THE JOURNAL OF BIOLOQICAL CHEMISTRY Vol. 249, No. 18, Issue of September 25, pp. 5995-6003, 1974

Printed in U.S.A.

Golgi Apparatus of Rat Kidney

PREPARATION AND ROLE IN SULFATIDE FORMATION*

(Received for publication, January 28, 1974)

BECCA FLEISCHER AND FERNANDO ZAMBRANO~

From the Department of Molecular Biology, Vanderbilt University, Nashville, Tennessee 37235

SUMMARY

A procedure has been developed for the isolation of mor- phologically recognizable, largely intact, Golgi apparatus from rat kidneys. The fraction contains both uridine diphospho- galactose: IV- acetylglucosamine galactosyltransferase and 3’-phosphoadenosine 5’-phosphosulfate: cerebroside sulfo- transferase activities enriched 50- to 80-fold over the original homogenate. Comparison of the Golgi-rich fraction with plasma membranes, endoplasmic reticulum, mitochondria, and nuclei isolated from rat kidney indicates that the Golgi apparatus is the main locus of both of these enzymes in kidney cells. Golgi apparatus from rat liver, although rich in galactosyltransferase, appears to be devoid of the sulfo- transferase.

The cerebroside sulfotransferase of kidney Golgi forms sulfatide from added galactocerebrosides. Both hydroxy- fatty acid and normal fatty acid-containing cerebrosides are sulfated. The enzyme is activated almost maximally by the addition of as little as 0.1% Triton X-100. The presence of 60 mu Mn2+ stimulates the activity about 3-fold. The apparent K, for cerebrosides was 1.3 x 10e4 M, while that for 3’-phosphoadenosine 5’-phosphosulfate was 2.1 X low3 M.

These results show that the Golgi apparatus in kidney cells can function to modify glycolipids as well as glycoproteins.

One of the major advances in cell fractionation in recent years has been the development of isolation procedures for the Golgi apparatus of the mammalian hepatocyte (l-5). This has been followed by extensive enzymatic and chemical characterization (1, 3, 6, 7). The Golgi apparatus has been shown to be bio- chemically unique as compared to the other membranes of the liver cell in that it is the sole locus of UDP-galactose:N-acetyl- glucosamine galactosyltransferase activity (1, 3, 7, 8) and has a unique protein profile on polyacrylamide gel electrophoresis as compared to other cell organelles of liver (3, 9). It is involved

* This work was supported in part by United States Public Health Service, National Institutes of Health Grants AM14632 and AM 17223, and an United States Public Health Service Inter- national Fellowship to F.Z.

$ Permanent address, Universidad De Chile Facultad de Cien- cias, Casilla 6635, Correo 4, Santiago, Chile.

in the secretion of serum albumin (10, 11) and of serum very low density lipoproteins (12-14). It is presumed to modify glycoproteins that it secretes by catalyzing the addition of some of the terminal sugars of the carbohydrate chains (15).

The present study was undertaken with a number of objectives in mind. First, we wanted to see if methods for the isolation of the Golgi apparatus from liver which we had devised (1, 3, 16) could be applied to other tissues such as kidney where little or nothing is known about the role of the Golgi apparatus. Second, it is important to know how comparable the properties of Golgi apparatus are in various t,issues of the same species. Third, we wished to investigate the possible role of the Golgi apparatus in glycosphingolipid formation, and kidney was particularly suited for such a study.

Glycosphinogolipids are minor components of the lipids of most mammalian tissues other than brain. They are thought to be constituents primarily of plasma membranes and to be in- volved in the immunological specificity exhibited by various tissues (for a review of the biological significance of glycosphingo- lipids see Ref. 17). Most studies on the biosynthesis of sphingo- lipids have been carried out in brain, where the appearance of cerebrosides and sulfatides closely parallels the process of myeli- nation. Studies on the biosynthesis of galactosyl and glucosyl ceramides generally have used brain microsomes as a source of enzyme (for review see Ref. 18). Sulfatide can be formed in vitro by the transfer of sulfate from 3’-phosphoadenosine 5’- phosphosulfate to galactosyl ceramide. The enzyme involved is PAPSl:cerebroside sulfotransferase and has been described in rat brain (19) and rat kidney (20) microsomes. Sulfatides are not present to a significant extent in liver (21). In kidney, sulfatides have been shown to be enriched in the outer medulla where sodium transport is also enriched (22).

Little is known of the exact subcellular localization of any of the enzymes involved in cerebroside and sulfatide formation. Microsomes are a heterogeneous collection of vesicles which con- sist of various proportions of endoplasmic reticulum, Golgi apparatus, and plasma membrane depending on the tissue under study. Cerebroside sulfotransferase is highest in a light, smooth microsome fraction from brain and is not present to a significant extent in myelin (19). These results led us to investigate the localization of this enzyme in the Golgi apparatus. A prelim- inary report of this work has been published (23).

i The abbreviations used are: PAPS, 3’-phosphoadenosine 5’- phosphosulfate; Hepes, N-2-hydroxyethylpiperazine-N’-2-ethane- sulfonic acid; VLDL, serum very low density lipoproteins.

5995

by guest on January 5, 2021http://w

ww

.jbc.org/D

ownloaded from

5996

EXPERIMENTAL PROCEDURES

Materials-Male Holtzman rats, 200 to 250 g, fed ad l&turn were used. They were killed by decapitation, exsanguinated, and the kidneys quickly removed, decapsulated, and collected in ice-cold 0.25 aa sucrose: [36S]PAl% (approximaiely 1 Ci per mmole) was purchased from New England Nuclear Corp. and diluted with carrier to a specific activity of approximately 1 mCi per mmole Carrier PAPS was prepared from ATP and NazSOd by the method of Hodson and Schiff 124) usine: extracts of Chlorella vurenoidosa as a source of sulfate-a&ivating enzyme.

_”

UDP-Gal was obtained from Calbiochem, Los Angeles, Calif. UDP-Gal uniformly la- beled with’% in the sugar moiety was obtained from New England Nuclear Corp., Boston, Mass. It was diluted with carrier to a specific a&iv& of abodt 1 mCi per mmole.

For column chromatoeranhv of linid extracts. the silicic acid used was Unisil, 100 to 2Grn&&, from*Clarkson Chemical Co., Inc., Williamsport, Pa. For thin layer chromatography of lipid com- ponents, plain silica gel (Research Specialties Co., Richmond, Calif.) containing 10% magnesium silicate was used (25). “Sulf- spray” was obtained from Supelco, Inc., Bellefonte, Pa.

Preparation of Cell Fractions-All sucrose solutions were pre- pared using “Ultra-Pure” grade (Mann Research Laboratories). Sucrose gradients were adjusted to the desired concentration of sucrose (per cent sucrose -= w/w solution) using a Bausch and Lomb Abbe 3L refractometer at 25”. Centrifugations were carried out in a Spinco ultracentrifuge unless indicated otherwise and all operations carried out at @4”.

A Golgi apparatus fraction from rat kidneys was prepared by a modification of the method previously described for rat liver (16). Kidneys from 50 rats are collected, blotted, weighed, and minced with scissors. About 100 g of minced tissue are suspended in 2 volumes of 52yo sucrose containing 0.1 M sodium phosphate, pH 7.1, and homogenized with three full strokes at 1,000 rpm using a 50-ml Potter-Elvehjem type homogenizer (inside diameter of the glass vessel = 1.000 inch) and a Teflon pestle machined to a diame- ter of 0.974 inch. The homogenization is repeated using a pestle with diameter of 0.982 inch. The homogenate is filtered through four single layers of cheesecloth and adjusted to 43.770 sucrose with homogenizing medium. About 250 ml of homogenate is pumped into the periphery of a Ti-14 zonal rotor previously filled with a step gradient consisting of the following unbuffered sucrose solutions: 115 ml of 29%, 120 ml of 33%, 130 ml of 36yo,, and 150 ml of 38.70/,. The rotor is then filled with about 50 ml of homogeniz- ing medium and the contents centrifuged at 45,000 rpm for 30 min (including the acceleration time). The rotor is then decelerated to 4,000 rpm, the contents displaced by pumping 55yo sucrose into the periphery, and 20-ml fractions collected from the center. Fractions are combined as described in individual experiments and diluted with ~~ volume cold distilled water. The upper frac- tions are centrifuged at 30,000 rpm for 1 hour to recover the Golgi. Plasma membrane- and mitochondria-rich fractions are prepared from the zonal gradient at the same time as Golgi-rich fractions (see Fig. 1). Fractions richin plasma membranes are diluted with an equal volume of cold distilled water and sedimented at 10,000 rpm for 10 min in the JA-20 rotor of the Beckman J-21 centrifuge. The upper white portion of the pellet is resuspended in 0.25 M sucrose containing 0.01 M Hepes, pH 7.4, and 0.001 M EDTA using a Dounce homogenizer with an A pestle. The membranes are re- centrifuged as before and the entire washing process is repeated four times. Mitochondria-rich fractions from the zonal gradient are combined and the mitochondria are recovered by centrifuging in the JA-20 rotor at 16,000 rpm for 10 min. The upper light por- tion of the pellet is discarded and the lower brown portion is re- suspended in sucrose-Hepes-EDTA and recentrifuged as before. The washing procedure is repeated twice more. All pellets were suspended finally in 0.25 M sucrose. For purposes of comparison, nuclei, plasma membranes, mitochondria, and microsomes were prepared from rat kidneys by the method of Stein et al. (26) as modified by Fleischer and Kervina (27). The microsomes were further fractionated into a smooth and rough fraction by a modi- fication (27) of the method of Dallner (28).

Measurement of Enzyme Activities-PAPS:cerebroside sulfo- transferase was measured by our modification of the assay condi- tions used by Farrell and McKhann (19). The assay mixture (0.1 ml) contained the following (in micromoles) in the order in

which they are added: imidazole chloride, pH 7.0, 12; mercapto- ethanol, 6; manganese chloride, 6; Triton X-100, 0.5 mg; protein, 50 to 200 pgg; ATP, 3; mixed bovine cerebrosides, 20 pg; and [%I- PAPS, 0.27. Mixed bovine cerebrosides (Supelco Inc., Bellefonte, Pa.) were first suspended at 10 mg per ml in water containing 1% (w/v) Triton X-100 using a small Potter-Elvejhem homogenizer. Incubations were carried out at 37” for 1 hour. The reaction was stopped by the addition of 3.0 ml of chloroform-methanol 2:l (v/v). The contents were mixed by vortexing and the tubes were &lowed to stand 15 min at room temperature.- Non-lipid radioac- tivitv was removed bv washing once with 0.6 ml of 0.034Vo MgClt followed by three wa.&es cons&ina of 1 ml each of “uppeiphase” (29). The final lower phase was removed, taking c&e to leave behind anv material at the interface. and dried in uucuo. Ten milliliters of scintillation fluid consist&g of a mixture of 6.7 ml of toluene containing 3.76 mg of 2,5diphen$loxazole (PPO), 0.376 mg of 1.4-bis12-(5-nhenvloxazolvl)lbenzene (POPOP). and 3.3 ml of Tritbn X-iO0; was aided to I&& vial. The sampi& were counted finally in a Packard Tri-Garb liquid scintillation spectrometer.

Galactosyltransferase (16) and succinate-cytochrome c reduc- tase (30) were determined as described previously. Glucose 6- phosiha&e was determined according to-the metgod of Swanson (31) except that incubat,ion.. were made for 5 and 10 min and the phosphate released measured by the method of Chen et al. (32). %otai ATPase was measured & described previously (30) except that the assav mixture also contained 100 moles of NaCl and 10 eoles of KCi. Ouabain-sensitive ATPasewas taken as the differ- ence between total ATPase and the same activity measured in the presence of 1 mM ouabain and 2 rnM KCl.

Other Analyses-Protein was estimated by the Lowry procedure (33) using crystalline bovine serum albumin as a standard. Phos- phorus was determined by the method of Chen et al. (32).

The [3B]labeled sulfolipid formed as a product of the sulfotrans- ferase was characterized using column and thin layer silicic acid chromatography. The lower phase containing the lipid products was evaporated to dryness. taken up in 0.3 ml of chloroform, and fractionated using a &cic &id column (1.2 X 3 cm). The column was eluted with 30 ml of chloroform (neutral lipid fraction), fol- lowed by 60 ml of acetone (glycolipid fraction), and finally with 30 ml of methanol (phospholipid fraction). Aliquots of each frac- tion were evaporated to dryness and counted as described for the sulfotransferase assay. The dvcolinid fraction w&s further frac- tionated by two-dimensional &in layer chromatography according to the method of Rouser et al. 05). Before chromatoeranhy, 200 ~g of a cerebroside plus sulfatide fraction from bo&e brain (a gift of Dr. George Rouser) was added as carrier. The plates were run first in chloroform-methanol-28% ammonia. 65:25:5 (v/v), dried 10 min under nitrogen, and run in the second direction using chloroform-acetone-methanol-glacial acetic-water, 3:4:1:1:0.5 (v/v). After chromatography, the plates were air dried for 10 min and sprayed with “Sulfspray.” After about 5 min they were sprayed with water. The spots were scraped into vials, scintilla- tion fluid was added, and the samples counted as described above.

Chromatography of the reaction products of the galactosyl- transferase reaction was carried out on DEAE-paper, Whatman No. DE81, using 1-butanol, 1-propanol, water (3:l:l v/v/v) aa described previously (1). The sugars run on the chromatograms as standards were visualized by spraying with analine phthalate (260 mg/lO ml of water-saturated butanol) and heating at 110” for 10 to 15 min.

Electron iVicro.scopvFor best morphological preservation, ali- quota of the fractions containing Goigi apiarat& were fixed im- mediatelv after isolation from the zonal madient. before dilution and centiifugation to concentrate the Gilgi. Ode-half milliliter of 25yo glutaraldehyde made up in 0.2 I sodium cacodylate, pH 7.4, was added to 5 ml of the combined fractions (about 0.5 to 1.0 mg of protein). The other cell fractions were fixed by treating an ali- quot (containing about 1 mg of protein) with an equal volume of 5yo glutaraldehyde in 0.25 M sucrose and 0.2 M sodium cacodylate, pH 7.4. After standing overnight in the refrigerator, the samples were centrifuged at 10,000 rpm and the supernatant discarded. The pellets were washed by suspension in 1 ml of 0.25 Y sucrose and recentrifugation. The washing procedure was repeated twice. The final pellets were then fixed with 1% osmium tetroxide, de- hydrated, embedded, and sectioned as described previously (1).

by guest on January 5, 2021http://w

ww

.jbc.org/D

ownloaded from

RESULTS

Fractionation of rat kidney homogenates by procedures sim- ilar to those which have proved effective in isolating liver Golgi apparatus resulted in the distribution of protein and enzymic activities shown in Fig. 1 and Table I. The upper portion of the gradient is the Golgi-rich fraction. It is enriched in both cerebroside sulfotransferase and galactosyltransferase and poor in glucose 6-phosphatase, ouabain-sensitive ATPase, and suc- cinate-cytochrome c reductase. Farther down the gradient, this profile of enzymic activities is reversed. The ratio of galacto-

18-

I I I I

5 IO I5 20

FRACTIONS (20ml) FIG. 1. Fractionation of rat kidney homogenate by zonal ultra-

centrifugation. Twenty-milliliter fractions were collected. Fractions were combined as indicated by the bars. Analysis of the fractions is given in Table I. O-O, milligrams of protein per tube; O--O, per cent sucrose concentration of each frac- tion collected. PM, plasma membrane; MITO, mitochondria.

5997

syltransferase to the sulfotransferase in the upper parts of the gradient appears to be fairly constant, giving further support to the conclusion that both activities are present in the same sub- cellular organelle.

The presence of cerebroside sulfotransferase is not linked to the presence of galactosyltransferase in all tissues. In brain homogenate, a high sulfotransferase is present together with a low level of galactosyltransferase, and in liver a high galactosyl- transferase activity is present whereas no sulfotransferase is de- tectable (Table I). The Golgi apparatus isolated from rat liver also does not exhibit sulfotransferase activity (23).

Since the amount of protein in Fraction 1 was rather low, Frac- tions 1 and 2 were combined with some sacrifice of purity. The combined fractions, generally tubes 2 to 10 from the gradient, as well as the plasma membrane-rich and mitochondrial-rich fractions, were studied by electron microscopy. A typical field of the combined Fractions 1 + 2 which are rich in both galacto- syltransferase and in sulfotransferase is shown in Fig. 2A. The morphology of this fraction clearly identifies it as being derived predominantly from the Golgi apparatus. The plasma mem- brane-rich fraction consists of large smooth vesicles (Fig. 2C) and the mitochondrial fraction is morphologically identifiable although many of the mitochondria are swollen and fragmented (Fig. 20).

The Golgi appratus fraction illustrated in Fig. 2A appears to be about 70% pure as judged morphologically. From the glu- cose 6-phosphatase activity (Table II) it can be estimated to contain about 25% endoplasmic reticulum. From the ouabain- sensitive ATPase activity (Table I), it appears to be about 5% contaminated with plasma membranes. It also contains about 7% mitochondrial contamination as measured by succinate- cytochrome c reductase activity (Table II).

About 15% of the total sulfotransferase and about 10% of the galactosyltransferase of the homogenate is found in the com- bined Fractions 1 + 2. If these enzymes are only present in the Golgi apparatus it would mean the yield of Golgi by this procedure is 10 to 20% and that Golgi constitutes about 1% of the protein of the kidney homogenate.

TABLE I Enzymatic activities of rat kidney fractions from zonal rotor

EYactions

1

2

3

4

PM-rich

Miti-rich

Kidney Hang.

Brain Hauxj.

Liver Ikxcg.

mtalPnkein ong)

6.0

5.6

11.1

12.3

20.4

73.0

6,150

c I

29.4

28.7

27.9

25.9

20.7

11.7

-I-

Cerebroside SLilfo-

Transf erase*

Galactosyl ransferaze*

28.0 579

22.0 467

12.1 347

7.9 158

0.5 14

0.0 0.5

0.26 9.7

0.41 1.9

0.00 8.4

_ . * Nanomoles per hour per mg of protein at 37"; all other activities expressed as micromoles per min per mg of protein at 32” except

glucose 6-phosphatase which was carried out at 37”. Fractions obtained as described under “Experimental Procedures” and Fig. 1. Rat brain and liver homogenates, prepared in the same manner as the kidney homogenate, are included for purposes of comparison.

Glu-6 P'ase

.055

.077

.102

.151

.079

.008

.065

-039

-109

-126

.209

-311

-816

.256

I Total

WP’ase

0.35

0.61

0.82

1.16

1.42

0.25

0.20

, -- Sensitive ATP'ase

0.00

0.05

0.13

0.26

0.52

0.02

by guest on January 5, 2021http://w

ww

.jbc.org/D

ownloaded from

by guest on January 5, 2021http://w

ww

.jbc.org/D

ownloaded from

5999

TABLE II

Distribution of “marker enzymes ” in puri$ed subcellular fractions of Tat kidney

Average of four preparations from pooled rat kidneys f standard deviations of the mean. Supernatant and nuclei are the mean of two experiments. Enzyme activities expressed as in Table I. Supernatant is the fraction of the homogenate which is not sedimentable at 104,060 X g for 1 hour. The Golgi apparatus fractions used were equivalent to Fractions 1 + 2 shown in Fig. 1 and Table I.

Fraction

Homogenate. Nuclei . . . Rough microsomes. Smooth microsomes.. Golgi apparatus. Mitochondria. Plasma membranes. Supernatant

mount phosphorur per mg protein

fig 31.0 f 1.5 41.2 =I= 4.0 36.2 f 0.9 29.4 f 0.4 32.5 f 3.1 13.8 f 0.3 19.1 f 2.9 15.5 f 0.5

Cerebroside sulfotransferase

0.31 f 0.16 0.19 f 0.19 0.39 f 0.18 2.6 f 0.6

25.6 f 3.4 0.16 f 0.24 0.32 f 0.12

0.00

Galactosyl transferase

10.9 & 5.1 3.4 f 0.6 8.2 f 4.1

54.5 f 5.0 553 rt 17

1.7 f 0.9 6.1 f 5.4

0.0

TABLE III

Identification of products formed by cerebroside-sulfotransferase of rat kidney Golgi

A Go&$-rich fraction (206 pg of protein), equivalent to those described in Table II, was incubated as described under “Experi- mental Procedures” for the measurement of PAPS:cerebroside sulfotransferase activity. After washing with “upper phase” (29) to remove water-soluble radioactive substrates and products, the lower phase was first chromatographed on silicic acid columns. The “glycolipid” fraction was further fractionated on silicic acid thin layer chromatography (TLC) as described under “Experi- mental Procedures.” Before thin layer chromatography, 200 pg of a cerebroside plus sulfatide fraction from bovine brain was added to the “glycolipid” fraction to enable the visualization of the cerebroside and sulfatide areas of the chromatogram.

A. Silicic acid column chromatography

Fraction Counts per min Per cent of total

Chloroform (NL). . 0 0 Acetone (GL). . . . , _ 8910 98 Methanol (PL). 210 2

Recovery........................... 97%

B. Two-dimensional TLC of glycolipid fraction

FM2ti0n

Origin.............................. Hydroxy FAa cerebrosides. . Normal FA cerebrosides. . . Hydroxy FA sulfatides.. . . Normal FA sulfatides.. .

Recovery...........................

D FA, fatty acid.

Counts per min Per cent of total

11 1 59 2

0 0 1058 41 1443 56

97%

In order to verify that cerebroside sulfotransferase activity was localized predominantly in the Golgi apparatus from rat kidney, we compared the activities of purified subcellular or- ganelles prepared by more conventional means with our Golgi apparatus-rich fraction. The results are summarized in Table II. In all the fractions studied, only smooth microsomes had

GhWOSe 6-phosphatase

0.055 f. 0.017 0.103 f 0.023 0.191 f 0.018 0.309 f 0.064 0.077 f 0.017 0.013 f 0.009 0.031 * 0.010

Total ATPase

0.172 f 0.054 0.052 f 0.025 0.255 i 0.013 0.968 f 0.035 0.391 f 0.139 0.393 f 0.045 1.084 f 0.250

-

S

_-

-

uccinate-cytochrome c reductase

0.272 i 0.079 0.043 f 0.018 0.013 f 0.066 0.008 f 0.004 0.052 f 0.017 0.746 f 0.142 0.129 f 0.075

significant levels of both sulfotransferase and galactosyltransfer- ase activities. Both activities appeared to be about 10% of the level found in the Golgi apparatus fraction. This is most likely due to contamination of this fraction with fragments of the Golgi apparatus. Rough microsomes, which are more clearly repre- sentative of endoplasmic reticulum, contain only about 1% of the level of activity found in the Golgi apparatus. The plasma membrane fraction prepared by the more conventional methods (26, 27), like the plasma membrane-rich fraction prepared from our zonal gradients, is poor in both sulfotransferase and galac- tosyltransferase activities. The only major cell fraction not studied with regard to the localization of galactosyltransferase and sulfotransferase was the lysosomal fraction.

Table III summarizes the results obtained when the product of the sulfotransferase reaction was chromatographed on silicic acid columns and thin layer plates. Chromatography of the lipid extract on silicic acid columns yielded 98% of the initial counts in the glycolipid fraction. Two-dimensional thii layer chromatography of the glycolipid fraction to which carrier brain cerebrosides and sulfatides were added revealed that 97% of the counts chromatographed with the added sulfatides, and that both normal fatty acid sulfatides and hydroxyfatty acid sulfa- tides are formed in about equal amounts.

The effect of pH and the type of buffer used on the activity of the sulfotransferase in the purified Golgi fraction is shown in Fig. 3. Maximum activity is observed at pH 7 to 7.25 using imidazole. There are sharp falloffs in activity above and below this pH range. The use of Hepes instead of imidazole results in a slightly lower activity.

Addition of divalent cations to the enzyme results in stimula- tion of activity, as shown in Fig. 4. Manganese, magnesium, or calcium stimulate the sulfotransferase, but manganese gives a maximum rate at a lower concentration than magnesium or cal- cium. The level of manganese required for the maximum rate is 60 mu, clearly not a physiological level.

Triton X-100 must be added to the assay mixture in order to measure the sulfotransferase activity (Fig. 5). Near-maximal stimulation is obtained with the addition of as little as 0.1% Triton to the assay mixture.

Varying cerebroside concentration at a fixed [W]PAPS con- centration (2.7 mM) and a fixed amount of protein (66 pg) re- sulted in a typical saturation curve. A Lineweaver-Burk plot of the data yielded an apparent Em of 1.3 x 10m4 M for cerebro- side, assuming a molecular weight of 800 for mixed brain cerebro-

by guest on January 5, 2021http://w

ww

.jbc.org/D

ownloaded from

PH mM DIVALENT CATION TRITON X-lOO(mg/ml)

I I I I

20 40 60 80 100

FIG. 3 (left). Effect of varying pH of the buffer used on the PAPS:cerebroside sulfotransferase activity of isolated kidney Golgi apparatus. ~---a, imidazole-HCl; A-A, K-Hepes; O-0, Tris-HCl. The assay conditions are described under “Experimental Procedures.” In all cases, the final buffer concen- trations were 0.12 M.

FIG. 4 (center). Effect of varying concentrations of divalent cations on PAPS:cerebroside sulfotransferase activity of isolated kidney Golgi apparatus. The cations were added in the form of

TABLE IV

Analysis of reaction products formed after incubation of rat kidney fractions with UDP-[W]galactose

In the Dowex 2 method, galactose released is obtained by meas- uring the amount of radioactivity which passes through Dowex 2 when the sample is incubated with UDP-[i4C]Gal in the absence of added N-acetylglucosamine. Galactose transferred is taken to be the increase in the radioactivity passing through Dowex 2 when N-acetylglucosamine (3 Nmoles/75 ~1 of assay mixture) is present in the assay mixture (1). It is assumed that galactose released is the same f acceptor. In the chromatography method, the products of the reaction are separated by chromatography on Whatman DE-81 paper (see “Experimental Procedures” section for details). In this assay radioactivity migrating as a monosac- charide is considered “released” while radioactivity migrating as a disaccharide is “transferred.”

Fraction Assay method

Golgi Dowex 2

Plasma membrane- rich

Rough endoplasmic reticulum

Chromatography

Dowex 2

Chromatography

Dowex 2

Chromatography

N-Acetyl- lucosamin

- + - +

- + - +

- + - +

-

158 166

40

48 35

247

309 306

532 44

665

53 3

49

8 105

83

I-

>-

I-

, I 2 3 4 5

their chloride salts: A-A, Mn*; O--O, Mgz+; Ca*+. Other constituents added as described under “Experimental Procedures.” The amount of protein used was 61 pg.

FIG. 5 (right). Effect of varying Triton X-199 concentration on PAPS:cerebroside sulfotransferase activity of isolated kidney Golgi apparatus. Other constituents added as described under “Experimental Procedures.” The amount of protein used was 66 rg.

sides. The apparent K, for PAPS using 25 mM cerebrosides and 66 pg of protein was 2.1 X 1(Ya M.

The products of the galactosyltransferase reaction were in- vestigated further by chromatography. After incubation of the kidney fractions with UDP-[r%]Gal in the presence and absence of added N-acetylglucosamine as acceptor for the transferase enzyme, the reaction mixture was chromatographed to separate monosaccharide and disaccharide reaction products. The re- sults are summarized in Table IV (see also Fig. 6). In the ab- sence of added acceptor, some hydrolysis of UDP-Gal occurs. in the Golgi fraction but a much higher level of hydrolysis occurs in the endoplasmic reticulum fraction, as well as a second pe’ak migrating in the disaccharide region. The level of this second peak formed by the endoplasmic reticulum fraction is not stim- ulated by the addition of N-acetylglucosamine as it is in the Golgi fraction. Good agreement is obtained between the chromato- graphic estimation of the amount of disaccharide formed by the Golgi or by the plasma membrane fraction and the values esti- mated by difference using Dowex 2 columns. In the case of endoplasmic reticulum, however, the high rate of hydrolysis makes it difficult to measure small rates of transfer by differ- ence, as is required in the Dowex 2 method. The presence of a second peak migrating in the disaccharide region both in the absence or presence of acceptor also makes the estimation of transferase activities difficult using the chromatographic method. It was recently reported by Bergeron et al. (7) that homogenates and endoplasmic reticulum fractions of liver, when assayed for galactosyltransferase activity using chromatography on DEAE- paper, exhibited an unknown peak intermediate in mobility between free galactose and N-acetyllactosamine. This peak disappeared when the incubated reaction mixture was extracted with 2 volumes of chloroform-methanol, 2: 1 (v/v), before chro- matography of the aqueous phase. We investigated the effect

by guest on January 5, 2021http://w

ww

.jbc.org/D

ownloaded from

6001

N-ACETYL GALACTOSE I I I I

LACTOSAMINE

A.

a.00 16.00 211.00 32.00 CENTIMETERS

FIG. 6. Chromatography of the reaction mixture on DEAE- paper after incubation of: A, kidney Golgi fraction, or B, kidney rough endoplasmic reticulum with UDP-[%]Gal in the galactosyl- transferase assay. 13-m) without added N-acetylglucosa- mine; O-O, with added N-acetylglucosamine; A-A, with added N-acetylglucosamine but extracted with 2 volumes of a of removal of such organic solvent-soluble components on the chromatography of the reaction products of the galactosyltrans- ferase assay on both the Golgi apparatus fraction and rough endoplasmic reticulum. The results are illustrated in Fig. 6. In the presence of N-acetylglucosamine as acceptor, the main reaction product of the galactosyltransferase reaction in the Golgi apparatus is a peak which migrates with the same relative mobility as N-acetyllactosamine. Extraction of the reaction mixture with 2 volumes of chloroform-methanol, 2:1, does not affect the level of this reaction product. Rough endoplasmic reticulum, on the other hand, when incubated with UDP-[‘%I- Gal, forms a reaction product which migrates in the dissaccharide region. The amount of this product is not enhanced by the presence of N-acetylglucosamine in the reaction mixture. Fur- thermore, extraction of the mixture with chloroform-methanol, 2:1, eliminates this product from the aqueous phase. A small amount of radioactive material remains in the N-acetyllactos- amine region of the chromatogram. It amounts to only 0.6% of the activity per mg of protein found in the purified Golgi apparatus fraction, and is probably due to contamination of the fraction with Golgi.

DISCUSSION

The procedures we have developed for the isolation of purified Golgi apparatus from rat liver (3, 16) are applicable, with minor modifications, to the isolation of Golgi apparatus from rat kid- ney. Our method has the following useful features: (a) it can be used on animals in a normal metabolic state; (5) it yields practically intact Golgi apparatus rather than fragments from different parts of the apparatus; and (c) it can be applied to other tissues with minor modifications and thus allows the com- parison of Golgi apparatus from different cell types within the same organism.

The isolated Golgi apparatus of kidney cells, like that of liver,

N-ACETYL GALACTOSE t I t I

LACTOSAMINE

B.

.oo

mixture of chloroform and methanol, 2:l (v/v), to remove lipid- soluble material. Galactose (1.5 moles) and N-acetyllactosa- mine (3 pmoles) were chromatographed simultaneously as stand- ards. After chromatography they were located by spraying with analine phthrdate (see “Experimental Procedures”).

mizing tubules. The major morphological difference appears to be the absence in kidney Golgi of particle-filled large vesicles associated with the Golgi apparatus of liver. These vesicles are involved in the secretion of very low density serum lipopro- teins (12-14, 34, 3.9, a function probably not present in kidney Golgi. The original morphology of the Golgi apparatus is pre- served to a remarkable degree during our isolation procedure. This fact was not recognized in our early work (3) because the isolated Golgi was partially disrupted during the last step of the isolation procedure where dilution with distilled water was neces- sary to decrease the density of the suspending medium so that the membranes could be recovered. Dilution with the reagents of lower osmolarity used for fixation must also be avoided if Golgi morphology is to be preserved (16).

A number of procedures are available for the isolation of Golgi apparatus from liver (l-5). All of these depend to some extent upon the lower density of the Golgi apparatus membranes as compared to the other cell organelles. This difference is en- hanced in the method of Ehrenreich et al. (5) in which the Golgi apparatus of the liver is stimulated to secrete VLDL by ethanol intoxication of the rats. .The difference in density, however, also exists between the Golgi apparatus of kidney and the other cell organelles, although no evidence exists so far that this or- ganelle secrets lipoproteins.

Rat kidney Golgi is functionally similar to rat liver Golgi in that it is the main locus of UDP-Gal:N-acetylglucosamine galac- tosyltransferase activity. This enzyme together with other sugar transferases present in liver Golgi apparatus (15) appear to he involved in the modification of glycoproteins destined for secretion. In kidney, glycoproteins such as renin, which is thought to be formed in juxtaglomerular cells and stored there in the form of granules (3638), may be secreted through the Golgi apparatus of these cells. When released into the blood stream, renin acts on a serum protein substrate to form angio-

consists of a number of flattened cisternae with attached anasto- tensin I, a key step in the homeostatic control of blood pressure

by guest on January 5, 2021http://w

ww

.jbc.org/D

ownloaded from

6002

by the kidney (39). Erythropoietin is another glycoprotein of physiological significance (40) which may be secreted by kidney Golgi apparatus.

The finding that PAPS : cerebroside sulfotransferase is localized mainly in the Golgi apparatus of rat kidney, brings to light a new functional aspect of the Golgi apparatus in mammalian cells. It is the first demonstration that Golgi can modify not only mucopolysaccharides and glycoproteins, but glycolipids as well. It is also interesting to note that the sulfatides formed there are probably destined not for secretion, but to be used as constituents of the cell membranes, themselves.

Sulfatides occur in all the cell fractions of kidney but are most abundant in Golgi apparatus and plasma membranes.2 Little is known about how the sulfatides, which are not freely soluble in the cell cytoplasm, are transported from the Golgi apparatus to the plasma membranes of the cell. Some evidence of a mech- anism involving a carrier protein in the cytoplasm which picks up newly formed sulfatide from the Golgi and transfers it to the plasma membrane has been found in brain. In these experi- ments, the transfer in vitro of newly formed sulfatide from micro- somes to myelin prepared from rat brain was investigated. It was shown that a lipoprotein fraction isolated from the soluble fraction of brain accepts [35S]sulfatide from microsomes (41). Similar results have been reported for chick embryo sciatic nerves in tissue culture (42). These results, however, do not rule out the possibility that the lipoprotein recovered in the soluble fraction was originally present in the cisternae of the Golgi apparatus, and appeared in the cytoplasmic fraction due to disruption of the Golgi during the homogenization procedures employed.

The PAPS-cerebroside sulfotransferase of kidney Golgi appa- ratus appears to be similar to the enzyme described in some de- tail by Farrell and McKhann from brain microsomes (19). Both enzymes have a sharp pH optimum of 7.0 to 7.2 in imidazole buffers, and are stimulated by addition of ATP, Triton X-100, divalent cat.ions, and exogenous cerebroside. Both enzymes will transfer sulfate to added psychosine but not to added gluco- cerebrosides (19, 23). The enzyme from kidney Golgi is stimu- lated by manganese > magnesium > calcium. The response of the brain enzyme to added manganese was not studied by Farrell and McKhann, but the response to magnesium was also greater than the response to calcium for this enzyme. The concentration of divalent cations necessary for maximum activ- ity of the two enzymes appear to be quite different, however. Maximum activity of the Golgi preparation occurred at 60 my Mn*, or at 100 mM Mg*, whereas the brain preparation showed maximum activity at 20 mM Mg*.

Although the sulfotransferase in brain and kidney appears to have very similar properties in in vitro assays, important differ- ences are apparent in the two tissues, possibly in the control mechanisms involved in its expression. For example, the de- velopmental pattern of the sulfotransferase in rat is different in the brain and the kidney (20). The enzyme in brain rises sharply to a peak of activity at about 20 days postpartum, which corre- lates well with the time of maximum myelination in rat brain. After this point, the activity drops almost as rapidly to about one-third the maximum value which persists in the adult. The kidney enzyme, however increases steadily with kidney weight and remains at a fairly constant level in the adult. A second indication of different control mechanisms is the report that the sulfotransferase is lower than normal in the brains of mouse

neurological mutants “Jimpy” and “Quaking” whereas the ac- tivity in the kidney is unimpaired (43).

Acknowledgments--We gratefully acknowledge the advice and help of Dr. S. Fleischer and the technical assistance of Mrs. Ikuko Ishii in the course of this work. Electron micrographs were prepared by Mr. Akitsugu Saito. We would also like to thank Dr. Jerone Schiff for his gift of [a5S]PAPS and for a start- ing culture of Chlorella pyrenoidosa, and Dr. Mary C. Glick, University of Pennsylvania, for her gift of N-acetyllactosamine.

REFERENCES 1. FLEISCHER, B., FLEISCHER, S., AND OZAWA, H. (1969) J. Cell

Biol. 43, 59-79 2. MORRI$ D. J., HAMILTON, R. L., MOLLENHAUER, H. H.,

MAHLEY, R. W., CUNNINGHAM, W. P., CHEETHAM, R. D., AND LEQUIRE, V. S. (1970) J. Cell Biol. 44, 484-491

3. FLEISCHER, B., AND FLEISCHER, S. (1970) Biochim. Biophys. Acta 219, 301-319

4. LEELAVATHI, D. E., ESTER, L. W., FEINGOLD, D. S., AND LOMBARDI, B. (1970) Biochim. Biophys. Acta 211, 124138

5. EHRENREICH, J. H., BERGERON, J. J. M., SIEKEVITZ, P., AND PALADE, G. E. (1973) J. CeEZ Biol. 69,45-72

6. CHEETHAM, R. D., MORRE, D. J., AND YUNGHANS, W. N. (1970) J. Cell Biol. 44, 492-500

7. BERGERON, J. J. M., EHRENREICH, J. H., SIEKEVITZ, P., AND PALADE, G. E. (1973) J. Cell Biol. 69,73-f%

8. MORF&, D. J., MERLIN, L. M., AND KEENAN, T. W. (1969) Biochem. Biophys. Res. Commun. 37. 813-819

9. ZAHLER, W. L., F~EISCHER, B., AND FLEISCHER, S. (1970) Bio- chim. Biophvs. Acta 203. 283-290

10. PETERS, T.; JAR., FLEISC&R, B., AND FLEISCHER, S. (1971) J. Biol. Chem. 246.240-244

11. GLAUMANN, H.. AND ERICSSON. J. L. E. (1970) J. Cell Biol. 47, 555-567 ’ ’

.

12. MAHLEY, R. W., HAMILTON, R. L., AND LEQUIRE, V. S. (1969) J. Livid Res. 10. 433-439

13. MAHLE;, R. W., I&RSOT, T. P., LEQUIRE, V. S., LEVY, R. I., WINDMUELLER, H. G:, AND BROWN, W. V. (1970) Science 168, 380-382

14. CHAPMAN, M. J., MILLS, G. L., AND TAYLAUR, C. E. (1972) Biochem. J. 128, 779-787

15. SCHACHTER, H., JABBAL, I., HUDGIN, R. L., PINTERIC, L., MCGUIRE, E. J., AND ROSEMAN, S. (1970) J. Biol. Chem. 245, 1090-1100

16. FLEISCHER, B. Methods Enzymol. 31, 180-191 17. STOFFEL, W. (1971) Annu. Rev. Biochem. 40.57-82 18. MORELL; P., AND I&AUN, P. (1972) J. Lipid Res. 13,293-310 19. FARRELL. D. F.. AND MCKHANN. G. M. 119711 J. Biol. Chem.

246, 4694-47Oi \ I

20. MCKHANN, G. M., AND Ho, W. (1967) J. Neurochem. 14, 717- 724

21. ROUSER, G., NELSON, G. J., AND FLEISCHER, S. (1968) in Biological Membranes (CHAPMAN, D., ed) pp. 5-69, Academic Press, New York

22. KARLSSON, K.-A., SAMUELSSON, B. E., AND STEEN, G. 0. (1968) Acta Chem. Stand. a&2723-2724

23. FLEISCHER, B. AND ZAMBRANO, F. (1973) Biochem. Biophys. Res. Commun. 62, 951-958

24. HODSON, R. C., AND SCHIFF, J. (1969) Arch. Biochem. Biophys. 132, 151-156

25. ROUSER, G., KRITCHEVSKY, G., AND YAMAMOTO, A. (1967) in Lipid Chromatographic Analysis (MARINETTI, G. V., ed) Vol. 1, pp. 99-162, Marcel Dekker, Inc., New York

26. STEIN, Y., WIDNELL, C., AND STEIN, 0. (1968) J. Cell Biol. 39, 185-192

27. FLEISCHER, S., AND KERVINA, M. Methods Enzymol. 31, 6- 41

28. DALLNER, G. (1963) Acta Pathol. Microbial. Stand. Suppl. 166, l-94

29. FOLCH, J., LEES, M., AND SLOANE-STANLEY, G. H. (1957) J. Biol. Chem. 226, 497-509

30. FLEISCHER, S., AND FLEISCHER, B. (1967) Methods Enzymol. 10. 406-433

t B. Fleischer and F. Zambrano, manuscript in preparation. 31. SWA&ON, M. A. (1955) Methods Enzymol. 2.541-543

by guest on January 5, 2021http://w

ww

.jbc.org/D

ownloaded from

32. CHEN, P. S., TORIBARA, T. Y., AND WARNER, H. (1956) Anal. Chem. 28, 1756-1758

33. LOWRY, 0. H., ROSEBROUGE, N. J., FARR, A. L., AND RANDALL, R. J. (1951) J. Biol. Chem. lQ3,265-275

34. JONES, A. L., RUDERMAN, N. B., AND HERRERA, M. G. (1967) J. Lipid Res. 8, 429446

35. HAMILTON, R. L., REGEN, D. M., GREY, M. E., AND LEQUIRE, V. S. (1967) Lab. Invest. 18,30%319

36. EDELMAN, R., AND HARTCROFT, P. M. (1961) Circ. Res. 9, 1069-1077

37. TSUDA, N., NICKERSON, P. A., AND MOLTENI, A. (1971) Lab. Invest. I, 644-652

38. MORIMOTO, S., YAMAMOTO, K., AND UEDA, J. (1972) J. Appl. Physiol. 33, 306-311

39. DAVIS, J. 0. (1971) in Kidney Hormones (FISHER, J. W., ed) pp. 173-205, Academic Press, New York

40. JACOBSON, L. O., GURNEY, C. W., AND GOLDWASSER, E. (1969) Advan. Intern. Ned. 10,297-327

41. HERSCHKOWITZ, N., MCKHANN, G. M., SAXENA, S., SHOOTER, E. M., AND HERNDON, R. (1969) J. Neurochem. 16,1@49-1057

42. PLEASURE, D. E., AND PROCKOP, D. J. (1972) J. Neurochem. 19, 283-295

43. SARLIEVE, L. L., NESKOVIC, N. M., AND MANDEL, P. (1971) Fed. Eur. Biochem. Sot. Lett. 19, 91-95

by guest on January 5, 2021http://w

ww

.jbc.org/D

ownloaded from

Becca Fleischer and Fernando ZambranoFORMATION

Golgi Apparatus of Rat Kidney: PREPARATION AND ROLE IN SULFATIDE

1974, 249:5995-6003.J. Biol. Chem. 

  http://www.jbc.org/content/249/18/5995Access the most updated version of this article at

 Alerts:

  When a correction for this article is posted• 

When this article is cited• 

to choose from all of JBC's e-mail alertsClick here

  http://www.jbc.org/content/249/18/5995.full.html#ref-list-1

This article cites 0 references, 0 of which can be accessed free at

by guest on January 5, 2021http://w

ww

.jbc.org/D

ownloaded from