Article

• The EMBO Journal (2008) 27, 948 - 955
• doi:10.1038/emboj.2008.36

Published online: 6 March 2008

• Subject Categories:

  • Membranes and Transport | Cell Cycle

Spatial separation of Golgi and ER during mitosis protects SREBP from unregulated activation

René Bartz^1,3, Li-Ping Sun², Blaine Bisel¹, Jen-Hsuan Wei¹ and Joachim Seemann¹

1. Department of Cell Biology, University of Texas Southwestern Medical Center, Dallas, TX, USA
2. Department of Molecular Genetics, University of Texas Southwestern Medical Center, Dallas, TX, USA

Correspondence to:

Joachim Seemann, Department of Cell Biology, University of Texas Southwestern Medical Center, 5323 Harry Hines Blvd, Dallas, TX 75390-9039, USA. Tel.: +1 214 648 0317; Fax: +1 214 648 8694; E-mail: joachim.seemann@utsouthwestern.edu

³Present address: Merck & Co. Inc., RNA Therapeutics, 770 Sumneytown Pike, West Point, PA 19486, USA

Received 30 November 2007; Accepted 13 February 2008

• Keywords:

  • cholesterol,
  • Golgi,
  • mitosis,
  • SREBP

Introduction

Proliferation of animal cells requires the growth and partitioning of all cellular components during the cell cycle. The growth and integrity of organelles rely on the coordinated synthesis of lipids and proteins and must be tightly regulated. Sterol regulatory element-binding proteins (SREBPs) are key regulators of lipid and membrane biogenesis. The proteins are membrane-bound transcription factors that reside as silent precursors in the membrane of the endoplasmic reticulum (ER) (Espenshade and Hughes, 2007). The activation of SREBP is achieved by regulated intramembrane proteolysis after transport from the ER to the Golgi apparatus (Brown et al, 2000; Rawson, 2002). This process is highly regulated and crucially depends on the integrity and the spatial separation of the ER and the Golgi apparatus. After stimulation, SREBP is transported from the ER to the Golgi apparatus, where the proteins are sequentially cleaved by site-1 and site-2 proteases (S1P and S2P, respectively). Cleavage leads to the release of the active transcription factors from the membrane into the cytosol, which then enter the nucleus to initiate transcription. In the nucleus, SREBP binds to sterol regulatory element sequences in promoter regions to induce genes that regulate cellular levels of cholesterol and phospholipids.

SREBP activation is regulated by cellular levels of sterols. SREBP is bound to its escort protein Scap (SREBP cleavage activation protein), and in the presence of high sterol levels in the ER membrane the complex is retained by binding to the ER-resident protein Insig (Radhakrishnan et al, 2007; Sun et al, 2007). As sterol levels drop, Scap dissociates from Insig, causing a conformational change in Scap that exposes the binding site for the COPII coat proteins. The SREBP–Scap complex then leaves the ER in COPII vesicles and moves to the Golgi apparatus, where SREBP is cleaved by S1P and S2P to release the active transcription factor (Sun et al, 2005).

Previous experiments showed that the presence of S1P and SREBP in the same membrane compartment can induce cleavage and SREBP activation independent of sterol levels in the membrane (DeBose-Boyd et al, 1999; Hampton, 2000). Instead of transporting SREBP to the Golgi apparatus, the relocation of S1P into the ER is sufficient for cleavage. This has been shown by two different approaches. Treatment of cells with brefeldin A (BFA) disrupts the Golgi structure and induces parts of the Golgi apparatus to merge with the ER. This causes SREBP to be released from the ER membrane even in the presence of high cellular cholesterol levels, which normally regulate and inhibit the transport and processing of SREBP. BFA-induced activation of SREBP is dependent on S1P, because SREBP remained intact in a mutant cell line lacking active S1P. This suggests that BFA causes S1P to be translocated from the Golgi into the ER to induce SREBP cleavage (DeBose-Boyd et al, 1999). In a second approach, S1P was tagged with a KDEL sequence and expressed in cells. ER-resident proteins with a KDEL sequence (e.g. BiP) that escape the ER are bound in the Golgi apparatus by the KDEL receptor and subsequently transported back to the ER (Munro and Pelham, 1987). Expression of S1P–KDEL induces the processing and activation of SREBP, suggesting that active S1P in the ER is sufficient to activate SREBP even if S2P is unaltered (DeBose-Boyd et al, 1999; Hampton, 2000).

By mechanisms that resemble BFA treatment, the structural integrity of the Golgi apparatus is also compromised during mitosis as part of its normal inheritance process (Shorter and Warren, 2002). We wanted to explore this possibility, as it is currently unknown whether this process leads to the activation of SREBP, which would be independent of the well-characterized sterol-regulated activation. At the beginning of mitosis, the Golgi apparatus breaks down and, after successful cell division, the fragments fuse to re-form a functional Golgi in both daughter cells (Shorter and Warren, 2002). The underlying mechanisms that organize the division of the Golgi during mitosis are still not well understood. Similar to the nuclear envelope, which fragments in prophase and then fuses with the ER, components of the Golgi apparatus (e.g. Golgi enzymes) have also been described to merge with the ER during mitosis to facilitate Golgi partitioning (Zaal et al, 1999; Altan-Bonnet et al, 2006). However, other works using transiently expressed chimaeric Golgi enzymes found no evidence that Golgi and ER membranes did combine during mitosis (Pecot and Malhotra, 2004).

Previous approaches have not addressed the physiological relevance of why some Golgi components may be divided either independently or together with the ER. Here, we show that SREBP is not activated during mitosis, suggesting that the Golgi-resident enzyme S1P remains separated from the ER. This suggests that the distinct boundaries between the Golgi and the ER are preserved not only during interphase, but also during mitosis, to prevent unspecific activation of the transcription factor SREBP.

S1P is a Golgi-resident enzyme

We first determined the distribution and properties of mammalian S1P. Western blot analysis with affinity-purified antibodies against S1P demonstrated that the enzyme is highly enriched in Golgi membranes when compared to rat liver homogenate, membrane-enriched fraction and cytosol (Figure 1A) (Wang et al, 2006). The protein was highly enriched in the Golgi fraction and co-purified with the Golgi matrix protein GM130, a well-established Golgi marker protein (Nakamura et al, 1995; Seemann et al, 2000). The luminal ER-resident chaperone GRP78/BiP was not detectable in the Golgi fraction, but was present in both total lysate fraction and membrane homogenate (lanes 1 and 3).

Figure 1.

S1P associates with Golgi membranes and recycles through the ER. (A) S1P is enriched on Golgi membranes. Rat liver homogenate, cytosol, intermediate membrane fraction and Golgi membranes were separated by 10% SDS–PAGE (40 g protein for homogenate (homog.), cytosol and intermediate (interm.); 10 g for Golgi fraction) and analysed by immunoblotting with antibodies against S1P, GM130, Grp78/BiP and tubulin. (B) S1P is found in the Triton X-100-insoluble Golgi matrix fraction. Golgi membranes ( 300 g) were extracted with Triton X-100 and pelleted by centrifugation. Equal amounts of protein were subjected to 10% SDS–PAGE and immunoblotted with the indicated antibodies. (C) BFA separates S1P from GM130. NRK cells were incubated for 45 min with 5 g/ml BFA or carrier control. Cells were then fixed, permeabilized and stained with mouse anti-GM130 and rabbit anti-S1P antibodies followed by fluorescently conjugated secondary antibodies (anti-mouse AlexaFluor 594 and anti-rabbit AlexaFluor 488). (D) S1P is retained in the ER after Sar1^DN microinjection. NRK cells were injected with 1 mg/ml purified Sar1^DN protein, fixed after 3 h incubation, permeabilized and stained with an antibody against GM130 together with polyclonal antibodies against S1P or ManII. Injected cells were identified by coinjected cascade-blue-conjugated BSA (asterisks). (E) The Golgi apparatus re-forms without S1P. NRK cells were incubated for 45 min with 5 g/ml BFA and injected with recombinant Sar1^DN protein as in (D). BFA was washed out and after 2 h the cells were fixed, permeabilized and stained with an antibody against GM130 together with antibodies against S1P and ManII. Scale bars, 15 m.

View full figure (150 KB)

The Golgi apparatus is composed of at least two different parts: Golgi matrix proteins including peripheral Golgi membrane proteins of the GRASP and golgin protein families and Golgi-resident proteins that associate with the Golgi matrix, such as glycosylation enzymes (Seemann et al, 2002; Gillingham and Munro, 2003; Short et al, 2005). To characterize the association of S1P with the Golgi apparatus, we tested if the protein can be extracted from intact Golgi membranes with detergent. Golgi membranes were extracted with Triton X-100 and the insoluble Golgi matrix was recovered by centrifugation (Slusarewicz et al, 1994). Immunoblotting revealed that S1P was Triton X-100 resistant (Figure 1B). The Golgi glycosylation enzyme mannosidase II (ManII, right panel), which has been shown to bind the Golgi matrix, and also GM130 and the Golgi SNARE GOS-28 (middle panel) were also detergent resistant. In contrast, the transmembrane proteins p24 and Golgin84 that are incorporated into COPI transport vesicles were solubilized and did not associate with the Golgi matrix fraction (left panel).

S1P cycles between the Golgi apparatus and the ER

In intact cells, the treatment of BFA leads to the separation of Golgi matrix proteins from the Golgi enzymes. In the presence of BFA, the Golgi enzymes move back to the ER, leaving behind Golgi remnants, which are tubular-vesicular structures that contain Golgi matrix proteins (Seemann et al, 2000). Immunofluorescence analysis showed that S1P colocalized with GM130 in the perinuclear Golgi ribbon of NRK (normal rat kidney) cells (Figure 1C, upper panel). After treatment with BFA for 45 min, the characteristic Golgi structure was dissolved and GM130 was present in dispersed Golgi remnants. However, S1P did not colocalize with GM130, suggesting that it was transported along with Golgi-resident glycosylation enzymes into the ER (compare upper panel with lower panel).

In contrast to Golgi matrix proteins, Golgi enzymes slowly but continuously cycle during interphase between the Golgi complex and the ER (Cole et al, 1996; Storrie et al, 1998). To test whether S1P behaves like other Golgi enzymes, we inhibited proteins from leaving the ER by microinjection of a dominant-negative form of Sar1 protein (Sar1^DN; Figure 1D). Under these conditions, the enzymes still recycle back into the ER but over time become trapped. After 4 h, staining for GM130 showed an intact Golgi ribbon that was depleted of ManII (Figure 1D). The staining pattern of S1P was comparable to that of ManII, demonstrating that it is relocalized into the ER when cycling back to the Golgi is inhibited (lower panel).

Figure 1E shows an experiment designed to determine whether S1P is required for Golgi re-formation after BFA treatment. For this purpose, BFA incubation was combined with Sar1^DN injection: first, the Golgi was disrupted with BFA and enzymes were relocated into the ER. The cells were then injected with Sar1^DN and BFA was washed out for 2 h. The Golgi remnants containing GM130 re-formed the characteristic Golgi ribbon in the perinuclear region of the cells independent of the Golgi enzyme ManII (lower panel) and S1P (upper panel), which were retained in the ER. Together, the results suggest that S1P behaves like a Golgi-resident enzyme that cycles between the Golgi apparatus and the ER.

S1P colocalizes with Golgi matrix proteins during mitosis

Similar to BFA, the structural integrity of the Golgi apparatus is also compromised during mitosis as part of its normal inheritance process. Glycosylation enzymes have been described to relocate into the ER during mitosis (Zaal et al, 1999; Altan-Bonnet et al, 2006), but Golgi matrix proteins remain distinct from the ER (Seemann et al, 2000). Therefore, we analysed the localization of S1P during mitosis to determine in which compartment it resides (Figure 2). Exponentially growing NRK cells were fixed and stained with the DNA dye Hoechst and with an antibody against tubulin to identify cells at different stages of mitosis. They were also stained with antibodies against S1P (Figure 2A) or with anti-S1P and anti-GM130 antibodies together (Figure 2B) to test whether the proteins colocalize. During interphase, GM130 and S1P were both located on the Golgi ribbon in the perinuclear region of the cell. During prometaphase, the Golgi cisternae unstacked and fragmented into clusters of tubules and vesicles. In the course of M-phase, the Golgi membranes (Figure 2B) and S1P (Figure 2A and B) were seen to be separated into similar amounts on each side of the equatorial plate. The majority of Golgi membranes were found concentrated around astral microtubules and the two spindle poles and a smaller pool was dispersed throughout the cytoplasm. Finally, the Golgi apparatus re-formed in each of the daughter cells (telophase II). During the entire M-phase, GM130 and S1P showed a high degree of colocalization, suggesting that S1P remained associated with Golgi membranes throughout the cell cycle and was not present in the ER (Figure 2B).

Figure 2.

S1P colocalizes with Golgi proteins during mitosis. (A) Localization of S1P during mitosis. Exponentially growing NRK cells were fixed, permeabilized and triple stained with antibodies for S1P (green), tubulin (red) and DNA (blue). The different mitotic stages were identified by the organization of the microtubule network and the chromatin. (B) GM130 and S1P colocalize during mitosis. Cells were prepared and mitotic cells were identified as in (A) after staining with anti-GM130 and anti-S1P antibodies. Scale bars, 10 m.

View full figure (254 KB)

S1P remains distinct from the ER during mitosis

To biochemically analyse the distribution of S1P during mitosis, synchronized mitotic CHO (Chinese hamster ovary) cells, either control-treated or grown in the presence of BFA (3 h), were harvested by shake-off and post-chromosomal supernatants (PCSs) were separated on a 5–30% glycerol gradient (Jesch and Linstedt, 1998; Jesch et al, 2001; Seemann et al, 2002). After centrifugation, fractions were collected and membranes were pelleted by centrifugation. In control-treated cells, the majority of S1P was found in the top fractions, revealing a pattern similar to GM130. The ER protein GRP78/BiP, however, was found only in the bottom fractions of the gradient (Figure 3A, right panel). A portion of mitotic Golgi membranes was present in the same fraction as BiP at the bottom of the gradient. Previous work, however, showed that these Golgi membranes consist of larger breakdown products of the Golgi, which remain distinct from the ER (Jesch and Linstedt, 1998; Jesch et al, 2001; Seemann et al, 2002). In cells incubated with BFA, GM130 continues to be found in the top fractions, comparable to control cells. S1P, however, is shifted to the bottom fractions, consistent with ER being shown to contain GRP78/BiP (left panel). These results confirm that during mitosis S1P remains separated from ER membranes, although it was possible to force a translocation of the protease into the ER with BFA treatment.

Figure 3.

S1P is distinct from the ER during mitosis. Gradient fractionation of mitotic membranes. CHO cells treated with or without 3 g/ml BFA were mitotically arrested with 2 mM S-trityl-L-cysteine. Mitotic cells were collected by shake-off and post-chromosomal supernatants were fractionated on a 5–30% glycerol gradient. Fractions were collected, separated by 12% SDS–PAGE and analysed by immunoblotting for S1P, GM130 and GRP78/BiP.

View full figure (33 KB)

SREBP remains uncleaved during mitosis

As our results found no evidence for the relocation of S1P into the ER during mitosis, we investigated whether its substrate, SREBP, is cleaved. SREBP is retained in the ER membrane in the presence of high sterol levels by binding to the ER-resident protein Insig (Yang et al, 2002). As the level of sterols drops, the SREBP-binding protein Scap dissociates from Insig and the SREBP–Scap complex is transported to the Golgi apparatus. There, SREBP is cleaved by S1P and S2P to release the active transcription factor. Independent of sterol levels in the ER, the treatment of interphase cells with BFA redistributes S1P from the Golgi apparatus into the ER (Figure 1B), leading to cleavage and activation of SREBP (DeBose-Boyd et al, 1999). To test whether SREBP is cleaved during mitosis, unsynchronized cells grown in the presence of cholesterol were stained with antibodies against SREBP (Figure 4A, upper panels). If S1P is located in the ER during mitosis, one would expect to detect the cleaved transcription factor portion of SREBP in the nucleus once cell division is complete. However, daughter cells that just finished mitosis, apparent by the remnants of the central spindle connecting the daughter cells, showed an ER-like staining pattern similar to interphase cells. In contrast, cells depleted of cholesterol that just finished cell division showed a strong nuclear staining for the cleaved active transcription factor portion of SREBP. The results show that S1P stays distinct from the ER, and SREBP remains unprocessed during mitosis (Figure 4A, lower panels).

Figure 4.

SREBP remains uncleaved in mitosis. (A) SREBP is not localized in the nuclei of post-mitotic cells. CHO cells were grown in sterol-depleting medium in the absence or presence of 1 g/ml 25-hydroxycholesterol for 16 h. Cells were fixed, permeabilized and stained for SREBP and tubulin. DNA was labelled with Hoechst. Scale bar, 10 m. (B) Biochemical analysis reveals SREBP processing in mitotic and interphase cells only after the addition of BFA. CHO cells treated with 3 g/ml BFA or carrier control were arrested in mitosis for 5 h in the presence of 2 mM S-trityl-L-cysteine. Mitotic cells (M) were harvested by shake-off and the remaining cells (interphase, I) were collected by trypsinization. Equal amounts of protein (30 g) were separated by 10% SDS–PAGE and immunoblotted with the indicated antibodies.

View full figure (83 KB)

To corroborate these results, we analysed whether SREBP is cleaved during mitosis by immunoblotting. Control cells or cells growing in the presence of BFA were enriched for 5 h in mitosis and harvested by shake-off (mitotic cells, M). Lysates from the remaining attached cells (interphase cells, I) and mitotic cells were probed with antibodies against endogenous SREBP (Figure 4B). BFA-treated cells showed the cleaved form of SREBP (left panel, lanes 3 and 4, lower band), as well as a reduced amount of full-length SREBP (upper band). Furthermore, in BFA-treated mitotic cells, an increased amount of SREBP was cleaved (compare lanes 3 and 4). In untreated mitotic and interphase cells, however, cleaved SREBP was not detectable and only unprocessed full-length SREBP was observed (lanes 1 and 2). Blots were reprobed with an antibody against the phosphorylated form of histone 3, which is specifically phosphorylated in mitotic cells (Dai et al, 2005). The difference in SREBP cleavage was not due to an altered expression level of S1P during the cell cycle or BFA treatment, as shown by immunoblots with an antibody against S1P (Figure 4B).

S1P is active in mitosis

One possible explanation for the absence of detectable cleavage products of SREBP in post-mitotic cells could be that S1P might be inactivated during mitosis. To address this possibility, we probed for S1P activity by using ATF6–GFP as a reporter. In contrast to SREBP, which is cleaved by the relocation of S1P into the ER alone (Figure 4B), ATF6 remains intact until ER stress is induced (Shen et al, 2002; Nadanaka et al, 2007). Similar to SREBP, the cleaved transcription factor then enters the nucleus, leading to the activation of ER stress response elements (Ron and Walter, 2007). Therefore, the induction of ER stress by DTT addition to mitotic cells should reveal whether S1P is active in mitosis. We induced the relocation of S1P into the ER before the cells entered mitosis in stably transfected CHO cells that express human ATF6 tagged with GFP under an inducible promoter, as antibodies against endogenous processed ATF6 were not sensitive enough. ATF6–GFP-expressing cells were treated with BFA together with nocodazole to arrest cells in mitosis and monitored by phase-contrast time-lapse microscopy (Figure 5). Cells arrested in mitosis were identified and DTT was added for 45 min to induce ER stress and thereby ATF6–GFP cleavage. The drugs were washed out to allow the cells to exit mitosis (0 min). At the end of the experiments (80 min), the cells were fixed and stained for the nuclear localization of ATF6–GFP. Cells that were treated with DTT during mitosis showed a strong nuclear localization of ATF6–GFP compared to cells that were not exposed to DTT (Figure 5). However, when BFA was omitted and cells were incubated with DTT alone, no nuclear form was observed (middle panel). This suggests that S1P remained active during mitosis. In summary, we conclude that SREBP is protected from unregulated activation because the separation of the ER from the Golgi apparatus is maintained during mitosis.

Figure 5.

S1P remains active in mitotic cells. Stably transfected CHO cells were induced to express ATF6–GFP by induction with 5 M ponasterone A for 16 h. Cells were then arrested in mitosis with 0.2 g/ml nocodazole in the presence or absence of 3 g/ml BFA. The arrested cells were then treated for 45 min with 2 mM DTT or control. The drugs were removed and the cells were followed by time-lapse microscopy (t=0 min) until mitosis was complete (t=80 min). The cells were fixed and stained with antibodies to GFP to detect ATF6–GFP. Note that only cells that entered mitosis in the presence of BFA that were further treated with DTT in mitosis showed cleaved ATF–GFP in the nucleus (+BFA +DTT), whereas in control cells treated with only BFA (+BFA) or DTT (+DTT), ATF6–GFP was restricted to the ER. Scale bar, 10 m.

View full figure (83 KB)

In this study, we explored whether the Golgi-resident protease S1P activates the ER-localized transcription factors SREBPs during cell division. We show that S1P has similar properties as Golgi-resident glycosylation enzymes. Under steady-state interphase conditions, S1P is localized to Golgi membranes, where it binds to the Triton X-100-insoluble Golgi matrix in vitro. After BFA treatment or blocking proteins from leaving the ER, glycosylation enzymes are relocated into the ER. Similarly, S1P slowly but continuously cycles between the Golgi complex and the ER, which is in contrast to matrix proteins that remain on the Golgi apparatus.

SREBPs are transcription factors that are the key regulators of genes for the biogenesis of cholesterol, lipids and membranes. The proteins reside as silent transmembrane precursors in the ER. After stimulation, the proteins leave the ER and are translocated to the Golgi complex by vesicular transport. In the Golgi, they are sequentially cleaved by two proteases S1P and S2P, leading to the release of the active transcription factor from the membrane into the cytosol and subsequent transport into the nucleus. This regulation of activation requires distinct boundaries between the Golgi complex and the ER. Indeed, it is sufficient to process SREBP to the active, nuclear form by relocating S1P into the ER membrane with a KDEL retrieval signal or by relocating the protease into the ER after BFA treatment (DeBose-Boyd et al, 1999).

Previous work analysed the distribution of Golgi-resident glycosylation enzymes by immunofluorescence, showing that under steady-state conditions about 80–90% of the signal was detectable in the Golgi complex and the remaining 10–20% was present in the ER (Rhee et al, 2005; Altan-Bonnet et al, 2006). As S1P shows similar dynamics as glycosylation enzymes, and therefore a significant percentage of the protein would be in the ER, a constant low level of SREBP activation during interphase would be expected. However, we did not see this small amount of cleaved SREBP. This could be explained by a short half-life of the activated transcription factor. Nevertheless, a potential low level of SREBP activation caused by cycling of S1P through the ER might be desirable for cell growth so that a basal activation level of genes involved in lipid biogenesis induced by SREBP might remain constant. This might be especially true, as the SREBP pathway is essential for growth and cell cycle progression (Goldstein and Brown, 1990; Goldstein et al, 2006).

At the beginning of mitosis, the Golgi apparatus disassembles into tubular vesicular fragments, which fuse after successful partitioning to re-form a new Golgi apparatus in both daughter cells (Shorter and Warren, 2002). Similar to the nuclear envelope that is reabsorbed into the ER during mitosis, it has been suggested that Golgi-resident enzymes also relocate into the ER. Because Golgi enzymes are mobile and cycle between the ER and the Golgi complex (Zaal et al, 1999; Altan-Bonnet et al, 2006), and because at the beginning of mitosis exit from the ER is inhibited (Featherstone et al, 1985; Collins and Warren, 1992), Golgi enzymes may accumulate in the ER. Although the inhibition of ER exit during interphase caused the relocation of S1P into the ER, we found no evidence for the presence of S1P in the ER membranes during mitosis. Moreover, S1P showed a similar staining pattern as GM130, decorating mitotic Golgi fragments in proximity to the spindle poles during all stages of mitosis and a smaller pool was found dispersed throughout the cytoplasm. The mitotic Golgi signal has been proposed to be generated by dispersed Golgi membranes that are below the optical resolution (Jesch et al, 2001; Jokitalo et al, 2001; Axelsson and Warren, 2004; Pecot and Malhotra, 2004), whereas other reports suggested that the signal represents Golgi enzymes that are dispersed throughout the ER (Zaal et al, 1999; Altan-Bonnet et al, 2006). We directly addressed this issue by gradient fractionation of mitotic membranes. The majority of S1P co-migrated with GM130 and remained distinct from the ER-containing fraction. S1P was found in the ER fraction when the Golgi enzymes were shifted into the ER by BFA before the cells entered mitosis. Second, only the intact precursor of SREBP was detected by immunoblotting in interphase and mitotic cells under high sterol conditions, whereas the processed and cleaved form was detected only after incubation with BFA. If cells enter mitosis in the presence of BFA, however, an increased amount of SREBP was cleaved, arguing that S1P remains active during mitosis.

Previous studies to determine whether Golgi enzymes move into the ER during mitosis gave opposite conclusions. The approaches relied on the overexpression of chimaeric proteins (Pecot and Malhotra, 2004) in combination with temperature alterations to trap the molecules in the ER during mitosis (Altan-Bonnet et al, 2006). Our approach addressed the physiological relevance of whether the Golgi enzymes remain independent of the ER. In principle, relocation of S1P into the ER membrane could cleave and activate SREBP already during cell division. This would ensure rapid transcriptional activation of genes for cholesterol, lipid and membrane biogenesis to allow the growth of organelles in G1. The fact that S1P remained distinct from the ER and that we did not detect activation of SREBP during mitosis suggest a mechanism that specifically activates the transcription factors during G1 when the cells need to grow in preparation for the next cell division.

From these experiments, we showed that the spatial separation between S1P in the Golgi apparatus and its substrate SREBP remains intact in the ER during mitosis to prevent an inappropriate activation of these transcription factors. This mechanism seems sufficient to block SREBP activation, because mixing of S1P with SREBP alone leads to SREBP cleavage.

Reagents and antibodies

The following antibodies were used (kindly provided by those listed below): monoclonal antibodies against actin and -tubulin (Sigma, St Louis, MO), GM130, Gos28, Golgin84, GRP78/BiP (BD Biosciences, San Jose, CA) and SREBP-2 (7D4; DeBose-Boyd et al, 1999). Polyclonal antibodies: affinity-purified S1P (DeBose-Boyd et al, 1999), SREBP-2 (raised in rabbits against amino acids 32–250 of hamster SREBP-2), phospho-histone H3 (Ser10) (Millipore, Billerica, MA), GM130 (Martin Lowe), Golgin84 (Ayano Satoh), anti-ManII (Graham Warren), p24 (Tommy Nielsson), GFP (Pelletier et al, 2002). AlexaFluor 488-conjugated goat anti-rabbit antibodies, AlexaFluor 594-conjugated goat anti-mouse antibodies, Hoechst 33342, cascade-blue-conjugated BSA and ponasterone A were from Invitrogen (Carlsbad, CA). BFA was from Epicenter Technologies (Madison, WI), and Mowiol 4-88 and protease inhibitor cocktail 3 were obtained from Calbiochem (La Jolla, CA). Fortified cosmic calf serum and fetal bovine serum were from Hyclone (Logan, UT). DTT, nocodazole and cholesterol were obtained from Sigma. All other reagents were from Calbiochem, Invitrogen and Sigma. Recombinant His-tagged Sar1^DN protein was expressed in Escherichia coli and isolated on Ni-NTA agarose (Qiagen, Valencia, CA) as described (Rowe and Balch, 1995). The buffer was exchanged by dialysis against 25 mM Hepes pH 7.2, 125 mM potassium acetate, 1 mM MgCl₂, 1 mM glutathione, 10 M GDP and 50 M EGTA.

Cell culture

NRK cells were maintained in Dulbecco's modified Eagle's medium supplemented with 10% fortified cosmic calf serum (Hyclone), 2 mM L-glutamine, 100 U/ml penicillin and 100 g/ml streptomycin at 37°C in a standard tissue culture incubator with 5% CO₂. CHO cells were grown in medium A (1:1 mixture of Ham's F-12 and Dulbecco's modified Eagle's medium supplemented with 100 g/ml streptomycin sulphate and 100 U/ml penicillin) supplemented with 5% fetal calf serum (Hyclone). Sterol-depleting medium contains medium A supplemented with 5% newborn calf lipoprotein-deficient serum, 50 M sodium mevalonate and 10 M compactin (prepared as described by Sun et al, 2005).

To establish a CHO cell line that stably expresses GFP–ATF6 gene in an inducible system, we used pIND and pVgRXR vectors, which are components of an ecdysone-inducible mammalian expression system (Invitrogen). pIND-GFP-ATF6 encodes full-length human ATF6 fused with GFP at the N terminus driven by the ecdysone-inducible promoter (G418 resistant) and pVgRXR encodes ecdysone receptor (zeocin resistant). Expression of ATF6–GFP was induced with 5 M ponasterone A for 14–16 h before the experiments.

Microinjection and immunofluorescence analysis

Capillary microinjection was performed with a Transjector 5246 and a Micromanipulator 5171 (Eppendorf, Westbury, NY). NRK cells grown on glass coverslips were injected into the cytoplasm with 1 mg/ml Sar1^DN together with 5 mg/ml cascade-blue-conjugated BSA to mark injected cells. For immunofluorescence analysis, cells were fixed and permeabilized for 10 min in methanol at -20°C or first fixed in 3.7% formaldehyde in PBS followed by permeabilization in methanol at -20°C. The cells were then incubated with the indicated primary antibodies followed by secondary antibodies conjugated to Alexa fluorophores (goat anti-mouse AlexaFluor 594 and goat anti-rabbit AlexaFluor 488). DNA was stained with Hoechst 33342 and the cells were mounted in Mowiol 4-88. Fluorescence analysis was performed with an LD Plan-Neofluar 40/1.3 DIC or a Plan-Apochromat 63/1.4 DIC objective (Zeiss, Thornwood, NY). For phase-contrast time-lapse microscopy, cells were observed at 37°C in CO₂-independent medium (Invitrogen) supplemented with 10% fortified cosmic calf serum (Hyclone) with an A-PLAN 10/0.25 PH1 objective (Zeiss) and an Axiovert 200M microscope equipped with a temperature incubation chamber (Zeiss). Fluorophores were exited with an XBO75 xenon lamp (Zeiss) in combination with the following fluorescence filter sets from Chroma Technology (Rockingham, VT): 31000 for Hoechst and cascade-blue illumination, 41001 for AlexaFluor488 and 41004 for AlexaFluor594. Images were captured with an Orca-285 camera (Hamamatsu, Hamamatsu City, Japan) and the software package Openlab 4.02 (Improvision, Lexington, MA).

Cell fractionation

For preparation of total lysates, cells of one 150 mm dish were washed twice with PBS on ice and then scraped into 5 ml of ice-cold PBS. The cells were pelleted (500 g for 5 min at 4°C) and resuspended in 1 ml of buffer A (250 mM sucrose, 20 mM tricine pH 7.8, proteinase inhibitors). After incubation for 20 min on ice, cells were homogenized with a Dounce homogenizer (20 strokes). The protein concentrations of the lysates were determined and proteins were dissolved in SDS sample buffer containing 2% -mercaptoethanol, separated by SDS–PAGE and processed for immunoblotting. Triton X-100 extraction experiments were performed as described previously (Slusarewicz et al, 1994). Golgi membranes were purified from rat liver as described (Wang et al, 2006).

Preparation of mitotic cells

Cells were seeded at day 0 in growth medium at a confluence of 70%. At day 1, the medium was changed to growth medium supplemented with either 5 M S-trityl-L-cysteine (Acros Organics, Morris Plains, NJ) or 0.2 g/ml nocodazole to enrich cells in mitosis. After 4–6 h, the mitotic cells were collected by shake-off and the remaining attached cells were washed three times with growth media and then collected by trypsinization (interphase cells).

Gradient centrifugation

Mitotic cells from two confluent T175 flasks were collected by shake-off, pelleted by centrifugation (5 min at 500 g) and resuspended in 12 ml of buffer B (10 mM Hepes pH 7.4, 250 mM sucrose, 1 mM EDTA, protease inhibitors) and pelleted again. The cells were resuspended in 1 ml of buffer C (10 mM Hepes pH 7.4, 50 mM NaCl, 1 mM EDTA) and broken using a ball bearing homogenizer with a clearance of 10 m (Isobiotec, Heidelberg, Germany). The homogenate was centrifuged for 5 min at 1000 g at 4°C and the PCS was recovered. For gradient centrifugation, 250 l of an 80% sucrose solution was applied to the bottom of an SW50.1 tube. The sucrose was overlaid with 4.3 ml of a 30–5% glycerol solution in buffer D (10 mM Hepes pH 7.4, 1 mM EDTA) and finally topped with PCS (900 l). The gradients were centrifuged for 20 min in an SW50.1 rotor at 114 562 g_av and 4°C. A total of 12 fractions (450 l each) were collected from the top, diluted in buffer D and membranes were recovered by centrifugation at 71 680 g_av for 30 min at 4°C. The membrane pellets were dissolved in 2 SDS sample buffer (2% -mercaptoethanol) and analysed by immunoblotting.

We are indebted to Michael Brown and Joseph Goldstein for advice, support and reagents. We thank Graham Warren for critical reading of the manuscript, Jin Ye for providing ATF6–GFP-expressing CHO cells, and Martin Lowe, Tommy Nilsson and Ayano Satoh for antibodies. JS is a Virginia Murchison Lithicum Scholar in Medical Research.

Altan-Bonnet N, Sougrat R, Liu W, Snapp EL, Ward T, Lippincott-Schwartz J (2006) Golgi inheritance in mammalian cells is mediated through endoplasmic reticulum export activities. Mol Biol Cell 17: 990–1005 | Article | PubMed | ChemPort |

Axelsson MA, Warren G (2004) Rapid, endoplasmic reticulum-independent diffusion of the mitotic Golgi haze. Mol Biol Cell 15: 1843–1852 | Article | PubMed | ChemPort |

Brown MS, Ye J, Rawson RB, Goldstein JL (2000) Regulated intramembrane proteolysis: a control mechanism conserved from bacteria to humans. Cell 100: 391–398 | Article | PubMed | ISI | ChemPort |

Cole NB, Sciaky N, Marotta A, Song J, Lippincott-Schwartz J (1996) Golgi dispersal during microtubule disruption: regeneration of Golgi stacks at peripheral endoplasmic reticulum exit sites. Mol Biol Cell 7: 631–650 | PubMed | ISI | ChemPort |

Collins RN, Warren G (1992) Sphingolipid transport in mitotic HeLa cells. J Biol Chem 267: 24906–24911 | PubMed | ISI | ChemPort |

Dai J, Sultan S, Taylor SS, Higgins JM (2005) The kinase haspin is required for mitotic histone H3 Thr 3 phosphorylation and normal metaphase chromosome alignment. Genes Dev 19: 472–488 | Article | PubMed | ISI | ChemPort |

DeBose-Boyd RA, Brown MS, Li WP, Nohturfft A, Goldstein JL, Espenshade PJ (1999) Transport-dependent proteolysis of SREBP: relocation of site-1 protease from Golgi to ER obviates the need for SREBP transport to Golgi. Cell 99: 703–712 | Article | PubMed | ISI | ChemPort |

Espenshade PJ, Hughes AL (2007) Regulation of sterol synthesis in eukaryotes. Annu Rev Genet 41: 401–427 | Article | PubMed | ChemPort |

Featherstone C, Griffiths G, Warren G (1985) Newly synthesized G protein of vesicular stomatitis virus is not transported to the Golgi complex in mitotic cells. J Cell Biol 101: 2036–2046 | Article | PubMed | ISI | ChemPort |

Gillingham AK, Munro S (2003) Long coiled-coil proteins and membrane traffic. Biochim Biophys Acta 1641: 71–85 | Article | PubMed | ISI | ChemPort |

Goldstein JL, Brown MS (1990) Regulation of the mevalonate pathway. Nature 343: 425–430 | Article | PubMed | ISI | ChemPort |

Goldstein JL, Debose-Boyd RA, Brown MS (2006) Protein sensors for membrane sterols. Cell 124: 35–46 | Article | PubMed | ISI | ChemPort |

Hampton RY (2000) Cholesterol homeostasis: ESCAPe from the ER. Curr Biol 10: R298–R301 | Article | PubMed | ChemPort |

Jesch SA, Linstedt AD (1998) The Golgi and endoplasmic reticulum remain independent during mitosis in HeLa cells. Mol Biol Cell 9: 623–635 | PubMed | ISI | ChemPort |

Jesch SA, Mehta AJ, Velliste M, Murphy RF, Linstedt AD (2001) Mitotic Golgi is in a dynamic equilibrium between clustered and free vesicles independent of the ER. Traffic 2: 873–884 | Article | PubMed | ISI | ChemPort |

Jokitalo E, Cabrera-Poch N, Warren G, Shima DT (2001) Golgi clusters and vesicles mediate mitotic inheritance independently of the endoplasmic reticulum. J Cell Biol 154: 317–330 | Article | PubMed | ISI | ChemPort |

Munro S, Pelham HR (1987) A C-terminal signal prevents secretion of luminal ER proteins. Cell 48: 899–907 | Article | PubMed | ISI | ChemPort |

Nadanaka S, Okada T, Yoshida H, Mori K (2007) Role of disulfide bridges formed in the luminal domain of ATF6 in sensing endoplasmic reticulum stress. Mol Cell Biol 27: 1027–1043 | Article | PubMed | ChemPort |

Nakamura N, Rabouille C, Watson R, Nilsson T, Hui N, Slusarewicz P, Kreis TE, Warren G (1995) Characterization of a cis-Golgi matrix protein, GM130. J Cell Biol 131: 1715–1726 | Article | PubMed | ISI | ChemPort |

Pecot MY, Malhotra V (2004) Golgi membranes remain segregated from the endoplasmic reticulum during mitosis in mammalian cells. Cell 116: 99–107 | Article | PubMed | ISI | ChemPort |

Pelletier L, Stern CA, Pypaert M, Sheff D, Ngo HM, Roper N, He CY, Hu K, Toomre D, Coppens I, Roos DS, Joiner KA, Warren G (2002) Golgi biogenesis in Toxoplasma gondii. Nature 418: 548–552 | Article | PubMed | ISI | ChemPort |

Radhakrishnan A, Ikeda Y, Kwon HJ, Brown MS, Goldstein JL (2007) Sterol-regulated transport of SREBPs from endoplasmic reticulum to Golgi: oxysterols block transport by binding to Insig. Proc Natl Acad Sci USA 104: 6511–6518 | Article | PubMed | ChemPort |

Rawson RB (2002) Regulated intramembrane proteolysis: from the endoplasmic reticulum to the nucleus. Essays Biochem 38: 155–168 | PubMed | ISI | ChemPort |

Rhee SW, Starr T, Forsten-Williams K, Storrie B (2005) The steady-state distribution of glycosyltransferases between the Golgi apparatus and the endoplasmic reticulum is approximately 90:10. Traffic 6: 978–990 | Article | PubMed | ChemPort |

Ron D, Walter P (2007) Signal integration in the endoplasmic reticulum unfolded protein response. Nat Rev Mol Cell Biol 8: 519–529 | Article | PubMed | ChemPort |

Rowe T, Balch WE (1995) Expression and purification of mammalian Sar1. Methods Enzymol 257: 49–53 | PubMed | ISI | ChemPort |

Seemann J, Jokitalo E, Pypaert M, Warren G (2000) Matrix proteins can generate the higher order architecture of the Golgi apparatus. Nature 407: 1022–1026 | Article | PubMed | ISI | ChemPort |

Seemann J, Pypaert M, Taguchi T, Malsam J, Warren G (2002) Partitioning of the matrix fraction of the Golgi apparatus during mitosis in animal cells. Science 295: 848–851 | Article | PubMed | ISI | ChemPort |

Shen J, Chen X, Hendershot L, Prywes R (2002) ER stress regulation of ATF6 localization by dissociation of BiP/GRP78 binding and unmasking of Golgi localization signals. Dev Cell 3: 99–111 | Article | PubMed | ISI | ChemPort |

Short B, Haas A, Barr FA (2005) Golgins and GTPases, giving identity and structure to the Golgi apparatus. Biochim Biophys Acta 1744: 383–395 | Article | PubMed | ISI | ChemPort |

Shorter J, Warren G (2002) Golgi architecture and inheritance. Annu Rev Cell Dev Biol 18: 379–420 | Article | PubMed | ISI | ChemPort |

Slusarewicz P, Nilsson T, Hui N, Watson R, Warren G (1994) Isolation of a matrix that binds medial Golgi enzymes. J Cell Biol 124: 405–413 | Article | PubMed | ISI | ChemPort |

Storrie B, White J, Röttger S, Stelzer EH, Suganuma T, Nilsson T (1998) Recycling of Golgi-resident glycosyltransferases through the ER reveals a novel pathway and provides an explanation for nocodazole-induced Golgi scattering. J Cell Biol 143: 1505–1521 | Article | PubMed | ISI | ChemPort |

Sun LP, Li L, Goldstein JL, Brown MS (2005) Insig required for sterol-mediated inhibition of Scap/SREBP binding to COPII proteins in vitro. J Biol Chem 280: 26483–26490 | Article | PubMed | ChemPort |

Sun LP, Seemann J, Goldstein JL, Brown MS (2007) Sterol-regulated transport of SREBPs from endoplasmic reticulum to Golgi: Insig renders sorting signal in Scap inaccessible to COPII proteins. Proc Natl Acad Sci USA 104: 6519–6526 | Article | PubMed | ChemPort |

Wang Y, Taguchi T, Warren G (2006) Purification of rat liver Golgi stacks. In Cell Biology: A Laboratory Handbook, Celis J (ed), Vol. 2, 3rd edn, pp 33–39. San Diego: Elsevier Science (USA)

Yang T, Espenshade PJ, Wright ME, Yabe D, Gong Y, Aebersold R, Goldstein JL, Brown MS (2002) Crucial step in cholesterol homeostasis: sterols promote binding of SCAP to INSIG-1, a membrane protein that facilitates retention of SREBPs in ER. Cell 110: 489–500 | Article | PubMed | ISI | ChemPort |

Zaal KJ, Smith CL, Polishchuk RS, Altan N, Cole NB, Ellenberg J, Hirschberg K, Presley JF, Roberts TH, Siggia E, Phair RD, Lippincott-Schwartz J (1999) Golgi membranes are absorbed into and reemerge from the ER during mitosis. Cell 99: 589–601 | Article | PubMed | ISI | ChemPort |

Main navigation

• Search EMBO Journal

  Search

• Advanced search

• The EMBO Journal

  • Home
  • Aims and scope
  • Current issue
  • Advance online publication
  • Web focuses
  • Archive:
  • Browse by issue
  • Browse by subject
  • Browse by category
  • Free online sample issue
  • Press releases

• Authors and Referees

  • Editorial process
  • Guide for authors
  • Submit an article
  • Guide for referees
  • Editors and Editorial Board
  • Contact editorial office

• Customer Service

  • Subscribe
  • Order sample copy
  • Purchase articles
  • Reprints and permissions
  • Contact NPG
  • Advertising