THE JOURNAL OF BKXOGICAL CHEMISTRY cc) 1990 by The American Society for Biochemistry and Molecular Biology, Inc.

Vol. 265, No. 17, Issue of June 15, pp. 10109-10117, 1990 Printed m U.S. A.

Purification of Three Related Peripheral Membrane Proteins Needed for Vesicular Transport*

(Received for publication, January 19, 1990)

Douglas 0. ClarySg and James E. Rothmanj From the SDepartment of Biology, Princeton University, Princeton, New Jersey 08544 and the *Department of Biochemistry, Stanford Uniuersity, Stanford, California 94305

We report conditions under which Golgi membranes depleted of peripheral membrane proteins can be re- constituted for intra-cisternal vesicular transport. Analysis of the reconstitution reveals requirements for N-ethylmaleimide-sensitive fusion protein, a purified peripheral protein involved in the fusion stage of ve- sicular transport, as well as other peripheral protein activities which can be provided by mammalian cytosol but not yeast cytosol. The restorative activity in bovine brain cytosol is found in two broad and complementing fractions, of average native molecular masses of about 500 and 40 kDa, termed Frl and Fr2, respectively. This resolved transport system was used to develop a purification scheme for Fr2. Three proteins of appar- ent molecular masses of 35, 36, and 39 kDa (Fr2-a, -8, and -7, respectively) were found to be responsible for Fr2 activity and were purified to homogeneity. Each Fr2 protein has activity by itself in the reconsti- tuted in vitro Golgi transport assay, although each exhibits a different specific activity and plateau value. No synergy of the three Fr2 proteins was observed during mixing experiments. The three Fr2 proteins seem to be closely related based on size, in vitro ac- tivities, chromatographic properties, and peptide maps and may comprise a new family of proteins involved in vesicular transport.

Our understanding of the process of intracellular vesicular transport has been advanced by the analysis of in vitro assays which faithfully reproduce transport events (1). One of the best characterized systems reconstitutes vesicular protein trafficking between the cis and medial cisternae of the Golgi stack (2, 3). In this assay, the movement of vesicular stoma- titis virus-encoded glycoprotein (VW-G)’ from “donor” cis- ternae lacking the glycosylation enzyme N-acetylglucosamine transferase I to “acceptor” cisternae containing this enzyme is measured by the incorporation of radiolabeled N-acetylglu- cosamine. This transport reaction has several well character- ized kinetic stages which correspond to the morphological events of coated vesicle budding, transfer and targeting, un- coating, and fusion (4-6). Manipulation of the assay condi- tions has uncovered roles for acyl coenzyme A and GTP in

*This research was supported by National Institutes of Health Grant DK27044 (to J. E. R.). The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked “aduertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

’ The abbreviations used are: VSV-G, vesicular stomatitis virus- glycoprotein; CHO, Chinese hamster ovary cells; DTT, dithiothreitol; Frl, Fraction 1; Fr2, Fraction 2; GlcNAc, N-acetylglucosamine; NSF, N-ethylmaleimide-sensitive fusion protein; SDS-PAGE, sodium do- decyl sulfate-polyacrylamide gel electrophoresis.

the transport reaction (7-9) and has led to conditions under which the coated vesicle transport intermediate could be accumulated and purified (9, 10).

Not as much progress has been made in identifying and purifying the individual proteins involved in each stage of the transport process. One round of vesicular transport involves several sequential biochemical events each likely to involve multiple components. The study and purification of proteins acting during transport can be accomplished by reducing this complex series of events to an assay measuring a single component. This may be achieved in one of two ways: either the process can be reduced to a sub-reaction which involves fewer components and which can be measured independently (11, 12) or the whole reaction can be depleted of one activity, such that the transport reaction as a whole becomes the assay for the missing activity (7, 13).

An example of the latter approach involves the selective removal of a transport activity by chemical inactivation. Treatment of the membrane fractions used in the intra-Golgi assay with the alkylating agent N-ethylmaleimide causes a drastic loss of transport activity; transport was restored by addition of a supernatant prepared by incubating Golgi frac- tions with ATP and salt (7). This assay led to the purification of NSF, the N-ethylmaleimide-sensitive fusion protein, as well as an initial characterization of its function in the trans- port process (13, 14). The introduction of assays measuring yeast cytosol transport activity in homologous (15) and het- erologous (16) systems provides another method of monitor- ing a single activity through complementation of genetic defects in transport components.

The classic approach of biochemical resolution and recon- stitution of total transport activity has been difficult in this system, due to the large number of components and the relatively crude membrane fractionation scheme employed. Golgi membranes are isolated in a very gentle fashion in order to maintain them in a transport competent state (3); this leads to high concentrations of cytosol-derived transport ac- tivities associated with the membranes. Many such compo- nents can be expected to cycle between cytosol and mem- branes. This has confounded attempts to develop by fraction- ation an assay devoid of a particular transport activity; although a factor may be removed from the cytosol by frac- tionation, it is readded by addition of the membranes. Here we present conditions under which Golgi membrane fractions, stripped of many peripheral transport proteins by a high salt wash, can be reconstituted for transport using four cytosolic fractions. Analysis of one of these cytosolic pools, called fraction 2 (Fr2), reveals a novel set of proteins involved in intracellular transport.

10109

by guest on August 24, 2020

http://ww

w.jbc.org/

Dow

nloaded from

I Purification of a Family of Vesicular Transport Proteins

K Golgl K Golgl + NSF

FIG. 1. Yeast cytosol is defective in reconstituting intra- Golgi transport with KCl-treated membranes. Yeast cytosol (31.5 pg, bars 2, 5, and S), CHO cytosol (18 pg, bars 3, 6, and 9), or buffer (bars 1, 4, and 7) were tested for their ability to reconstitute transport with KC1 treated donor and acceptor Golgi (K Golgi) mem- branes (bars 13), K Golgi supplemented with 5 ng of purified CHO NSF (bars 4-6) or untreated Golgi membranes (bars 7-9). Incubations were performed under the standard K Golgi conditions; transport was measured by the incorporation of tritiated GlcNAc into VSV-G protein.

MATERIALS AND METHODS’

RESULTS

Yeast Cytosol Is Defective in Reconstituting Vesicular Trans- port with KCL-treated Golgi (K Golgi) Membranes-We have modified the standard inter-cisternal Golgi transport assay to examine proteins peripherally associated with the membrane fraction. Treatment of the membranes with 1 M KC1 (37 “C for 10 min, termed K Golgi) leads to a drastic loss of activity when assayed with CHO cytosol (Fig. 1, compare bars 3 and 9). Addition of purified NSF, a membrane-associated trans- port protein which has been inactivated in standard CHO cytosol by depletion of ATP (13), restores activity to these membranes when assayed with CHO cytosol (Fig. 1, bar 6). Thus, it appears that any peripherally located transport com- ponents that are removed or inactivated by salt treatment can be restored to the Golgi membranes by cytosolic comple- mentation. Yeast cytosol has been shown to substitute for CHO cytosol in the standard transport assay using untreated membrane fractions (Ref. 16 and Fig. 1, bar 8). Yeast cytosol assayed with K Golgi membranes apparently lacks activity either in the presence or absence of purified NSF (bars 2 and 5). The most likely explanation is that one or more peripheral membrane transport proteins removed by the salt wash can- not contributed by yeast cytosol. These activities, missing from yeast cytosol but present in CHO cytosol, could also be contributed by the CHO donor membrane fraction, the CHO acceptor membrane fraction, or by mammalian cytosolic frac- tions and could not be attributed to inhibition by the yeast cytosol based on mixing experiments (data not shown).

Yeast cytosol can provide all of the soluble (Ref. 16 and Fig. 1) but apparently not all of the peripheral protein com- ponents needed to reconstitute Golgi transport with animal cell membranes. This provides an opportunity to use yeast cytosol as a “biochemical mutant” to develop a complemen- tation assay for the activities that are lacking. Specifically, incubation of K Golgi membranes with yeast cytosol and

’ Portions of this paper (including “Materials and Methods” and Figs. 10-13) are presented in miniprint at the end of this paper. Miniprint is easily read with the aid of a standard magnifying glass. Full size photocopies are included in the microfilm edition of the

purified NSF provides an assay that will reveal only those components present in animal cell cytosol not provided in a functional form by yeast cytosol (other than NSF).

The activities in bovine brain cytosol that complement the yeast cytosol/K Golgi/NSF assay were examined by gel per- meation chromatography on Sephacryl S-300. When column fractions were assayed individually, little activity was seen (Fig. 2, filled squares). However, when fraction 21 was mixed across the column profile, a broad peak of activity appeared, eluting with an average native molecular mass of 40 kDa (filled triangles). Mixing fraction 28 across the profile revealed the complementary peak with an average native molecular mass of 500 kDa (#led circles). Thus, at least two mammalian membrane-associated transport activities (other than NSF) are needed for transport with K Golgi membranes and are not provided by yeast cytosol. These two peaks are designated fraction 1 (Frl) and fraction 2 (Fr2).

In sum, the transport activity of K Golgi membranes can be reconstituted by the addition of four resolved cytosol- derived pools: yeast cytosol, brain cytosol Frl, brain cytosol Fr2, and NSF. Each of these pools is required, as omission of any one of the four leads to a drastic reduction of overall transport activity (Fig. 3A, bars 2-5). This provides an oppor- tunity to analyze the components in each pool which contrib- ute to the transport process. We decided to focus on the activity present in the Fr2 pool by using the K Golgi transport system as a Fr2 activity assay. Fig. 3B shows that transport is linearly dependent on addition of Fr2 when it has been omitted. The assay reaches a plateau equal to about 80% of that reached by saturating bovine brain cytosol (not shown).

Purification of a 35kDa Protein Providing Fr2 Activity-A purification protocol was developed for the transport activity

Fraction number

FIG. 2. Sephacryl S-300 chromatography reveals two com- plementing fractions which restore transport activity to the K Golgi/yeast cytosol assay. A 20-ml Sephacryl S-300 column equilibrated in 50 KTD was loaded with 250 ~1 (3 mg) of bovine brain cytosol; 0.4.ml fractions were collected. Four ~1 of each fraction were tested for activity alone (squares), mixed with fraction 21 (triangles), or with fraction 28 (circles) in incubations containing 21 pg of yeast cytosol, 2.2 rg of K Golgi (donor and acceptor mixed), and 5 ng of CHO NSF under standard K Golgi assay conditions. Fractions 21 and 28 are indicated with arrows. Transport was measured by the incor- poration of tritiated GlcNAc into VSV-G protein. Background counts/min subtracted were 145, 164, and 265, respectively. Protein concentration (open squares) was determined with the Bio-Rad pro- tein assay, using bovine serum albumin as a standard and a correction

Journal that is available from Waverly Press. factor (actual concentration = 2.5 X measured concentration).

by guest on August 24, 2020

http://ww

w.jbc.org/

Dow

nloaded from

Purification of a Family of Vesicular Transport Proteins

B.

FIG. 3. Development of an assay for Fraction 2 (Fr2) activ- ity. A, transport activity in the reconstituted K Golgi assay is dependent on four cvtosol-derived protein f’ractions. The complete reaction (bar I) contains 2.2 pg K Golgi (mixed donor and acceptor), 21 pg of yeast cytosol, 5 ng of CHO NSF, 23 fig of fraction 1 (Frl) pool, and 5 ~1 of a partially purified Fr2 pool. Reactions in which Frl (bar 2). Fr2 (bar 3), yeast cytosol (bar 4), or CHO NSF (bar 5) is omitted are also shown. Reactions were performed using the standard K Golgi assay conditions, and transport was measured by incorpora- tion of tritiated GlcNAc into VSV-G protein. B, the K Golgi assay is linearly dependent on addition of the Fr2 pool. Increasing amounts of a partially purified Fr2 pool were added to K Golgi reactions lacking Fr2 using conditions described in A. Background subtracted for A and B was 95 cpm, the result of a complete reaction incubated on ice.

in Fr2 from bovine brain, described in detail under “Materials and Methods.” The initial step utilized the intrinsic affinity of this activity for membranes. It was found that extracting a crude bovine brain membrane fraction with 1 M KC1 (4 “C) yielded a homogenate with 2-3-fold higher specific activity than that of standard bovine brain cytosol. This homogenate was subjected to a 30-50% ammonium sulfate precipitation step, followed by chromatography on DEAE-cellulose. Fr2 activity was found to elute from this column in a single peak centered at 210 mM KC1 (Fig. 10). The peak was concentrated by ammonium sulfate precipitation, dialyzed into buffer con- taining 100 mM KCl, and applied to a column of Reactive Red 120 coupled to Sepharose CL-6B equilibrated in the same buffer; the column was then washed with a buffer containing 200 mM KCl. The activity eluted in a broad peak after the flowthrough peak of protein (Fig. 11). These fractions were pooled and loaded onto a phenyl-Sepharose column, equili- brated in the 200 mM KC1 buffer. The column was washed with 10% ethylene glycol and eluted with a 10-75s ethylene glycol gradient. Fr2 activity eluted very broadly (Fig. 4A), appearing in almost all fractions of the gradient elution; by contrast, most polypeptides were resolved by this procedure (Fig. 4B).

We focused on the fractions comprising the early part of the ethylene glycol gradient, because they had higher specific activity and lower polypeptide complexity. Fractions 3-18 were pooled (termed the N pool) and subjected to chromatog- raphy on Mono Q using potassium phosphate buffer (pH 7.0). A single peak of activity eluted at 250 mM KC1 (Fig. 12A). This peak was applied to a Superose 12 column and eluted with an apparent native molecular mass of 35-40 kDa (Fig. 13). At this stage in the purification, the predominant protein species had a relative molecular mass of 35 kDa on an SDS gel and coeluted with the Fr2 activity (not shown). The peak of activity was pooled and subjected to two rounds of chro-

A 2 a -- pods B Y

0 10 20 30 40

Fraction number

B Fraction number

---=w _ H *,l4 --. “‘ _- ----ww-- -- -43

- 31

--

--- a pool P Pool Y Pool

FIG. 4. Chromatography of Fr2 activity on phenyl-Sepha- rose. The Fr2 pool from the Reactive Red 120-Sepharose column (95 ml, 36 mg of protein) was loaded onto a 15 x l&cm phenyl-Sepharose column equilibrated in 200 KTDG. The flow rate was 33 ml/h. After loading the protein pool, the column was washed with 1 column volume of 50 KTD + 10% ethylene glycol. Four-ml fractions were collected starting at the 11th ml of the wash. The column was developed with a lo-75% ethylene glycol gradient. A, the fractions were analyzed for Fr2 activity (filled c&es) and protein concentration (AL”,); solid line). Background activity in the absence of Fr2 (869 cpm) has been subtracted. Ethylene glycol concentration is shown as a dotted line. The black bars labeled U, 8, and y refer to pools used in succeeding purification steps (see text). B, SDS gel electrophoresis of the protein population in the phenyl-Sepharose eluate fractions. Fraction numbers are shown at the top; 30 ~1 of each fraction was electrophoresed on a 10% SDS-PAGE gel and stained with Coomassie Blue R-250. The relative molecular weights of the marker proteins are shown at left: phosphorylase b (97,000), bovine serum albumin (66,000), ovalbumin (43,000). and carbonic anhydrase (31,000). Bars at the bottom refer to the CY, 8, and y pools used in succeeding chromatographic steps. The positions of the 35., 36-, and 39 kDa proteins are marked (see text).

matography on Mono Q using Tris buffer (pH 7.8); the activity peak eluted at 250 mM KC1 and corresponded to the major Azxo peak as well as the 35-kDa band (Fig. 6, lane 1). Mono Q chromatography using both Tris (pH 7.8) and phosphate (pH 7.0) conditions was essential for the purification; although the 35-kDa polypeptide elutes at the same position in the KC1 gradient, contaminating polypeptides change their elution positions and are thus removed (see next section). Fig. 5 shows aliquots of the pools from each step of the purification. A summary of the purification is shown in Table I. The Fr2

by guest on August 24, 2020

http://ww

w.jbc.org/

Dow

nloaded from

10112 Purification of a Family of Vesicular Transport Proteins

activity was enriched 400-fold over the starting membrane KC1 extract, which corresponds to approximately 1000-fold over standard bovine brain cytosol. This is likely to be an underestimate, given the low yield of activity (2%). The protein yield was 144 pg from a starting extract containing 3.1 g of protein.

Three Protein Species Have Fr2 Activity-Reexamination of the protein profile of the phenyl-Sepharose column (Fig. 4B) shows that the 35-kDa protein is present only in the first

-21.5

-14.4

12 3 4 5 6 7 6

FIG. 5. SDS gel electrophoresis of Fr2-a purification pools.

-97

-65

-43

The protein pools from Fr2-n purification were analyzed on a 12.5% SDS-PAGE gel in the following amounts and stained with Coomassie Blue R-250: Innc I. membrane KC1 extract, 94 pg; lane 2, 30-50% ammonium sulfate cut, 92 rg; lane 3, DEAE-cellulose pool, 26 pg; (ant 4, Reactive Red 120.Sepharose, 15 pg; lane 5, phenyl-Sepharose pool. 4.4 pg; lane 6‘, Mono Q (phosphate) pool, 3.0 pg; lane 7, Superose 12 pool, 2.1 ~8; Iane 8, Mono Q (Tris) pool, 1.8 pg. The molecular weight markers shown at r&t are those described in Fig. 4R, together with soybean trypsin inhibitor (21,500) and lysozyme (14,400). The nrrou indicates the position of the 35.kDa band.

part of the elution profile; yet activity is present throughout the elution, likely due to other active Fr2 species. The Mono Q and Superose 12 purification steps described above were repeated with the p and y pools shown in Fig. 4, A and B.

The Fr2 activity derived from the p pool eluted from the Mono Q (phosphate) column (Fig. 12B) and Superose 12 column in a similar manner to that described above for the cy pool. However, the activity eluted from the Mono Q (Tris) step at a slightly higher KC1 concentration (260 versus 250 mM). The species coeluting with this activity has an apparent molecular mass of 36 kDa (Fig. 6, lane 2), and corresponds to the most persistent contaminant in the Fr2-a purification (Fig. 5, compare lanes 6 and 7 to lane 8). This explains the observation that removal of this contaminant led to almost no increase in specific activity (Table I).

The Fr2 activity present in the y pool eluted from the Mono Q (phosphate) column as two peaks (Fig. l2C). The first peak eluted at about 250 mM KC1 and corresponded to the 36-kDa protein (not shown). The second broad peak of Fr2 activity unique to the y pool was centered at 270 mM KCl, a higher concentration than that seen for the 35 or 36-kDa proteins. The 270 mM KC1 pool of FrS-y activity chromatographed as a 35-40 kDa protein on Superose 12 and was then applied to the Mono Q (Tris) column, eluting at 270 mM KCl. The purified Fr2 activity from the y pool consisted of a single major protein with an apparent molecular mass of 39 kDa (Fig. 6, lane 3).

Each of the Fr2 proteins has a different retention time when chromatographed on the Mono Q column in Tris buffer (pH 7.8). This procedure was used to confirm that each purified protein species coelutes with Fr2 activity. The LY, fl, and y pools from the last step of purification were chromat- ographed on Mono Q in Tris buffer and each assayed for activity (Fig. 7). The peak of Fr2 activity from each pool eluted from the column at a unique position. Samples from the peak fractions were analyzed by SDS gel electrophoresis (not shown). In each case, the peak of activity coeluted with the protein species (e.g. 35,36, or 39 kDa) present in the pool. Trace amounts of Fr2-0 contaminating Fr2-oc and Fr2-a con- taminating Fr2-fl eluted at the proper retention times and serve as an internal control that the Mono Q column was separating these proteins reproducibly. Because the activity peaks coelute with the protein peaks and have different reten- tion times, another protein contaminating all three pools could not be responsible for their activities.

Although any one of the three purified Fr2 proteins could reconstitute the transport activity, their specific activities and plateau values varied. Fig. 8 shows a titration of the three proteins into the Fr2-dependent transport assay. FrB-cu has

TABLE I

Purificatwn of Transport FrZ-tu

Fractmn Total protem Total actlwty”

SpC!dk actwty Yield

ml mb’ units unitsjmg % Membrane salt extract 1660 3100 5.18 x IO” 1.67 x lo:’ (1) (100) N-hO% (NH.,),SO., 600 1320 5.71 x 10” 4.34 x lo:’ 2.6 110 DEAE-cellulose 136 144 2.53 x 10” 1.76 x 10’ 10.5 49 Red agarose 9.5 36.1 1.96 x 10” 5.43 x 10’ 32.5 38 Phenyl-sepharose 6’2 6.77 6.00 x lo” 8.87 x lo” 53.1 12 Mono Q (phosphate) 2.t i 0.938 4.47 x lo” 4.76 x lo” 285 8.6 Superose 12 6 0.642 4.08 x lo” 6.36 x lo” 381 7.9 Mono Q (Tris) 2 0.144 9.93 x 10” 6.90 x 10” 413 1.9

” The activity in each pool was determined by assaying dilutions of the pool in the Fr2 transport assay such that the activity value fell into the initial linear portion of the Fr2 titration curve. Units are defined as 1000 cpm of [“H]GlcNAc incorporated into VSV-G protein. All activity determinations were done in the same experiment in duplicate.

by guest on August 24, 2020

http://ww

w.jbc.org/

Dow

nloaded from

Purification of a Family of Vesicular Transport Proteins

a e Y

hl,x10-3

97 -

SO-

43-

- .-

31 -

21.5 -

:. -39

1 2 3 FIG. 6. SDS-PAGE analysis of the three Fr2 protein species.

1.0 pg of each Fr2 protein was subjected to electrophoresis on a 10% SDS-PAGE gel and stained with Coomassie Blue R-250. Lane 1 contains Fr2-cu (35 kDa), lane 2 contains Fr2-/3 (36 kDa), and lane 3 contains Fr2-y (39 kDa). Molecular weight markers are as in Figs. 4B and 5.

0 FrZ-a

A Fr2- Y

37 39 41 43 45 47 Fraction number

FIG. 7. Each of the Fr2-a, +, and -y proteins has Fr2 activ- ity and elutes at a unique position. The FrZ-cu, -p, and -y activity pools after the first Mono Q (Tris) chromatography step (see text) were diluted 2-fold and each reapplied to the Mono Q column. The load was 60, 140, and 120 pg, respectively. In each case the column was developed with a 30-ml 50-400 mM KTDG gradient; 0.5-ml fractions were collected. The fractions were assayed for Fr2 activity (0.07, 0.03, and 0.07 ~1, respectively) using the standard Fr2 assay. Transport is measured by incorporation of tritiated GlcNAc into VSV-G protein. Background activity in the absence of added Fr2 (610 cpm) has been subtracted.

the highest specific activity and plateau value, whereas Fr2-0 is somewhat lower and Fr2-7 is much less active. Assays containing up to 100 ng of either Fr2-/3 or FrS-7 did not attain the plateau of Fr2-oc (not shown). In another experiment, the three Fr2 protein species from two different purifications were assayed for specific activity (Table II). Although the two preparations differed in specific activity, the ratios of specific activities (e.g. c&y) of the Fr2 proteins were quite similar.

FIG. 8. Titration of the Fr2 proteins in the Fr2-dependent transport assay. The Fr2 proteins were titrated from 0 to 20 ng in the FrB-dependent transport assay. Background activity in the ab- sence of added Fr2 protein (685 cpm) has been subtracted. Also shown is the titration of an equal weight mixture of FrZ-n, -8, and -y assayed at the same total concentration of added protein. This experiment was performed twice; the average is shown.

TABLE II Relative specific activities of Fr2 preparations

Specific activitf

Ratio of specific

activities

urutslml: X IO-*’

Preparation 1 FrB-tu 1.25 (1) Fr2-8 0.55 0.44 FrZ-7 0.35 0.28

Preparation 2 FrB-cu 1.75 (1) Fr2-8 0.98 0.56 Fl-3--V 0.54 0.31

* The Fr2 proteins were titrated in the Fr2 transport assay, and the specific activity was derived from the initial rate of the reaction. Units are defined as 1000 cpm of [“H]GlcNAc incorporated into VSV- G. All titration points were determined in duplicate and measured in the same assay.

The three Fr2 proteins might form a hetero-complex during the transport reaction, and their individual activities would then be the result of complementation with Fr2 proteins left on the Golgi membranes after extraction. If so, then mixing should yield a higher specific activity and/or plateau value than that observed when assaying Fr2+, -p, or -y singly. Fig. 8 shows that a mixture of the three proteins acts approxi- mately as an average of their individual activities. Also, ad- dition of Fr2-P and/or Fr2--y to saturating concentrations of Fr2-a did not raise the plateau value (not shown). The lower activity of the Fr2 protein mixture compared with Fr2-a is not due to an artifact of purification, as a crude pool of Fr2 activity prepared from bovine brain cytosol by a single anion exchange chromatography step reached an activity plateau which was slightly lower than that seen with purified Fr2-oc protein (not shown), which is identical to the behavior of the Fr2 protein mixture. The assay data therefore do not suggest that more than one of these proteins is needed for intra-Golgi transport, although that possibility cannot be ruled out (see “Discussion”).

The three proteins seem to be related to each other based on similar in vitro activities, molecular weights, and chromat- ographic properties. To gain some information on the relat- edness of their polypeptide sequences, they were subjected to partial proteolysis mapping. Fig. 9 shows the results of partial digestion of the three purified Fr2 proteins with Staphylococ-

by guest on August 24, 2020

http://ww

w.jbc.org/

Dow

nloaded from

10114 Purification of a Family of Vesicular Transport Proteins

time 45' 135

Fr2 -uPvaPuaPu

va +-- -++++++

12345676910

FIG. 9. Partial proteolysis mapping of Fr2-a, $3, and -y proteins. Partial proteolytic digests of the three Fr2 proteins were obtained by the method of Cleveland (19) and electrophoresed on a 15% SDS-PAGE gel. 300 ng of each protein was digested with 60 ng of S. a~rc~s V8 protease for 45 min (lanes 5-7) or 135 min (lanes 8- 10) in solution, boiled, and loaded on the gel. Lanes 2-4 contain undigested protein, and lane 1 contains the protease alone. FrB-n is shown in lanes 2, 5, and 8; FrP-0 is shown in lanes 3, 6, and 9, FrZ-y is shown in lanes 4, 7, and 10. Protein and peptide bands were visualized by silver staining. The relative molecular weight of the marker proteins is shown on the right: carbonic anhydrase (31,000), soybean trypsin inhibitor (21,500), and lysozyme (14,400).

cus aureus V8 protease. Although there is some similarity in the peptides produced from FrB-cu and Fr2+, there is no obvious similarity in the pattern of FrZ-y with the other two. Therefore, although Fr2-cu and Fr2-P might be related to each other by post-translational modification, it is unlikely that Fr2-cu or Fr2-/3 are proteolytic products of Fr2-7. Treatment of the three polypeptides with an equal weight of bacterial alkaline phosphatase for 4 h did not change their gel mobility, which suggests that the mobility difference between FrZ-Lu and Fr2-P is not due to phosphorylation (data not shown).

DISCUSSION

A molecular understanding of the events which occur during the budding, targeting, and fusion of inter-cisternal transport vesicles requires reconstitution of the process using purified components. Here we introduce an assay which uses KCl- extracted Golgi membranes (K Golgi) and four cytosolic pools to reconstitute vesicular transport. The required soluble pro- tein fractions in this assay are yeast cytosol, brain cytosol Frl and Fr2, and purified NSF. This assay was used to purify the active species in brain cytosol Fr2 pool. Three distinct pro- teins with similar biochemical properties were purified, each of which can restore transport activity to reactions from which the Fr2 pool was omitted.

Transport Assay with Salt-washed Golgi-One of the diffi- culties in devising fractionation schemes for Golgi transport components seems to have been the varying presence of the same transport activities in both the cytosolic and membrane fractions, frustrating many attempts to produce an assay dependent upon one transport component. Evidence of this was obtained when whole cytosol was fractionated on DEAE- cellulose and assayed for activity (11, 16). Contrary to expec- tation, many fractions had considerable activity when assayed individually, although the set of cytosolic transport compo- nents had been well resolved. Much of this can be attributed to cross-complementation of an individual cytosolic transport factor by a mixture of the others, present on Golgi membranes in significant (and likely variable) amounts. We present here conditions under which the Golgi membrane fractions can be stripped of many peripherally associated proteins and recon-

stituted for transport using whole and fractionated cytosolic preparations. These assay conditions have confirmed require- ments for the transport protein NSF and led as well to the discovery of at least two other membrane associated proteins not detected previously using standard membrane fractions.

The existence of a transport protein in both soluble and membrane-associated forms is likely a consequence of cycling between these states during rounds of activity. The transport factor NSF can be isolated in a membrane-bound form and released to a soluble form upon incubation with ATP and salt (7, 13); purified NSF rebinds to membranes in a specific fashion (12). As another example, signal recognition particle, which is required for protein translocation across the endo- plasmic reticulum, was purified on the basis of complemen- tation of salt extracted microsomes in a translocation assay (17). Fr2 activities, which are also present in the cytosol fraction, require relatively harsh conditions (1 M KC1 at 37 “C) for removal from membranes. Frl activity is also extracted from membranes by salt, although treatment of the membrane fraction with 500 mM KC1 at 0 “C is sufficient to remove this activity (not shown). In principle, such KC1 treatments could either release peripheral proteins from the membranes or inactivate them in situ. Evidence that Fr2 activity is released by KC1 treatment comes from cytosol preparation experi- ments; when bovine brain tissue is used as a source of activity, the specific activity of Fr2 in the resulting cytosol increased 3.1-fold when the homogenate was prepared in 1 M KC1 compared with 0 M KC1 (not shown). The overall activity of bovine brain cytosol assayed with K Golgi membranes is also much higher when it is prepared using high salt versus low salt conditions.

When substituted for mammalian cytosol in assays with unextracted Golgi membranes, yeast cytosol has significant transport activity, yielding plateau values up to 70% of those reached by CHO cytosol (Ref. 16 and Fig. 1, bar 8). However, when salt-washed CHO Golgi are assayed with yeast cytosol, transport does not occur unless additional cytosolic transport activities (NSF, Frl, and Fr2) are provided from an animal cell or tissue source, such as CHO cells or bovine brain. This may mean that some of the transport components in yeast are incompatible with what should be the complementary set from animal cells. Yeast and mammalian cells could have essentially identical transport mechanisms, but incompatibil- ity might express itself when protein complexes are required to form from heterologous components; an aberrant complex would be most likely to occur at the membranes, where the mammalian membrane proteins are interacting with yeast peripheral transport factors. If such incompatibilities exist, they must be few, as only a handful of animal components are required to restore transport, a fact which we have used to identify them. Alternatively, the absence of a function in yeast cytosol may be due to the lack of the relevant component in the soluble fraction from which the yeast cytosol is pre- pared, resulting from instability or a principle association with the yeast membrane fraction under the conditions of cell breakage. Functional assays as well as DNA homology have shown the yeast homolog of mammalian NSF to be SEC18p (18); given the striking similarities in sequence of this fusion protein in yeast and animals, it is reasonable to assume that yeast have functions homologous to Frl and Fr2 activities. Assays to reveal Frl or Fr2 activities in yeast cytosol may require a change in conditions or use of a homologous yeast transport assay.

Fr2 Transport Proteins-The purification of the Fr2 activ- ity pool found in bovine brain cytosol led to the surprising finding that at least three protein species were responsible for

by guest on August 24, 2020

http://ww

w.jbc.org/

Dow

nloaded from

Purification of a Family of Vesicular Transport Proteins 10115

the activity. They appear to be related to each other, based on similar molecular weights, chromatographic properties and in uitro activities. Fr2-a and Fr2-P also share some similarity in their protease digestion products. FrB--y has a different protease digestion pattern, as well as a much lower specific activity. The similarity of Fr2-a and Fr2-P could be explained by some sort of post-translational modification, such as pro- teolysis (although the higher specific activity of the smaller protein might be considered unusual), but it is very unlikely that the larger FrB-7 protein could be related in the same fashion. It seems more likely that the three species constitute a family of proteins involved in intracellular transport. How- ever, proof that each of these proteins has endogenous Fr2 activity awaits the development of antibodies or cDNA expression systems which can directly link the polypeptides to the in uitro activities.

The Fr2 requirement can apparently be filled by any of at least three different protein species, FrB-a, -& and -7, albeit to different plateau values. The three proteins might be func- tioning by substituting for each other at a certain transport step. Alternately, two or three of them might function to- gether; for example, they might form a complex. If so, the activity of an individual Fr2 protein would probably be acti- vated by addition of the other proteins. The three Fr2 species were examined alone and in combination for specific activity (Fig. 8) and for plateau value (not shown) in the FrB-depend- ent transport assay, and such a response was not apparent. A mixture of the three Fr2 proteins behaved as an average of the three assayed separately, and the plateau value of FrB-ol did not increase after addition of Fr2-P and -7. Thus, there is no evidence to suggest that more than one of these proteins at a time is needed in the process of intra-Golgi vesicular transport. However, there is a major caveat to this conclusion. In the absence of antibodies or other probes for the Fr2 polypeptides, we cannot be certain to what extent the mem- branes or Frl might be contaminated with endogenous Fr2. Even small amounts of cross-contamination between the Fr2 protein preparations may be enough to eliminate the synergy between the Fr2 proteins to be expected if the Fr2 proteins acted in concert. Also, it is possible that yeast cytosol could contribute a homolog of Fr2 which participates in complex formation, but would itself be insufficient to reconstitute the entire transport reaction.

Development of a fractionated Golgi transport assay has allowed the identification of a family of transport-related proteins. Using activity assays and reagents derived from

these pure proteins, we hope to elucidate their role in vesicle formation, targeting, or fusion and to understand their actions at the molecular level.

Acknowledgments-We would like to thank M. Block for advice on the purification of NSF and Lyne Piquet and Barbara Devlin for excellent technical assistance.

REFERENCES

1. 2.

3.

4.

5.

6.

7. 8.

9.

10.

11.

12.

13.

14.

15.

16.

17.

18.

19.

20.

21. 22.

Goda, Y., and Pfeffer, S. R. (1989) FASEB J. 3, 2488-2495 Fries, E., and Rothman, J. E. (1980) Proc. N&l. Acad. Sci. U. S.

A. 77, 3870-3874 Balch, W. E., Dunphy, W. G., Braell, W. A., and Rothman, J. E.

(1984) Cell 39.405-416 Balch, W. E., G&k, B. S., and Rothman, J. E. (1984) Cell 39,

525-536 Wattenberg, B. W., Balch, W. E., and Rothman, J. E. (1986) J.

Biol. Chem. 261,2202-2207 Orci, L., Malhotra, V., Amherdt, M., Seratini, T., and Rothman,

J. E. (1989) Cell 56, 357-368 Glick, B. S., and Rothman, J. E. (1987) Nature 326,309-312 Pfanner, N., Orci, L., Glick, B. S., Amherdt, M., Arden, S. R.,

Malhotra, V., and Rothman, J. E. (1989) Cell, 95-102 MelanCon, P., Glick, B. S., Malhotra, V., Weidman, P. J., Serafini,

T., Gleason, M. L., Orci, L., and Rothman, J. E. (1987) Cell 51, 1053-1062

Malhotra, V., Serafini, T., Orci, L., Shepherd, J. C., and Rothman, J. E. (1989) Cell 58, 329-336

Wattenberg, B. W., and Rothman, J. E. (1986) J. Biol. Chem. 261,2208-2213

Weidman, P. J.. Melancon, P., Block. M. R., and Rothman, J. E. (1989) i. Cell’Biol. 1&,‘1589-1596

Block. M. R.. Glick. B. S.. Wilcox. C. A.. Wieland. F. T.. and Rothman, j. E. (1988) Pk. Natl. kad. f&i. U. S. A. 85, i852- 7856

Malhotra, V., Orci, L., Glick, B. S., Block, M. R., and Rothman, J. E. (1988) Cell 54, 221-227

Baker, D., Hicke. L.. Rexach. M.. Schleter. M.. and Schekman. R. (i988) Cell 54,335-344

Dunphy, W. G., Pfeffer, S. R., Clary, D. O., Wattenberg, B. W., Glick, B. S., and Rothman, J. E. (1986) Proc. Natl. Acad. Sci. U. S. A. 83,1622-1626

Walter, P., and Blobel, G. (1980) Proc. Natl. Acad. Sci. U. S. A. 77,7X-7116

Wilson. D. W.. Wilcox. C. A.. Flvnn. G. C.. Chen. E.. Kuana. W.- J., Henxel, W. J., Block, Ml R:, Uilrich, A., and RdthmanyJ. E. (1989) Nature 339,355-359

Cleveland, D. W., Fischer, S. G.. Kirschner, M. W.. and Laemmli. U. K. (1977) J.’ Biol. Chem. 2j2, 1102-1106

Scopes, R. K. (1987) in Protein Purification, 2nd Ed., pp. 146- 147, Springer-Verlag, New York

Laemmli. U. K. (1970) Nature 227.680-685 Merril, C. R., Goldman, D., Sedman, S. A., and Ebert, M. H.

(1981) Science 211, 1437-1438

Continued on next page.

by guest on August 24, 2020

http://ww

w.jbc.org/

Dow

nloaded from

10116

by guest on August 24, 2020

http://ww

w.jbc.org/

Dow

nloaded from

Purification of a Family of Vesicular Transport Proteins 10117

4

by guest on August 24, 2020

http://ww

w.jbc.org/

Dow

nloaded from

D O Clary and J E Rothmantransport.

Purification of three related peripheral membrane proteins needed for vesicular

1990, 265:10109-10117.J. Biol. Chem. 

  http://www.jbc.org/content/265/17/10109Access 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/265/17/10109.full.html#ref-list-1

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

by guest on August 24, 2020

http://ww

w.jbc.org/

Dow

nloaded from