Isolation and Characterization of a Novel Low Molecular Weight Protein Involved in Intra-Golgi Traffic*

Analysis of the cytosolic requirements for in vitro in-tra-Golgi transport led to the characterization of three proteins: N -ethylmaleimide-sensitive fusion protein (NSF), soluble NSF attachment protein (SNAP), and p115, all involved in the docking and fusion of transport vesicles to their target membranes. In the course of de-termining the minimal cytosolic requirements for intra-Golgi transport in vitro , we identified three additional factors that are sufficient to replace crude cytosol. We describe here the purification and characterization of one of these factors, a novel 16-kDa protein, p16, an essential factor for intra-Golgi protein transport. Based on transport activity, this purification procedure resulted in ; 1,400-fold enrichment of p16 to apparent ho-mogeneity. The activity of p16 could be observed in the absence of vesicle formation, suggesting that it may par-ticipate in the docking and fusion processes. Transport of proteins between membrane-bound organelles in eukaryotic cells is a multistage process utilizing soluble and membrane The molecular machinery mediating this process has been explored biochemically the identification and characterization of numerous transport factors (1–3).

Transport of proteins between membrane-bound organelles in eukaryotic cells is a multistage process utilizing soluble and membrane proteins. The molecular machinery mediating this process has been explored biochemically and genetically, leading to the identification and characterization of numerous transport factors (1)(2)(3).
p115 was isolated as a cytosolic factor required for intra-Golgi transport in vitro (6) and was suggested to act together with NSF and SNAP in direct Golgi-Golgi fusion (17). p115 is a peripheral membrane protein localized predominantly in the Golgi apparatus (6) but has also been identified as a component of transcytotic vesicles (15). Recently, p115 has been impli-cated, together with NSF and SNAP, in the process of reassembly of post-mitotic Golgi fragments into Golgi cisternae (18,19). Uso1p, the yeast homolog of p115, is required for assembly of the endoplasmic reticulum-Golgi SNARE complex (20). Other cytosolic factors such as Rab proteins and their effectors were also shown to be involved in this process (21)(22)(23). It appears, however, that the amount of Rab proteins present on the membrane is sufficient to promote the transport reaction in vitro.
It was demonstrated originally by Clary and Rothman (24) that in addition to NSF and the SNAPs, several other cytosolic factors were required for reconstituting the SNAP-dependent transport assay. The need for additional cytosolic transport factors was demonstrated further for the p115-dependent assay (6) as well as for direct fusion between Golgi stacks (17). Thus, identification and isolation of these novel factors are essential for understanding the exact molecular machinery of intracellular protein traffic. In the present study we describe the purification of a novel 16-kDa transport factor, p16, from bovine brain on the basis of the in vitro intra-Golgi transport assay. Our data suggest that p16 is a transport factor that participates in the docking/fusion reaction.

General Procedures
Vesicular stomatitis virus (VSV)-G protein-containing donor Golgi membranes from 15B cells and acceptor membranes from wild-type Chinese hamster ovary cells were prepared as described (25). Protein concentration was determined with the Bio-Rad protein assay. The pH values were determined at room temperature. All fractionations were performed at 4°C. All cytosolic fractions tested in this assay were dialyzed to 25 mM Tris-HCl, pH 7.4, 50 mM KCl, and 1 mM dithiothreitol (dialysis buffer) before their addition to the transport assay.

Cis-to Medial-Golgi Transport Assay
The standard assay mixture (25 l) contained 0.4 Ci of UDP-N-[ 3 H]acetylglucosamine (America Radiolabeled Chemical), 5 l of a 1:1 mixture of donor and acceptor Chinese hamster ovary cellular Golgi membrane, and crude bovine brain cytosol as described (25).

Preparation of Cytosolic Factors
Bovine brain cytosol was prepared by the method of Malhotra et al. (26). Recombinant His 6 -NSF and His 6 -␣SNAP were prepared as described (27). Fraction I␤ was obtained by chromatography of 500 ml of 40% ammonium sulfate precipitate on a 400-ml Fast Flow Q column (Pharmacia Biotech Inc.) equilibrated with 25 mM Tris-HCl, pH 7.4, 100 mM KCl, 10 mM ␤-mercaptoethanol, and 10% glycerol. The column was washed with 800 ml of the same buffer and then eluted with a 0.1-0.5 M KCl gradient in 1,200 ml. Nine-ml fractions were collected, dialyzed to reduce the KCl concentration to 50 mM, and assayed for transport activity in the presence of 0.5 g of p115, 5 ng of His 6 -NSF, 60 ng of His 6 -␣SNAP, and 5 l of Golgi membranes. Typically, a peak of transport activity was eluted at 0.22-0.30 M KCl, and the peak fractions were pooled and concentrated by ultrafiltration using Amicon PM-10 filter. The concentrated pool, designated ␤, was dialyzed against dialysis buffer and had a protein concentration of about 30 mg/ml. p115 was purified from bovine liver cytosol as described previously (6).

Development of an Intra-Golgi Cell-free Transport Assay for Novel Cytosolic Components
The assay is a modification of the one described by Waters et al. (6). The 25-l assay contained 0.4 Ci of UDP-N-[ 3 H]acetylglucosamine, 5 l of a 1:1 mixture of donor and acceptor Chinese hamster ovary cellular Golgi membrane, 100 g of I␤, 0.5 g of p115, 5 ng of recombinant NSF, 60 ng of recombinant SNAP, 10 M palmitoyl-coenzyme A, ATP and UTP regeneration systems, and 10 l of the various cytosolic fractions as indicated in the figure legends. The transport reactions were incubated at 30°C for 2 h. N-[ 3 H]Acetylglucosamine incorporated into VSV-G protein was determined as described previously (25).

Purification of p16
Preparation of Bovine Brain Cytosol-Bovine brains were obtained immediately after slaughtering and placed on ice-cold 25 mM Tris-HCl, pH 7.4, 340 mM sucrose. The tissue (600 g) was placed on a glass Waring blender, which was then filled with 800 ml of homogenization buffer containing 25 mM Tris-HCl, pH 7.4, 500 mM KCl, 250 mM sucrose, 2 mM EGTA, 1 mM dithiothreitol, 1 mM phenylmethylsulfonyl fluoride, 0.5 M 1,10-phenantroline, 2 M pepstatin A, 2 g/ml aprotonin, and 0.5 g/ml leupeptin. The homogenate was centrifuged at 7,500 ϫ g in a Sorvall GS3 rotor for 1 h at 4°C. The supernatants were pooled and centrifuged at 120,000 ϫ g in a Beckman Ti-45 rotor for 1 h at 4°C. The supernatants were pooled and dialyzed against 25 mM Tris-HCl, pH 7.4, 100 mM KCl, and 10 mM ␤-mercaptoethanol (buffer A). The dialyzed material was collected and clarified by centrifugation at 7,500 ϫ g in a Sorvall GS3 rotor for 1 h at 4°C. This material was designated bovine brain cytosol and had a protein concentration of about 8 mg/ml.
Polyethylene Glycol Precipitation-The protein concentration of bovine brain cytosol was adjusted to 7.5 mg/ml by dilution with fresh dialysis buffer, and KCl was added to a final concentration of 0.5 M. Polyethylene Glycol 4000 was added slowly to a final concentration of 12.5%. The solution was stirred for 30 min at 4°C and then centrifuged in a Sorvall GS3 rotor at 8,500 rpm for 30 min at 4°C. The supernatant was discarded, and the pellets were resuspended in a 100-ml Dounce homogenizer. The insoluble material was removed by centrifugation at 120,000 ϫ g for 6 min at 4°C, the protein concentration was adjusted to 7 mg/ml and the salt concentration to 0.1 M KCl. This material was termed polyethylene glycol precipitate.
Fast Flow Q Chromatography-The polyethylene glycol precipitate was loaded onto a 400-ml Fast Flow Q column equilibrated with buffer A containing 10% glycerol at 3 ml/min. The column was washed with 400 ml of the equilibration buffer and then eluted with a 1,200-ml gradient of KCl at a range of 0.1-0.5 M. Two active fractions were detected, one in the unbound material and the other in fractions eluted from 0.2 to 0.5 M KCl. All unbound protein was pooled and termed flow-through Q; it had a protein concentration of 0.4 mg/ml. The active fractions eluted from the column were termed ␤ (see above).
Isoelectric Precipitation-The flow-through Q pool was transferred to Spectra/por 3 dialysis bags and dialyzed against 10 mM potassium phosphate, pH 6.6, and 10 mM ␤-mercaptoethanol. The insoluble material was removed by centrifugation in a Sorvall SS34 rotor at 7,800 ϫ g for 20 min at 4°C. The supernatant, which had a protein concentration of about 0.21 mg/ml, was concentrated about 10-fold by ultrafiltration on an Amicon PM3 filter.
Superdex 75 Chromatography-The concentrated isoelectric supernatant was centrifuged for 10 min in a microcentrifuge to remove insoluble material, and the supernatant was chromatographed on a 24-ml Superdex 75 HR 10/30 column (Pharmacia) equilibrated in 25 mM Tris-HCl, pH 7.4, 150 mM KCl, 10 mM ␤-mercaptoethanol, and 10% glycerol at 0.3 ml/min. Fractions of 0.7 ml were collected and tested for transport activity. A single peak of activity was detected at an elution volume corresponding to a molecular mass of 14 -16 kDa as determined by low molecular weight calibration kit (Pharmacia). This material had a protein concentration of about 0.08 mg/ml.
Mono S Chromatography-The Superdex 75 pool was adjusted by dialysis to 10 mM phosphate buffer, pH 6.5, 10 mM ␤-mercaptoethanol, and 10% glycerol and loaded onto a 1-ml Mono S HR 5/5 (Pharmacia) column equilibrated by the same buffer at 0.5 ml/ml. The column was washed with the equilibration buffer and eluted with 25 ml of a 0 -400 mM KCl gradient. Fractions (1 ml) were collected, and aliquots were analyzed for transport activity and by electrophoresis followed by Coomassie Blue staining. The fraction purified in this way was referred to as p16.

Identification of Novel Cytosolic Factors Involved in Intra-
Golgi Transport-The cell-free system that reconstitutes intra-Golgi transport has been used in recent years for the characterization and isolation of several factors involved in intracellular trafficking (2). Proteins such as NSF, SNAPs, and p115 were isolated by this transport assay in which the activity of each factor could be assessed specifically. One such an experimental design is a complementation assay where saturating levels of different crude cytosolic fractions are added to the transport assay, and a signal is observed only upon the addition of the protein of interest. To identify novel soluble factors required to reconstitute intra-Golgi transport in vitro, we analyzed the transport activity of different cytosolic fractions in the presence of saturating levels of NSF, SNAP, and p115. Fig.  1A describes the fractionation of 12.5% polyethylene glycol precipitate of bovine brain cytosolic protein on a Q-Sepharose anion exchange column. This chromatography step separated two soluble factors with considerable transport activity (Fig.  1B). A peak of transport activity could be detected in the unbound material and was tentatively termed ␣; another factor that showed transport activity was eluted as a single peak between 0.22 and 0.30 M KCl and was tentatively termed ␤ (Fig. 1B). Each of these cytosolic factors reconstituted only part of the transport activity observed in the presence of crude cytosol, whereas both factors together recovered the full transport activity (Fig. 1C).
We then followed the transport activity present in the flowthrough material of the Fast Flow Q column (␣, see Figs. 1 and 2D). Fig. 2 describes the chromatography of factor ␣ on a CM-Sepharose cation exchange column. The transport activity of the different fractions eluted from this column was deter- mined in the presence or absence of ␤. Transport activity was clearly detected in the unbound material and was termed I␣ (Fig. 2B). Significant transport activity was also observed in fractions eluted at 0.15-0.25 M KCl, tentatively termed II␣ (Fig. 2B). When the different fractions obtained from the CM-Sepharose column were tested in the presence of ␤, the trans-port activity of both bound and unbound materials was increased significantly (Fig. 2B).
We next tested whether all three factors, I␣, II␣, and ␤, are required to reconstitute the transport assay to a level comparable to that observed with crude cytosol. As shown in Fig. 2C, factors I␣, II␣, and ␤ all showed low transport activity when each was added separately to the transport assay containing NSF, ␣SNAP, and p115. When either I␣ or II␣ was added together with ␤, a synergistic effect was obtained leading to about 80% of the signal observed with the cytosol. When both I␣ and II␣ were added in the absence of ␤ only a small increase in transport activity was observed. Together, I␣, II␣, and ␤ reconstituted the full transport activity. These experiments demonstrate the involvement of at least three different cytosolic factors (as illustrated in Fig. 2D) in intra-Golgi transport in addition to NSF, SNAP, and p115.
Characterization of a II␣-dependent Assay-In the present study we focused on the purification of the transport factor present in II␣. To assure that II␣ acts as part of the known transport machinery, we tested whether the signal obtained with the different factors was NSF-, SNAP-and p115-dependent. Golgi membranes treated with N-ethylmaleimide were tested for NSF-dependent transport activity. II␣-dependent signals could be observed only in the presence of recombinant NSF (Fig. 3A). Furthermore, in the absence of either ␣SNAP or p115, no significant II␣-dependent transport activity could be observed, indicating that II␣ is acting in conjunction with the known transport factors.
Increasing concentrations of II␣ in a transport reaction containing NSF, ␣SNAP, p115, and ␤ resulted in a saturable signal (Fig. 3B). This signal could be observed in the presence of 150 M brefeldin A, a drug that prevents budding of transport vesicles and promotes uncoupled fusion (17,28) (Fig. 3B). Apparently, II␣ is involved in docking and fusion and not in vesicle budding. The II␣-dependent assay was used further for the purification of the protein responsible for this transport activity.
Purification of a Novel 16-kDa Protein from Pool II␣-Using the II␣-dependent assay a 16-kDa protein, p16, was purified from bovine brain cytosol by the following steps: (i) ammonium sulfate precipitation; (ii) Q-Sepharose anion exchange chromatography; (iii) CM-Sepharose cation exchange chromatography; (iv) gel filtration on a Superdex 75; and (v) Mono S cation exchange chromatography. Fractions obtained from each chromatography step were tested in the II␣-dependent assay. The quantitation of p16 purification is summarized in Table I. This procedure resulted in the purification of a 16-kDa polypeptide (on SDS-polyacrylamide gel electrophoresis) to apparent homogeneity. Protein profile and the transport activity of the last purification step are depicted in Fig. 4. A typical purification resulted in about 0.3 mg of the 16-kDa protein with an ϳ1,400fold increase in specific activity and 4% activity yield. We have tentatively termed this protein p16.
Involvement of p16 in Fusion-It has been demonstrated previously that the signal in the intra-Golgi transport system could be mediated by coated vesicles (29), and as such, it is sensitive to GTP␥S. In the absence of the coat proteins, the Golgi cisternae fuse directly with each other, and the assay becomes resistant to GTP␥S (17,28). Fig. 5 demonstrates that transport observed in the presence of crude cytosol is inhibited by GTP␥S, whereas p16-dependent assay is not. This result is consistent with the finding that the II␣-dependent assay is resistant to brefeldin A. Taken together, these results suggest that p16 is likely to be involved in docking or fusion rather than in vesicle production. The possibility that p16 affects glycosylation of VSV-G protein rather than the transport process per se was ruled out because we found no difference in the rate of [ 3 H]UDP-GlcNAc uptake into the Golgi lumen (30) in the presence versus the absence of purified p16 (2 Ϯ 0.06 pmol/min/mg compared with 1.9 Ϯ 0.04 pmol/min/mg, respectively). DISCUSSION We have utilized the well characterized intra-Golgi cell-free transport assay to detect yet unidentified cytosolic transport factors. This work describes the identification of three cytosolic factors and the purification of a novel 16-kDa protein required for intra-Golgi transport.
In the process of identifying novel cytosolic factors required for intra-Golgi protein transport in vitro, we characterized three different protein pools each exhibiting low transport activity in the presence of NSF, SNAP, and p115, yet together they reconstituted the full transport activity observed with crude cytosol. Previously, Waters et al. (6) identified two crude cytosolic factors required to reconstitute intra-Golgi transport in vitro in addition to NSF, SNAP, and p115. Factors ␣ and ␤ described in the present study could be related to these factors. In this study we were able to separate further the ␣ factor into two distinct activities. Pool ␤ significantly stimulated the signal of either I␣ or II␣, whereas combining I␣ and II␣ in the absence of ␤ resulted in only a slight increase in the assay signal (Fig. 2C). Conceivably, both ␣ factors could share the active protein component. However, further purification of I␣ based on its transport activity revealed that the active component in this pool was a 56-kDa polypeptide, 2 whereas in the present study we show that the transport activity of II␣ was attributed to p16. It could still be possible that both crude fractions in the early stages of the purification might include small amounts of reciprocal activity.
We described here a purification procedure based on a functional intra-Golgi transport assay, which led to the isolation of a novel low molecular weight protein, p16. Using this purification procedure, p16 was enriched by about 1,400-fold to apparent homogeneity. Similar enrichment of transport activity was required for the other soluble transport factors, such as ␣SNAP  and p115, indicating that the activity of p16 in the cytosol is comparable to these other transport factors. Amino acid sequence analysis of five different tryptic peptides derived from the pure 16-kDa protein indicates that it is a novel protein (data not shown). We have recently cloned the p16 encoding cDNA from bovine brain, and all sequences of the peptides obtained from the endogenous protein are represented in the putative cDNA p16 amino acid sequence. 3 This clearly demonstrates that the 16-kDa protein in the pure fraction is a single polypeptide. We also found that a recombinant p16 was active in the intra-Golgi transport cell-free assay, strongly supporting the notion that the 16-kDa polypeptide described here is the active component of the pure fraction. 4 Several well defined steps are required for intracellular vesicular protein traffic, including budding, targeting, docking, and fusion of vesicles with their target membranes, each requiring a different set of cytosolic factors. It has been demonstrated previously that ARF-1 and the coatomer are the only cytosolic factors required for the production of Golgi-derived COPI (31). Vesicle targeting involves interaction between integral membrane proteins, the v-SNAREs located on vesicles and the t-SNAREs present on the target membrane (32,33). The v-SNARE⅐t-SNARE complex then binds SNAP and NSF which, in turn, catalyze the disassembly of the SNARE complex, thus initiating fusion (32). This core machinery of protein transport probably requires the participation of additional accessory factors. Our data suggest that all three factors described in this study, I␣, ␤, and p16, are part of the targeting and fusion machinery. Analysis of the intra-Golgi transport assay revealed that in the absence of the coat proteins the assay measures mainly fusion between the Golgi cisternae (17). This uncoupled fusion reaction requires known components such as NSF, SNAP, and p115 and is resistant to GTP␥S. We demonstrated here that the p16-dependent assay was not inhibited by GTP␥S or brefeldin A, indicating that it represents an uncoupled fusion. Based on these results we suggest that p16 plays a role in docking or fusion of vesicles rather than being part of the budding apparatus. FIG. 5. GTP␥S inhibition of the intra-Golgi assay. Transport activity of crude cytosol, measured by the standard transport assay (see "Materials and Methods") (panel A) or by the p16-dependent assay (after the Mono S chromatography described in Fig. 4) (panel B), was tested in the presence or absence of 25 M GTP␥S as indicated. Assays were carried out in duplicate, and the mean is plotted with the error bar representing the higher value.