Distribution of membrane glycoproteins among the organelles of a single identified neuronm of aplysia. II. isolation and characterization of a glycoprotein associated with vesicles

Brain Research, 207 (1981) 33~,8 © Elsevier/North-Holland Biomedical Press

33

DISTRIBUTION OF MEMBRANE GLYCOPROTEINS AMONG THE ORGANELLES OF A SINGLE IDENTIFIED N E U R O N OF APLYSIA. II. ISOLATION AND CHARACTERIZATION OF A GLYCOPROTEIN ASSOCIATED WITH VESICLES

R I C H A R D T. AMBRON, A. A. SHERBANY and JAMES H. SCHWARTZ

Departments of Anatomy, Physiology and Neurology, Division of Neurobiology and Behavior, College of Physicians and Surgeons, Cohlmbia University, New York, N. Y. 10032 (U.S.A.} (Accepted July 17th, 1980)

Key words: vesicles - - membrane glycoprotein - - identified Aplysia neuron - - anisomycin

SUMMARY

Glycoprotein-I (mol.wt. 180,000) is associated with a vesicle fraction from the cytoplasm of R2, the giant cholinergic neuron in the abdominal ganglion. Electron microscopy has shown that R2 contains both lucent and compound vesicles. We have used anisomycin, an inhibitor of protein synthesis, to provide evidence that Component-I is a constituent of somatic compound vesicles. In the presence of the drug, [3H]N-acetylgalactosamine injected into R2 is incorporated almost exclusively into Component-I. Quantitative electron microscopic radioautography of treated cells shows a marked increase in the proportion of silver grains over compound vesicles and a decrease in labeling of other organelles compared with untreated cells. Analysis of Component-I, isolated using the chaotropic agent lithium diiodosalicylate, shows it to contain fucose, N-acetylgalactosamine, and N-acetylglucosamine. Proteolytic digestion with pronase yields a complex pattern ofglycopeptides. The proportion of [aH]N-acetylgalactosamine in these glycopeptides is altered in the presence of anisomycin. These results together with radioautographic analyses suggest that large carbohydrate chains are elaborated within the endoplasmic reticulum and that smaller chains are added in the Golgi region or on the membrane of the compound vesicle.

INTRODUCTION

In the companion paper 4 evidence was presented that glycoprotein-I might be a component of vesicles in the cell body of R2. This glycoprotein has a molecular weight of 180,000 Daltons and migrates anomalously during gel electrophoresis in detergent,

34 suggesting that the molecule has a large proportion of carbohydrate residues. In this paper, we have isolated and partially characterized Component-I labeled with [3H]Nacetylgalactosamine and have also attempted to determine whether it is characteristic of a particular type of vesicle. Since R2's cytoplasm contains more than one kind of vesicle, it would require the isolation of each type before Component-I could be assigned definitively. This has not yet been accomplished, but we have taken another approach. We have used anisomycin, an inhibitor of protein synthesis2, 24. Exposure of R2 to the drug prior to injection results in a marked change in the distribution of labeled glycoprotein within the cell: Component-I becomes essentially the only glycoprotein labeled. Using subcellular fractionation and quantitative radioautography we have been able to determine the distribution of Component-1 within R2 and to show that this glycoprotein is associated with a specific population of somatic vesicles. MATERIALS AND METHODS The techniques of intracellular injection, subcellular fractionation, glass bead filtration, and gradient centrifugation are described in the companion paper 4. In experiments with anisomycin, nervous tissue was exposed to the drug at a concentration of 18 #M 4 h prior to injection, and for the entire period thereafter2, z4. Abdominal ganglia were prepared for radioautography 4 and were analyzed by the morphometric procedure of Williams a0 and Salpeter, et al. 2z. Discontinuous polyacrylamide gel electrophoresis in sodium dodecyl sulphate (SDS) was carried out as described 4. Since we have observed unaccountable differences in the migration of the membrane glycoproteins when using different batches of gel reagents, the identity of any isolated glycoprotein was always verified by running a sample containing all of the [aH]glycoproteins on a parallel gel. Radiochemicals were purchased from New England Nuclear, Boston, Mass.

Extraction of [3H]glycoprotein with lithium diiodosalicylate [3H]glycoproteins in the lipid-depleted 105,000 × g membrane pellet from an injected R2 were extracted in 0.5 ml 0.3 M lithium diiodosalicylate (LIS) (Eastman Kodak, Rochester, N.Y.), 0.2 M 2-mercaptoethanol, in 0.05 M Tris-HCl (pH 7.6) z°. The homogenate was heated at 70 °C for 15 rain, diluted with an equal volume of water, and particulate material was removed by centrifugation at 105,000 × g for 1 h. The supernatant was removed, mixed at 4 °C with an equal volume of 50 ~ (v/v) aqueous freshly-distilled phenol, and phase separation was expedited by low speed centrifugation. The aqueous (upper) phase was removed and filtered on a column (1.5 × 10 cm) of Sephadex G-200 to separate [3H]glycoprotein from phenol. Alternatively, [3H]glycoprotein in the aqueous phase was recovered by acetone precipitation 1. Proteolytic digestion. Acetone-precipitated [3H]glycoproteins were digested with pronase (Calbiochem, San Diego, Calif.; 45 U/mg) for 72 h at 37 °C in a volume of 0.5 ml containing approximately 0.25 mg substrate protein, 0.5 mg enzyme, and a drop of

35 toluene. Digests and controls were examined by gel filtration on a column (1.5 x 25 cm) of Sephadex G-50 (Pharmacia, Piscataway, N.J.) in 0.1 M pyridine-acetate (pH 4.7). The pronase was incubated at 37 °C for 30 min prior to the digestion to destroy any other enzyme activities which might contaminate the enzyme. Paper chromatography. Compounds were chromatographed on Whatman No. 1 in the descending direction using butanol-ethanol-HzO (10:1:2, solvent system A), ethyl acetate-pyridine-acetic acid-water (5:5:1:3, solvent sytem B) of 0.1 M ammonium acetate (pH 7.5)-ethanol (3:7, solvent system C). Sugar standards were visualized with silver nitrate 21. Acid hydrolysis. [SH]glycoprotein was hydrolized in 4 N HC1 at 105 °C for 8 h and chromatographed in solvent B, a system that separates galactosamine from glucosamine 11. Under this condition of hydrolysis, all of the radioactivity becomes water-soluble and N-acetylgalactosamine is deacylated to galactosamine.

Digestion with chondroitanase ABC, and hyaluronidase The effects of the mucopolysaccharidase, chondroitanase ABC (Proteus vulgaris type, Miles Laboratories, Elkhart, Ind.) on the [3H]glycoproteins were tested as described by Yamagata, et al. 31. Membrane [3H]glycoproteins precipitated with acetone were suspended in 50 mM Tris, 60 mM sodium acetate (pH 8.0) containing 5 U of enzyme and 0.50 mg of chondroitan sulfate. Control tubes contained either chondroitan sulfate (from umbilical cord, Sigma Chemicals, St. Louis, Mo.) or chondroitan sulfate together with enzyme. Incubation was at 37 °C for 1 h, followed by addition of ice-cold acetone to all the samples. Although the enzyme was in great excess, all of the treated [ZH]glycoprotein was precipitated, indicating that the enzyme had no effect. This was confirmed by comparison of treated and untreated [aH]glycoprotein on gel electrophoresis in SDS. The effect of hyaluronidase was determined using the same protocol: 60 U bovine testis hyaluronidase (Sigma Chemicals, St. Louis, Mo., Type IV) in 0.5 ml 0.15 M NaC1, 0.1 M sodium phosphate buffer (pH 5.3) containing 0.1 mg umbilicus hyaluronic acid (Sigma) was added to a sample of acetone-precipitated membrane [3H]glycoprotein. The suspension was incubated at 37 °C for 1 h followed by acetone precipitation. All of the [3H]glycoprotein was precipitated after enzyme treatment. Examination of the precipitate by gel electrophoresis showed no difference between samples that were treated with enzyme and the controls. This series of tests indicates that the labeled macromolecules are glycoproteins and not glycosaminoglycans. RESULTS

Isolation of Component-I Component-I can be separated from the other [aH]glycoproteins with the chaotropic agent LIS. Treatment of the membrane pellet with hot LIS extracted between 45-60 % of the total [aH]glycoprotein (Table I). When this extract was partitioned between phenol and water, approximately half of the radioactivity entered the aqueous phase. Analysis by polyacrylamide gel electrophoresis in SDS showed that

36 TABLE I

Extraction of membranes from R2's cell body with lithium diiodosalicylate 22 h after injection of [aHJN-acetylgalactosamine The abdominal ganglion containing an R2 injected either in the presence or absence of anisomycin was homogenized and fractionated at 105,000 × g to isolate membranes4. After lipid was removed, the pellet was extracted with hot LIS and the extract partitioned with phenol as described in Methods. Values include mean ~ S.E.M.

Untreated Anisomycin

3 4

% Totalparticulate [3H]glycoprotein extracted

% extracted [aHJglycoprotein in aqueous phase

45.7 zk 8.9 68.0 ~: 7.2

58.2 i 5.8 68.8 ~ 3.9

most of the [3H]glycoprotein from the aqueous phase was in Component-I (Fig. 1). Similar results were obtained when the protein was labeled after injection of [3H]fucose, providing evidence that the glycoprotein contains both sugars. The other labeled glycoproteins were found in the phenol phase. Presumably Component-I partitions into the aqueous phase after LIS extraction because it contains a large number of carbohydrate moieties 2°.

The effects of anisomycin in glycosylation of Component-I Anisomycin profoundly inhibited incorporation of injected sugar into membrane glycoprotein (Table II); in the presence of the drug most of the [3H]fucose or [3H]N-acetylgalactosamine in glycoprotein was associated with Component-I (Fig. 2 900

'1'

II

II

M

750

BPB

600 450 rl.

300 150 /

~b

20

30

40

50

60

~o

DISTANCE(mrn) Fig. 1. Discontinuous polyacrylamide gel electrophoresis in SDS of the aqueous phase after LIS extraction. 15 h after intrasomatic injection of [aH]N-acetylgalactosamine into R2, the abdominal ganglion was homogenized and membranes isolated by centrifugation at 105,000 × g. The membranes were extracted with LIS and the extract partitioned with phenol as described in Methods. The aqueous phase was treated with acetone and the precipitate was subjected to electrophoresis on a 7.5 % gel. Radioactivity in mm segments of the gel was then determined. The position of [3H]Component-I from an SDS extract of R2's membranes, run on a parallel gel, is indicated by an arrow.

37 TABLE Ii

Effect of anisomycin on incorporation of [aH] N-acetylgalactosamine into glycolipid and glycoprotein in R2' s cell body 22 h after injection in the presence or absence of anisomycin, the abdominal ganglion containing R2 was homogenized and fractionated at 105,000 × g. Soluble glycoprotein was precipitated from the supernatant and the pellet was sequentially extracted with chloroform-methanol to remove glycolipid, SDS to remove glycoprotein, and formic acid to remove the small amount of residual pH]glycoprotein 4. All of the radioactivity in the cell is recovered by this procedure. Labeled particulate glycoprotein is the sum of the radioactivity in the SDS and formic acid extracts.

Distribution of radioactivity ( % total radioactivity in cell body) Control (n = 7) Soluble glycoproteins Particulate macromolecules Glycolipid Glycoprotein

M 300

Anisomycin (n =3)

Inhibition(%)

9.0 ~_ 1.8

6.2 d_ 2.0

31

14.3 :k 3.0 40.4 i 3.6

8.0 ± 3.0 7.7 -- 3.3

44 81

BPB

A

250

200 150 I00

i >,_

50-

300 _o 250

20C 150 IOC

10

20

30

40

50

60

DISTANCE ( m m )

Fig. 2. Discontinuous polyacrylamide gel electrophoresis in SDS of membrane glycoproteins from R2 injected in the presence of anisomycin. The central nervous system was removed from the animal and placed in artificial seawater containing 18 ~ M anisomycin for 4 h. R2 was then injected with (A) [SH]N-acetylgalactosamine or (B) [ZH]fucose. After 15 additional hours in the presence of the drug, the abdominal ganglion was homogenized and fractionated as described in the legend to Table II, except that lipid was not removed. The membrane pellet was extracted with sample buffer and electrophoresed. The running gel was 7 ~ . all-lipid ran with the dye marker. D, dimer; M, monomer of beef serum albumin; BPB, Bromophenol blue. Arrow indicates position of Component-I from membranes of untreated R2.

38 and ref. 2). As expected, LIS extracted a greater proportion of [SH]glycoprotein from R2s treated with anisomycin, and a greater percentage of the extracted material was found in the aqueous phase (Table I). Even in anisomycin-treated cells, however, not all of the glycoprotein could be extracted. Attempts to increase the yield by repeated extraction, by lengthening the duration of the extraction, or by raising the temperature were not successful. Because we have not been able to solubilize the labeled glycoprotein which remains in the pellet after LIS treatment, the composition of this fraction is unknown.

Association of Component-I with vesicles in presence of anisomycin The distribution of Component-I among the organelles of injected R2s treated with the drug was examined by subcellular fractionation and glass bead filtration 4. Labeled cytoplasm from anisomycin-treated cells was first separated f r o m the nucleus and external membrane by dissection. The cytoplasm was then added to a homogenate of Aplysia nervous tissue and mitochondria, terminal lysosomes, and other large organelles were removed by differential centrifugation. Membranes in the supernatant ($15) f r o m this procedure contained Component-I as the only labeled glycoprotein. When filtered on a column of glass beads, $15 gave an elution profile containing two distinct peaks of radioactivity (Fig. 3). Approximately 10 ~ of the total radioactivity was partially included. This is less than half the proportion found in untreated cells. Otherwise the distribution of radioactivity from the column was similar to that

36,000

30,000

PsB

24,000

18,OO0

o~

C3 12,000

VOLUME (rnl)

Fig. 3. Filtration of the mitochondrial supernatant (Szs) from the cytoplasm of an R2 injected with

[aH]N-acetylgalactosamine in the presence of anisomycin. Cytoplasm obtained by dissection 24 h after injection and containing 147,000 cpm, was combined with the low-speed supernatant from a homogehate of 30 Aplysia ganglia, was subjected to differentiation centrifugation and the supernatant (Sz~) was filtered on a column of glass beads (0.9 × 55 cm, nominal pore size 200 nm)4. The column was standardized with polystyrene beads (PSB), 10, 60 and 90 nm colloidal gold particles (arrows), [3H]serotonin-containing vesicles (SV), and blue dextran (BD). The dotted line shows the position of the membrane fraction from untreated cells4. This is a representative profile obtained from one of four individual experiments.

39 TABLE III Distribution of silver grains over R2's cell body 15 h after injection of !3HJ N-acetylgalactosamine in the presence of anisomycin

For the analysis, 652 silver grains and 2197 area circles were counted. Grain distribution (%)

Labeled* Golgi membranes Multivesicular bodies Peroxisomes Lucent vesicles Compound vesicles Not labeled* * Large lysosomes Cytoplasm Mitochondria Smooth endoplasmic reticulum Rough endoplasmic reticulum Nucleus Outside (glial cells, connective tissue)

Effective area (%)

Observed grains

Expected grains

Relative specific activity* * *

5.8 4.0 10.0 2.9 22.9

1.0 1.0 ! .5 1.5 3.6

38 26 65 19 149

6.6 6.5 16.2 9.8 23.4

5.8 4.0 6.7 1.9 6.4

4.8 34.2 2.1 9.5 0.6 0

4.4 61.4 1.9 16.7 0.7 0.4

31 223 14 62 4 0

28.4 398.2 12.7 108 4.7 0

1.1 0.6 1.1 0.6 0.9 0

3.2

6.0

21

38.9

0.5

* Items labeled to a greater extent than expected from a random distribution of grains. Significance (P < 0.05) was determined by a Z2 test. ** Items labeled to a lesser extent than expected from a random distribution of grains (P < 0.05). * * * Relative spec. act. is the ratio of the frequency of grains to the effective area. previously obtained using the cytoplasm f r o m untreated, injected cells 4. The peak o f partially included radioactivity coincided with that o f serotonin-labeled vesicles and also with particles o f gold (90 n m mean diameter) which were used to calibrate the column. M o r e than 80 ~ o f the radioactivity in this fraction was associated with membranes and, on the average, these membranes contained 34 % of the total C o m p o n e n t - I in the cytoplasm. Labeled material in the totally-included fraction was all soluble glycoprotein and low molecular weight precursors. Using a 3-step gradient, we further analyzed the vesicle fraction f r o m the glass bead column. T w o fractions o f radioactive glycoprotein were obtained, b o t h in the region o f the gradient that has been previously shown to contain vesicles (see Fig. 6 in a c c o m p a n y i n g paper 4) and which was also occupied by serotonin-labeled vesicles. W h e n the radioactive material f r o m the gradient was centrifuged at high speed, and the pellet washed with alkaline, hypotonic buffer 1, approximately 60% of the radioactivity was f o u n d to be in membranes, one-half in glycoprotein, the remainder in glycolipid. R a d i o a u t o g r a p h y . The distribution o f silver grains in electron microscopic radioautographs o f R2's cell b o d y injected after treatment with the drug was quite different f r o m that in untreated cells (Tables III, IV, and Table I, a c c o m p a n y i n g paper4). Thus, the s m o o t h and r o u g h endoplasmic reticulum were significantly labeled in untreated

40 TABLE IV Distribution o f silver grains over membrane organelles after injection in the presence or absence o f anisomycin Values for the grain distribution were calculated from data in Table 1II by considering only those grains associated with membranous organelles within R2. Values for relative specific activities (RSA) presented in Table III were adjusted by neglecting the silver grains that did not lie over the neuron: new RSAs were calculated for each item using the adjusted values for effective area and for per cent of grains. In each experiment, one mean combined RSA was then calculated for all of the unlabeled items ( ÷ anisomycin, 0.54; - - anisomycin, 0.65) only if that item was not labeled in both experiments (e.g. mitochondria). The adjusted RSA was then obtained by normalizing the specific activity of unlabeled items to a value of 1.0. Organdie

Golgi membranes Multivesicular bodies Peroxisomes Lucent vesicles Compound vesicles Smooth endoplasmicreticulum Rough endoplasmic reticulum External membrane Total grains

**

% o f grains

Adjusted relative specific activity

+ anisomycin

- - anisomyein

+ anisomycin

- - anisomycin

9.3 6.4 15.9 4.6 36.5 15.2 1.0 *

12.2 5.0 ** 3.5 4.2 42.3 16.4 6.6

9.1 6.2 10.8 3.2 10.5 1.0 1.5 *

13.2 6.3 ** 3.4 4.6 4.5 3.8 3.9

408

286

* Not analyzed. Occupied too small an effective area to be statistically significant.

~;iiiiL :~i~iii!i~!i;iiiii:~:~i!!il ~/i~i i ! ~

! ii dCI~ ¸ ! ~

Fig. 4. High magnification view of compound vesicles (arrow) in cytoplasm of R2. x 62,000.

41

-g

525

T

450 575 500 225 0 cl

150

75


0~. . .IO . . 20 3 0' " 04 . .50 . . 60 ORIGIN DISTANCE (mm)

Fig. 5. Discontinuous polyacrylamide gel electrophoresis in SDS of total membranes obtained from the giant cerebral neuron injected with [3I-I]N-acetylgalactosaminein the presence of anisomycin. [3H]lipid ran at the front. Arrow indicates position of Component-I from R2. cells, but not after anisomycin treatment. Among all the labeled organelles, only compound vesicles showed an increase in relative specific activity in the presence of anisomycin (Table IV). There was also a marked increase in the percentage of total silver grains associated with these vesicles. Compound vesicles in R2 consist of a small vesicle contained within a larger vesicle (90 nm mean diameter) (Fig. 4). Three consecutive serial sections containing 12 vesicles have shown that these are indeed vesicles and not tubes. Similar vesicles are also seen in other neurons of Aplysia, for example, in the serotonergic giant cerebral neuron14,zs, 27. It is interesting, therefore, that this neuron also has a rapidly labeled glycoprotein similar to Component-I. This glycoprotein, which partitions into the aqueous phase after LIS extraction, also continues to be glycosylated in the absence of protein synthesis, and has a similar mobility on polyacrylamide gels as the Component-I from R2 (Fig. 5).

Characterization of Component-Iformed in the absence of anisomycin Analysis of glycopeptides. Prolonged exposure of Component-I labeled with [3H]N-acetylgalactosamine to pronase reproducibly degraded the glycoprotein to 3 size classes of [3H]glycopeptides. The major proportion of the radioactivity was associated with [aH]glycopeptides that were excluded or partially-included during gel filtration on Sephadex G-50 (peaks A and B, Fig. 6A). A small amount of radioactivity was totally included (peak C, Fig. 6A). Prolonging the digestion of the excluded or partially-included material did not further reduce its molecular weight, nor did it result in the production of totally-included glycopeptides. When the excluded radioactivity was further analyzed on a column of Sephadex G-100, it was partially included. Since intact Component-I is excluded on this column this material must consist of large glycopeptides and not undigested Component-I*. Glycopeptides of this s]ze are * When [SH]glycoproteinsin the phenol phase from the LIS extraction were exposed to pronase, more than 90 700of the radioactivity appeared in [:~HlglycopeptidessmaUenough to be included by Sephadex G-50. The presence of excluded glycopeptides after digestion is therefore not a general property of Aplysia glycoproteins.

42 unusual in nervous tissue but large glycopeptides were recently found in digests of glycoproteins from the garfish olfactory tract 10. Incorporation of [35S]methionine into Component-L When R2's membranes, labeled during incubation of the excised nervous system for 17 h with [35S]methionine (224 Ci/mMol) 9, were extracted with LIS, less than 10 ~ of the particulate radioactivity was found in the aqueous phase. Three main peaks of radioactivity were resolved by polyacrylamide gel electrophoresis of this material; one of these had a migration similar to that of Component -6 (not shown). Treatment with Mucopolysaccharidases. None of the glycoproteins in R2 are affected by treatment with chondroitanase ABC or hyaluronidase (see Methods). After treatment the labeled glycoproteins remained acetone-precipitable and their migration on polyacrylamide gels was unaltered. Moreover, cetylpyridinium bromide, which forms an insoluble complex with mucopolysaccharides 23, did not precipitate Component-I purified by LIS, even though the chondroitan sulfate added to Component-I as carrier was precipitated from the solution. Acid hydrolysis. When Component-l, isolated after injection of [3H]N-acetylgalactosamine, was hydrolyzed under conditions that released all of the radioactive sugar, both pH]galactosamine and pH]glucosamine were found after paper chromatography (Fig. 7). An additional component which had approximately 30 ~ of the radioactivity was also present on the chromatogram. Its migration indicated that it might be a trisaccharide, but it was not further characterized.

Characterization of UDP-[3H] N-acetylgalactosamine We have tentatively identified the nucleotide form of N-acetylgalactosamine in the soluble fraction of injected R2s~. Paper chromatography in solvent system C of the

A

B

1200 • "~ I000 8OO

BO

b- 6OO

2OO

5

I0

15

20

i5

~

o

i

5

i

,o

15

i

r

20

25

i

FRACTION NUMBER

Fig. 6. Pronase digestion of pH]Component-I from R2 injected with pH]N-acetylgalactosamine in the absence (A) or presence (B) of anisomycin, pH]Component-I was obtained from the aqueous phase after LIS extraction. A sample was removed for polyacrylamide electrophoresis to be sure no other labeled glycoproteins were present. The remaining material was digested with pronase for 72 h at 37 °C as described in Methods. The digest was filtered on a column (1.5 × 25 cm) of Sephadex G-50 using 0.1 M pyridine-acetate (pH 4.7). The column was standardized with blue dextran (BD) and pH] N-acetylgalactosamine (arrow).

43 GluNFI5 o. u

GalNH3 ?

240

NAG01 "

C)

)-

160 F--

o 0

47 - -

/ Origin

5

I0

15

20

25

30

35

DISTANCE (cm)

Fig. 7. Acid hydrolysis of [sH]Component-I isolated from R2 after injection of [3H]N-acetylgalactosamine. [3H]Component-I was obtained from the aqueous phase after LIS extraction and was hydrolyzed in 4 N HCI for 8 h at 100 °C. Under these conditions, all of the [SH]sugar is released from glycoprotein. The acid was removed by lyophilization and the residue transferred to a sheet of Whatman No. 1 for development in the descending direction in solvent system B. The chromatogram was cut into 1 cm segments for 22 cm and 0.5 cm for the next 6 cm. Each segment was then counted. Sugar standards were visualized with the silver nitrate reagent. [aH]N-acetylgalactosamine (NAGal) hydrolyzed under identical conditions was converted to [3H]galactosamine (GalNH3) and glucosamine, (GluNH3).

UDP-

c~

glu .

v

1200 > I-0

_o C3 OC

800 40O 0

-

5

IO

15

20

25

3"0

Origin

O o

B

Man. NAqlu. GIuNH3,

u

,cI

NAgol.

90C

).I--

0

"

6001300

0 r~

Gal. r~

nr

0 5 Origin

I0 15 20 DISTANCE (crn)

25

:30

Fig. 8. Characterization of UDP-N-acetylgalactosamine from the soluble fraction of R2 after injection of [3H]N-acetylgalactosamine. Soluble radioactivity (68,000 cpm) free of membranes and soluble glycoprotein, was placed on a column (1 x 5 cm) of AG-1-2 × (acetate form). The column was eluted, in sequence, with water, 0.1 M pyridine-acetate (pH 4.7), 1 M KCI, and finally with 6 M urea. All of the radioactivity was recovered. The material eluted by 1 M KCI (32,000 cpm) was treated with activated charcoal as previously described 4. All of the radioactivity was adsorbed and all was released with dilute ammonia in 50 70 aqueous ethanol. A: a sample of the released radioactivity was chromatographed in solvent system C. A single peak is present with a mobility relative to standard UDP-glucose of 1.1. B: the remaining material was hydrolyzed in 0.25 M HCI at 100 °C for 15 rain, lyophilized, and chromatographed in solvent system B. Most of the radioactivity migrated with standard N-aeetylgalactosamine but was not completely separated from N-acetylglucosamine. Glucosamine, GluNH3; Mannose, Man galactose, Gal.

44 isolated material gave a single radioactive peak with a migration of 1.2 relative to standard UDP-glucose (Fig. 8A). Mild acid hydrolysis (0.25 N HC1, 100 °C, 15 min) released 80 ~ of the radioactivity as a single component which migrated with N-acetylgalactosamine in solvent system B (Fig. 8B). Since this peak was not completely resolved from N-acetylglucosamine, we cannot exclude the possibility that this labeled sugar was present also. Two minor components on the chromatograph were not further identified.

Comparison of Component-Iformed in the presence and absence of anisomycin In contrast to the normal distribution of labeled glycopeptides from Component-I (see above, Fig. 6A) the profile of digested Component-I from anisomycintreated cells showed that most of the radioactivity was associated with the totallyincluded peak (Fig. 6B). Excluded and partially included material was also present, but contained only a small proportion of the radioactivity. Thus, the shorter oligosaccharide moieties are preferentially labeled in the presence of anisomycin. Although there were quantitative differences in the proportion of radioactivity associated with the [ZH]glycopeptides from the two sources of Component-I, these are apparently all derived from a single species of glycoprotein. We have analyzed the included material (Fig. 6A, B) from digests of Component-I obtained from normal and anisomycin-treated R2s. Each yielded two [ZH]glycopeptides which co-migrated on paper chromatography. Preliminary analyses of the larger material (Fig. 6A, B) from both sources indicated that it too consists of similar [3H]glycopeptides. A plausible explanation for these results (see Discussion) is that Component-I isolated from both untreated cells and those exposed to anisomycin have common labeled oligosaccharide moieties, but that the proportion of [3H]N-acetylgalactosamine introduced into shorter and longer branches is selectively altered by the drug. DISCUSSION

Association of Component-I with compound vesicles Treatment of R2 with anisornycin 4 h prior to injecting [3H]N-acetylgalactosamine resulted in a marked increase in the proportion of labeled Component-I within the cell. At the same time, radioautographic analyses have shown that the proportion of radioactive glycoprotein associated with compound vesicles is greatly increased (Table IV). The most reasonable explanation for these results is that Component-I is a constituent of compound vesicles. This idea is supported also by our subcellular fractionation studies, since in both anisomycin-treated and untreated cells the glycoprotein is present in an organelle that emerges from the glass bead column in the same fraction as serotonergic vesicles and 90 nm gold particles (Fig. 3). Moreover, when the labeled membranes were analyzed by centrifugation on an iso-osmolar sucrose-Ficoll gradient, [~H]Component-I was found to be associated only with the vesicle fraction above the 0.7 M interface that has been shown to contain vesicles 4. We have explicitly made the assumption that the association of Component-I with compound vesicles in the presence of anisomycin is not an artifact resulting from

45 the action of the drug. Although this assumption has not been proved, evidence in its support is of two kinds, physiological and biochemical. Under the conditions in which the drug was used, anisomycin has been shown not to change the neuron's electrophysiological properties 24, nor does it reduce the concentration of somatic ATPL Moreover, neither the export of serotonergic vesicles from the cell body of the giant cerebral neuron 13 nor the rapid transport of serotonin la or [3H]glycoprotein2 along axons is affected by the drug. Biochemically, the glycoprotein produced in the presence of anisomycin appears to be identical to Component-I made in the untreated neuron: (1) both glycoproteins co-migrate during electrophoresis in several polyacrylamide gel systems in SDS; (2) both are extracted with LIS and partition into the aqueous phase, an unusual property characteristic of glycoproteins that contain large numbers of carbohydrate residues2°; and (3) both glycoproteins yielded similar classes of labeled glycopeptides after pronase digestion (Fig. 6). Although the proportion of [3H]glycolipid to [aH]glycoprotein produced in treated neurons was increased, the 5 [3H]glycolipids from both treated and untreated cells were the same 26. It has not yet been determined whether any of the [3H]glycolipids are associated with compound vesicles. Radioautography is of little help since glycolipids are extracted during fixation and dehydration 16.

The biosynthesis of Component-I Correlation of quantitative radioautography and biochemical analyses of [3H]Component-I has provided some insights as to how this glycoprotein is processed. Initial characterization of the labeled molecule shows that it contains fucose, N-acetylgalactosamine, and N-acetylglucosamine (Figs. 2, 7). N-acetylglucosamine is probably added during processing of the nascent polypeptide on membranes of the rough endoplasmic reticulumn,7,17. Consistent with this idea is the observation that the reticulum is labeled after injection (Table IV). Fucose, on the other hand, is likely to be added to Component-I in the Golgi region 18,2~. We are not certain where in the cell Nacetylgalactosamine is added, although recent studies have implicated both the smooth and rough endoplasmic reticulum in this processT, 28. Glycoproteins labeled with radioactive sugars would normally contain label incorporated at all of these sites. In cells treated with anisomycin, however, glycosylation would be expected to occur primarily on those polypeptides already synthesized, but not completely processed prior to introduction of [3H]precursor. Labeling, therefore, should preferentially reflect later steps in the sequential glycosylation of protein. Thus, it is likely that the reduction in labeling of the large chain(s) of Component-I by anisomycin (Fig. 6) indicates that these oligosaccharides are normally assembled on the rough endoplasmic reticulum, since radioautographic labeling of these structures is also reduced relative to controls (Table IV). In contrast, treatment with anisomycin results in a relative increase in the labeling of short chains with N-acetylgalactosamine, and a parallel increase in the labeling of Golgi membranes. These observations are consistent with the idea that this sugar is added in the Golgi apparatus. Labeling of the short chains of Component-I might also occur after

46 the molecule is incorporated into compound vesicles, since this has already been demonstrated in Aplysia axons 5. Further radioautographic studies at various times after injection of the labeled precursor would clarify where the glycosylation occurs in the neuron.

Populations of compound vesicles with different functions Compound vesicles similar to those in Aplysia have been observed in the pars intermedia of Xenopus 15, in the mouse cerebrum s, and elsewhere (see ref. 27). It has been suggested that in these diverse tissues they may play a role in the restructuring or reorganization of membranes. The association of specific glycoproteins with vesicles is not restricted to Aplysia, since characteristic glycoproteins have also been found in membranes of the chromaffin granulelL Compound vesicles labeled with ComponentI in R2 are not likely to be transmitter storage vesicles, however, since R2 is cholinergic and the neurotransmitter is thought to be associated with lucent vesicles (see ref. 29). Moreover, only a small proportion of [3H]Component-I appears in the axon.

Compound vesicles are also found in the axon25,27 indicating that there are two populations of vesicles with similar morphology. One type, containing Component-I is destined to remain in the cell body, and the other, containing little or no Component-I, is exported into the axon. We have previously provided evidence that there are also two types of compound vesicles in the serotonergic GCN (Refs. 3, 4, 27). One type contains serotonin; the other may be homologous to the somatic compound vesicle of R2. Studies of GCN's glycoprotein composition have shown that it also has a glycoprotein that is glycosylated in the presence of anisomycin and which has similar properties to the Component-I of R2 (Fig. 5). Since somatic compound vesicles in GCN are rapidly labeled after injection of sugar precursor 8, it is likely that they also contain glycoprotein-I. Neither of the two vesicles types in R2, however, take up [3H]serotonin after intrasomatic injection of the transmitter25. The function of somatic compound vesicles associated with Component-I is not known in either R2 or GCN. They may be precursors of neurotransmitter vesicles, however. If so, the reduced amount of [zH]Component-I in the axon may result from conversion of Component-I to a lower molecular weight glycoprotein that takes place during transport of the vesicles along the axon. Processing of proteins during transport has been described in neurosecretory vesicles in Aplysia19 and in other animals. Maturation of aminergic synaptic vesicles occurs both in vertebrates and invertebrates (see ref. 27). ACKNOWLEDGEMENTS

The authors thank Ms. Lise Castellucci for expert technical assistance and Drs. M. Gershon, D. J. Goldberg, J. E. Goldman, and E. R. Kandel for critical reading of the manuscript. This research was supported by National Institutes of Health Research Grants NS 12066 (to J.H.S.) and by NS 14555 and Career Development Award NS 00350 (to

47 R . T . A . ) . D u r i n g p a r t o f this work, R . T . A . was a Senior I n v e s t i g a t o r o f the N e w Y o r k H e a r t Association.

REFERENCES 1 Ambron, R. T., Goldman, J. E. and Schwartz, J. H., Axonal transport of newly synthesized glycoproteins in a single identified neuron of Aplysia californica, J. Cell BioL, 61 (1974) 665-675. 2 Ambron, R. T., Goldman, J. E. and Schwartz, J. H., Effect of inhibiting protein synthesis on axonal transport of membrane glycopr oteins in an identified neuron of Aplysia, Brain Research, 94 (1975) 307-323. 3 Ambron, R. T., Goldman, J. E., Shkolnik, L. J. and Schwartz, J. H., Synthesis and axonal transport of membrane glycoproteins in an identified serotonergic neuron of Aplysia, J. Neurophysiol., 43 (1980) 929-944. 4 Ambron, R. T., Sherbany, A. A., Shkolnik, L. J. and J. H. Schwartz, Distribution of membrane glycoproteins among the organelles of a single identified neuron of Aplysia. I. Association of a [aH]glycoprotein with vesicles, Brain Research, 207 (1981) 17-32. 5 Ambron, R. T. and Treistman, S. N., Glycoproteins are modified in the axon of R2, the giant neuron of Aplysia californica, after intra-axonal injection of [3H]N-acetylgalactosamine, Brain Research, 121 (1977) 287-309. 6 Blobel, G. and Dobberstein, B., Transfer of protein across membranes. II. Reconstitution of functional rough microsomes from heterologous components, J. Cell Biol., 67 (1975) 852-862. 7 Cook, G. M., Biosynthesis of plasma membrane glycoproteins. In G. Poste and G. L. Nicholson (Eds.), The Synthesis, Assembly, and Turnover of Cell Surface Components, Elsevier/North Holland, 1977, pp. 85-136. 8 Eckenhoff, M. F. and Pysh, J. J., Double walled coated vesicle formation: evidence for massive and transient internalization of plasma membranes during cerebellar development, J. Neurocyt., 8 (1979) 623-638. 9 Eisenstadt, M., Goldman, J. E., Kandel, E. R., Koike, H., Koester, J. and Schwartz, J. H., Intrasomatic injection of radioactive precursors for studying transmitter synthesis in identified neurons of Aplysia californica, Proc. nat. Acad. Sci. (Wash.), 70 (1973) 3371-3375. 10 Elam, J. S. and Peterson, N. W., Axonal transport of glycoproteins in the garfish olfactory nerve: isolation of high molecular weight glycopeptides labeled with [SH]fucose and [aH]glucosamine, J. Neurochem., 33 (1979) 571-573. 11 Fischer, F. G. und Nebel, H. J., Nachweis und Bestimmung von glucosamin und galaktosamine auf papierchromatogrammen, Hoppe-Seylers Z. physiol. Chem., 302 (1955) 10-19. 12 Geissler, D., Martinek, A., Margolis, R. U., Margolis, R. K., Skrivanek, J. A., Leeden, R., Konig, R. and Winkler, H., Composition and biogenesis of complex carbohydrates ofox adrenal chromaffin granules, Neuroscience, 2 (1977) 685-693. 13 Goldberg, D. J., Sherbany, A. A. and Schwartz, J. H., Kinetic properties of normal and perturbed axonal transport of serotonin in a single identified neuron, J. Physiol. (Lond.), 281 (1978) 559-579. 14 Goldman, J. E., Kim, K. S. and Schwartz, J. H., Axonal transport of [3H]serotonin in an identified neuron of Aplysia californica, J. Cell Biol., 70 (1976) 304-318. 15 Hopkins, C. R., Studies on secretory activity in the Pars intermedia of Xenopus laevis, Tissue and Cell, (1970), 59-70. 16 Johnson, P. V. and Roots, B. L, Nerve Membranes, Pergamon Press, New York, 1972, p. 122. 17 Katz, F. N., Rothman, J. E., Lingappa, V. R., Blobel, G. and Lodish, H., Membrane assembly in vitro: synthesis, glycosylation, and asymmetric insertion of a transmembrane protein, Proc. nat. Acad. Sci. (Wash.), 74 (1977) pp. 3278-3282. 18 Leblond, C. P. and Bennett, G., The role of the Golgi apparatus in terminal glycosylation. In B. R. Brinkly and K. R. Porter (Eds.), International CellBiology 1976-1977, Rockefeller Univ. Press, New York, 1977, pp. 326-336. 19 Loh, ¥. P. and Gainer, H., Low molecular weight specific protein in identified molluscan neurons. II. Processing, turnover and transport, Brain Research, 92 (1975) 193-206. 20 Marchesi, V. T. and Andrews, G. P., Glycoproteins : isolation from cell membranes with lithium diiodosalicylate, Science, 174 (1971) 1247-1248. 21 Mayer, H. and Rapin, A., In K. Randerath, Thin Layer Chromatography~ 2nd edn., Academic Press, New York, 1966, p. 236,

48 22 Salpeter, M. M., Bachman, L. and Salpeter, E. G., Resolution in electron microscope radioautography, J. Cell BioL, 41 (1969) 1-20. 23 Schiller, S., Slover, G. A. and Dorfman, A., A method for the separation of mucopolysaccharides: its application to the isolation of heparin from the skin of rats, J. bioL Chem., 236 (1961) 983-987. 24 Schwartz, J. H., Castellucci, V. F. and Kandel, E. R., Functioningof identified neurons and synapses in the abdominal ganglion of Aplysia in the absence of protein synthesis, J. NeurophysioL, 34 (1971) 939-953. 25 Schwartz, J. H., Goldman, J. E., Ambron, R. T. and Goldberg, D. J., Axonal transport of vesicles carrying [3H]serotonin in the metacerebral neuron ofAplysia californica, ColdSpring Harb. Symp. Quant. BioL, 40 (1976) 83-92. 26 Sherbany, A. A., Ambron, R. T. and Schwartz, J. H., Membrane glycolipids: regional synthesis and axonal transport in a single identified neuron ofAplysia californica, Science, 203 (1979) 78-81. 27 Shkolnik, L. J. and Schwartz, J. H., Genesis and maturation of serotonergic vesiclesin the identified giant cerebral neuron of Aplysia, J. Neurophysiol., 43 (1980) 945-967. 28 Strous, G. J. A. M., Initial glycosylation of proteins with acetylgalactosaminylserine linkages, Proc. nat. Acad. Sci. (Wash.), 76 (1979) 2694-2698. 29 Thompson, E. B., Schwartz, J. H. and Kandel, E. R., A radioautographic analysis in the light and electron microscope of identified Aplysia neurons and their processes after intrasomatic injectionof L-[ZH]fucose, Brain Research, 112 (1976) 251-281. 30 Williams, M. A., The assessment of electron microscopic autoradiographs, Advanc. opt. electr. Micros., 3 (1969) 219-272. 31 Yamagata, T., Saito, H., Habuchi, O. and Suzuki, S., Purification and properties of bacterial chondroitinases and chondrosulfatases, J. biol. Chem., 243 (1968) 1523-1535.