哺乳動物細胞におけるゴルジ体の形態維持機構に関する研究

(1)

九州大学学術情報リポジトリ

Kyushu University Institutional Repository

哺乳動物細胞におけるゴルジ体の形態維持機構に関する研究

山口, 智広

Graduate School of Pharmaceutical Sciences, Kyushu University

https://doi.org/10.11501/3150796

(2)

Regulatory Mechanis� of Organization of Golgi Appar!ltuE;

1999

'romob

^-� ^.... J...&. 1ro "�a� "=' ...l. ..1...&.-Ub

aurhi

-

(3)

Regulatory MechanisQJ of Organization of Golgi Apparatus

1999

I

Tomohiro Yamaguchi

(4)

CONTENTS

GENERAL IN'I'RODUCilON · · · · · · · · · · · ·· · · · · · · · · · · · · · · · · · · · · · · · . . . · · · 4

PART I. INVOLVE:MENT OF HETERDTRIMERIC G PROTEINS IN TilE ORGANIZATION OF TilE GOLGI APPARATUS . . . ... ... ... ... ... 8

1. INTRODUCfiON ^{· · ·}· · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · 8

2. EXPERIMENTAL PROCEDURES . . . ...^.^{. . . .}. ^·^{. . .}^{. . .}^{. . ·}. . . .. . · . . . · . . . · . . · 10

2. 1 Materials ^{· · ·}· · · · · · · · · · · · · · · · · · · · · · · · · · · · 10

2.2 Cell Culture··· . . · · · . . · · · . . · . . · . . · · . . · · · ·· · · · · · · · · ·· · · 10

2. 3 Permeabilization and Incubation Conditions . . . · . . . · . . . .. . . · 11 2. 4 Pertussis and Cholera Toxin Treatment . . . . . . . · . . . .. . . · . . .. . . · 11

2. 5 Immunofluorescence and Electron Microscopic Analysis··· · · · · · · · · · · · · · · · 11

2. 6 Expression Vectors and Transfection . . . . . . . ..^{. .}. . . .. . · . . . .. . · 12

3. RESULl"S · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · 13

3.1 NDGA Causes Disassembly of the Golgi Apparatus in a Microtubule-independent Manner · · · · · · · · · · · · · · · · · · · · · 13

3. 2 �-COP Is not Released by NDGA . . . .. . . · . . .. .. .. .. .. .. .. .. .. .. .. .. .. .. 14

3. 3 AlF ^4'or Mastoparan Markedly Prevents the Disassembly of the Golgi Apparatus . . . .. . .. . .. . .. . . . . . . .. . . . . .. . .. . .. . .. . .. . . .. . .. . . . . . . 14

3.4 GTPyS Blocks the NDGA-promoted Disassembly of the Golgi Apparatus ... 20

3. 5 Effect of �y-Subunits of Heterotrimeric G proteins · · · · · · · · · · · · · · · 21

3. 6 Effect of Pertussis and Cholera Toxins . . . . · . . . . · . . . . .. · .. . .. . .. . . · . . . .. . . . . . · 24

3. 7 Overexpression of a-subunits of heterotrimeric G proteins 3. 8 Overexpression of �y-subunit of heterotrimeric G proteins 24 25 4. DISCUSSION · · · · · · · · · · · · · · · · · · · · · · · · 32

PART II. ADP-RIBOSYIATION FACfOR I IS SENSITIVE TO N-ETIIYLMALEIMIDE · · · · · · · · · · · · · 37

1

(5)

1. INTRODUCTION · · · 37

2. EXPERIMENTAL PROCEDURES .. · .. · .... · .... · .. · ... .. · ... · .. · ... · .. .. · 38

2. 1 Materials · · · ^·· · · ^·· · · 3 8 2. 2 Cell Culture and Immunofluorescence Analysis · · ^·· · · ^·· · · ^·· · ^·· · · 38

2. 3 Permeabilization and Incubation Conditions · .. · .. · .. · .. · .. · .. · .... · ... · .... · .. · 38

2.4 Preparation of ARF-rich and Coatomer-rich Fractions · · · ·· · · ·· ·· · ·· · ·· · ·· · · 38

2. 5 Binding of Coatomer Proteins to Isolated Golgi Membranes · · · 39

2. 6 Determination of the Sulfhydryl Content ... ... · .... · · ... · 39

3. RESULTS AND DISCUSSION · .... · ... · .. · .. · .. · .. · .. · .. · .. · ... ... · .. ·.. 40

3.1 �-COP is Released from the Golgi Apparatus by NEM-Treatment · ·· · · 40

3.2 Activation of Heterotrimeric G Proteins Markedly Prevents the NEM-Induced Release of �-COP from the Golgi Membranes ... .. ^·... ... · .. .. .. 41

3.3 Cytosolic Proteins Are NEM-Sensitive ... · .. ... · ... · .. ·.. ... .. .... ... 41

3. 4 ARF Is Sensitive to NEM · · · ^·· · · 45

3. 5 Cys-159 Is the Only Cysteine Residue in ARF · .. · .. .... · .... · .... · ... ... ·.. 45

SUMMARY ... ... ... ··· ··· ··· 54

ACKNOWI..EDGEMENTS · · · ^·· · · ^·· · · ^·· · · 56

REFERENCES··· ··· ··· ··· ··· ··· 57

(6)

ABBREVIATIONS

ARF, ADP-ribosylation factor

ATPyS, adenosine 5'-3-0-(thio) triphosphate BFA, brefeldin A

COP, coat protein

DT NB, 5,5'-dithiobis (2-nitrobenzoic acid) ER, endoplasmic reticulum

Ga�y, a�y

subunits of heterotrimeric

G

proteins GTPyS, guanosine 5'-3-0-(thio) triphosphate man II, mannosidase II

NDGA, nordihydroguaiaretic acid NEM, N-ethylmaleimide

NSF, N-ethylmaleimide-sensitive factor NRK, normal rat kidney.

SNAP, soluble NSF attachment protein SNARE, SNAP receptor

3

(7)

General Introduction

In eucaryotic cells, the Golgi apparatus is usually located in the perinuclear region via association with the microtubule organizing center. It contains a collection of flattened cisternae resembling a stack of plates. These Golgi stacks usually consist of four or six cisternae, each typically about 1�m in diameter. The Golgi apparatus plays an important role in glycosylation of glycoproteins and sorting of the secretory proteins made in the endoplasmic reticulum (1,

2).

The intracellular transport pathway of newly synthesized secretory proteins is estimated as follows. Secretory proteins, which are synthesized on polyribosomes and translocated to the endoplasmic reticulum, exit from the ER via COPII vesicle. This vesicle serves to shuttle the proteins to the ER-Golgi intermediate

compartment (ERG I C) or vesicular-tubular clusters (VTCs ). From there, they could be transported to the cis-Golgi network (CGN) and cis compartment, then move to the medial compartment, and finally move to the trans compartment, where glycosylation is

completed. From the trans compartment, the proteins move to the trans Golgi network (TGN). In this tubular reticulum, they are segregated into different transport vesicles and delivered toward plasma membrane along the regulated secretory pathway via secretion granules or along the constitutive pathway (Fig. 1)

(3).

Transport of proteins between cictema is most likely mediated by transport vesicles, termed COPI coated vesicles ( ... so nm in diameter), which are also involved in the

retrograde transport, from the Golgi apparatus to the ER ( 4,

5).

Since the Golgi apparatus is the product of a dynamic balance between anterograde and retrograde membrane flow, it is considered that the structure of Golgi is very fragile. Many different drugs or mutations cause a loss of recognizable Golgi structures. For instance, brefeldin A, a reagent that causes a rapid release of COPI from Golgi membranes and blocks anterograde transport, causes an imbalance in membrane traffic and therefore leads to redistribution of Golgi components into the ER (

6).

Golgi stacks are also broken down during vertebrate mitosis.

At the onset of mitosis, the Golgi apparatus is disassembled into small vesicles and

(8)

stacks. But this mechanism is still controversial

(7

-9).

How does the Golgi apparatus maintain the steady-state membrane system under the dynamic membrane flow? In this study, I investigated the factors that are involved in the organization of the Golgi apparatus. As a result, I revealed that heterotrimeric G proteins are involved in the maintenance of the Golgi structure (Part I). And during the course of that study, I revealed that ADP-ribosylation factor-1 is sensitive to N-ethylmaleimide (Part II).

s

(9)

PM�

T Clathrin

9 COP II

• COP I

'

^Ribosome ^ER

(10)

Fig.

1.

Diagram depicting the major routes of vesicular traffic to and through the Golgi apparatus along the exocytic and endocytic pathways.

Exocytic pathways: Secretory proteins, membrane glycoproteins and lysosomal enzymes are synthesized on polyribosomes and translocated to the endoplasmic reticulum (ER)

(1)

where they undergo cotranslational and posttranslational processing. They exit from the ER via COPII vesicles

(2),

which serve to shuttle them to the ER-Golgi intermediate

compartment (ERGIC, or vesicular-tubular clusters, VTCs). From there, they could be transported to the cis-Golgi network (CGN) via COPI-coated vesicles

(3),

but this is still controversial. COPI-coated vesicles also function in retrograde transport (Golgi to ER) (

4).

Subsequently, either the proteins traverse the Golgi cisternae one by one via vesicular carriers

(5)

or transport occurs by cisternal maturation. Retrograde transport is also

assumed to take place between the stacked cisternae ( 6

)

. Sorting occurs in trans cisterna or trans-Golgi network (TGN). Lysosomal enzymes bind to mannose 6-phosphate receptors in the Golgi, are packaged into clathrin-coated vesicles in the TGN and deliverd to late (or early) endosomes (7). Membrane and secretory proteins are also sorted in the TGN and delivered by exocytosis along the regulated secretory pathway via secretion granules

(8)

or along the constitutive pathway

(9).

Endocytic pathways: The best-characterized endocytic pathway is receptor-mediated endocytosis through clathrin-coated vesicles budding from plasma membrane (PM)

(10).

Many receptors recycle back to the plasma membrane from early endosomes

(11

), whereas many ligands are transported from early to late endosomes to reach lysosomes

(12).

7

(11)

PART I.

INVOLVEMENT OF HETEROTRIMERIC G PROTEINS IN THE ORGANIZATION OF THE GOLGI APPARATUS

1. INTRODUCTION

Although the principle of the organization of the Golgi apparatus is not fully understood, recent progress has revealed that a common fusion machinery works for the formation of cisternae from vesiculated Golgi membranes. Warren and colleagues (10, 11) developed a cell-free system that reconstitutes the reassembly of vesiculated Golgi membranes that have been formed upon incubation of purified Golgi stacks with mitotic cytosol (12, 13), and demonstrated that N-ethylmaleimide-sensitive factor (NSF) and p97 independently mediate the fusion of vesiculated Golgi membranes. A 20S NSF complex consisting of NSF, SNAPs and SNAREs is known to mediate the fusion of transport vesicles with the target membranes in the exocytotic and endocytotic pathways (14, 15). p97 (16), alternatively named VCP (17), is a mammalian homologue of cdc48 protein of Saccharomyces

cerevisiae, which mediates the fusion of ER membranes during karyogamy (9, 10). Since the structures of assembled Golgi cisternae by NSF and p97 are different from each other, it was postulated that the two proteins have different roles in the assembly of the Golgi apparatus (11 ).

Malhotra and colleagues also discovered the involvement of NSF and p97 in the assembly of vesiculated Golgi membranes (20, 21). They found that ilimaquinone, which is one of sea sponge metabolite, causes disassembly of the Golgi apparatus to form small vesicles with 60-90 ^nmdiameter (22), and reconstituted the assembly of Golgi stacks from

vesiculated Golgi membranes (20). In the process of the assembly, NSF acts first to form larger vesicles of 300 ^nmdiameter from small vesicles of 60-90 nm, and then p97 play a

(12)

GTPyS nor rab-GDI, inhibitors for intra-Golgi vesicular transport (23, 24), blocked stack formation, suggesting that the mechanism of the assembly of vesiculated Golgi membranes is not exactly the same as that of intra-Golgi vesicular transport.

Tagaya and colleagues screened transport inhibitors using an intra-Golgi transport assay reconstituted in a cell-free system (25), and found that nordihydroguaiaretic acid (NDGA), known as a dual inhibitor for lipoxygenases and cyclooxygenase, inhibits the transport assay (26) (Structure of NDGA is descrived below). A recent study using BHK-21 cells showed that NDGA without affecting in the organization of microtubules blocks protein transport from the ER to the Golgi apparatus and from the Golgi apparatus to the plasma membrane, and retards the brefeldin A-mediated retrograde transport of man II from the Golgi apparatus to the ER (27).

In this part, I examined the effect of NDGA on the dynamic membrane-flow system of the Golgi apparatus in normal rat kidney (NRK) cells by immunofluorescence analysis. I found that NDGA inactivates heterotrimeric G proteins, thereby causing to the disassembly of the Golgi apparatus (28).

nordihydroguaiaretic acid (NDGA)

9

(13)

2. EXPERIMENTAL PROCEDURES

2.1 Maten'als ----^- NDGA was obtained from Biomol Research Laboratories and freshly dissolved in dimethylsulfoxide before use. Cycloheximide, AIC� and NaF were obtained from Wako Chemicals. Digitonin was obtained from Merck. C6-NBD-ceramide was obtained from Molecular Probes. Cholera toxin was purchased from List Biological Laboratories Inc. Pertussis toxin was purchased from Kaken Pharma. Plastic coverslips (Celldesk FL1) were purchased from Sumitomo Bakelite. A polyclonal antibody against the amino-terminal fragment (residues 1-10) of bovine G� was obtained from Upstate

Biotechnology, Inc. Monoclonal antibodies against a-tubulin (clone DM1A), Mannosidase II (Man II) (clone 53FC3) and protein disulfide isomerase were purchased from BioMakor, BAbCo and Fuji Chemical Industry, respectively. Poly clonal antibodies against man II and

�-COP were kindly donated from Dr. K. Moremen at University of Georgia and Dr. Y.

lkehara at Fukuoka University, respectively. Mastoparan and MAS17 were purchased from Sigma. Polyclonal antibodies against Ga13, Gas' Gaq/11, G�, Gaz were purchased from Santa Cruz Biotech. Inc .. Ga�y were prepared from bovine brain as described (29). �y subunits were prepared as described (30). Briefly, purified Ga�y were incubated with 20

�M AIC�, 10 mM NaF and 10 mM MgC12 at 30oC for 1 h, and then G�y were isolated on a Phenyl Superose column. To prepare detergent-free G�y, the isolated G�y were

extensively dialyzed against 25 mM HEPES (pH 7.0) containing 125 mM potassium acetate, 2.5 mM magnesium acetate, 1 mM dithiothreitol and 1 mglml glucose (buffer A).

No significant precipitation was observed during dialysis. Immunoblotting confirmed no significant loss of the proteins during dialysis without detergents. G�1y2 in 20 mM HEPES (pH 8.0) containing 1 mM EDTA, 3 mM MgC12, 3 mM dithiothreitol, 100 mM NaCl and 0.7% CHAPS were kindly donated from Drs. Y. Kaziro and H. Itoh. Bovine brain cytosol was prepared as described (31).

2.2 Cell Culture--- NRK cells obtained from the Riken Cell Bank were grown on glass coverslips in a-minimum essential medium supplemented with 50 IU/ml penicillin,

10

(14)

50

J!g/ml streptomycin,

10 %

fetal calf serum at 37oC in a

5%

C02 incubator. COS7 cells obtained from the Riken Cell Bank were grown on glass coverslips in Dulbecco's modified Eagle medium, supplemented with

50

IU/ml penicillin,

50

f.tg/ml streptomycin,

10%

_fetal

calf serum at

37oC

_{in a}

5%

C02 incubator.

2.3 PenneabJ1ization and Incubation Conditions ---- ^-NRK cells were washed twice with buffer A and then permeabilized with

40

J!g/ml digitonin in the same buffer at ooc _for

10

min. After washing with buffer A twice, the cells were incubated in

1

ml of the reaction mixture at

32oC

_for

40

min. The reaction mixture contained buffer A, an ATP-regenerating system (final concentration

5

mM creatine phosphate,

0.25

_{mM UTP,}

0.05

mM ATP and

12

IU/ml creatine kinase), and bovine brain cytosol (final concentration

1.5-2

_mglml).

2. 4 Pertussis and Cholera Toxin Treatment --- NRK cells were pretreated with pertussis

(0.1

mglml) or cholera

(0.1

mglml) toxin for

16

_h.

2.5 Immunofluorescence and Electron Microscopic Analyses--- For immunofluore

scence analysis, cells were fixed in PBS containing

4%

paraformaldehyde and

0.15M

sucrose, treated with PBS containing

0.2%

_Triton

X-100,

and then incubated in PBS

2%

containing bovine serum albumin. After incubation with primary antibodies, the cells were incubated with fluorescein-conjugated goat anti-mouse· lgGs and/or rhodamine-conjugated goat anti-rabbit IgG. The coverslips were mounted with PBS containing

90%

_glycerol,

1

mglml p-phenylenediamine, and

10

mM sodium azide and then observed under an Olympus

BX50

microscope

(32).

For electron microscopic analysis, cells were cultured on plastic coverslips and fixed in

2%

glutaraldehyde in

0.1

M phosphate buffer (pH 7.4) at

4oC

overnight. The cells were washed in

0.1

M cacodylate buffer (pH 7.4), and were post-fixed in

1%

_{Os04 in}

0.1

M cacodylate buffer for

1

h. After washing in distilled water, the cells were dehydrated in

50%

ethanol for

10

min, and block-stained with

2%

uranyl acetate in

70%

ethanol for

2

h. They were further dehydrated with a graded series of ethanol, and were embedded in epoxy resin. Ultrathin sections were doubly stained with uranyl acetate

11

(15)

and lead citrate, and observed under a Hitachi H7000 electron microscope.

2. 6 Expression Vectors and Transfection --- cDNAs for rat Gail' Gai2, Gai3, Gaz, mouse Ga11,and bovine Gas (Gas4) inserted into pCMVS (33-36), and cDNAs for constitutively activated GTPase-deficient mutants of Gas' Gaz, and Ga11 (a8Q213L, az0204L, and a11Q209L) inserted to pCMVS (37), and Gf31 and Gy2 cDNAs inserted into pCMVS (38) were kindly donated by Dr. H. ltoh at Tokyo Institute of Technology. In COS7 cells grown on 12-well plate (Falcon), transient transfection was carried out using LipofecatAMINE Plus (Gibco BRL) following this operation manual. For transient transfection in NRK cells grown on CELLocate Coverslips (Eppendorf), plasmids

(0.1mg/ml in PBS) were microinjected to intra-nucleus of the cells using micromanupirator 5171 (Eppendorf).

(16)

3. RESULTS

3.1 NDGA Causes Disassembly of the Golgi Apparatus in a Microtubule-independent Manner --- NRK cells were incubated with NDGA at

37 oC

for 30 min, and the distribution of the Golgi apparatus was analyzed with immunofluorescence microscopy using an antibody against man II, a medial Golgi marker protein. As shown in Fig. 2, NDGA caused dispersion of man II from the perinuclear region into the cytoplasm. At 20 1-1M NDGA, the dispersed pattern of man II was observed in more than 95% of cells.

Microtubules play an important role in the perinuclear localization of the Golgi apparatus

(39).

However, the morphology of microtubules was not affected by NDGA (Fig.

3).

When the Golgi apparatus was stained with C6-NBD-ceramide ( 40) or rhodamine

conjugated WGA, which is known to bind to the trans-Golgi ( 41 ), similar dispersed patterns were observed, indicating the whole region of the Golgi apparatus is disassembled by NDGA (data not shown).

The disassembly of the Golgi apparatus started at about 5 min after addition of NDGA, and was completed at about 40 min. Upon removal of NDGA, the disassembled Golgi membranes started to reassemble, and complete Golgi stacks were formed at about 1 h. The disassembly of the Golgi apparatus by NDGA did not occur in the presence of an

ATP-depleting system (50 mM 2-deoxyglucose and 0.05% azide), implying the requirement of an energy for this process (data not shown).

Electron microscopic analysis confirmed the results of the immunofluorescence study. At 15 min after addition of NDGA, Golgi membranes were converted into aggregates of small vesicles and tubulovesicular structures. At 50 min, no aggregates of vesicles were observed in the perinuclear region in many cells, instead small vesicles appeared to be dispersed throughout the cytoplasm. Electron microscopic analysis also showed that no significant morphological changes occurred in mitochondria, the nucleus or the plasma membrane. A slight dilation of the ER was observed, but this phenomenon was generally observed when vesicular transport was inhibited. These results suggest that NDGA specifically affects the Golgi apparatus and the vesicular transport pathways (data not shown).

13

(17)

3.2

f3

-COP Is not Released by NDGA ^--

-

^-

- Brefeldin A (BFA) causes disassembly of the Golgi apparatus and promotes the retrograde transport of Golgi components to the ER ( 6).

The effect by brefeldin A detectable within 30 s is the release of �-COP ( 42), a component of Golgi-derived transport vesicles ( 43-45). I therefore examined whether or not �-COP is rapidly released from the Golgi apparatus by NDGA. When disassembly of the Golgi apparatus started at 5 min after addition of NDGA, �-COP was still associated with it. In addition, �-COP was co-localized with man 11-containing vesicles even after complete disassembly occurred (Fig. 4). These results suggest that the mechanism of Golgi disassembly by NDGA is different from that of BFA.

3.3 AJF4- or Mastoparan Markedly Prevents the Disassembly of the Golgi Apparatus by

NDGA

--- Recent studies revealed the involvement of heterotrimeric G proteins in the membrane traffic processes ( 46). I examined whether or not heterotrimeric G proteins are involved in the disassembly of the Golgi apparatus. For this purpose, NRK cells were preincubated with or without AIF4' (50 llM AIC� and 30 mM NaF) at 37 _'C for 10 min, and then NDGA was added. AIF4- is an activator of the heterotrimeric but not the small

molecular weight G proteins (47, 48). It is believed that AIF4• acts as a mnemonic for the y-phosphate and activates the GDP-bound form of Ga.

As

shown in Fig. 5, the

preincubation of cells with AIF4' markedly blocked the disassembly of the Golgi apparatus

by NDGA, whereas 50 llM AIC� alone did not exhibit such an effect. Mastoparan is an

amphiphilic peptide derived from wasp venom and directly activating G proteins by a

mechanisman analogous to that of G protein-coupled receptors. When intact NRK cells

were incubated with NDGA in the presence of mastoparan, disassembly of Golgi apparatus

was markedly suppressed (Fig. 6). In contrast to this, MAS17 which is inactive analog of

mastoparan did not suppress the Golgi disassembly (Fig. 6). These results raise the

possibility that heterotrimeric G protein(s) are involved in the disassembly of the Golgi

apparatus by NDGA.

(18)

Fig. 2. Effect of NDGA on the morphology of the Golgi apparatus. NRK cells were incubated without

(A)

^{or with}

(B)

20 �-tM NDGA at 37 oC for 20 min. Man II was immunostained.

(19)

Man II a-Tubulin

control

NDGA

Fig. 3. Effect of NDGA on the morphology of the Golgi apparatus and microtubules. NRK cells were incubated without

(Upper panel)

^{or with}

(Lower panel)

20 �M NDGA at 37 oC for 20 min. Man II

(left panel)

and a-tubulin

(right panel)

^were

double-immunostained.

(20)

Manti p-COP

control

NDGA

Fig.

^4.

Effect of NDGA on the association of �-COP with the Golgi apparatus.

^NRKcells were incubated without

(Upper panel)

or with

(Lower panel)

NDGA at 37 oC for 30 min. Man II

(left panel)

and �-COP

(Right panel)

were double

immunostained. Bar represents 20 IJ.ffi.

17

(21)

Fig. 5. Preincubation of cells with AIF; blocks the NDGA-promoted Golgi disassembly. NRK

cells were preincubated in the presence of AIC� (C) or AIF4- (D) at

^{37 aC}

for

¹⁰

min, and then incubated without (A) or with (B,

^C

and D)

^NDGA

in the presence of AIC� (C) or AIF4- (D) for

¹⁵

min.

^{Man II}

was immunostained. Bar represents

20�m.

(22)

control NDGA

·-..,,

·t:

^o'

^�

^..

�

lt�

llf1j�

NDGA

+

mastoparan NDGA

+

MAS17

Fig. 6. Activation of heterotrimeric G proteins by mastoparan blocks disassembly of the Golgi apparatus induced by NDGA in intact NRK cells.

Cells were incubated without or with 20 11M NDGA in the presence of 50 1-1M mastoparan or 50 11M MAS17 at 37 oc for 20 min. Man II was immunostained.

(23)

3. 4 GTPyS Blocks the NDGA-promoted Disassembly of the Golgi Apparatus in

Digitonin-penneabilized Cells ----^-To provide a more direct line of evidence for the involvement of heterotrimeric G proteins in the disassembly of the Golgi apparatus by NDGA, we reconstituted the disassembly event in permeabilized NRK cells, and examined the effect of membrane nonpermeable effectors for heterotrimeric G proteins. Digitonin was used for permeabilization because it selectively permeabilize the cholesterol-rich plasma membrane but not other organelle membranes. This selectivity has allowed the

reconstitution of a variety of membrane traffic events (

49-51).

I permeabilized NRK cells with

40

J.tg/ml digitonin at

0

^ocfor

10

min because under this condition the transport of vesicular stomatitis virus-encoded protein from the ER to the Golgi apparatus could be reconstituted by the addition of an ATP-regenerating system and cytosol

(51).

When permeabilized NRK cells were incubated with

60

^f.!MNDGA in the absence of an ATP-regenerating system or cytosol at 32 oc _for

40

min, the Golgi apparatus was considerably dispersed (Fig. 7B). It is not certain whether or not this dispersion is relevant to the disassembly of the Golgi apparatus in intact cells because ATP is required in the latter case. When an ATP-regenerating system was included in the assay, the Golgi apparatus was not significantly disassembled by NDGA (Fig. 7C). When permeabilized cells were incubated with

60

^J.!MNDGA in the presence of an ATP-regenerating system and bovine brain cytosol, the Golgi apparatus became dispersed throughout the cells (Fig.

70). The dispersion was remarkably suppressed when permeabilized cells were incubated in the presence of an ATP-depleting system and cytosol (data not shown), suggesting that the requirement of ATP hydrolysis for this process. The dispersion pattern of the Golgi apparatus in the presence of an ATP-regenerating system and cytosol was somewhat different from that in intact cells, and rather similar to the ER pattern. Double-staining revealed that the distribution of man II partly coincides with that of protein disulfide

isomerase, an ER marker protein, suggesting that a part of man II is redistributed to the ER.

�-COP was not co-localized with man II in NDGA-treated permeabilized cells, and rather seemed to be redistributed on vesicle-like structures at the perinulcear region. This

(24)

present we can not reconstitute the disassembly of the Golgi apparatus without the redistribution of man II to the ER. However, the dispersion of the Golgi apparatus in permeabilized cells was also blocked by AlF4- (data not shown), suggesting that initial stage of the disassembly is relevant to that in intact cells. Redistribution of man II to the ER is probably due to a low efficiency in the recruitment of coat proteins from added bovine brain cytosol. Perhaps, disassembled Golgi membranes were fused with the ER when coat attachment is not enough.

When 20 11M GTPyS was added in the complete mixture containing an ATP-regenerating system, cytosol and 60 f.1M NDGA, the disassembly of the Golgi apparatus was blocked (Fig. 7E). A similar result was observed when the concentration of GTPyS was 2 f.1M. In contrast, addition of 20 11M AMPPNP (Fig. 7F) or 211M ATPyS (data not shown) showed little effect. It should be noted that the concentration of ATP in the assay mixture is 50 f.1M.

When 20 11M ATPyS was added, the disassembly of the Golgi apparatus was significantly suppressed. AMPPNP might be less effective than ATPyS with respect to the inhibition of ATP hydrolysis. These results suggest that the activation of GTP-binding proteins blocks the NDGA-induced disassembly of the Golgi apparatus.

3.5 The Effect of {3y-Subunits of Heterotrimeric G Proteins ^----- To exclude the possibility that small GTP-binding proteins are involved in the NDGA-promoted Golgi disassembly in permeabilized cells, we next examined the effect of �y subunits of

heterotrimeric G proteins on the disassembly of the Golgi apparatus. G�y bind to Ga but not to ras-related small GTP binding proteins (52, 53). In the resting state, the GDP-bound form of Ga is associated with G�y. When the bound GDP is replaced with GTP, Ga is activated and released from G�y. When excess G�y are present, Ga is associated with them, which results in the inactivation of Ga. If this scheme is also applicable to the case for the disassembly of the Golgi apparatus, it is expected that the addition of G�y would prevent the effect of GTPyS and cause disassembly of the Golgi apparatus. As shown in Fig. 8D, addition of 111M G�y to the assay mixture containing GTPyS resulted in

disassembly of the Golgi apparatus. In contrast, addition of buffer alone had no effect (data

21

(25)

A

^' ^.�^..^.

\._ . •

.....

0

�

� *•

.,_ ^• I

,

.. �

;J

^"

•

B ..

^'

�·

^I^• ^·-.

J

^fr. ⁱ ^,.

��

I

l

� _._!*' ·��

t

'-

c F

^...

�

I ..

I \..,.

��

�

Fig. ^7. Effect of GTPyS on the NDGA-promoted disassembly of the

Golgi apparatus in digitonin-permeabilized cells. Permeabilized NRK cells were incubated at 32

oC

_for⁴⁰_min.

(A)

Cells were incubated without NDGA in the absence of an ATP-regenerating system or cytosol.

(B)

Cells were incubated with 60 !J.M NDGA in the absence of an ATP-regenerating system or cytosol. (C) Cells were incubated with 60 tJ.M NDGA in the presence of an ATP-regenerating system but not cytosol.

(D)

Cells were incubated in the presence of an ATP-regenerating system and cytosol (complete assay mixture) with 60 tJ.M NDGA. (E) Cells were incubated in the complete assay mixture with 60 !J.M NDGA plus 20 tJ.M GTPyS. (F) Cells were incubated in the complete assay mixture

(26)

Fig. 8.

Gpy

reverses the effect of

GTPyS. Permeabilized cells were incubated

in

the complete assay mixture without NDGA (A), with

60

f.A.M NDGA (B),

60

f.A.M NDGA and

20

f.A.M GTPyS (C), with

60

f.A.M NDGA and

20

f.A.M GTPyS plus

1

f.A.M G�y (D) or with

1

f.A.M G�y (E). Bar represents

20 f.lffi·

23

(27)

not shown). Furthermore, G�y alone did not significantly cause disassembly of the Golgi apparatus in the absence of NDGA (Fig.

8E).

180

_nM

G�1y2 were used instead of G�y . In this case, the assay mixture contained

0.014%

CHAPS, which was derived from a solution containing G�1y2• This concentration of CHAPS without G�1y2 had no significant effect on the Golgi morphology.

3. 6 Effect of Pertussis and Cholera Toxins --- Pertussis and cholera toxins are useful

tools for determining which type of heterotrimeric G proteins are involved in signal transduction

(54, 55).

Pertussis toxin catalyzes the ADP-ribosylation of a-subunit of Gi, which results in the uncoupling of the Gi from activation. Cholera toxin, on the other hand, catalyzes the ADP-ribosylation of Gs' which results in constitutive activation of the G,. I next examined the effect of pertussis and cholera toxins on the NDGA-promoted

disassembly of the Golgi apparatus. When pertussis toxin- or cholera toxin-treated cells were incubated with NDGA, the Golgi apparatus was disassembled by NDGA and GTPyS blocked the effect of NDGA as in the case of nontreated cells (data not shown), suggesting that heterotrimeric G protein(s) involved in the process of the disassembly are different from pertussis toxin- or cholera toxin-sensitive ones.

3. 7 Overexpression of a- subunits of heterotrimen·c G proteins--- To identify the sub-class of a-subunits involved in the NDGA-induced Golgi disassembly reaction, I examined the effect of overexpression of a -subunit in NRK cells. In order to express a-subunit transiently, plasmids were microinjected to the intra-nucleus of NRK cells. Gas' Ga11,one of the Gaq family, Gaz, Ga11, Ga12, or Ga13 was overexpressed and then the cells were treated with NDGA. ^Asshown in Fig.

9,

Golgi disassembly was suppressed in

30%

cells overexpressing Gaz or Ga12• Since Golgi disassembly occured in about

95%

cells by NDGA, the suppressive effects of other a -subunits were at background levels. I next examined the effect of overexpression of active mutant a -subunits

(QL

_mutants,

constitutinely activated GTPase-deficient mutants). In the case of active mutant, similar

(28)

disassembly was suppressed in some cells, and expression of other subunits had no effect (Fig.

10).

One possibility to explain the suppressive effect of Gaz and Gai2 on Golgi disassembly is that these two proteins just enhance the junction between Golgi membranes, but not affect specific processes. If this is the case, it is expected that the Golgi apparatus is generally resistant against reagents that cause Golgi disassembly. To test this possibility, cells overexpressing Gaz were incubated with nocodazole to disassemble microtubules

(39)

or brefeldin A to promote redistribution into the ER (

6).

As shown in Fig.

11,

the disassembly of the Golgi apparatus by nocodazole or brefeldin A occurred in cells overexpressing Gaz.

This result may rule out the possibility that Gaz is specifically· involved in the action of NDGA.

3. 8 Overexpression of

{31

y2 -subunit of heterotrimeric G proteins --^--- In a recent report, Malhotra and colleagues showed that free �y-subunits per se triggers Golgi vesiculation in permeabilized NRK cells

(56).

In order to investigate how �y-subunits affect the Golgi apparatus in intact cells, I transfected �1 y2-subunit to NRK cells and observed the morphology of the Golgi apparatus. As shown in Fig. 12, no morphological change of Golgi apparatus was observed in cells overexpressing �1 y2 -subunits. Furthermore, the effect of NDGA did not change in cells overexpressing �1 y2 -subunits. These results indicate that NDGA-induced disassembly of the Golgi 'apparatus is regulated by the pathway including a -subunbit of trimeric G proteins.

25

(29)

Man II Ga

Gs

G11

Gz

Gi1

Gi2

Gi3

(30)

B

50 Cl) =

...

... �

� a.

a. 40

c(

·-

9J 0

"

... CJ 30

... �

.E

.c ...

"i

.!! 20

"ii 0 ....

0 a>

0) 10

... � c a>

CJ ..

a>

Q. 0

Gas Ga11 Gaz Gai1 Gai2 Gai3

^non

Fig. ^9. Overexpression of wild-type-Ga, or -Ga12 inhibits the effect of NDGA in NRK cells. A, Wild-type a-subunits were transiently overexpressed, after incubation for 4 h at 37 °C, they were incubated with 20 J.A.M NDGA at 37 °C for 10 min.

Man II and a -subunit were double-immunostained. B, Quantitation of cells with intact Golgi after NDGA treatment. The vertical axis shows the percentage of the cells with the intact Golgi apparatus. More than 200 cells on each (up to 600 cells) were scored.

27

(31)

Q213L Gas

Ga11 Q209L

Q205L Guz

Man II Ga

(32)

B

50 Cl) .. � .. cu cu Q.

Q. 40

<(

!1) 0 CJ ..

(,) 30

.. cu -c

.c ..

"i

!l ²⁰

u Gi

....

0 G)

0) 10

.. cu c G) ... (,) G) D.

0

Gas Ga11 Gaz

^non

Fig. 10. Effects of overexprcssion of active-mutant-a-subunits on the NDGA-induced disassembly of Golgi apparatus. A, In order to express a-subunit transiently, plasmids encoding a-subunits were microinjected to the intra

nucleus of NRK cells, after incubation for 4 h at 37 °C, they were incubated with 20 f.!M NDGA at 37 oc for 10 min. Man II and a-subunit were double-immunostained. B, Quantitation of cells with intact Golgi after NDGA treatment. The vertical axis shows the percentage of the cells with the intact Golgi apparatus. More than 300 cells on each (up to 800 cells) were scored.

29

(33)

Man II Gaz

BFA

Nocodazole

Fig. 11. Ovcrexpression of Gaz docs not affect Brcfcldin-A (BFA)- or Nocodazolc-induccd Golgi disassembly in NRK cells. In order to express wild-type Ga.z transiently, plasmids were microinjected in the intra-nucleus of NRK cells, and then incubated for 4 h at 37 °C. Then the cells were incubated with 5 �M BFA at 37 oc for 10 min or with 15 �glml nocodazole at 37 oc for 20 min. Man II and Ga.z were

double-immunostained.

(34)

NDGA Man II G�

-

+

Fig.

^12.

Overexpression of p1y1-subunit in

^NRK

cells. Plasmids encoding

�1- and y2-subunits were microinjected in the intra-nucleus of NRK cells and then incubated for

⁴

h at

^{37 °C.}

They were incubated without (A, B, Cand D) with (Band F)

²⁰

f.LM NDGA at

^{37 oc}

for

¹⁰

min. The cells were double-immunostained with Man II (A, Cand E) and �-subunit (B, D and F). A, B: Intact Golgi apparatus as control.

31

(35)

4. DISCUSSION

In this study, I showed that NDGA causes disassembly of the Golgi apparatus of NRK cells in a dose-, time- and energy-dependent manner. A kinetic analysis using electron microscopy revealed that the Golgi apparatus is first converted to aggregates of small vesicles, and then dispersed. Although microtubules were not disassembled by NDGA, disassembled Golgi vesicles seemed to be dispersed throughout the cytoplasm. In contrast to brefeldin A ( 42), NDGA did not promote the rapid release of �-COP, a component of CO PI-coated vesicles ( 43-45). Since no significant morphological changes were observed in mitochondria, the nucleus, or the plasma membrane, it is obvious that NDGA interacts with a limited number of components involved in the organization of the Golgi apparatus.

NDGA causes the Golgi disassembly in a wide variety of cell lines such as PC12, COS, HeLa, CHO, BHK, L-929, although sensitivities to the compounds may be different in

types of cell lines (data not shown).

As

I described before, at the onset of mitosis, the Golgi apparatus is converted into small vesicles in both COP-1-dependent and -independent pathways (12, 13), at telophase, the vesiculated membranes are equally dispersed in the mitotic cytoplasm ( 66-68). I do not know whether or not the mechanism of the NDGA-induced disassembly of the Golgi apparatus is exactly the same as that of the break down of the Golgi apparatus at the onset of mitosis. Further studies will be required to address this problem.

Tagaya and colleagues previously suggested that the target molecule(s) for NDGA may be phospholipase� (26). This conclusion was based on the observation that

phospholipase � inhibitors such as unsaturated fatty acids and analogous of arachidonic

acid including NDGA inhibit intra-Golgi protein transport (26). If this was true, it is

expected that the structure of the Golgi apparatus may be maintained by phospholipase �'

therefore, inhibition of this activity by NDGA would cause disassembly of the Golgi

apparatus. So I examined whether the addition of arachidonic acid or hog pancreas

phospholipase� reverses the effects of NDGA in digitonin-permeabilized cells.

As

the

(36)

(data not shown). Therefore, I do not prefer this possibility.

In the experiment using intact and digitonin-penneabilized NRK cells, I obtained several lines of evidence that heterotrimeric G protein( s) are involved in the NDGA-promoted disassembly of the Golgi apparatus. First, AlF4- or mastoparan markedly prevented this disassembly in intact cells and digitonin-penneabilized cells. Second, GTPyS had the same effect in digitonin-penneabilized cells, and G�y suppressed the effect of GTPyS. These results strongly suggest that NDGA causes the disassembly of the Golgi apparatus via a pathway involving heterotrimeric G protein(s).

Recent reports indicate that heterotrimeric G proteins are present on the Golgi apparatus and regulate vesicular transport. The first suggestion came from the finding that AlF4- inhibits intra-Golgi protein transport in a cell-free system (23). A later study revealed that Gai3 is present on the Golgi apparatus and regulates the secretion of heparan sulfate proteoglycan in LLC-PKl cells (57). In MDCK cells, Gas and Gai regulate the apical and basolateral transport pathways, respectively (58). In PC12 cells, the formation of large dense core vesicles is inhibited by Gai and stimulated by Gas (59). Denker et al. (60) demonstrated that Gas' Gai and Gaq/11 are present on the Golgi apparatus. Our present finding point to a new role of heterotrimeric G proteins; regulation of the organization of the Golgi apparatus.

In the latter part of this study, I examined the effect of overexpression of a -subunits of trimeric G proteins in order to determine the sub-class 'of a -subunits involved in the action of NDGA. I found that Golgi disassembly was suppressed in some cells when Gaz or Gai2 was overexpressed. Previously, Stow and colleagues reported that overexpression of Gau causes accumulation of secretory proteins in the Golgi apparatus (57). Taken together these results suggest that different sub-classes of a-subunits in the Golgi apparatus have each different roles.

Gaz, formally a member of the Gai family, was found to be inactivated by arachidonate and related unsaturated fatty acids (61). Interestingly, this protein is insensitive to pertussis toxin (62). This is reasonable, because our results suggest that heterotrimeric G protein(s) involved in the process of the disassembly are different from pertussis toxin- or cholera

33

(37)

toxin-sensitive ones. Although Gail is sensitive to purtussis toxin, it is expect that Gaz could compensates for the effect of Gail even when Gail was affected by pertussis toxin.

Although Gaz is predominately expressed in platelets and neurons, this does not necessarily mean that thls protein is not expressed in other types of cells. NSF, which consists of general fusion machlnery for vesicular transport in cells, is also predominantly expressed in neurons

( 63, 64).

Malhotra and colleagues reported that the �y-subunits per se are required and sufficient for the disassembly of the Golgi apparatus using a permeabilized cell system. I examined the effect of overexpression of �1y2-subunits in intact cells. However, the present result did not show any significant morphological change of the Golgi apparatus. One possibility to explain this discrepancy is that �1y2-subunits do not have enough efficiency to cause the Golgi disassembly, and different species of �y-subunits besides �1y2 have some roles to cause the disassembly. Another explanation is that some factors against the effect of

�y-subunits exist in intact cells. Regulators of activation of trimeric G proteins such as RGS family proteins

(65)

may have a role to protect the Golgi apparatus from disassembly.

As a conclusion, I would like to propose a scheme of the action of NDGA as follows. I speculate that trimeric G proteins in the Golgi are active in the absence of signals. When signals such as NDGA bind to putative receptors, the signal-receptor complex promotes the hydrolysis of GTP bound to trimeric G proteins, which results in their inactivation, and thereby causes Golgi disassembly (Fig.

13).

It should.be noted that the Golgi is assembled without signals, and activators such as A1F4-, mastoparan, GTPyS protect the Golgi

apparatus from disassembly. I speculate that trimeric G proteins in the Golgi are the prototype of trimeric G proteins, although plasma membrane trimeric G proteins were discovered first. In unicellular organisms, perhaps trimeric G proteins have roles to

maintain the structure of organelles such as the Golgi. After the appearance of multicellular organisms, trimeric G proteins acquired the reverse regulatory mechanism for

communication between cells.

What kind of cascades regulates the organization of the Golgi apparatus? One possible

(38)

included in the classical MAP kinase cascade is required for Golgi fragmentation during mitosis (7). However, Warren and colleagues suggest that Cdc2 is directly involved in Golgi fragmentation but not signaling via

MEKl (8).

In order to clarify the cascade in the maintenance of Golgi structure, it is required to identify the putative factors which exist in the upstream or downstream of the heterotrimeric G proteins in the Golgi apparatus.

35

(39)

NDGA --...- Rec eptor

' G protein

G protein

az

and

a12

(active)

^...^�^__

______ ...-_ az

and

^a12

(inactive)

Golgi Stacks Vesiculation

c=:::> 0

c::> 0

c �

c::::: � ⁰₀^c::>_c=:> ⁰ ⁰

0

c::::: ---=:::>

0

⁰ ⁰

Fig. 13 Scheme of action of NDGA. When signals such as NDGA bind to putative receptors, the signal-receptor complex promot

e

s the hydrolysis of GTP bound to trimeric G proteins, which results in their inactivation, and thereby causes Golgi

disassembly.

(40)

PART II

ADP-RIBOSYLATION FACTOR I IS SENSITIVE TO N-ETHYLMALEIMIDE

2. INTRODUCTION

COPI-coated vesicles mediate protein transport in both the anterograde direction through the Golgi stack and the retrograde direction from the Golgi apparatus to the endoplasmic reticulum

( 4, 5).

COPI consists of coatomer, comprising

7

subunits, and a small

GTP-binding protein, ARF

( 45, 69, 70).

These cytosolic proteins are required and enough for the formation of COPI-coated vesicles

(71).

ARF is recruited onto Golgi membranes in the presence of a nonhydrolyzable GTP analogue, GTPyS, in vitro

(69).

The recruitment of ARF onto Golgi membranes in vivo occurs as a consequence of the replacement of bound GDP with GTP in ARF, which is mediated by ARF-specific guanine nucleotide exchange factor

(72,73).

^AlF4- also promotes the recruitment of ARF onto Golgi membranes by activating heterotrimeric G proteins

(74, 75)

or directly inhibiting the GTPase activity of the ARF associated with ARF GAP

(76).

ARF is a prerequisite for the binding of coatomer to Golgi membranes

(77, 78),

although its role is a matter of controversy

(69, 76, 79).

In Part I, I showed that NDGA causes heterotrimeric G proteins to be in an inactive state, and thereby causes disassembly of the Golgi apparatus in NRK cells

(28).

During the course of that study, I found that the treatment of NRK cells with

1

mM NEM blocks the NDGA-induced Golgi disassembly and promotes the release of �-COP, a component of coatomer, from the Golgi apparatus. In the present study, I examined which component is sensitive to NEM among the proteins involved in the association of COPI with Golgi

membranes. I found that the treatment of ARF with NEM results in loss of its ability to bind to Golgi membranes

(80).

37

(41)

2. EXPERIMENTAL PROCEDURES

2.1 Materials--- NEM, AIC�, NaF, and DTNB were obtained from Wako Chemicals.

Monoclonal antibodies against Man II (clone 53FC3), y-adaptin, and ARF were purchased from BAbCo, Transduction Laboratories, and Affinity Bioreagents I nc., respectively. A polyclonal antibody against �-COP was raised against a synthetic peptide (residues 496-513 of �-COP). A polyclonal antibody against man II was raised against its lumenal 42-kDa region close to the transmembrane domain. Recombinant myristoylated ARF1 (

8

1) was kindly donated by Dr. K. Nakayama at Tsukuba university. Rat liver Golgi membranes were isolated as described previously (25).

2.2 Cell Culture and Immunofluorescence Analysis --- NRK cells obtained from the ATCC were grown on glass coverslips in ^a-minimum essential medium supplemented with 50 IU/ml penicillin, 50 mg/ml streptomycin and 10 %fetal calf serum at 37 °C in an

incubator containing 5 % C02 and 95% air. Immunofluorescence analysis was performed as described in EXPERIMENTAL PROCEDURES in Part I (32).

2.3 Permeabilization and Incubation Conditions --·--NRK cells were washed twice with KHM buffer (25 mM HEPES/KOH (pH 7.0) containing 110 mM potassium acetate and 2.5 mM magnesium acetate), and then penneabilized with

40

mg/ml digitonin in KHM buffer at

0

°C for 5 min. After washing with KHM buffer, the cells were incubated with KHMG buffer (25 mM HEPES/KOH (pH 7 .0) containing 110 mM potassium acetate, 2.5 mM magnesium acetate, and 1 mg/ml glucose) in the presence of an A TP regenerating system ( 5 mM creatine phosphate, 0.25 mM UTP, 0.05 mM ATP, and 12 IU/ml creatine

phosphokinase) and bovine brain cytosol

(0.8

- 1mg/ml).

2. 4 Preparation of ARF·n'ch and Coatomer-rich Fractions ^--^--- Bovine brain cytosol was prepared as described previously (31). ARF-rich and coatomer-rich fractions were prepared

(42)

applied to a Sephacryl S-300 HR column in a FPLC system (Pharmacia Biotechnology Inc.), and then the column was developed at the flow rate of 0.5 ml/min with 25 mM Tris-HCl (pH 8.0) containing 0.1 M KCI. Fractions of 1.25 ml each were collected, and the fractions containing ARF or coatomer were examined by immunoblotting. Appropriate fractions were collected, concentrated 5 times, and stored at -80 °C until use.

2.5 Binding of Coatomer Proteins to Isolated Golgi Membranes -^--^-- Mixtures containing rat liver Golgi membranes (10 �g), bovine brain cytosol (3 mglml protein), and an ATP regenerating system were preincubated at 37 °C for 10 min with or without 1 mM NEM in 100 �1 of buffer A (25 mM HEPES/KOH (pH 7 .0) containing 25 mM KCl and 2.5 mM MgClJ. After preincubation, GTPyS was added at a final concentration of 100 �M, and then the mixtures were further incubated at 37 oc for 15 min. They were layered on top of 200 �1 of buffer A containing 0.25 M sucrose, and then centrifuged at 4 °C for 15 min at 15,000 rpm. The membrane pellets were suspended in 12 ml of SDS sample buffer, and then subjected to SDS-polyacrylamide electrophoresis. Membrane-bound �-COP was detected by immunoblotting with an ECL system. To remove the probe, the blot after detection was washed in phosphate-buffered saline containing 0.02% Tween 20, and then incubated with 62.5 mM Tris-HCl (pH6.7) containing 2 % SDS and 100 mM

2-mercaptoethanol at 50 oc for 30 min. It was then incubated with the antibody against man II, and then developed with an ECL system.

2.6 Detennination of the Sulfhydryl Content --- The amount of sulfhydryl groups was determined using DTNB. The purified recombinant ARF1 in KHM buffer was incubated with or without 1 mM NEM at 32 °C for 10 min. To remove excess NEM, the reaction mixture was passed through a spin column (Bio-Rad P6-DG ) equilibrated with 0.1 M Tris-HCl (pH 8.0) containing 6 M guanidine-HCl and 0.01 M EDTA. DTNB was added to the flow-through fraction, and then the absorbance at 412 ^nmwas measured with a

Beckman DU640 spectrophotometer. The molar absorption coefficient at 412 ^nmof the reaction product of DTNB in 6 M guanidine-HCl was assumed to be 13,380 M1cm·1(82).

39

(43)

3. RESULTS AND DISCUSSION

3.1 f3-COP is Released from the Golgi Apparatus by NEM-Treatment

--- In a previous study, I showed that NDGA causes disassembly of the Golgi apparatus by affecting

heterotrimeric G proteins ( 2 8 ). During the course of that study, I found that the treatment of NRK cells with 1 mM NEM at 37

^oc

for 10 min blocks NDGA-induced Golgi disassembly (data not shown), and in addition, promotes the release of �-COP, a subunit of coatomer, from the Golgi apparatus (Fig. 14C). The release of �-COP started at 5 min after the addition of NEM, and once released it did not return to the Golgi apparatus after extensive washing to remove NEM, implying that the effect of NEM is irreversible. The effect of NEM was temperature-dependent. No release of �-COP was observed when cells were incubated with NEM at 1 mM at 0 oC for 20 min. �-COP released upon NEM treatment at 37

^t

was not distributed equally throughout the cytoplasm, but rather in a scattered punctate pattern (Fig. 14C). This may suggest that the released �-COP is associated with membrane structures or accumulates in aggregates. BFA is known to cause the rapid release of �-COP from the Golgi apparatus, which is followed by redistribution of Golgi

components into the endoplasmic reticulum (6). Although BFA treatment causes �-COP ^to be distributed equally throughout the cytoplasm in most cells, �-COP accumulates in large cytoplasmic aggregates in several types of cells (83-85). A scattered punctate distribution of �-COP observed upon NEM treatment may correspond to �-COP aggregates, but I did not investigate further.

In contrast to

�-COP,

the localization of man II, a medial Golgi-resident protein, was not affected by NEM treatment (Fig. 14D), implying that the reagent does not cause

disassembly of the Golgi apparatus. In this regard the effect of NEM is different from that of BFA. However, this difference can be partly explained by the fact that NEM treatment at

37

^tcauses inactivation of not only proteins involved in the binding of

COPI

^{but also} N-ethylmaleimide-sensitive factor (NSF) , whereas BFA may affect only the former proteins. A recent study showed that NSF activity is required for Golgi disassembly

(44)

NEM at 37 oc must inactivate NSF because it is sensitive to 1 mM NEM at 0 'C (87).

To clarify the mechanism underlying the NEM-induced release of �-COP from the Golgi apparatus, I first examined the effect of taxol on this reaction. �-COP was originally discovered as a protein in the Golgi apparatus that interacts with microtubules (88) . Since microtubules are sensitive to NEM, �-COP may be released from the Golgi apparatus as a consequence of their depolymerization. To examine this possibility, NRK cells were

preincubated with 10 JJ.g/ml taxol for 1 h, to stabilize the microtubules, and then treated with NEM. Although the microtubules detected with anti-a-tubulin remained intact, �-COP was released by the NEM-treatment as observed for cells without taxol (data not shown).

3.2 Activation of Heterotrimen'c G Proteins Markedly Prevents the NEM-Induced Release of {3-COP from the Golgi Membranes --- AIF4• activates ARF by modulating heterotrimeric G proteins (74, 75), or inhibiting the GTPase activity of the ARF associated with ARF GAP (76). If the redistribution of �-COP induced by NEM is due to the

inactivation of proteins involved in the binding of CO PI, AIF4• may affect this

redistribution. When NRK cells were preincubated with AIF4• at 37 oc for 10 min, the release of �-COP was almost completely blocked upon incubation with 1 mM NEM at 37

°C for 10 min (Fig. 15C).

To confirm that this blockage occurs via GTP-binding proteins, I examined the effect of GTPyS, a nonhydrolyzable analogue of GTP, on the release of �-COP in permeabilized NRK cells. When NRK cells were permeabilized with 40 JJ.g/ml digitonin, and then incubated with 1 mM NEM in the presence of an ATP-regenerating system and cytosol at 32 °C for 30 min, �-COP was released from the Golgi apparatus (Fig. 15E). The release of

�-COP was markedly suppressed when the permeabilized cells were incubated with NEM in the presence of 10 mM GTPyS (Fig. 15F). These results suggest that NEM modifies specific protein( s) involved in the association of coatomer.

3.3 Cytosoh'c Proteins Are NEM-Sensitive --- To determine which protein(s) involved in the association of COPI with the Golgi apparatus are sensitive to NEM, I examined the

41

(45)

Fig. 14. NEM causes the release of

p-COP

from the Golgi apparatus without affecting the Golgi morphology in NRK cells. NRK

cells were

incubated without (A and B) or with (

C

and D)

1

mM NEM at

37

°C for

10

min. �-COP (A

and C) and a

medial

Golgi-resident protein, man

II

(Band D), were immunostained.

(46)

Fig. 15. Activation of GTP-binding proteins prevents the NEM-induced release of p-COP from the Golgi apparatus.

NRK cells were preincubated in the absence (A and B) or presence of AlF4- (30

mM

NaF, 50 J.!M AlCIJ (C) at 37

_oc

for 10 min, and then incubated without (A) or with 1

mM

NEM (Band C) at 37° _{C for} 10 _min.

Alternatively, NRK cells were permeabilized with 40 f.!g/ml digitonin at 0

_oc

for 5 _{min, and} then incubated at 32 ° _{C for} 30 min in the presence of an ATP-regenerating system and bovine brain cytosol without NEM (D), _with 1 mMNEM (E), or with 1

mM

NEM plus 10 J.!M GTPyS (F). �-COP was immunostained.

43

(47)

binding of COPI to the Golgi apparatus using a cytosol fraction and permeabilized cells.

^I

first treated NRK cells with 1 mM NEM at 37

^oc

for 10 min to promote the release of COPI, and then permeabilized them with digitonin (Fig. 16A and B). When the NEM-treated cells were incubated with bovine brain cytosol in the presence of an ATP-regenerating system at 32

^oC

for 20 min, a significant amount of �-COP became associated with the Golgi apparatus (Fig. 16C and D), suggesting that NEM-sensitive factor(s) are not present in membranes, but in the cytosol. To confirm this, I examined the effect of NEM treatment on the cytosol. In this case, cells without NEM treatment were permeabilized, and then preincubated in the absence of an ATP-regenerating system at 32

'C

for 20 min to allow the release of COPI from the Golgi apparatus (Fig. 17 A and B).

^As

shown in Fig. 17C and D, �-COP was recruited onto the Golgi apparatus when the cells were incubated with bovine brain cytosol in the presence of an A TP-regenerating system at

32

^oc

for 30 min. In contrast, �-COP was not recruited onto the Golgi apparatus when NEM-treated cytosol was used (Fig. 17E and F).

Similar results were obtained for Golgi membranes isolated from rat liver. When isolated Golgi membranes were incubated with an ATP-regenerating system and bovine brain cytosol in the presence of GTPyS, a significant amount of �-COP bound to the membranes (lane 2, Fig. 18). GTPyS was included in this assay because it is required for the effective binding of COPI ( 42, 89, 90). In the absence of GTPyS, the amount of �-COP associated with Golgi membranes upon incubation with ATP and.cytosol was low (lane 1, Fig. 18) and comparable to that endogenously associated with rat liver Golgi membranes (data not shown), suggesting that the recruitment of �-COP onto isolated Golgi membranes occurs little in the absence of GTPyS. Endogenous �-COP was tightly associated with isolated Golgi membranes, and not released upon incubation without ATP or cytosol (data not shown), although the reason for this is unclear at present. When Golgi membranes were pre incubated with A TP and cytosol in the presence of 1 mM NEM at 3 7

^oC

for 10 min, no significant association of �-COP with Golgi membranes was detected above the

endogenous level (lane 3, Fig. 18).

(48)

3. 4 ARF Is Sensitive to NEM

--- It is known that coatomer and ARF1 are the only cytoplasmic proteins needed for the assembly of COPI-coated vesicles (71). ARF1

promotes the binding of coatomer to the Golgi apparatus by functioning as a complex with coatomer ( 69), an activator of phospholipase D (79), or an activator that produces high affinity binding sites for coatomer (76). I therefore prepared ARF-rich and coatomer-rich fractions, and examined which fraction contains NEM-sensitive factor(s). As shown in Fig.

19, the NEM-treatment of coatomer did not affect its binding to the Golgi membranes, whereas the recruitment of coatomer did not occur when ARF was treated with NEM.

Similar results were obtained using the purified recombinant ARF1 protein (Fig. 20).

Furthermore, significant recruitment of coatomer was observed when purified ARF1 was added back to the NEM-treated cytosol (data not shown). These results imply that ARF1 is an NEM-sensitive factor among cytosolic proteins involved in the association of COPI with the Golgi apparatus.

ARF regulates not only the binding of coatomer to the Golgi apparatus but also the recruitment of the AP-1 complex onto the

trans

-Golgi network (76, 91, 92). If ARF is sensitive to NEM, it is expected that the AP-1 complex would also be released by NEM-treatment. This is indeed the case. ARF and y-adaptin, a component of the AP-1 complex, both of which are located in the Golgi region in intact cells (Fig. 21A and C), were released and dispersed throughout the cytoplasm when cells were treated with 1 mM NEM at 37

oc

for 10 min (Fig. 21B and D).

3.5 Cys-159 Is the Only Cysteine Residue in ARF

--- To confirm that ARF is modified by NEM, the sulfhydryl content of ARF was determined before and after NEM-treatment.

Before NEM-treatment, the sulfhydryl content of ARF was 0.91

+

0.04 moVmol protein.

Concomitant with NEM treatment, the sulfhydryl content decreased to 0.17

±

0.02 moVmol protein (the mean

±

SD of duplicate samples). Since the ARF family contains the only cysteine residue at position 159 (93), this result implies that NEM modified Cys-159.

Cys-159, which is located in the CAT motif in loop 110, and is conserved not only in the ARF family (94) but also in heterotrimeric GTP-binding proteins (95), participates in the

45

(49)

formation of the roof of the guanine pocket (96). This cysteine residue is replaced by a homologous amino acid, a serine residue, in small GTP-binding proteins such as Ras-p21 and Rab proteins, and this serine residue forms a hydrogen bond with the side chain of a highly conserved aspartic acid residue, which determines the specificity for guanine of GTP-binding proteins (95). The presence of Cys-159 in the guanine pocket of ARF1 explains why AIF4• and GTPyS block the release of �-COP from the Golgi apparatus induced by NEM. Probably, GDP-AIF4• or GTPyS occupies the guanine pocket so that NEM cannot gain access to the side chain of Cys-159.

哺乳動物細胞におけるゴルジ体の形態維持機構に関 する研究