heterogeneous distribution of glycosyltransferases in the

Plant Physiol. (1983) 72, 547-5520032-0889/83/72/0547/06/$00.50/0

Heterogeneous Distribution of Glycosyltransferases in theEndoplasmic Reticulum of Castor Bean Endosperm'

Received for publication August 24, 1982 and in revised form February 7, 1983

MICHAEL J. CONDER AND J. MICHAEL LoRDDepartment of Biological Sciences, University of Warwick, Coventry, CV4 7AL England

ABSTRACT

Endoplasmic reticulum membranes stripped of attached ribosomes wereisolated from homogenates of germinating castor bean (Ricinus commuwusL.) endosperm by sucrose density gradient centrifuation. The isolatedendoplasmic reticulum fraction was further separated into two majormembrane subfractions by centfifugation on a flotation gradient. Bothsubfractions appeared to be derived from the endoplasmic reticulum inas-much as they share several enzymk markers including cholinephospho-transferase, NADH-cytochrome c reductase, and glycoprotein fucosyl-transferase and phase separation of membrane pobpptides using TritonX-114 revealed a stiking similarity in both their hydrophilk and hydro-phobic protein components. The endoplasmic reticulum membrane subfrac-tions contain glycoproteins which were readily labeled by incubating intactendosperm tissue with radioactive sugars prior to fractionation.

Castor bean eadosperm endoplasmic reticulum apparently exhibits adegree of enzymk heterogeneity, however, since the enzymes responsiblefor the synthesis of dolicholpyrophosphate N-acetylglucosamine and doli-cholmonophosphate mannose together with their incorporation into theoligosaccharde-Jipid precursor of protein N-glycosylation were largelyrecovered in a single endoplasmic reticulum subfraction.

The morphological heterogeneity of the ER in most eukaryoticcells is well known. Cellular fractionation studies, predominantlyperformed using animal tissues, have refined a variety of ap-proaches for separating ER vesicles bearing attached ribosomes(rough microsomes) from those not associated with ribosomes(12). Analytical studies on these isolated membrane fractions havealso revealed that the ER is biochemically heterogeneous in certainrespects (1). The studies have largely focused on the distributionof ER marker enzymes among ER subfractions. In general, al-though it seems that qualitative differences, if present at all, arerare, quantitative differences in the relative distribution of certainenzymes may well occur (1). The possibility of enzymic hetero-geneity within particular ER subfractions, membranes derivedfrom either rough surfaced or smooth surfaced ER, has also beensuggested (32).

In contrast to the situation with mammalian tissues, very fewattempts have been made to isolate and characterize ER mem-branes from plant cells. ER membranes have been isolated fromcastor bean endosperm cells (20). Sucrose density gradient cen-trifugation of endosperm homogenates effectively separates theER membranes from the other major organelle fractions. Thesemembranes, stripped of attached ribosomes, form a discrete bandon such gradients at a mean buoyant density of 1.12 g/ml (20).

'Supported by the Science and Engineering Research Council (U. K.)through Grant GR/B43753.

In the present study, we have subjected ER membranes isolatedfrom germinating castor bean endosperm as described above, toa second fractionation on a flotation gradient. In this way, twomajor membrane subfractions are obtained. Whereas it is clearthat on the basis of several shared enzymic and structural featuresboth membrane subfractions are derived from the ER, differencesin their capacity to synthesize mono- and oligosaccharide lipidintermediates involved in protein glycosylation were observed.

MATERIALS AND METHODS

Materials. [35S]Methionine (750-1000 Ci/mmol), 14C-labeledamino acid mixture (>50 mCi/milliatom of carbon), CDP-[methyl-14C]choline (50 Ci/mol), N-acetyl-D-[1-3H]glucosamine(11 Ci/mmol), D-[1-3H]mannose (5 Ci/mmol), L-[1-3Hlfucose (5.4Ci/mmol), D-[6-3Hjglucose (12.8 Ci/mmol), UDP-N-acetyl-D-[U-'4Cjglucosamine (282 Ci/mol), GDP-[U-'4CJmannose (199 Ci/mol), GDP-L-[I-'4Cjfucose (24.5 Ci/mol), and UDP-D-[U-'4C]glu-cose (293 Ci/mol) were from Amersham International Inc. (Amer-sham, U. K.) and Triton X-1 14 was from Sigma (Poole, U. K.).

Plant Material. Dry castor bean (Ricinus communis L.) seedswere soaked overnight and germinated in moist vermiculite indarkness at 30°C. After 3 d, developing seedlings were selectedfor uniformity and the endosperm tissue was excised.

Organeile Isolation. Endosperm homogenates were preparedand fractionated as described previously (20). After centrifugation,gradients were either collected as 1-ml fractions using an ISCOmodel 185 gradient fractionator or the visible organelle bandswere removed directly using a syringe.

Subfractionation of ER Vesicles. Three ml ofan ER membranefraction isolated as described above (in approximately 30% [w/w]sucrose) were mixed with 2.5 ml of 70% (w/w) sucrose. Themixture was placed in a 17.5-ml centrifuge tube and was succes-sively overlaid with 3 ml of40% (w/w) sucrose, 6.5 ml of 30%o (w/w) sucrose, and 2 ml of 20% (w/w) sucrose. Flotation gradientswere centrifuged at 24,000 rpm (75,000gav) and 2°C for 17 h in anSW 27 rotor on a Beckman L5-40 centrifuge. After centrifugation,the gradients were collected as 1.0-ml fractions.

Labeling in Vivo of Intact Tissue. Endosperm halves wereplaced abaxial surface down on moistened filter paper in a Petridish. Depending on the required labeling, each half was treatedwith (a) 25 ,uCi of ImS]methionine, (b) 2 ,uCi of '4C-labeled aminoacids, (c) 10 ,uCi of 3H-labeled sugars, or (d) 25 nCi of CDP-[14C1choline which were added directly to the adaxial surface involumes of 10 to 25 id. For the subsequent determination ofincorporated radioactivity present in isolated fractions, the labeledtissue was incubated at 30°C for 2 h, whereas for samples ulti-mately subjected to electrophoresis and fluorography the tissuewas incubated at 20°C for 16 h.Enzyme Assays. Cholinephosphotransferase, fumarase, isocit-

rate lyase, and NADH-Cyt c reductase were assayed as describedpreviously (9, 20).

Glycoprotein Fucosyltransferase. The reaction mixture con-

547

Dow

nloaded from https://academ

ic.oup.com/plphys/article/72/2/547/6081468 by guest on 18 January 2022

CONDER AND LORD

4A1-

z

Ul

z

0

z

LU

0

0L

0)

E

x

0

(-

4/S0U

n

Vo0o10

FRACTION

FIG. 1. Sucrose density gradient fractionation of castor bean endo-sperm homogenates. Collected gradient fractions were assayed for (a)cholinephosphotransferase (0), nmol/h.fraction x 20; fumarase (A),pmol/min-fraction x 10; and isocitrate lyase (0), ,umol/minfraction x20; and (b) for sucrose concentration (0).

tained 900 pl of collected gradient fraction and 100 mm Tris-HCl,pH 7.4, 10 mM MgCl2, 2 mM ,-mercaptoethanol, and 0.05% TritonX-100 in a final volume of 1.0 ml. After adding 0.05 ,uCi (2 ,ul) ofGDP-[14CJfucose, the mixture was incubated at 30°C for 30 mi.The reaction was stopped by adding 1 ml of methanol followedby cold TCA (10% [w/v] final concentration). Precipitated mate-rial was filtered on a Whatman GF/A disc, washed in 10%o TCA,and counted after immersing the disc in 10 ml of scintillant (6).Controls established that radioactivity incorporated into the re-action product was not reduced by prior extraction with chloro-form/methanol (2:1) or chloroform/methanol/water (10:10:3).

Glycolipid Glycosyltransferases. Two 400-pl aliquots were re-moved from each collected gradient fraction. These were added toa reaction mixture (final volume, 500 pl) containing either 100lmMTris-HCl, pH 7.5, 10 mm MgCl2, 2 mM ,B-mercaptoethanol, and0.1% (v/v) Triton X-100, or these components supplemented with15 ug dolichol monophosphate. Reactions were initiated by adding0.05 ,uCi UDP-N-acetyllf C]glucosamine, 0.05 ,uCi GDP-['4C]man-nose, or 0.1 ,uCi UDP-1[4CJglucose and were incubated at 300Cfor 30 min. The reactions were stopped by adding 4 ml ofchloroform/methanol (2:1). Incorporation of radioactive sugarsinto (a) product soluble in chloroform/methanol (2:1) (monosac-chande lipid) and (b) product soluble in chloroform/methanol/water (10:10:3) (oligosaccharide lipid) was determined as describedpreviously (23).

Phase Separation of Membrane Proteins. Washed membraneswere treated essentially as described by Bordier (5) in order toseparate amphipathic integral membrane proteins from hydro-

arE-x

I0

-

*_

Plant Physiol. Vol. 72, 1983

LUJ

0

L)

IC,)

Lu

I

FRACTION

FIG. 2. Subfraction of isolated ER vesicles by flotation. After centri-fuging the flotation gradient, 1.0-ml fractions were collected and assayedfor protein (A) and sucrose (0) content (a). (b), The distribution of TCA-insoluble radioactivity in fractions obtained using ER vesicles derivedfrom "4C-labeled amino acid (0) or N-[3Hjacetylglucosamine-labeled en-

dosperm tissue (0). (c), The distribution of cholinephosphotransferase(0) and NADH-Cyt c reductase (0).

philic proteins. A solution ofTriton X- 1 14 is homogeneous at 0°Cbut separates into an aqueous phase and a detergent phase above20°C; this is the basis of the method. Sucrose gradient fractionscontaining either glyoxysomes or ER vesicles were diluted with 50mM Tricine, pH 7.5, and the membranes were recovered bycentrifugation. The membrane pellets were washed by resuspen-sion in 50 mM Tricine, pH 7.5. An aliquot of this suspension was

transferred to a Microfuge tube and the membranes were collectedby centrifugation (15 min in an Eppendorfmodel 5412 centrifuge).The pellet was resuspended in 200 pl of 10 mm Tris-HCl, pH 7.5,0.5 M KCI, and 14 id of Triton X-1 14 was added. The Triton X-114 had been precondensed exactly as described (5) except thatNaCl had been replaced by an equivalent concentration of KCI.The mixture was placed on ice for 5 min to form a single phasewhich was then sonicated and allowed to stand on ice for a further1 h. Insoluble material was removed by 15-min centrifugation ina chilled Eppendorf model 5412 centrifuge. The clear supernatantrepresented the total proteins extracted from the membrane prep-arations.

548

Dow



HETEROGENEOUS DISTRIBUTION OF GLYCOSYLTRANSFERASES

Iux

3

0

b

u

I

I

1uU

0

U)

FRACTION

FIG. 3. Distribution of glycolipid N-acetylglucosaminyltransferase be-tween ER subfractions. Aliquots of collected flotation gradient fractionswere incubated with UDP-N-acetyl[4C]glucosamine at 30°C in the pres-ence (0) or absence (0) of exogenous dolicholmonophosphate. After 30min, the incorporation of radioactivity into chloroform/methanol (2:1)-soluble monosaccharide lipid (a), or chloroform/methanol (2:1)-solublemonosaccharide lipid (a), or chloroform/methanol/water (10:10:3)-solubleoligosaccharide lipid (b) was determined.

To separate the proteins, the clear supemnatant was layered overa 300-,ul cushion of 6% (w/v) sucrose, 10 mm Tris-HCI, pH 7.4,150 mm KCI, and 0.06% Triton X-114 contained in a 1.5-mlMicrofuge tube. The tube was incubated at 30°C for 3 min whenclouding of the upper phase occurred. The tube was centrifugedfor 30 to 60 s at room temperature. After centrifugation, thedetergent phase formed an oily droplet beneath the sucrose cush-ion. The upper aqueous phase was transferred to a clean Microfugetube and was re-extracted by adding 8 ,l of Triton X-1 14. Thetube was cooled on ice to allow the formation of a single phasewhich was then reloaded onto the original sucrose cushion, incu-bated at 30°C for 3 min, and then centrifuged to combine thedetergent phases. The aqueous phase was again removed and itwas re-extracted in exactly the same manner after adding a further30 ,ul of Triton X-1 14; the detergent phase from this extractionwas discarded. The hydrophilic protein solution was mixed withan equivalent volume of cold 20%1o (w/w) TCA. After standing onice for 30 min, the precipitated proteins were collected by centrif-ugation and retained for electrophoretic analysis.

After removing the sucrose cushion, the detergent phase was

549

tU5C0

0

0

5

25®

x 15

5-

4 8 12 16FRACTION

FIG. 4. Distribution of glycolipid mannosyltransferase between ERsubfractions. Aliquots of collected flotation gradient fractions were incu-bated with GDP-[14CJmannose at 30°C in the presence (0) or absence(e) ofexogenous dolicholmonophosphate. After 30 min, the incorporationinto chloroform/methanol (2:1)-soluble monosaccharide lipid (a), or chlo-roform/methanol/water (10:10:3)-soluble oligosaccharide lipid (b) wasdetermined.

mixed with 300 jil of 100 m, Tris-HCI, pH 7.4, 0.5 M KC1 andallowed to stand on ice until a single phase had formed. Afterincubation at 30°C for 3 min, the detergent phase (which con-tained integral membrane proteins with an amphipathic nature)was recovered by centrifugation and the aqueous phase wasdiscarded.

Prior to electrophoresis, the detergent was extracted with chlo-roform. The Triton X-1 14 extract (20-25 td) was mixed with 80td of 50 mm Tris-HCl, pH 7.4, containing 40 jig/ml PMSF.2 Thissolution was placed on ice until a single phase had formed andthen 1 ml of chloroform was added. After centrifugation, theupper aqueous phase plus interface material were removed andmixed with an equivalent volume of cold 20%o (w/v) TCA. Afterstanding on ice for 30 min, the precipitated proteins were re-covered by centrifugation.

Dodecyl Sulfate/Polyacrylamide Gel Electrophoresis. Proteinswere separated on a 10%o acrylamide slab gel, fixed, and stainedwith Coomassie blue as described previously (29).

Other Methods. Protein was determined by the Lowry proce-dure (21) using BSA as the standard and sucrose concentrationwas determined refractometrically.

2Abbreviation: PMSF, phenylmethylsulfonylfluoride.

Dow




X

w1~1AA

4 8 12 16FRACTION

FIG. 5. Distribution of glycolipid glucosyltransferase between ERsubfractions. Aliquots of collected flotation gradient fractions were incu-bated with UDP-[14CJglucose at 300C in the presence (0, A) or absence(0, A) of exogenous dolicholmonophosphate. After 30 min, the incorpo-ration of radioactivity into chloroform/methanol (2:1)-soluble monosac-

charide lipid (0, 0) or chloroform/methanol/water (10:10:3)-soluble oli-

gosaccharide lipid (A, A) was determined.

RESULTS AND DISCUSSION

Ribosome-denuded ER membranes were purified from castorbean endosperm homogenates by sucrose density gradient centrif-ugation (20). The distribution oforganelle marker enzymes amongfractions collected from such gradients suggested that the ERvesicles Oocated using the exclusive marker cholinephosphotrans-ferase) were not contaminated by, and did not themselves contam-inate, the mitochondria and glyoxysomes (located using fumaraseand isocitrate lyase as respective markers, Fig. la).

Salt-washed membranes prepared from organelle fractions sep-arated from [Slmethionine-labeled tissue were subjected to do-decyl sulfate-polyacrylamide gel electrophoresis. ER and mito-chondrial and glyoxysomal membranes contained characteristic35S-labeled polypeptide profiles which also indicated that cross-contamination between these fractions was insignificant (data notshown).The isolation gradients used here contain EDTA and the ER is

isolated as smooth surfaced vesicles. These vesicles are thought toderive from rough surfaced ER since (a) electron micrographs of

germinating castor bean endosperm cells show that the ER ispredominantly rough surfaced (33), and (b) over 90%o of theprotein, phospholipid, and marker enzyme activity present in thesmooth vesicle band (fractions 9 to 11, Fig. la) is recovered athigher sucrose densities under conditions which maintain ribo-some-membrane attachment (20).The ribosome-stripped ER membranes recovered from the ini-

tial fractionation gradient, suspended in approximately 30%o (w/w) sucrose, were further subfractionated by flotation. The sucrosedensity of the collected membrane preparation was increased toapproximately 50%o (w/w) sucrose by the addition of 70Yo (w/w)sucrose solution. This suspension was overlaid by the successiveaddition of 40%, 30%o, and 20%1o (w/w) sucrose layers. After cen-trifugation, two membrane bands were visible, at the 40 to 30%oand the 30 to 20%o sucrose interfaces (Fig. 2a). The uppermostband contained twice as much protein as the lower band. Proteinspresent in both membrane fractions were readily labeled byincubating intact tissue with radioactive amino acids or radioactiveN-acetylglucosamine (Fig. 2b), mannose, glucose, or fucose (datanot shown).Both membrane subfractions contained the characteristic ER

marker enzymes cholinephosphotransferase and NADH-Cyt creductase (Fig. 2c) together with glycoprotein fucosyltransferase,which in the nondividing endosperm cell is confined to the ERrather than the Golgi apparatus (30) (data not shown). Bothmembrane subfractions accumulated ["C]phosphatidylcholinesynthesized in vivo by incubating intact tissue with CDP-[methyl-14CJcholine. The extent to which newly synthesized ['4Cjphospha-tidylcholine accumulated in these membrane subfractions wasdirecly proportional to their relative cholinephosphotransferaseactivity (data not shown).

Castor bean endosperm membrane glycoproteins contain N-glycosidically linked oligosaccharide chains and the incorporationof labeled glucosamine into TCA-insoluble membrane compo-nents is inhibited by prior treatment ofthe tissue with tunicamycin(Ref. 4 and M. J. Conder, unpublished data). The synthesis ofdolichol-linked intermediates containing N-acetylglucosamineand mannose, their incorporation into a mannose-rich oligosac-charide lipid and the transfer of the oligosaccharide from its lipidcarrier to asparagine residues in acceptor proteins are all catalyzedby enzymes located in the ER membrane (22, 24, 26). The ERsubfractions separated in the present study were assayed for thepresence ofglycosyltransferases catalyzing the transfer ofthe sugarmoiety from nucleotide diphosphate derivatives to the lipid accep-tor and the assembly of the monosaccharide lipids into an oligo-saccharide lipid. The transferases catalyzing the synthesis of lipid

Table I. Enzyme Distribution between ER SubfractionsThe distribution of enzymic activity between the lower gradient ER band and the upper gradient ER band is

compared. With the exception of NADH-Cyt c reductase, for which activity is expressed as pmol/min, all otherenzyme activities are given as pmol/h.

Lower ER Band Upper ER Band

ConstituentActivitytActivity/mg %Total Activity Activity/mg % To1protein protein

Protein 380pg 34 730.ug 66Cholinephosphotransferase 30 78.9 30 70 95.9 70NADH-Cyt c reductase 4.3 11.3 29 10.7 14.6 71Fucosyltransferase 9.4 24.7 35 17.4 23.8 65Glucosyltransferase 8.2 21.7 31 18.4 25.2 69MannosyltransferaseMonosaccharide lipid 68.2 179.5 10 600 821.9 90Oligosaccharide lipid 5.4 14.2 9 54.8 75.1 91

N-AcetylglucosaminyltransferaseMonosaccharide lipid 75.5 198.7 9 724.5 992.5 91Oligosaccharide lipid 6.4 16.8 7 61.3 83.9 93

550 CONDER AND LORD

Dow



HETEROGENEOUS DISTRIBUTION OF GLYCOSYLTRANSFERASES

1 2 3 4 5 67 84F:a

v

FIG. 6. Dodecyl sulfate-polyacrylamide gel electrophoresis of organelle membrane proteins. ER and glyoxysomal membranes were isolated and theER membranes were subfractionated by flotation. Membrane proteins were partitioned into water-soluble hydrophilic components and detergent-soluble amphipathic integral membrane proteins by phase separation in Triton X-l 14 solution before electrophoretic analysis. Lanes I and 4, water-soluble glyoxysomal membrane proteins; lane 2, water-soluble proteins from the lower ER subtraction; lane 3, water-soluble proteins from the upperER subtractions; lanes 5 and 8, detergent-soluble glyoxysomal membrane proteins; lane 6, detergent-soluble proteins from the lower ER subtraction;lane 7, detergent-soluble proteins from the upper ER subtraction. Protein bands were visualized using Coomassie blue.

intermediates containing either N-acetylglucosamine (Fig. 3) ormannose (Fig. 4) were predominantly located in the upper ERsubtraction. The monosaccharide lipid intermediates have previ-ously been identified as dolicholpyrophosphate N-acetylglucosa-mine and the dolicholmonophosphate mannose in castor beanendosperm (25) as they have been in other plant tissues (2, 13,14). Both sugars are probably incorporated into the same dolichol-linked oligosaccharide (14, 15). The involvement of dolichol asthe lipid carrier was confirmed by its stimulatory effect whenadded exogenously to the assay systems; the synthesis of mono-saccharide lipids in particular was increased 20- to 30-fold byadding dolicholmonophosphate to a final concentration of 30 Iug/ml (Figs. 3a and 4a). In contrast, the enzymic activity catalyzingglucose transfer from UDP-glucose to chloroform/methanol-sol-uble glucolipid was present in both ER subtractions (Fig. 5). Atpresent we have not characterized the glucolipid(s) formed in thisreaction. Recent experimental findings have shown that the core

oligosaccharide involved in the initial N-glycosylation reaction ofplants, like its counterpart in animal cells (8, 27), contains glucosein a characteristic (glucose)3(mannose)5(N-acetylglucosamine)2structure (17, 31). The synthesis of dolicholmonophosphate glu-cose in plant tissues has been described and it apparently functionsas an intermediate in both protein glycosylation (28) and polysac-charide synthesis (7). Glucolipid synthesis observed here was notstimulated by exogenous dolicholmonophosphate (Fig. 5) and theglucolipid itself is stable during acid hydrolysis, suggesting that itis not dolicholmonophosphate glucose (M. J. Conder, unpublisheddata). It is possible that the glucolipid synthesized by the castorbean enzyme is a sterylglucoside. In any event, glucose is clearlynot incorporated into the oligosaccharide lipid (Fig. 5).The heterogeneity in the distribution of N-acetylglucosaminyl-

and mannosyltransferases between the two ER subtractions isemphasized by the data in Table I. The upper fraction containedapproximately twice as much protein as the lower fraction. Like-

551

Dow




wise, it contained 65 to 70%Yo ofthe total cholinephosphotransferase,NADH-Cyt c reductase, and fucosyl- and glucosyltransferaseswhich therefore have similar activities in each subfraction. Incontrast, over 90% of the total N-acetylglucosaminyl- and man-nosyltransferases were present in the uppermost subfraction atspecific activities some 5-fold higher than those of the lowersubfraction.The quantitative differences in the distribution of the glucosyl-

transferases between the subfractions (Table I) probably repre-sents a corresponding difference in the distribution of activeenzymes. The total activity measured was not significantly reducedwhen aliquots of the two subfractions were mixed prior to assay.This suggests that the reduced activity present in the lower ERsubfraction cannot be attributed to destruction of nucleotidesugars by microsomal pyrophosphatase or glycosidase (3).The contention that the two microsomal fractions obtained in

the present study represent genuine ER subfractions wasstrengthened by characterizing and comparing their constituentpolypeptides. Prior to dodecyl sulfate-polyacrylamide gel electro-phoresis, membrane-associated polypeptides were separated intohydrophilic and hydrophobic classes by phase separation in TritonX-1 14 according to the procedure of Bordier (5). The ER subfrac-tions were strikingly similar in both their water-soluble and TritonX-l 14-soluble polypeptide profiles (Fig. 6). Whereas some simi-larity in the hydrophilic polypeptides associated with each subfrac-tion could conceivably be due to nonspecific adsorption of thesame contaminating soluble proteins, the identical profiles ex-

hibited by the subfractions (Fig. 6, lanes 2 and 3) strongly suggestthat they also share identical peripherally bound or relativelyhydrophilic membrane proteins. The subfractions also sharedvirtually identical Triton X-1 14-soluble hydrophobic integralmembrane protein profiles (Fig. 6, lanes 6 and 7) although some

possibly significant differences were noted (polypeptides predom-inantly confined to a particular subfraction are denoted by thearrows in Fig. 6, lanes 6 and 7). Both the water-soluble and Triton-soluble polypeptides present in the ER subfractions were clearlydifferent from those present in corresponding fractions preparedfrom glyoxysomes (Fig. 6).

Demonstrations of biochemical heterogeneity in microsomaland submicrosomal vesicles have previously been restricted largelyto results obtained using rat liver. Numerous quantitative differ-ences in enzyme distribution between rough and smooth ER andin subfractions derived therefrom have been reported (1, 11, 32).This heterogeneity suggests that there may be distinct regions ofthe ER which are specialized for the performance of a givenfunction. In castor bean endosperm ER, one specialized regionmay be enriched in dolichol-lined glycosyltransferases involvedin assembling the core oligosaccharide of N-glycosylated proteins.We may further speculate that this region may be derived fromrough ER since (a) it is thought that the core sugars are added tonascent proteins before polypeptide chain termination (18, 19)and (b) rough ER has been identified as the intracellular locationof dolichol utilizing glycosyltransferases (10, 16, 26).

Acknowledgments-We thank Dr. L. K. Evans (Croda Premier Oils, Hull, U. K.)for generously providing the castor beans and the Science and Engineering ResearchCouncil (U. K.) for the award of a research studentship to M. J. C.

LITERATURE CITED

1. BEAUFAY H, A AMAR-COSTESEC, D TEINES-SEMpoux, M WIBo, M ROBBI, JBERTHtr 1974 Analytical study of microsomes and isolated subcellular mem-branes from rat liver. Ill. Subfractionation of the microsomal fraction byisopycnic and differential centrifugation in density gradients. J Cell Biol 61:213-231

2. BEEvRS L, RM MENSE 1977 Glycoprotein biosynthesis in cotyledons of Pisumsativum L. Involvement of lipid-linked intermediates. Plant Physiol 60: 703-708

3. BERGMAN A, G DALLNER 1978 Incorporation ofN-acetylglucosamine into proteinand lipid intermediates in microsomal and Golgi membranes from rat liver.Biochim Biophys Acta 512: 123-135

4. BERGNER U,W TANNER 1981 Occurrence ofseveral glycoproteins in glyoxysomalmembranes of castor beans. FEBS Lett 131: 68-72

5. BoRDiER C 1981 Phase separation of integral membrane proteins in Triton X-114 solution. J Biol Chem 256: 1604-1607

6. BRAY GA 1960 A simple efficient liquid scintillator for counting aqueoussolutions in a liquid scintillation counter. Anal Biochem 1: 270-285

7. CHADwIcK CM, DH NORTHCOTE 1980 Glucosylation of pyrophosphoryl-polyisoprenol and sterol at the plasma membrane of soya bean (Glycine max)protoplasts. Biochem J 186: 411-421

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

9. CooPER TG, H BEEvERs 1969 Mitochondria and glyoxysomes from castor beanendosperm. Enzyme constituents and catalytic capacity. J Biol Chem 244:3507-3513

10. CzIcm U,WJ LENNARZ 1977 Localization ofthe enzyme system for glycosylationofproteins via the lipid-linked pathway in rough endoplasmic reticulum. J BiolChem 252: 7901-7904

11. DALLMAN PR, G DALLNER, A BERGSTRAND, L ERNSTER 1969 Heterogeneousdistribution of enzymes in submicrosomal membrane fragments. J Cell Biol41: 357-377

12. DALLNER G 1974 Isolation of rough and smooth microsomes-general. MethodsEnzymol 31: 191-201

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

14. ERiCSON MC, DP DELmi 1977 Glycoprotein synthesis in plants. 1. Role of lipidintermediates. Plant Physiol 59: 341-347

15. FoRsEE WT, AD ELBEIN 1975 Glycoprotein biosynthesis in plants. Formation oflipid-linked oligosaccharides of mannose and N-acetylglucosamine by mungbean seedlings. Arch Biochem Biophys 174: 469-478

16. HANOVER JA, WJ LENNARZ 1980 N-linked glycoprotein assembly. Evidence thatoligosaccharide attachment occurs within the lumen of the endoplasmic retic-ulum. J Biol Chem 255: 3600-3604

17. HoRi H, DW JAMES, AD ELBEIN 1982 Isolation and characterization of theGlc3Man9GIcNAc2 from lipid-linked oligosaccharides of plants. Arch BiochemBiophys 215: 12-21

18. KATZ RN, JE ROTHMAN, VR LINGAPPA, G BLOBEL, HF LODISH 1977 Membraneassembly in vitro: synthesis, glycosylation and asymmetric insertion of atransmembrane protein. Proc Natl Acad Sci USA 74: 3278-3282

19. KiEy ML, GS McKNIGHT, RT ScHInAKE 1976 Studies on the attachment ofcarbohydrate to ovalbumin nascent chains in hen oviduct. J Biol Chem 251:5490-5495

20. LoRD JM, T KAGAWA, TS MooRE, H BEEVERS 1973 Endoplasmic reticulum asthe site of lecithin synthesis in castor bean endosperm. J Cell Biol 57: 659-667

21. LOwRY OH, NJ ROSEBROUGH, AL FARR, RJ RANDALL 1950 Protein measure-ment with the Folin phenol reagent. J Biol Chem 193: 265-275

22. MAmuuoTT KM,WTANNER 1979 Dolichylphosphate-dependent glycosyl transferreactions in the endoplasmic reticulum ofcastor bean endosperm. Plant Physiol64:445-449

23. MELLOR RB, JM LoRD 1979 Formation of lipid-linked mono- and oligosaccha-rides from GDP-mannose by castor bean endosperm homogenates. Planta 146:91-99

24. MELLOR RB, LM ROBERTS, JM LoRD 1979 Glycosylation of exogenous proteinsby endoplasmic reticulum membranes from castor bean (Ricinus communis)endosperm. Biochem J 182: 629-631

25. MELLOR RB, LM ROBERTS, JM LoRD 1980 N-Acetylglucosamine transfer reac-tions and glycoprotein biosynthesis in castor bean endosperm. J Exp Bot 31:993-1003

26. NAGAHASH J, L BEEVERS 1978 Subcelular localization of glycosyltransferasesinvolved in glycoprotein biosynthesis in the cotyledon of Pisum sativum L.Plant Physiol 61: 451-459

27. PARODI AJ, LF LELOIR 1979 The role of lipid intermediates in the glycosylationof proteins in the eukaryotic cell. Biochim Biophys Acta 559: 1-37

28. PoNT-LEZICA R, PA RoMERO, HE Hopp 1978 Glucosylation ofmembrane boundproteins by lipid-linked glucose. Planta 140: 177-183

29. RoBERTs LM, JM LoRD 1981 The synthesis of Ricinus communis agglutinin.Cotranslational and posttranslational modification of agglutinin polypeptides.Eur J Biochem 1I 9: 31-41

30. ROBERTS LM, RB MELLOR, JM LoRD 1980 Glycoprotein fucosyltransferase inthe endoplasmic reticulum of castor bean endosperm cells. FEBS Lett 1I 3: 90-94

31. STANELONI RJ, ME TOLmASKY, C PETRUELLA, LF LELOIR 1981 Transfer ofoligosaccharide to protein from a lipid intermediate in plants. Plant Physiol 68:1175-1179

32. SVENSSON H, G DALLNER, L ERNSTER 1972 Investigation of specificity in mem-brane breakage occurring during sonication of rough microsomal membranes.Biochim Biophys Acta 274: 447-461

33. VIGIL EL 1970 Cytochemical and developmental changes in microbodies (glyox-ysomes) and related organelles of castor bean endosperm. J Cell Biol 46: 435-454

552 CONDER AND LORD

Dow



heterogeneous distribution of glycosyltransferases in the

Documents