GCP60 Preferentially Interacts with a Caspase-generated Golgin-160 Fragment*

Golgin-160, a ubiquitous protein in vertebrates, localizes to the cytoplasmic face of the Golgi complex. Golgin-160 has a large coiled-coil C-terminal domain and a non-coiled-coil N-terminal (“head”) domain. The head domain contains important motifs, including a nuclear localization signal, a Golgi targeting domain, and three aspartates that are recognized by caspases during apoptosis. Some of the caspase cleavage products accumulate in the nucleus when overexpressed. Expression of a non-cleavable form of golgin-160 impairs apoptosis induced by some pro-apoptotic stimuli; thus cleavage of golgin-160 appears to play a role in apoptotic signaling. We used a yeast two-hybrid assay to screen for interactors of the golgin-160 head and identified GCP60 (Golgi complex-associated protein of 60 kDa). Further analysis demonstrated that GCP60 interacts preferentially with one of the golgin-160 caspase cleavage fragments (residues 140–311). This strong interaction prevented the golgin-160 fragment from accumulating in the nucleus when this fragment and GCP60 were overexpressed. In addition, cells overexpressing GCP60 were more sensitive to apoptosis induced by staurosporine, suggesting that nuclear-localized golgin-160-(140–311) might promote cell survival. Our results suggest a potential mechanism for regulating the nuclear translocation and potential functions of golgin-160 fragments.

The Golgi complex is an essential organelle of the secretory pathway. In higher eukaryotic cells, it is composed of polarized stacks of cisternal membranes (reviewed in Ref. 1). In mammalian cells, these stacks are organized into a larger ribbon structure located next to the nucleus at the microtubule organizing center (2). This highly organized Golgi ribbon is not required for cargo processing and sorting but may have other functions, including a role in signal transduction (reviewed in Ref. 3).
The structure of the Golgi is very dynamic. During mitosis the Golgi reversibly disassembles into small vesicular and tubular compartments, which contain some structural components of the Golgi. This process is regulated by mitotically active kinases (reviewed in Ref. 4). The Golgi complex also undergoes irreversible disassembly during apoptosis. This process is con-trolled by caspase cleavage of Golgi structural proteins rather than by phosphorylation. Some caspase substrates with a role in establishment and maintenance of the Golgi complex are GRASP65 (5), GM130 (6), giantin (7), p115 (8), and golgin-160 (9) (reviewed in Ref. 10). Cleavage of these proteins is necessary for efficient apoptotic disassembly of the Golgi. When a caspase-resistant form of GRASP65, p115, or golgin-160 is expressed, Golgi disassembly during apoptosis is delayed. Interestingly, when overexpressed, a potential C-terminal caspase cleavage product of p115 translocates to the nucleus and can induce apoptosis (8). This suggests that the cleavage of Golgi proteins during apoptosis is important not only for Golgi disassembly but might also play a role in apoptotic signaling.
Golgin-160, like GM130 and giantin, is a member of the golgin family. Golgins were initially identified by antibodies from patients with autoimmune disease (reviewed in Ref. 11). This diverse family is characterized by localization to the cytoplasmic face of the Golgi and a large coiled-coil region that can form a rodlike structure (reviewed in Ref. 12). Golgin-160 is composed of a large C-terminal coiled-coil region and an N-terminal non-coiled-coil ("head") domain (13). The N-terminal head of golgin-160 can be cleaved by caspases at aspartates 59 (caspase-2), 139 (caspase-3), and 311 (caspase-2, -3, or -7) (9). Additionally, the golgin-160 head contains a Golgi targeting domain and a cryptic nuclear localization signal (NLS) 3 (14). The NLS is cryptic because the full-length molecule is not localized to the nucleus, but fragments corresponding to caspase cleavage products that contain the NLS are translocated and accumulate in the nucleus. Thus, it is possible that caspase cleavage exposes the NLS that is otherwise hidden in the fulllength protein (14). The function of these golgin-160 fragments is still not known, but cells expressing a non-cleavable mutant of golgin-160 are resistant to a subset of pro-apoptotic stimuli (15). These results support the idea of a signaling function for the Golgi complex during apoptosis.
We performed a yeast two-hybrid screen to identify proteins that interact with the golgin-160 N-terminal head domain. We found that the Golgi complex-associated protein of 60 amino acids (GCP60) binds golgin-160 and, interestingly, shows preferential binding for a fragment of golgin-160 that is generated by caspase cleavage. This binding can prevent nuclear accumulation of the golgin-160 fragment when GCP60 is overexpressed and alter apoptosis in response to staurosporine.

EXPERIMENTAL PROCEDURES
Yeast Two-hybrid Assay-A Matchmaker pretransfomed HeLa cDNA library (Clontech) was screened using the Matchmaker Gal4 two-hybrid system 3 and the golgin-160 head (amino acids 1-393) as bait, as previously described (16). The cDNA encoding amino acids 18 -528 from GCP60 was isolated. The full-length GCP60 cDNA was constructed by using the 5Ј-end from a commercial IMAGE EST clone (number 4813993, ATCC, Manassas, VA) and the two-hybrid cDNA.
Cells and Antibodies-HeLa cells were maintained in Dulbecco's modified Eagle's medium (Invitrogen, Carlsbad, CA) containing 10% fetal calf serum (Atlanta Biologicals, Norcross, GA) and 0.1 mg/ml normocin-O (InvivoGen, San Diego, CA) at 37°C in 5% CO 2 . HeLa cells stably expressing GFP-golgin-160 have previously been described (15). HeLa cells stably expressing GFP-GCP60 were selected after transient transfection (described below) in normal growth medium containing 400 g/ml Geneticin (Invitrogen). Colonies were screened for GCP60 expression by fluorescence. The anti-N-terminal golgin-160 antibody has been previously described (14). Polyclonal rabbit anti-GFP antibodies were from Molecular Probes (Eugene, OR), and anti-GFP mouse antibodies were from Roche Applied Science as were the monoclonal anti-Myc antibodies. The anti-human poly(ADP-ribose) polymerase (PARP) antibodies were from BD Biosciences (Pharmingen). The Texas Red-conjugated goat anti-rabbit IgG was from Jackson Immu-noResearch Laboratories, Inc. (West Grove, PA), and the Alexa 488-conjugated goat anti-mouse IgG was from Molecular Probes.
Indirect Immunofluorescence Microscopy-HeLa cells were cultured on coverslips (70 -80% confluent) and transfected with 1 g of DNA per 35-mm dish using FuGENE 6 (Roche Applied Science) as recommended by the manufacturer. At 18 h post-transfection, cells were rinsed in phosphate-buffered saline (PBS), fixed in 3% paraformaldehyde in PBS for 10 min, rinsed in PBS containing 10 mM glycine (Gly/PBS) for 5 min, permeabilized in 0.5% Triton X-100 in Gly/PBS for 3 min, and rinsed in Gly/PBS. The coverslips were then incubated in primary antibodies for 20 min, washed twice with Gly/PBS, and incubated with secondary antibodies. After washing, the cells were incubated with Hoechst 33258 (Sigma) for 3 min, washed, and mounted in glycerol containing 0.1 M N-propyl gallate. The images were collected on an Axioskop microscope (Zeiss, Thornwood, NY) equipped with epifluorescence and a Sensys CCD camera (Photometrics, Tucson, AZ) using IP Lab software (Signal Analytic, Vienna, VA). The scoring of cells for Fig. 3B was as follows. The nuclear classification was used for cells with exclusively nuclear staining or nuclear staining that was equal or stronger than Golgi staining. The Golgi classification was used when the Golgi staining was stronger than nuclear staining (see insets in Fig. 3B). Approximately 150 cells were counted for each transfection condition in four independent experiments.
Binding Assays-The GST constructs were expressed in Escherichia coli BL21-codon plus (Stratagene) and purified on glutathione-Sepharose 4B as recommended by the manufacturer (Amersham Biosciences). For the binding assays, 10 g of purified GST and GST fusion proteins were rebound to glutathione beads and incubated overnight with either transfected HeLacell lysates or [ 35 S]methionine-labeled protein. The 35 S-labeled protein was generated by coupled in vitro transcription and translation using the TNT T7 coupled reticulocyte lysate system (Promega, Madison, WI) programmed with a plasmid encoding the appropriate protein behind a T7 promoter and Redivue [ 35 S]methionine (Ͼ1000 Ci/mmol, Amersham Biosciences) following the manufacturer's instructions. In vitro translated protein was diluted in detergent solution (62.5 mM EDTA, 50 mM Tris-HCl, pH 8, 0.4% deoxycholate, 1.0% Nonidet P-40, and protease inhibitors (Sigma, catalog number P8340) before incubation with GST proteins. The transfected HeLa cell lysates were obtained as follows. Cells cultured in 60-mm plates (70 -80% confluent) were transfected using FuGENE 6 as described above. The cells were lysed in detergent solution, incubated on ice for 15 min, and spun at 14,000 ϫ g at 4°C for 15 min. Lysates (or in vitro translated protein) were incubated with the beads overnight and washed three times in detergent solution. Bound proteins were eluted in sample buffer, boiled, resolved by SDS-PAGE, and detected by immunoblotting or phosphor imaging. For immunoblotting, electrophoresed proteins were transferred to Immobilon-P membranes (Millipore, Bedford, MA), and the membranes were blocked in 5% milk-TBS-T (0.1% Tween 20, 150 mM NaCl, 10 mM Tris-HCl, pH 7.4). Incubation with the appropriate primary antibody was overnight at 4°C and with secondary antibody for 60 min at room temperature. The primary antibodies were diluted in TBS-T containing 4% bovine serum albumin and 0.02% sodium azide, whereas the horseradish peroxidase-conjugated goat anti-rabbit or antimouse secondary antibodies (Amersham Biosciences) were used at a 1:2000 dilution in TBS-T only. Membranes were analyzed using chemiluminescence (ECL, Amersham Biosciences) and measured by the VersaDoc Imaging System (Bio-Rad) using Quantity One software.
In Vitro Cleavage Assay-[ 35 S]Methionine-labeled proteins were generated by in vitro transcription and translation as described above. The labeled proteins were diluted with 100 l of either caspase-2 buffer (25 mM HEPES pH 7.5, 0.15 M NaCl, 10% sucrose, 1 mM ATP, 5 mM dithiothreitol) or caspase-3 buffer (50 mM HEPES pH 7.4, 100 mM NaCl, 1 mM EDTA, 10 mM dithiothreitol, 0.1% CHAPS, 10% sucrose). The diluted samples were split into two 60-l aliquots. One aliquot of each sample received the appropriate caspase: 0.159 g of caspase-2 (1.66 units/mg) or 0.116 g of caspase-3 (5.55 units/mg) (both from Sigma). The other aliquots served as mock-treated controls. All samples were incubated at 37°C for 1.5 h. Sample buffer was added, and the samples were boiled, resolved by SDS-PAGE, and visualized on a Molecular Imager FX Phospho-rImager (Bio-Rad) using Quantity One software.
Apoptosis Assay-The apoptosis assay has been previously described (15). Briefly, stable cell lines expressing golgin-160 (15) or GFP-GCP60 were plated in regular growth medium in six-well plates 1 day before treatment with 1 M staurosporine (Sigma). The drug was added in 1.5 ml of regular growth medium (without normacin). After treatment, cells were scraped into the medium and spun at 1000 rpm (150 ϫ g) for 20 min. The samples were rinsed in cold phosphate-buffered saline and respun. Cells were lysed in detergent solution (as described earlier) and incubated on ice for 15 min. An aliquot of the sample was used to determine protein concentration by bicinchoninic acid assay (Pierce Chemical, Rockford, IL). For analysis of PARP cleavage, 50 g of protein was electrophoresed in a 10% acrylamide gel. Proteins were transferred to Immobilon-P transfer membrane, and immunoblotting was performed using anti-PARP antibodies. Quantification of the chemiluminescent signal was performed using a VersaDoc imaging system (Bio-Rad). The percent of PARP cleavage was determined by measuring full-length and cleaved PARP bands, subtracting background, and then dividing the amount of cleaved PARP by the amount of total PARP (full-length plus cleaved).

RESULTS
Golgin-160 Interacts with GCP60 in Vitro-We previously showed that the N-terminal "head" domain of golgin-160 (amino acids 1-393) encodes important motifs (9,14). We thus searched for interactors with this region by performing a yeast two-hybrid screen using golgin-160 (1-393) as bait. The screen included ϳ1 ϫ 10 7 clones of a pretransformed HeLa cell library, and two of the isolates encoded the GCP60. GCP60 was initially identified in a yeast two-hybrid screen as a binding partner of the Golgi protein giantin (17). GCP60 was also reported by another group to interact with both the peripheral type benzodiazepine receptor and the cAMP-dependent protein kinase A regulatory subunit RI␣ at mitochondria, for which it was named PAP7 (18). The sequence of GCP60 predicts that the protein contains an acyl-CoA binding site, a coiled-coil region, and a Golgi dynamics (GOLD) domain (Fig. 1A) (17,19). To confirm the interaction of golgin-160 with GCP60 we performed an in vitro binding assay using GST fused to golgin-160-(1-393) or GST alone and lysates of HeLa cells expressing GFP-GCP60. GST proteins were prebound to glutathione beads and incubated with cell lysates, and after washing, GFP-GCP60 was detected by immunoblotting with anti-GFP antibodies. As shown in Fig. 1B, GST-golgin-160-(1-393) but not GST alone was able to precipitate GFP-GCP60. These data demonstrate that GCP60 can interact specifically with golgin-160.
GCP60 Prevents Accumulation of Golgin-160 Fragments in Nucleus-As mentioned earlier, the N-terminal domain of golgin-160 contains a cryptic NLS that is responsible for the nuclear localization of golgin-160 fragments (14). To determine whether overexpression of GCP60 could affect the targeting of golgin-160 fragments, we transfected HeLa cells with plasmids encoding GFP-GCP60 or GFP alone, in combination with the different Myc-tagged golgin-160-(1-393), -(60 -311), and -(140 -311) fragments. The localization of the expressed proteins was determined by immunofluorescence microscopy. Golgin-160-(1-393), -(60 -311), and -(140 -311) were mostly present in the nucleus when expressed in combination with GFP. However, when co-expressed with GFP-GCP60, the nuclear accumulation of these fragments was significantly reduced, and the Golgi localization increased (Fig. 3A). To quantify these results, we scored the expression pattern of each golgin-160 fragment as described under "Experimental Procedures." Fig. 3B shows that overexpression of GCP60 increased the Golgi localization of all the golgin-160 fragments relative to their nuclear localization. Interestingly, golgin-160-(140 -311), which was almost entirely targeted to the nucleus when expressed with GFP, showed an ϳ80-fold increase in Golgi localization when co-expressed with GFP-GCP60. These observations correlate with the in vitro binding assays that show an increased binding of GFP-GCP60 to golgin-160-(140 -311) relative to golgin-160-(1-393) or -(60 -311). Our results suggest that binding of GFP-GCP60 may prevent golgin-160 fragments from accumulating in the nucleus.
The C-terminal Portion of GCP60 Interacts with Golgin-160-To narrow down the GCP60 region responsible for binding golgin-160, plasmids encoding GFP-GCP60-(1-175) and GFP-GCP60-(328 -528) were expressed in HeLa cells. Binding assays were performed with GST-golgin-160 fragments. We observed that a region of GCP60 between residues 328 and 528 was responsible for binding to golgin-160 (Fig. 4A). Additionally, the truncated version of GCP60 bound best to the golgin-160-(140 -311) fragment similar to the interaction with full-length GCP60 ( Fig. 2A versus Fig. 4A, right). The slight difference in mobility of GFP-GCP60-(328 -528) observed in the lane with GST-golgin-160-(60 -311) compared with the other lanes was because of co-migration of the GST fusion protein and the GFP-GCP60 fragment. Because the C-terminal portion of GCP60 contains the giantin-binding domain that localizes the protein to the Golgi, we looked at co-transfected cells to see if this portion was sufficient to prevent golgin-160 fragments from translocating to the nucleus. Fig. 4B shows that golgin-160 fragments were preferentially targeted to the Golgi in the presence of GCP60-(328 -528) just as they were when the fulllength GCP60 protein was co-expressed (Fig. 3A). GCP60-(1-175), which does not bind to golgin-160, did not influence the targeting of golgin-160 fragments, as expected (Fig. 4B).
Caspase-3, but Not Caspase-2, Cleaves GCP60 in Vitro-We have previously mapped unique cleavage sites in golgin-160 for caspase-2 (Asp-59) and caspase-3 (Asp-139) and a site that can be cleaved with caspase-2, -3, or -7 (Asp-11) (9) (Fig.  2A). Because golgin-160-(140 -311) could be generated by caspase-3 alone or by combining cleavage of caspase-3 together with caspase-2 or -7 ( Fig. 2A), we investigated if the activated caspases responsible for generating this golgin-160 fragment could also cleave GCP60. All GCP60 aspartate residues are found within the N-terminal (15-156) and C-terminal (343-528) portions of the protein. Therefore, in vitro cleavage assays  GCP60 preferentially binds to golgin-160-(140 -311). A, schematic representation of full-length golgin-160 and the potential caspasegenerated golgin-160 fragments (C2, caspase-2; C3, caspase-3; C7, caspase-7). B, HeLa cells were transfected with a plasmid encoding GFP-GCP60. Aliquots of lysates were incubated with GST fused to different fragments of golgin-160 (as indicated by the residue numbers) or GST alone. Bound proteins were separated by SDS-PAGE, and GFP-GCP60 was detected by immunoblotting with anti-GFP antibodies. C, GCP60 was [ 35 S]methionine-labeled by in vitro transcription and translation and incubated with the same GST fusion proteins as in B. Bound proteins were resolved by SDS-PAGE, and GCP60 was detected by phosphor imaging. D, the GST fusion proteins used for the binding assays are shown in a Coomassie Blue-stained gel. Positions of the molecular mass markers (in kDa) are shown on the left.
were performed using plasmids encoding GCP60, GFP-GCP60- (1-175), and GFP-GCP60-(328 -528). The constructs were expressed and labeled with [ 35 S]methionine by in vitro transcription and translation, diluted in the appropriate buffer, and incubated in the presence or absence of caspase-2 or caspase-3 as described under "Experimental Procedures." Fig.  5A shows that caspase-2 does not cleave GCP60 although it did cleave Myc-tagged caspase-2 (used as positive control). On the other hand, caspase-3 cleaved GCP60 once (Fig. 5B). This cleavage site is located in the N-terminal portion of the protein because GFP-GCP60-(1-175) was cleaved but GFP-GCP60-(328 -528) was not (Fig. 5B). The fragment size generated after caspase-3 incubation led us to consider aspartate 15 and aspartate 22 as candidates for cleavage. When Asp-15, but not Asp-22, was mutated to glutamate, caspase-3 cleavage was prevented, indicating that VSVD 15 G is the cleavage site for caspase-3 (Fig. 5C). Thus, we predict that caspase-3 cleavage of GCP60 should not affect its localization to Golgi membranes or its interaction with golgin-160-(140 -311) (17) (Fig. 4A, right). We confirmed caspase cleavage in vivo in cells treated with staurosporine or anisomycin, where the N-terminal GFP tag disappeared over time, consistent with the caspase-3 cleavage of the N-terminal site in GCP60 (data not shown).
GCP60 Sensitizes HeLa Cells to Death in Response to Staurosporine Treatment-We have previously shown that a noncleavable form of golgin-160 impairs apoptosis induced by some pro-apoptotic stimuli (15). Because the above experiments show that GCP60 cleavage by caspase-3 should not affect its interaction with golgin-160 fragments, we investigated the effects of overexpressing GCP60 on apoptosis. We used HeLa cells stably expressing GFP-GCP60 to ensure that all cells expressed a similar level of the transgene. As a control, we used cells stably expressing another GFP-tagged peripheral Golgi protein, GFP-golgin-160 (15). The cell lines were treated with 1 M staurosporine, a broad spectrum kinase inhibitor that has been shown to promote cleavage of golgin-160 (15). Cell death was assessed by cleavage of PARP, a well characterized marker for apoptosis. Fig. 6 shows that after 2 h of staurosporine treatment, cells expressing GFP-GCP60 were significantly more sensitive to the drug. An increase in PARP cleavage was also observed when compared with untransfected HeLa cells (data not shown). Together, these results suggest that overexpression of GCP60 alters localization and the potential function of golgin-160-(140 -311) during apoptosis.

DISCUSSION
Golgin-160 is a peripheral membrane protein associated with the cytoplasmic face of the Golgi complex (13). It has a long coiled-coil region, characteristic of all golgins, and a non-coiled-coil region ("head") of 393 residues at the N terminus. We previously found that this head region has a cryptic nuclear localization signal (amino acids 232-239), a Golgi targeting domain (amino acids 172-257), and three aspartates targeted by caspases during apoptosis (Asp-59, Asp-139, and Asp-311) (9,14). In a yeast two-hybrid screen searching for interactors of the golgin-160 head domain, we identified GCP60.
GCP60 was previously identified as a giantin interactor (17). Besides a central predicted coiled-coil, GCP60 is predicted to contain an acyl-CoA binding site and a GOLD domain, which is found in other Golgi proteins and is predicted to mediate protein-protein interactions (17,19). The giantin-binding domain of GCP60 is contained between residues 373 and 528 and is responsible for its Golgi localization (17). GCP60 was proposed to be involved in structural maintenance of the Golgi because overexpression of wild type GCP60 or the giantin binding domain in COS-1 cells disrupted Golgi structure. However, we did not observe any changes in Golgi structure when GFP-GCP60 was overexpressed in HeLa cells, perhaps because expression levels were lower than in transfected COS-1 cells. Golgin-160 binds the C-terminal region of GCP60, similar to giantin. Further mutational analysis will be required to determine whether the binding sites overlap and if they are mutually exclusive.
A separate study searching for interactors of the peripheraltype benzodiazepine receptor and of cAMP-dependent protein kinase A regulatory subunit RI␣ identified PAP7 (20). Human PAP7 is 96% identical to human GCP60 (18). Although ubiquitous, GCP60/PAP7 expression is highest in steroid-producing tissues, including testis and ovary (20). PAP7 was detected in the Golgi region and in mitochondria in Leydig MA-10 cells (21). This observation and the finding that Golgi perturbation led to increased mitochondrial localization of PAP7 with increased steroidogenesis suggest that PAP7 may be involved in cholesterol trafficking from the Golgi to mitochondria, where the early steps of steroid hormone synthesis occur (18). However, GCP60/PAP7 appears to be localized exclusively at the Golgi complex in HeLa cells (Fig. 3A and Ref. 17).
Cleavage of golgin-160 during apoptosis generates different fragments (9), exposing a cryptic NLS that allows accumulation of some of these fragments in the nucleus (14). In this work, we found that the golgin-160 binding domain for GCP60 was included between residues 140 and 311. Interestingly, golgin-160 (140 -311) interacted much better with GCP60 than with either golgin-160-(1-393) or golgin-160-(60 -311). Perhaps the GCP60 binding domain in golgin-160 is mostly masked when the protein is in its full-length form, and it becomes exposed in the golgin-160-(140 -311) cleaved fragment. It is also possible that upon cleavage a conformational change in the golgin-160-(140 -311) fragment favors interaction with GCP60. In either case, this implies an interesting mechanism for regulation of this protein-protein interaction. By contrast, golgin-160 interaction with PIST (another peripheral Golgi membrane protein) requires the residues flanking Asp-139, such that caspase cleavage at this site is predicted to block interaction (16).
The expression of a caspase-resistant form of golgin-160 disrupts apoptosis induced by a subset of apoptotic signals, suggesting a potentially important role for the cleavage fragments (15). Here we showed that when GCP60 is overexpressed, the golgin-160 fragments that have an NLS tend to be retained at the Golgi. This Golgi retention correlates with the interaction between the golgin-160 fragments and GCP60. Golgin-160-(140 -311), which showed the strongest in vitro interaction, was affected most dramatically by increasing its Golgi localization over 80-fold in the presence of overexpressed GCP60. By contrast, Golgi retention of golgin-160-(60 -311) and golgin-160-(1-393) showed a much less dramatic effect.
To determine whether the interaction of golgin-160 fragments with GCP60 could occur in apoptotic cells, we investigated GCP60 as a target for the caspases that can generate golgin-160-(140 -311). GCP60 was not cleaved by caspase-2 but was cleaved once by caspase-3. We mapped the aspartate residue in GCP60 that was targeted by caspase-3 to Asp-15. Thus, the C-terminal caspase-3 product of GCP60 would still be predicted to bind strongly to golgin-160-(140 -311) and to be retained at the Golgi because its interaction with giantin would not be jeopardized. Giantin can also be cleaved by caspases, but the C-terminal fragment generated is still membrane-associated and contains the GCP60 binding region (7).
Because the preferential interaction exhibited by the golgin-160-(140 -311) fragment with GCP60 could take place during apoptosis, we investigated the effects of overexpressing GCP60 during cell death. When cells stably expressing GFP-GCP60 were treated with staurosporine, we observed an increase in apoptosis as compared with control cells. This surprising result implies that nuclear translocation of golgin-160-(140 -311) may be involved in survival signaling. We did not detect a difference in the steady-state levels of golgin-160-(140 -311) in GCP60-overexpressing cells compared with control cells (data not shown), suggesting that the localization rather than the stability of this fragment is important in regulating apoptosis. Modulation of GCP60 expression will thus be a useful tool to further study the role of this golgin-160 cleavage fragment in transducing apoptotic signals.