose area and one in the mannose area. However,in the presence of tunicamycin, the 2 fastest migrating peaks weregreatly diminished, but little effect was seen on the peaks migratingnear maltose and mannose. Also shown in this figure is the factthat tunicamycin had no inhibitory effect on the incorporation ofmannose into the CHC13:CH30H (1:1) soluble lipids.

Isolation of Five Major Oligosaccharldes from the LLO. Inorder to isolate sufficient amounts of the major oligosaccharidesfor characterization, the incubation mixtures described in Figure1 were scaled up 10- to 100-fold, and the LLO were isolated bysolvent extraction. The oligosaccharides were released by mildacid hydrolysis and chromatographed on a Biogel P4 column(Fig. 4). In addition to a large peak in the mannose area, six majoroligosaccharides were observed, and these were pooled and iden-tified as oligosaccharides A through F as shown by the brackets.Each of the pooled peaks was then rechromatographed on a

calibrated column of Biogel P4 in order to obtain a homogenousoligosaccharide and to determine the size of the oligosaccharide.Figure 5 shows the elution profiles of oligosaccharides A throughF and demonstrates that they were separated from each otherwith only slight overlap in a few cases. Each ofthe oligosaccharideswas again run on the column along with the standardsGlc3Man9GlcNAc2 (G3MoN2), Man9GlcNAc (M9N),Man7GlcNAc2 (M7N2), Man7GlcNAc (M7N), Man6GlcNAc(M6N), Man5GlcNAc2 (M5N2), stachyose, maltose, and mannose.The relative elution constant of each plant oligosaccharide and ofthe standard oligosaccharides was calculated according to Hub-bard and Robbins (16), and the Kd values are presented in TableI. According to this data, oligosaccharide A is a Man7GlcNAc2,oligosaccharide B is a Man5GlcNAc2, oligosaccharide C is aMan3GlcNAc2, oligosaccharide D is a Man2GlcNAc2, and oligo-saccharide E is a ManGlcNAc2. Also, as shown in Table I and asdiscussed below, endoglucosaminidase H digestion ofoligosaccha-ride A gave a new peak with the same Kd value as standardMan7GlcNAc.

Characterization of the Individual Oligosaccharides. Each ofthe oligosaccharides, A through E, was characterized using variousenzymatic treatments (a-mannosidase, endoglucosaminidase H)as well as chemical methods (reduction with NaB3H4, alkalinedigestion, methylation analysis, and acetolysis). These treatmentsare discussed in the various sections below.

Determination of the Reducing Terminus. In order to determinethe sugar at the reducing terminus, each oligosaccharide, A

0

a-L)

4z

2:1-

I-)0

cr

50 60 70FRACTION NUMBER

FIG. 2. Biogel PA column chromatography ofthe I3HJGlcNAc-labeledoligosaccharides from LLO isolated at various times of incubation. Incu-bations were as described in the text, and oligosaccharides were isolatedand chromatographed as described in Figure 1. Arrows marked 1, 2, and3 represent elution positions of mannose, maltose and stachyose, respec-tively.

14 Plant Physiol. Vol. 70, 1982



CHARACTERIZATION OF LIPID-LINKED OLIGOSACCHARIDESc:MS~~~~~ AC 40

3)0 40 5 0 0 6

FIG 3. p13heefecof6tuicmyi ontefrainoii-12ke

2 ~~~~~~~~~~~~~~8

exrcto 374H1:HO:20(01:) n heschrdswr

13 41 7.1

0-

10 40 50 60 4060FRACTION NUMBER

FIG. 3. The effect of tunicamycin on the formation of lipid-linkedsaccharides labeled from GDP-[l ClCmannose. Incubations were as de-

scribed in the text, but contained 10 or 50lg of tunicamycin. The controlwas without antibiotic. Lipids were isolated by CHCe3:CH30 H20 (1: 1:1)

extraction or by CHC13:CH30H:H20 (10:10:3), and the saccharides were

released by mild acid hydrolysis and chromatographed on Biogel Pa. F is

fraction F (see Fig. 4).

5 4 3 2

z

0 3

A B C DE F

a.

0)0

C 10 -4-0---- 50 60 70 80 90 100

FRACTION NUMBER

FIG. 4. BiogeIPA4column chromatography of large scale incubations

for isolation of [ "CJmannose-labeled oligosaccharides. Incubations were

scaled up 10- to 100-fold, and oligosaccharides were run on a 1.5 x 150

cm column of Biogel PA4. Aliquots of each fraction were analyzed for

radioactivity. Fractions were pooled as shown by the bars and labeled A

through E.

through E, was reduced with NaB3H4 and the reduced oligosac-charides were subjected to complete acid hydrolysis to release the

monosaccharides. Figure 6 shows the paper chromatograms used

to identify the sugars. In each case, the paper chromatograms were

cut into strips and counted to identify the radioactive sugars. Foreach oligosaccharide, a peak of tritium was detected that migratedwith glucosaminitol as well as a peak of '4C that corresponded tothe mannose standard. No evidence for [3H]mannitol or any otherreduced hexitol was seen. These data demonstrate that eacholigosaccharide has GlcNAc at the reducing end.

Digestion with a-Mannosidase. Each of the oligosaccharides Athrough D was treated with a-mannosidase, and the digests werepassed through columns of Biogel P4 (Fig. 7). For each of theoligosaccharides, this treatment gave rise to two new major peakscorresponding to the trisaccharide, Manf8-GlcNAc-GlcNAc (oli-gosaccharide E), and free mannose. However, in addition to thetwo major peaks, several small peaks of radioactivity were ob-served in several cases. For example, in the case of oligosaccharideA, a very small peak of undigested A is detectable as well as ashoulder on the peak eluting with oligosaccharide E. This isprobably a Man2GlcNAc2 which has not been completely digested.

C)

0

w

;7-

-j

cr

A r Aorta14res 77res.4;es i~\ 1 Sta Mal Man

B 50 60*.

IN60 70 4

E ~~~~70*80

F 70 80

70 80 90FRACTION NUMBER

FIG. 5. Further separation of oligosaccharides A through E on BiogelP4 columns. Each peak from Figure 4 was rechromatographed on BiogelP4. Standards shown are G3MqN2 (14 res), MIN2 (7 res), stachyose (Sta),maltose (Mal), and mannose (Man).

Table 1. Kd Values Calculated and Reported for VariousOligosaccharides

Plant Oligosac- Kd Value Ob- Standards KdReportedcharides served

G3MqN2 0.268 G3M9N2 0.279Peak A (M7N2) 0.411 M7N2 0.408Peak B (M5N2) 0.482 M5N2 0.474Peak C (M3N2) 0.571Peak D (M2N2) 0.607Peak E (M1N2) 0.661

M9Na 0.429Peak A-Endo H 0.518 M7Na 0.518

M5Na 0.571a These Kd values were determined in this laboratory.

15



Plant Physiol. Vol. 70, 1982

x 25t.- --cf t10

I.

4 0

E

0.4-

0.2-

0o 5 10 15 20DISTANCE OF DEVELOPMENT (cm)

FIG. 6. Determination of reducing sugar in oligosaccharides A throughE. Each oligosaccharide was reduced with NaB3H4. The reduced oligosac-charide was subjected to complete acid hydrolysis and chromatographedon paper in 1-butanol:pyridine:O. I N HCI (5:3:2). Papers were cut into l-cm strips and counted in the scintillation counters. The dashed linerepresents 3H and the solid line 14C. Standards are glucosaminitol(GlcNH2-ol), mannitol (Man-ol), and mannose (Man).

Also, in the case of oligosaccharide C, a small peak of undigestedC was seen in addition to the other peaks. However, these exper-

iments do indicate that oligosaccharides A through D have mostlya-linked mannose residues attached to Man,8-GlcNAc-GlcNAc.

Oligosaccharide E was also tested for its susceptibility to a-

mannosidase, but no free mannose was released by this treatmentnor could any change in the migration of E on Biogel PA beobserved (Fig. 8). However, treatment with f-mannosidase (pro-file C) released more than two-thirds the radioactivity as freemannose. Most of the radioactivity that was left in the trisacchar-ide after fB-mannosidase treatment could be released as free man-nose upon a second digestion with this enzyme. Thus, peak E ismost likely Man,8-GlcNAc-GlcNAc.

Digestion with Endoglucosaminidase H. The oligosaccharides, Athrough E, were treated with endoglucosaminidase H and theproducts were examined on Biogel P-4 (Fig. 9.). With oligosac-charide A, about two-thirds of the radioactivity was shifted to aslower migrating species whose elution pattern indicated that itcontained one less GlcNAc than the original oligosaccharide A.As shown in Table I, this endoglucosaminidase H-treated A hadthe same Kd value as an authentic sample of Man7GlcNAc.Although the digestion of peak A did not go to completion in thisexperiment, the undigested material could be digested further byincubation with more endoglucosaminidase H. However, completedigestion was not achieved, suggesting the possibility that theMan7GlcNAc2 peak contains small amounts of other isomers and

that some of these isomers are resistant to endoglucosaminidaseH. Nevertheless, the fact that the major portion of this peak wassusceptible indicates that it contains a branched mannose structureand that the mannose attached in a 1- 6 branch is further substi-tuted.When oligosaccharide B was treated with endoglucosaminidase

H for 48 h about 25% of the radioactivity was shifted to a slowerpeak which differed from the original B by one GlcNAc. Althoughthe Man5GlcNAc2 has been reported to be resistant to endoglu-cosaminidase H, it could be slightly susceptible during long incu-bation. Or the Man5GlcNAc2 could be heterogeneous and couldcontain small amounts of a susceptible isomer. Oligosaccharide Cwas digested to the extent of about 10 or 15% by the endoglucos-aminidase H as shown in scan 3 of Figure 9. It is not clear whatthis digestion means in terms of the oligosaccharide structure sincethe pentasaccharide would be expected to be resistant to endog-lucosaminidase H. Perhaps on long incubations with sufficientamounts of enzyme this oligosaccharide is slightly susceptible.Oligosaccharides D and E were found to be essentially resistant tothis enzyme, and only a very slight peak of unknown origin wasobserved in the case of E.

Methylation Analysis of Oligosaccharides. Oligosaccharides Athrough E were subjected to complete methylation and aftercomplete acid hydrolysis to liberate the monosaccharides, themethylated sugars were identified by TLC. The radioactive meth-ylated mannoses were compared to various standard methylatedsugars prepared from yeast mannan or ovalbumin. Figure 10shows the radioactive profiles of the thin-layer plates of oligosac-charides A through E. When [I 4Cjmannose-labeled oligosaccha-ride E was subjected to methylation (scan E), only one radioactive

1Z ,

2 A Man I

.0.C0 - 70 80 90 10

C4 2 EI B4v- 2x~~~~~~

3 C E

0*

4 D E

05 4..~~~~~~~~

50 60 70 80 90 100FRACTION NUMBER

FIG. 7. Effect of a-mannosidase treatment on oligosaccharides Athrough D. Each oligosaccharide was digested with a-mannosidase, andthe products were analyzed on Biogel P4. Numbers represent variousoligosaccharides: I = oligosaccharide A; 2 = B; 3 = C; 4 = D. Letters Athrough E represent elution positions of various intact oligosaccharides onthis column.

16 HORI AND ELBEIN




C.,

0

x

I-

0

cr

FRACTION NUMBERFIG. 8. Effect of 13-mannosidase and a-mannosidase digestion on oli-

gosaccharide E. This oligosaccharide was digested with either fi-mannos-idase (lower profile = C) or a-mannosidase (middle profile = B), and theproducts were examined on Biogel P4. The upper profile (A) is the controlwithout enzyme.

peak was detected which corresponded to 2,3,4,6-tetramethylman-nose. These results confirm the previous data indicating that peakE is Manj8-GlcNAc-GlcNAc.

Methylation of oligosaccharide D (scan D) gave rise to tworadioactive methylated sugars corresponding to 2,3,4,6-tetrame-thylmannose and 2,4,6-trimethylmannose. The amount of radio-activity in each of these peaks was almost the same. These resultscoupled with other data presented above indicated that oligosac-charide D is Manal- 3Man,8-GlcNAc-GlcNAc.

Oligosaccharide C, upon methylation and hydrolysis, gave riseto three radioactive methylated sugars which migrated with thestandards, 2,3,4,6-tetramethylmannose, 2,4,6-trimethylmannose,and 3,4,6-trimethylmannose. The absence of any radioactivity inthe area corresponding to 2,4-dimethylmannose suggests that thisoligosaccharide does not have a branched mannose structure.Thus, the likely structure for compound C is Manal--2Mana l-+

3Man,8-GlcNAc-GlcNAc.Oligosaccharide B also gave rise to three radioactive bands

upon methylation and hydrolysis. In this case, these radioactivebands corresponded to the standards, 2,3,4,6-tetramethylmannose,3,4,6-trimethylmannose, and 2,4-dimethylmannose. In this case,no radioactivity was found in the area corresponding to 2,4,6-trimethylmannose indicating that the 3-linked mannose was di-substituted. Therefore, the likely structure for oligosaccharide B isMana 1- 2Mana 1--2Mana 1--*3(ManaI--6)Manfl-GlcNAc-Glc-NAc. Further studies on this oligosaccharide, using acetolysis tocleave 1- 6 linkages (see below) are consistent with this structure.

Finally, oligosaccharide A yielded three radioactive bands aftermethylation and hydrolysis which corresponded to 2,3,4,6-tetra-

methylmannose, 3,4,6-trimethylmannose, and 2,4-dimethylman-nose. The absence of 2,4,6-trimethylmannose indicates that any 3-linked mannose is disubstituted and is compatible with either ofthe two structures shown in Figure 12. However, based on theresults of acetolysis (see below) and the susceptibility to endoglu-cosaminidase H, it seems likely that the more highly branchedstructure is the correct one.

Acetolysis of Oligosaccharides A and B. In order to decidebetween the alternative structures shown in Figure 12, oligosac-charide A was subjected to acetolysis (29) and the deacetylatedoligosaccharide products were chromatographed on Biogel P-4(Fig. 11). The upper profile shows the results obtained witholigosaccharide A. In this case, the major radioactive peak corre-sponded to the trisaccharide, raffmose, while the next major peakwas eluted at the position of the disaccharide, maltose. A smallerpeak of radioactivity was also found in the mannose area. Thesedata are consistent with the more highly branched structure (Fig-ure 12) for Man7GlcNAc2, having two 1- 6 linked mannoses.Thus, cleavage of the innermost 1--6 linkage would give a trisac-charide, whereas cleavage of the second 1-*6 linkage would givea disaccharide and a monosaccharide.The lower profile shows the acetolysis pattern obtained from

oligosaccharide B. Although several peaks were observed andsome of these could be artifacts since they are represented by asingle point, the major radioactive area corresponded to mannose.

2-1 A G2 G Mal Man

2B

4 0~~~~~~~~~~~~~~~2-

0~~~~~bo 30

-0 , .~. I-

1 F,

°~~~50 60 70 5

4-

-2 I4>~~~~

0~ ~ ~ 0

9 D~~~ .r

4I\0 -_---- - %W

50 60 70 80 9FRACTION NUMBER

FIG. 9. Treatment of oligosaccharides with endoglucosaminidase H.Each oligosaccharide was treated with endoglucosaminidase H and thetreated oligosaccharides were compared to untreated controls for theirmigration on Biogel PA. Numbers represent various oligosaccharides: 1= oligosaccharide A; 2 = B; 3 = C; 4 = D; 5 = E. Letters A through Erepresent elution position ofvarious intact oligosaccharides on this column.Standards are N,N'-diacetylchitobiose (G2), N-acetylglucosamine (G),maltose (Mal), mannose (Man).

17




o04

0.x

a-

I-

1-c)

0

0r

0O 5 10 15

DISTANCE OF DEVELOPMENT(cm)FIG. 10. Methylation analysis of oligosaccharides A through E. Each

oligosaccharide was subjected to complete methylation, and the products

were hydrolyzed and analyzed by TLC in benzene:acetone:

water:ammonium hydroxide (50:200:3:1.5). Standards shown by the num-

bers are I = 2,3-dimethylmannose; 2 = 2,4-dimethylmannose; 3 = 3,4,6-

trimethylmannose; 4 = 2,4,6-trimethylmannose; and 5 = 2,3,4,6-tetrame-thylmannose.

This is consistent with the structure for the Man5GlcNAc2 havinga single mannose residue in 1-*6 linkage.

DISCUSSION

Previous studies in plant systems have shown that mannose andGlcNAc are incorporated into LLO (1-3, 8, 13, 19). The oligosac-charide portion of these molecules have been isolated by paper

chromatography or by chromatography on columns of Biogel P-4and shown to range in size from trisaccharide to decasaccharide.However, except for some general studies on sugar composition,little is known about the structure of these oligosaccharides. It hasbeen shown that tissue slices from cotyledons form the same typeof lipid-linked saccharides as particulate enzyme fractions (2).

In the present study, five oligosaccharides were isolated fromthe LLO synthesized from GDP-['4C]mannose by a particulateenzyme from mung bean seedlings. These oligosaccharides were

separated from each other on columns of Biogel P-4 and were

further purified to apparent homogeneity by several passagesthrough long columns of Biogel P-4 (1.5 x 150 cm). Using a

number of oligosaccharide standards, the relative elution constant(Kd) of each of these oligosaccharide was calculated as well as theKd ofthe product resulting from endoglucosaminidase H digestion.Thus, the oligosaccharides were shown to be: A, Man7GlcNAc2;B, Man5GlcNAc2; C, Man3GlcNAc2; D, Man2GlcNAc2; and E,ManGlcNAc2.Each oligosaccharide was shown to have glucosamine at its

C

0I

0~x

a-.

I-

0

0

reducing terminus based on reduction with NaB3H4, and oligosac-charides A through D gave rise mostly to ['4CJmannose and thetrisaccharide, Man,8lcNAc, upon treatment with a-mannosidase.These results indicate that each oligosaccharide contains an NN'-diacetylchitobiose core at the reducing end and a number of a-linked mannose residues at the nonreducing end. OligosaccharideE was not susceptible to a-mannosidase but released free mannoseupon digestion with,B-mannosidase.

Oligosaccharide A was quite susceptible to endoglucosamini-dase H treatment and give rise to a Man7GlcNAc. Oligosaccha-rides B and C were somewhat susceptible to this enzyme especiallyupon long incubation but much less so than A. In this case, Bappeared to be more susceptible than C. With both oligosaccha-rides, a small peak of radioactivity was detected moving moreslowly than the original material, but these peaks were not avail-able in sufficient amounts for additional characterization. Basedon studies from other laboratories, the Man5GlcNAc2 and presum-ably the smaller oligosaccharides should not be susceptible toendoglucosaminidase H since the mannose in 1- 6 linkage is notsubstituted. Thus, the finding that the two plant oligosaccharidesare somewhat susceptible may indicate either that (a) the specific-ity of the endoglucosaminidase H is not so strict and that uponlong incubation these other oligosaccharides may show someactivity, or (b) the plant oligosaccharides may be composed ofseveral isomers such that one of the isomers that is present insmall amounts may be susceptible. Oligosaccharides D and E wereessentially resistant to the action of this enzyme.

Methylation analysis of oligosaccharides B through E indicatedthe following structures for these compounds: E, Manfl-GlcNAc-

50 60 70 80 90

FRACTION NUMBER100

FIG. 11. Acetolysis of oligosaccharides A and B. Compounds A and B

were subjected to acetolysis, and the products were analyzed on Biogel P-

4 columns. Standards shown by the arrows were: A to E = oligosaccharidesA to E; G2 = N,N'-diacetylchitobiose; Sta = stachyose; Raf = raffinose;

Mal = maltose; and Man = mannose.

A (D (2) )D ( (3f)2 -

1

3 :

2

1

1

o EI I-,

I Raf4

A B C D E G2 Sta . Mal Man

.00

0'~~~

7 0~~~~

*"*l4\ !;'i1! .L

0

0

0I.N ~ ~~. 0An .. ;i X z I~~0

18 HORI AND ELBEIN




Manul,3 \

Man M1 1

R(E)

Marnl,2

Maral ,3

Ian

R(D)

Manl1R

(C)

Mana1,2

Mana1,2

Man la1,3 \ /

Man- .

R(B)

Man Manat1,2 a1,2

Man Mana 1,2 a 1,2

Man Mancs1,3 \ / a1,6

Man1 R

/-1 R

Man/ a11,6

Mana1,2

Man Man Mana1,2 a1,3 \ /ol,6

Man Mana 1,3\x Q 1,6

Man

R(A)

FIG. 12. Proposed structures for oligosaccharides A through E. R = (G1cNAc)2-- P-- P-dolichol.

GlcNAc; D, Manal-3Manf8-GlcNAc-GlcNAc; C, Manal-+2Manald-3Man,8-GlcNAc-GlcNAc, and B, Manal-+2Manal- +

2Manal- '3(Manal-- 6)Man,8-GlcNAc-GlcNAc. Inasmuch as themajor radioactive peak detected upon acetolysis of B was man-nose, this would confirm a single mannose residue in a 1- 6linkage. Some of these structures are similar to those reported inanimal systems (4), but some are different indicating that thesequence of addition of mannose residues may be somewhatdifferent. For example, in animals, the pentasaccharideMan3GlcNAc2 has a 1-*6 linked mannose, whereas oligosaccha-ride C has a 1--2 linked mannose. Thus, in animal systems it hasbeen suggested that the addition of mannose residues to thetetrasaccharide is 1--6, then 1-*2, then 1- 2 to form theMan5GlcNAc2 (4), whereas in plants it may be 1--2, then 1--2 or1--6, then l--6 or l-+2.

It is interesting to note that the Man4GlcNAc2 andMan6GlcNAc2 species were not detected in these large scaleincubations. This may suggest that the two mannose residuesadded to the Man3GlcNAc2 are added very rapidly or at the sametime, as well as the two mannoses added to the Man5GlcNAc2.This could explain the type of heterogeneity observed in some ofthese oligosaccharides by Vijay and Fram (31) and Rearick et al.(23) who found evidence for several isomers in the various oligo-saccharide peaks. This might also explain some of the resultsobserved in the present study such as the susceptibility of theseoligosaccharides to endoglucosaminidase H and some ofthe minorpeaks observed with a-mannosidase digestion. However, the majorcompound present in oligosaccharide B from mung beans isapparently similar in structure to that formed in CHO cells (4)and in particulate and soluble extracts of aorta tissue (26).The largest oligosaccharide observed in this plant system is the

Man7GlcNAc2. The methylation analysis of this oligosaccharideshowed the absence of any 3-linked mannose (i.e. no 2,4,6-trime-thylmannose) which is also different from the Man7GlcNAc2 ofanimal cells (4). These data are consistent with either of thestructures shown in Figure 12. However, based on the fact thatacetolysis gave rise to a trisaccharide and a disaccharide, it seemslikely that the more highly branched structure is the correct one.This structure is consistent with the structure recently proposed

by Dorland et aL (9) for the oligosaccharide of soybean lectin. Itshould be mentioned that one difficulty with the methylationanalysis of these oligosaccharides formed in vitro is that themannose residues are probably not equally labeled. That is, thebiosynthesis involves the addition of mannose residues to endog-enous acceptors. Thus, we are not able to use the method for thequantitative determination of the number of mannose residuespresent in various glycosidic linkages. Nevertheless, the combi-nation of data from the various treatments used in this study dosuggest the structures presented in Figure 12. Although in thesestudies, the addition of UDP-glucose did not cause any alterationin the size of the mannose-labeled LLO, other workers have foundchanges in the nature of the oligosaccharides induced by UDP-glucose.

Acknowledgments-We are grateful to Drs. C. Hubbard and P. W. Robbins fortheir generous donation of oligosaccharide standards.

LITERATURE CITED

1. BAILEY DS, M DURR, J BURKE, GA MACLACHLAN 1979 The assembly of lipid-linked oligosaccharides in plant and animal membranes. J Supramol Struct I 1:123-138

2. BAILEY DS, V DELUCA, M DURR, DPS VERMA, GA MACLACHLAN 1980 Involve-ment of lipid-linked oligosaccharides in synthesis of storage glycoproteins insoybean seeds. Plant Physiol 66: 1113-1118

3. BEEVERS L, RM MENSE 1978 Glycoprotein biosynthesis in cotyledons of Pisumsativum L. Plant Physiol 60: 703-708

4. CHAPMAN A, E Li, S KORNFELD 1979 The biosynthesis of the major lipid-linkedoligosaccharide of Chinese hamster ovary cells occurs by the ordered additionof mannose residues. J Biol Chem 254: 10243-10249

5. CHEN WW, WJ LENNARZ 1978 Enzymatic synthesis of a glucose-containingoligosaccharide-lipid involved in glycosylation of proteins. J Biol Chem 253:5774-5779

6. CHEN WW, WJ LENNARZ 1978 Enzymatic excision ofglycosyl units linked to theoligosaccharide chains of glycoproteins. J Biol Chem 253: 5780-5785

7. DAs RC, EC HEATH 1980 Dolichyl diphosphoryl oligosaccharide-protein oligo-saccharyl transferase: solubilization, purification, and properties. Proc NatlAcad Sci USA 77: 3811-3815

8. DELMER DP, C KULOW, MC ERICSON 1978 Glycoprotein synthesis in plants. II.Structure of the mannolipid intermediate. Plant Physiol 61: 25-29

9. DORLAND L, H VAN HALBEEK, JFG VLIEGENTHART, H Lis, N SHARON 1981Primary structure of the carbohydrate chain of soybean agglutinin. A reinves-tigation by high resolution 'H NMR spectroscopy. J Biol Chem 256: 7708-7711

19



20 HORI AND ELBEIN

10. ELBEIN AD 1979 The role of lipid-linked saccharides in the biosynthesis ofcomplex carbohydrates. Annu Rev Plant Physiol 30: 239-272

1 1. ELBEIN AD, S ADYA, YC LEE 1977 Purification and properties of,B-mannosidasefrom Aspergillus niger. J Biol Chem 252: 2026-2031

12. ERICSON MC, JT GAFFORD, AD ELBEIN 1977 Tunicamycin inhibits GlcNAc-lipid formation in plants. J Biol Chem 252: 7431-7433

13. FORSEE WT, AD ELBEIN 1975 Glycoprotein biosynthesis in plants. Demonstra-tion of lipid-linked oligosaccharides. J Biol Chem 250: 9283-9293

14. HAKOMORI S 1964 A rapid permethylation of glycolipid and polysaccharidecatalyzed by methyl sulfinyl carbanion in dimethyl sulfoxide. J Biochem 55:205-208

15. Hsu AF, JW BAYNESS, EC HEATH 1974 The role of a dolichololigosaccharide asan intermediate in glycoprotein biosynthesis. Proc Natl Acad Sci USA 71:2391-2395

16. HUBBARD SC, PW ROBBINS 1980 Synthesis of the N-linked oligosaccharides ofglycoproteins. Assembly of the lipid-linked precursor oligosaccharide and itsrelation to protein synthesis in vivo. J Biol Chem 255: 11782-11793

17. KORNFELD R, S KORNFELD 1976 Comparative aspects of glycoprotein structure.Annu Rev Biochem 45: 217-237

18. LEHLE L 1981 Plant cells synthesize glucose-containing lipid-linked oligosaccha-ride similar to those found in animals and yeast. FEBS Lett 123: 63-66

19. LEHLE L, F FARTACZEK, W TANNER, H KAUSS 1976 Formation of polyprenol-linked mono- and oligosaccharides in Phaseolus aureus. Arch Biochem Biophys175: 419-426

20. LI E, I TABAS, S KORNFELD 1978 The synthesis ofcomplex-type oligosaccharides.I. Structure of the lipid-linked oligosaccharide precursor of the complex-typeoligosaccharides ofvesicular stomatitis virus G protein. J Biol Chem 253: 7762-7770

21. Lis S, N SHARON 1978 Soybean agglutinin-a plant glycoprotein, structure ofthe carbohydrate unit. J Biol Chem 253: 3468-3476

22. LIU T, B STETSON, SJ TURCO, SC HUBBARD, PW RoBBINS 1979 Arrangement ofglucose residues in the lipid-linked oligosaccharide precursor of asparaginyloligosaccharides. J Biol Chem 254: 4554-4559


23. REARICK JL, K FUJIMOTO, S KORNFELD 1981 Identification of the mannosyldonors involved in the synthesis of lipid-linked oligosaccharides. J Biol Chem256: 3762-3769

24. ROBBIN PW, SS KRAG, T LIu 1977 Effects of UDP-glucose addition on thesynthesis of mannosyl lipid-linked oligosaccharides by cell-free fibroblastpreparations. J Biol Chem 252: 1780-1785

25. SANFORD PA, HE CONRAD 1966 The structure of the aerogenes A3(S1) polysac-charide. I. A re-examination using improved procedures for methylationanalysis. Biochemistry 5: 1508-1517

26. SPENCER JP, AD ELBEIN 1980 Transfer ofmannose from GDP-mannose to lipid-linked oligosaccharide by soluble mannosyl transferase. Proc. Natl Acad SciUSA 77: 2524-2527

27. STANELONI RJ, ME TOLMASKY, C PETRIELLA, C, RA UGALDE, LF LELOIR 1980Presence in a plant of a compound similar to the dolichyl diphosphateoligosaccharide of animal tissue. Biochem J- 191: 257-260

28. TABAS I, S SCHLESINGER, S KORNFELD 1978 Processing of high mannose oligo-saccharides to form complex type oligosaccharides on the newly synthesizedpolypeptides of the vesicular stomatitis virus G protein and the IgG heavychain. J Biol Chem 253: 716-722

29. TAI T, K YAMASHITA, M OGATA-ARAKAWA, N KOIDE, T MURAMATSU, SIWASHITA, Y INOUE, A KOBATA 1975 Structural studies of two ovalbuminglycopeptides in relation to the endo-fl-N-acetylglucosaminidase specificity. JBiol Chem 250: 8569-8575

30. TREVELYAN WE, PROCTER, JS HARRISON 1950 Detection of sugars on paperchromatograms. Nature (Lond) 166: 444-445

31. VIJAY IK, SR FRAM 1977 Role of mannosyl-phosphoryl-polyisoprenol in biosyn-thesis of mammary glycoproteins. J Supramol Struct 7: 251-265

32. WAECHTER CJ, WJ LENNARZ 1976 The role of polyprenol-linked sugars inglycoprotein synthesis. Annu Rev Biochem 45: 95-112

33. WAECHTER CJ, JJ LUCAS, WJ LENNARZ 1974 Evidence for xylosyl lipids asintermediates in xylosyl transfers in hen oviduct membranes. Biochem BiophysRes Commun 56: 343-350



characterization oligosaccharides lipid-linked ... · biogel p-4 (200-400 mesh) was from bio-rad....

Documents