Participation of a fusogenic protein, glyceraldehyde-3-phosphate dehydrogenase, in nuclear membrane assembly.

We found an autoimmune serum, K199, that strongly suppresses nuclear membrane assembly in a cell-free system involving a Xenopus egg extract. Four different antibodies that suppress nuclear assembly were affinity-purified from the serum using Xenopus egg cytosol proteins. Three proteins recognized by these antibodies were identified by partial amino acid sequencing to be glyceraldehyde-3-phosphate dehydrogenase (GAPDH), fructose-1,6-bisphosphate aldolase, and the regulator of chromatin condensation 1. GAPDH is known to be a fusogenic protein. To verify the participation of GAPDH in nuclear membrane fusion, authentic antibodies against human and rat GAPDH were applied, and strong suppression of nuclear assembly at the nuclear membrane fusion step was observed. The nuclear assembly activity suppressed by antibodies was recovered on the addition of purified chicken GAPDH. A peptide with the sequence of amino acid residues 70-94 of GAPDH, which inhibits GAPDH-induced phospholipid vesicle fusion, inhibited nuclear assembly at the nuclear membrane fusion step. We propose that GAPDH plays a crucial role in the membrane fusion step in nuclear assembly in a Xenopus egg extract cell-free system.

We found an autoimmune serum, K199, that strongly suppresses nuclear membrane assembly in a cell-free system involving a Xenopus egg extract. Four different antibodies that suppress nuclear assembly were affinity-purified from the serum using Xenopus egg cytosol proteins. Three proteins recognized by these antibodies were identified by partial amino acid sequencing to be glyceraldehyde-3-phosphate dehydrogenase (GAPDH), fructose-1,6-bisphosphate aldolase, and the regulator of chromatin condensation 1. GAPDH is known to be a fusogenic protein. To verify the participation of GAPDH in nuclear membrane fusion, authentic antibodies against human and rat GAPDH were applied, and strong suppression of nuclear assembly at the nuclear membrane fusion step was observed. The nuclear assembly activity suppressed by antibodies was recovered on the addition of purified chicken GAPDH. A peptide with the sequence of amino acid residues 70 -94 of GAPDH, which inhibits GAPDH-induced phospholipid vesicle fusion, inhibited nuclear assembly at the nuclear membrane fusion step. We propose that GAPDH plays a crucial role in the membrane fusion step in nuclear assembly in a Xenopus egg extract cell-free system.
The nuclear envelope (NE) 1 of eukaryotes is composed of inner and outer nuclear membranes, nuclear pore complexes, and the nuclear lamina. NE breakdown during the mitotic prophase results in the dispersal of both nuclear membranes into the mitotic endoplasmic reticulum (ER) network. Inner nuclear membrane proteins such as lamin B receptor, laminaassociated polypeptide 2␤, and emerin reconcentrate on the surface of decondensing chromatin during the late anaphase, and then the NE is re-formed (1,2). These intrinsic membrane proteins and lamins are believed to play a critical role in NE reassembly (3)(4)(5). NE assembly can be studied in vitro using extracts of meiotic or mitotic cells (6,7). The assembly requires cytosolic factors and is inhibited by a non-hydrolyzable GTP analogue, N-ethylmaleimide, and a calcium chelator, BAPTA (1,2-bis(2-aminophenoxy)ethane-N,N,NЈ,NЈ-tetraacetic acid) (8). In vitro, NE assembly is initiated by the binding of membrane vesicles to decondensed chromatin in an energy-independent manner (9 -11). Once bound to chromatin, membrane vesicles fuse and flatten. The insertion of nuclear pore complexes and the expansion of the NE require both energy and cytosolic components (12). The fusion of vesicles on chromatin is inhibited by a non-hydrolyzable GTP analogue (8,13) and requires Ran GTPase and a regulator of chromatin condensation 1, RCC1/Ran GTP exchange factor (14 -16). Remarkably, agarose beads coated with Ran allow the assembly of a nuclear pore complex-containing NE (17,18). After the formation of a closed NE around chromatin, further NE growth requires nucleocytoplasmic transport and further membrane fusion (8,19). Analysis of membrane fractions of a Xenopus egg extract suggested that the population of vesicles required for NE assembly is not uniform and that different membranes might be required for different steps of the assembly process (9, 20 -22).
It was shown recently that p97, an ATPase previously implicated in the fusion of the Golgi and transitional endoplasmic reticulum membranes together with adaptor p47, has two discrete functions in NE assembly (23). The formation of a closed NE requires a p97⅐Ufd1⅐Npl4 complex and subsequent NE growth involves a p97⅐p47 complex (23). Current evidence suggests that trans-SNARE complexes implicating p97 might be not the terminal catalysts of membrane fusion (24). Rather, their role seems to be confined to an intermediate stage of the reaction, probably docking of the two lipid bilayers (24). Therefore, the components causing lipid bilayer fusion events remain unclear.
NE assembly, which includes the lipid bilayer fusion step, comprises complicated reactions, and many proteins should participate in these reactions. Therefore, if we could obtain specific antibodies against these proteins, we would have very powerful tools for analyzing the reaction mechanism. It is known that autoimmune disease sera react with many proteins in the NE (25)(26)(27) and cytosol fractions. We applied such sera to an in vitro Xenopus egg extract nuclear assembly system and found that most sera suppressed the assembly to various extents. Then we selected one serum, K199, that strongly suppresses the nuclear assembly and contains antibodies reactive with proteins in the Xenopus egg extract.
In this study, we purified five antibodies from K199 that suppress nuclear assembly. They reacted with GAPDH, fructose-1,6-bisphosphate aldolase, RCC1, and another protein.
GAPDH is known to be a fusogenic protein (i.e. a protein having the ability to merge phospholipid bilayers) in phospholipid vesicles containing phosphatidylserine (28 -32). Therefore, we examined the effects of a peptide comprising the phosphatidylserine-binding site of GAPDH, which inhibits GAPDH-induced phospholipid vesicle fusion, and authentic antibodies against GAPDH on the nuclear assembly in a Xenopus egg extract. Suppression of the nuclear membrane fusion step was clearly caused by the peptide and antibodies. From these results we propose that GAPDH participates in the step of fusion of nuclear membrane vesicles on chromatin.

MATERIALS AND METHODS
Autoimmune Serum, Antibodies, Peptides, and Enzymes-K199 is serum from a patient with Sjögren's syndrome (Sjs), rheumatoid arthritis, and systemic lupus erythematous overlap. The IgG fraction of anti-human GAPDH (rabbit) was purchased from Trevigen and used without further purification. Polyclonal antibodies against a 15-amino acid synthetic peptide (FNFRLKAGQKIRFGC, amino acid residues 390 -403) derived from Chinese hamster ovary cell phosphatidylserine decarboxylase (PSD) was generated in rabbits as previously described (28). The one thus-obtained polyclonal antibody, designated as aPSD-2, was further purified by affinity chromatography on a peptide column (28). Polyclonal antibodies against a 25-amino acid synthetic peptide, GAPDH70 (KPITIFQERDPVKIKWGDAGAEYVC, the sequence of amino acid residues 70 -94 of rat GAPDH), was generated in rabbits as previously described (28) and purified by affinity chromatography on a peptide column (28). Chicken GAPDH was obtained from Sigma (Tokyo).
Preparation of Demembranated Sperm Chromatin-Demembranated sperm chromatin was prepared as described (33) and stored at Ϫ80°C at a concentration of 40,000/l.
Preparation of a Crude Extract and Membrane and Cytosol Fractions of Xenopus Eggs-Xenopus eggs were collected, removed from the jelly, and then lysed to prepare an interphase extract essentially as described (22). The egg lysis buffer for the preparation of interphase extracts consisted of 50 mM HEPES-KOH (pH 7.7), 250 mM sucrose, 50 mM KCl, and 2.5 mM MgCl 2 supplemented with 2 mM 2-mercaptoethanol, 10 g/ml aprotinin and leupeptin immediately before use. Eggs were packed into tubes by brief centrifugation and then crushed by centrifugation at 15,000 ϫ g for 10 min. The crude extract was collected, mixed with 10 g/ml cytochalasin B, and then further separated into cytosol, membrane-rich, and gelatinous pellet fractions by ultracentrifugation at 200,000 ϫ g for 4 h. The cytosol fraction generated was then recentrifuged at 200,000 ϫ g for 30 min to remove residual membranes, divided into 30-l aliquots, frozen in liquid nitrogen, and stored at Ϫ80°C until use. The membrane fraction was resuspended in 10 volumes of cold lysis buffer supplemented with 1 mM dithiothreitol and 5 g/ml aprotinin and leupeptin and then centrifuged at 200,000 ϫ g for 1 h, with the resuspension and centrifugation steps repeated twice to remove residual cytosolic components. The membranes were resuspended in egg lysis buffer containing 500 mM sucrose to a final volume corresponding to ϳ10% of the original volume of crude extract (10ϫ membranes), divided into 10-l aliquots, frozen in liquid nitrogen, and stored Ϫ80°C until use.
Nuclear Assembly Assays-Assays were performed essentially as described by Smythe and Newport (34) and Sasagawa et al. (22). Frozen Xenopus egg extracts or cytosol and membrane fractions were rapidly thawed, supplemented with an ATP-regenerating system (10 mM phosphocreatine, 2 mM ATP (pH 7.0), and 5 g/ml creatine kinase), and then mixed with demembranated sperm chromatin. The standard reaction mixture consisted of 10,000 chromatins and 10 l of crude extract or 10 l of cytosol ϩ 1 l of membrane. After incubation at room temperature (23°C) for 1.5 h, a 2-l aliquot of the reaction mixture was removed and diluted with 2 l of Hoechst/DHCC buffer (15 mM PIPES-KOH (pH 7.4), 0.2 M sucrose, 7 mM MgCl 2 , 80 mM KCl, 15 mM NaCl, 5 mM EDTA containing 20 mg/ml bisbenzimide DNA dye (Hoechst 33342; Calbiochem-Novabiochem), a lipid dye, 3,3Ј-dihexyloxacarbocyanine iodide (DHCC; Aldrich), and 3.7% formaldehyde) on a glass slide. A fixed sample covered with a coverslip was examined by phase contrast and fluorescence microscopy. An Axioplan (Carl Zeiss, Germany)-equipped fluorescence microscope with exciter-barrier reflector combinations suitable for these fluorescent dyes was used. Nuclear assembly (%) represents 100 ϫ (number of nuclei)/(number of observed chromatins). Chromatin, of which the axial ratio is less than 2, was taken to be "nucleus-assembled." For confocal microscopy, propidium iodide DNA dye and DHCC lipid dye were used, and images were recorded with a Radiance 2000 confocal fluorescent microscopy system (Bio-Rad) mounted on a Nikon E600 fluorescence microscope (Nikon, Tokyo). In the case of Fig. 2A, assembled nuclei were subjected to electron microscopy after fixation with glutaraldehyde (22). The bars in the figure show the standard error.
Pretreatment of Crude Extracts and Membrane Fractions with Antibodies-Xenopus egg crude extracts were incubated for 90 min at 4°C with a 0.05 volume of K199 serum, eluted antibodies, or control phosphate-buffered saline (PBS) (137 mM NaCl, 8.10 mM Na 2 HPO 4 , 2.68 mM KCl, and 1.47 mM KH 2 PO 4 ) before the addition of sperm chromatin. Nuclear assembly assays were then performed, and "nuclear assembly" (100 ϫ (formed nuclei)/(observed chromatin) (%)) was measured at several time points by phase contrast and fluorescence microscopy.
SDS-PAGE, Silver Staining, and Immunoblotting-SDS-PAGE and silver staining were carried out according to the methods of Laemmli (35) and Morrissey (36), respectively. For immunoblotting, samples were subjected to SDS-PAGE (8% gel) and then electrophoretically transferred to polyvinylidene difluoride (PVDF) papers (Millipore Co.) at 2 mA/cm 2 for 2 h with a semidry blotting apparatus in transfer buffer (100 mM Tris, 192 mM glycine, 5% methanol, 0.005% SDS). The PVDF papers were then cut into strips. The strips were subjected to sequential 3-h incubations at room temperature with diluted K199 serum followed with horseradish peroxidase-conjugated anti-human IgG antibodies (Wako Pure Chemicals) in blocking buffer 1 (5% skim milk in PBS). The primary antibodies were used at a 1:500 dilution, and the secondary ones were used at a 1:2000 dilution. Bound antibodies were detected either by development with PBS containing 0.2 mM 3,3Ј-diaminobenzidine tetrahydrochloride and 0.03% H 2 O 2 or by means of the enhanced chemiluminescence method (SuperSignal West Pico kit; Pierce). The apparent molecular masses of proteins were estimated from their mobilities on SDS-PAGE using molecular mass standard proteins (myosin, phosphorylase b, bovine serum albumin, and ovalbumin).
Affinity Purification of Antibodies-The method used was originally developed by Olmsted (37). Cytosol fractions and the subfractions obtained with a DEAE-5PW anion exchange column (Tosoh, Tokyo) or a Superdex 200HR gel filtration column (Amersham Biosciences) were subjected to SDS-PAGE (8% gel) and then electrotransferred or transferred by diffusion to PVDF papers. The PVDF papers were then blocked at 4°C for 1 h with blocking buffer 2 (3% bovine serum albumin in PBS). The papers were then incubated with 1% K199 serum in blocking buffer 2 at 4°C overnight, washed three times in PBS and water, and then cut into several portions according to the molecular weight ranges. Antibodies were eluted by incubating these portions for 3 min at 4°C in 50 mM NaCl and 5 mM glycine (pH 2.8) followed by immediate neutralization with 500 mM Na 2 HPO 4 containing 5 g/ml bovine serum albumin. Each eluate was dialyzed for 8 h three times against 100 mM NH 4 OH-CH 3 COOH (pH 8.0), quickly frozen, and then lyophilized. The lyophilized affinity-purified antibodies were finally dissolved in PBS and used at 1:5 to 1:20 dilution to examine nuclear assembly inhibition.
DEAE-5PW Anion Exchange Chromatography-An aliquot of the Xenopus egg cytosol fraction (6 mg of protein) diluted with 6 ml of 25 mM ammonium acetate (pH 8.3) was applied to a DEAE-5PW anion exchange column (7.5 ϫ 75 mm, Tosoh, Japan) equilibrated with 25 mM ammonium acetate (pH 8.3) and then eluted with a 220-min stepwise gradient, from 0 to 100%, of 1 M ammonium acetate (pH 8.3). Analysis was performed at 0.5 ml/min and room temperature. The separated fractions were either lyophilized and solubilized with 10% SDS in 100 mM Tris-HCl (pH 6.8) for SDS-PAGE or subjected to Superdex 200HR gel filtration chromatography. The amount of protein derived from 6 mg of cytosol protein was defined as 1 unit.
Superdex 200HR Gel Filtration Chromatography-Fractions obtained with a DEAE-5PW column were dialyzed for 8 h three times against 100 mM NH 4 OH-CH 3 COOH (pH 8.0) and then concentrated by either re-application to a small DEAE-5PW column or lyophilization. The concentrated samples were applied to a Superdex 200HR 10/30 column (Amersham Biosciences) equilibrated with 0.1 M ammonium acetate (pH 8.3). Fractions were lyophilized and solubilized with 10% SDS in 100 mM Tris-HCl (pH 6.8) for SDS-PAGE.
Partial Amino Acid Sequencing-The Xenopus egg cytosol fraction (11 ml, 204 mg of protein) was separated on a DEAE-5PW column, and the fractions obtained were further separated by Superdex 200HR gel filtration chromatography. The fractions were electrophoresed, stained with Coomassie Brilliant Blue R-250, and gel slices containing protein bands of interest (about 8 g each) were subjected to trypsin digestion as described previously (38). The slices containing peptides were centrifuged after homogenization. Each supernatant was collected, concentrated by lyophilization, and then applied to a reversed-phase HPLC equipped with a silica base C8 column (4.6 ϫ 250 mm, Capcel Pak C8 column; Shiseido, Tokyo). Peptides were eluted with a linear gradient of 5-75% acetonitrile containing 0.1% trifluoroacetic acid at a flow rate of 0.5 ml/min. The amino acid sequences of the isolated peptides were determined with a Protein Sequencer 470A (Applied Biosystems).
Preparation of Fab Fragments of Anti-GAPDH-Anti-GAPDH was digested in 20 mM phosphate buffer (pH 7.0) and 10 mM EDTA containing 10 mM cysteine with immobilized papain (0.25 mg/ml papain crosslinked to 6% beaded agarose; Pierce) at 37°C for 3 h. The reaction time for complete conversion of all antibodies to Fab fragments was determined in preliminary experiments. The resulting digest was dialyzed against 50 mM ammonium bicarbonate, lyophilized, and then dissolved in PBS.

RESULTS
Characterization of Autoimmune Serum K199 -It is known that a subset of patients with autoimmune diseases have autoantibodies against nuclei or nuclear envelopes. Thus, we examined whether or not such sera inhibit nuclear assembly in a Xenopus egg extract. Aliquots of a Xenopus egg extract were mixed with autoimmune sera and then subjected to the nuclear assembly assay. Most sera inhibited nuclear assembly to various extents. Sera showing strong inhibition and adequate for use as probes for studying nuclear assembly were found in cases of Sjs and other diseases. 2 In this study, we used the thus found serum, K199, which strongly suppresses nuclear assembly.
A Xenopus egg crude extract was preincubated with control PBS, normal serum, or K199, and then nuclear assembly was examined with Xenopus sperm chromatin (Fig. 1A). When the egg extract was pretreated with K199 serum, nuclear assembly was strongly suppressed, although pretreatment with normal serum gave similar nuclear assembly activity to that of a control (Fig. 1A). When these samples were observed under a fluorescence microscope, chromatin was observed to have changed into spherical nuclei and to be surrounded by a continuous membrane stained with DHCC in egg extracts pretreated with control PBS and normal serum (Fig. 1B, 1 and 2, Lipid). However, chromatin did not change into spherical nuclei, although it became swollen in an egg extract pretreated with K199 (Fig. 1B, 3 and 4). The rim (Fig. 1B, 3, Lipid) and surface (Fig. 1B, 4, Lipid) of chromatin were stained discontinuously with DHCC. These data suggested that K199 inhibits the fusion of vesicles bound to chromatin. The thus-assembled nuclei and chromatin were observed under electron and confocal laser fluorescence microscopes (Fig. 2). Nuclei assembled with the control extract ( Fig. 2A, a and e) and the extract preincubated with normal serum ( Fig. 2A, b and f) were covered with a bi-layered nuclear membrane containing nuclear pore complexes. On the other hand, in the case of chromatin incubated with the K199-treated egg extract, the fusion of vesicles bound to chromatin was limited ( Fig. 2A, c, d, g, and h). These results show that K199 inhibits the fusion of vesicles on chromatin during nuclear membrane assembly. The binding of vesicles to chromatin may also be suppressed partially by the serum (Fig. 2A, c and d). Samples treated by the same method were also observed under a confocal laser fluorescence microscope (Fig. 2B). When we focused on the surface of chromatin incubated with the K199-treated extract, it was shown that the nuclear envelope was not enclosed, and thus, "vesicular" or "reticular" lipid-staining on chromatin could be determined by this method. Therefore, we used confocal microscopy to verify the inhibition of the closed nuclear envelope formation in the following experiments. Then we attempted to purify antigens that react with K199 and participate in nuclear assembly using the combination of K199 and the nuclear assembly assay method.
Fractionation of an Antigen(s) Related to Nuclear Assembly-To fractionate the antigen(s) responsible for the suppression, a crude extract was separated into cytosol and membrane fractions, and the membrane fraction was pretreated with either control PBS, normal serum, or K199. The membrane fractions were washed and combined with an untreated cytosol fraction, and nuclear assembly activity was examined. The membrane fractions treated with PBS, normal serum, and K199 showed very similar nuclear assembly activities (data not shown). These results suggest that the major antigens responsible for the nuclear assembly suppression are present in the cytosol but not in the membrane fraction.
To determine the molecular mass range of the antigen(s) participating in nuclear assembly, antibodies in K199 serum A, a Xenopus egg crude extract was mixed with control PBS (squares), a normal serum (triangles), or K199 serum (circles) and then treated as described under "Materials and Methods." Nuclear assembly assays were performed with the thus-treated crude extracts. Aliquots of the reaction mixtures were removed at the indicated times and then examined for nuclear assembly. B, a Xenopus egg crude extract was preincubated with control PBS (1), a normal serum (2), or K199 serum (3 and 4) at 4°C for 90 min. Chromatin was added to the pretreated extracts or egg lysis buffer (5)  were fractionated as follows. Xenopus egg cytosol proteins were electrophoresed and then transferred to PVDF papers. The papers were incubated with K199 serum. Antibodies bound to these papers were eluted after cutting each paper into five portions. Suppression of nuclear assembly by fractionated antibodies was examined (Fig. 3). The antibodies bound to portion 3, which contained 35-54-kDa proteins of the cytosol fraction, strongly suppressed nuclear assembly. These results suggested that proteins in the range of 35-54 kDa in the cytosol fraction react with K199 serum and participate in the vesicle fusion step of nuclear assembly.
To purify the antigen(s) participating in nuclear assembly, we performed DEAE-5PW anion exchange chromatography. A Xenopus egg cytosol fraction was applied to a DEAE-5PW anion exchange column and eluted with a stepwise gradient, which the fraction separated into six fractions (Fig. 4A). Each fraction was electrophoresed and then transferred to a PVDF paper. The paper was incubated with K199 serum, and the portions containing 35-54-kDa proteins were cut out. Suppression of nuclear assembly by antibodies eluted from these portions was examined (Fig. 4B). Marked suppression was observed with three fractions, i.e. the flow-through, 100 mM, and 1 M fractions.
To narrow the molecular mass range, antibodies were eluted from these papers after cutting them into three portions corresponding to three molecular mass ranges: 1, 47-54 kDa; 2, 40 -47 kDa; and 3, 35-40 kDa, respectively. Then the suppression of nuclear assembly with these antibodies was examined again. The antibodies that suppressed the nuclear assembly were those bound to portion 1 of the flow-through fraction, portion 3 of the 100 mM fraction, and portion 2 of the 1 M fraction (data not shown).
Identification of Cytosol Proteins Participating in Nuclear Assembly-To identify and purify the antigen(s) participating in nuclear assembly, we performed gel filtration chromatography. The flow-through, 100 mM, and 1 M fractions obtained on anion exchange chromatography were applied to a gel filtration column, Superdex 200HR. An aliquot of each eluted fraction was subjected to SDS-PAGE followed by staining with silver (Fig. 5, A, C, and E). Antibodies that bound to major protein bands within the 47-54 kDa (Fig. 5A), 35-40 kDa (Fig. 5C), and 40 -50 kDa (Fig. 5E) regions on SDS-PAGE were purified separately, and their abilities to suppress nuclear assembly were examined similarly as described in Fig. 3 (Fig. 5, B, D, and F). Antibodies purified from protein 6 (39 kDa), 7 (38 kDa), 13 (45 kDa), and 14 (48 kDa) markedly suppressed nuclear assembly. These results firmly suggested that these four proteins participate in nuclear assembly.
Then we analyzed the partial amino acid sequences of the 38-, 39-, 45-, and 48-kDa proteins. About 8 g of each these proteins was obtained from 34 units (204 mg) of Xenopus egg cytosol protein. These proteins were separated by SDS-PAGE and digested with trypsin. The resulting peptides were separated by reversed-phase HPLC (data not shown). Two proteins isolated from the 100 mM fraction, 38 and 39 kDa, exhibited very similar elution profiles (data not shown). All partial amino acid sequences obtained for the 38-and 39-kDa proteins completely matched parts of the Xenopus GAPDH amino acid sequence (Table I). The molecular mass of Xenopus GAPDH is 35,566 (Swiss-Prot accession number P51469), which is close to 39 kDa. Therefore, we concluded that the 39-kDa protein is Xenopus GAPDH. The 38-kDa protein may be a proteolytic fragment generated from the 39-kDa GAPDH during the purification procedures. Alternatively, the protein may be an isoform of GAPDH because GAPDH isoforms have been detected in Xenopus embryos under some conditions (39).
On the other hand, the amino acid sequences of the three peptides derived from the 45-kDa protein matched those of human fructose-1,6-bisphosphate aldolase A except for two amino acid substitutions, i.e. glycine in peptide 2 and glycine at the 11th residue from the amino terminus of peptide 3 are substituted by tyrosine and serine, respectively. These results suggest that the 45-kDa protein is Xenopus fructose-1,6bisphosphate aldolase. In the case of the peptides derived from the 48-kDa protein, the amino acid sequences completely matched those of Xenopus RCC1. The molecular mass of Xenopus RCC1 has been reported to be 45 kDa (40), which is close to 48 kDa. Thus, we concluded that the 48-kDa protein is Xenopus RCC1. The finding of RCC1 means that the method we used in this study was suitable for our purpose because RCC1 is known to participate in NE assembly (14). From these results, GAPDH, fructose-1,6-bisphosphate aldolase, and RCC1 were suggested to participate in nuclear assembly.
Verification of Participation of a Fusogenic Protein, GAPDH, in Nuclear Membrane Assembly-Interestingly, GAPDH is known to be a fusogenic protein. This fusogenic protein binds to phosphatidylserine in phospholipid vesicles and stimulates the fusion of lipid bilayers (28 -30). However, the participation of GAPDH in nuclear membrane assembly has not been demonstrated at all. Thus, we focused our attention on GAPDH. To confirm the above results, at first we examined whether or not authentic antibodies against GAPDH could suppress nuclear assembly. A commercially available anti-human GAPDH IgG fraction (rabbit) was first applied. The antibody cross-reacted selectively with a ϳ39-kDa protein in a Xenopus egg crude extract (Fig. 6A). The Xenopus egg crude extract was mixed with various concentrations of antibodies, and the nuclear assembly activity was examined. As can be seen in Fig. 6B, nuclear assembly was clearly suppressed by anti-GAPDH antibodies. When we used Fab fragments instead of the anti-GAPDH antibodies, clear inhibition was observed again (Fig.  6C). These results showed that the inhibition is not an artifact caused by bivalent antibody-mediated membrane aggregation. Then we attempted to recover the nuclear assembly activity by the addition of purified GAPDH to a Xenopus egg crude extract pretreated with anti-GAPDH antibodies. As can be seen in Fig.  7, the nuclear assembly activity suppressed by anti-GAPDH antibodies could be recovered with purified GAPDH to a level close to that in the control experiment. These results support the idea that GAPDH participates in the nuclear assembly.
To further characterize the suppression of nuclear assembly by anti-GAPDH antibodies, we used antibodies against specific sites of GAPDH. We previously demonstrated that GAPDH is a phosphatidylserine (PS)-binding protein and that its putative PS-binding site contains an amino acid sequence designated as GAPDH70 (28). It has been shown that the GAPDH70 peptide and anti-GAPDH70 antibodies inhibit GAPDH-induced phospholipid-vesicle fusion (28). Therefore, we purified the anti-GAPDH70 antibodies with a peptide affinity column, and the reactivity with Xenopus GAPDH was examined. The purified antibodies cross-reacted with a band corresponding to GAPDH in a Xenopus egg crude extract on immunoblotting (Fig. 8A). Then a Xenopus egg crude extract was pretreated with the purified antibodies, and the effects on nuclear assembly were examined. As can be seen in Fig. 8B, the antibody against the PS-binding site of GAPDH strongly suppressed nuclear assembly at a low concentration, whereas the control IgG had no effect. When the affinity-purified antibodies were premixed with the GAPDH70 peptide, the effects of the antibodies on nuclear assembly disappeared (data not shown). We also employed an affinity-purified polyclonal antibody against a 15amino acid synthetic peptide, amino acid residues 390 -403 of phosphatidylserine decarboxylase of Chinese hamster. The antibody, designated as aPSD-2, reacted with a site comprising amino acid residues 70 -94 of rabbit GAPDH (28). The antibody also reacted with GAPDH in the Xenopus egg cytosol (Fig. 8A) and strongly suppressed the nuclear assembly (Fig. 8B). These results more clearly suggested the participation of GAPDH in nuclear assembly. These results also suggested that anti-GAPDH antibodies inhibit the vesicle fusion step in nuclear assembly because anti-GAPDH70 has been shown to inhibit the GAPDH-induced fusion of phospholipid vesicles (28).
We showed previously that the synthetic peptide of GAPDH70 itself inhibits GAPDH-induced vesicle fusion (28). Thus, we applied this peptide to the nuclear assembly assay. As control peptides, ␥-endorphin and parathyroid hormone, which are similar to the GAPDH70 peptide in length and charge, were used. As can be seen in Fig. 9, the GAPDH70 peptide, but not the control ones, inhibited nuclear assembly. From these re-  figure (dotted lines). The elution was monitored as to absorption at 280 nm. The eluate was separated into six fractions, which were designated according to the buffer concentrations. This chromatography was repeated eight times, and corresponding fractions were combined. B, about 3 units of each fraction was subjected to 8% SDS-PAGE and electrotransferred to a PVDF paper. The PVDF papers carrying the six fractions were incubated in a buffer containing K199 serum, and the portions corresponding to 35-54-kDa proteins were cut out. Inhibition of nuclear assembly by antibodies eluted from these portions was examined as in Fig. 3. PBS, Normal serum, and K199 are negative and positive controls. Empty and filled bars indicate the results of independent two experiments.
sults obtained with the GAPDH70 peptide, various antibodies, and purified GAPDH protein, we conclude that GAPDH definitely participates in nuclear assembly.
The GAPDH-participating Step in Nuclear Assembly-To further characterize the participation of GAPDH in the membrane fusion in NE assembly, we observed chromatin, of which nuclear assembly was suppressed by anti-GAPDH and anti-GAPDH70 antibodies, and the GAPDH70 peptide itself by means of confocal fluorescent microscopy. As can be seen in Fig.  10A, continuous rim-staining with DHCC, which stains phospholipids, was observed in nuclei assembled with chromatin and a Xenopus egg crude extract pretreated with a buffer or control non-immune IgG. However, chromatin, of which nuclear assembly was suppressed by the anti-GAPDH and anti-GAPDH70 antibodies, and GAPDH70 peptide showed discontinuous lipid-staining of the rims (Fig. 10A, Anti-GAPDH, Anti-GAPDH70, and GAPDH70 peptide). When we focused on the surface of chromatins, it was more clearly seen that the nuclear envelope is not enclosed, and reticular lipid-staining was frequently observed on this chromatin (Fig. 10, B and C). These results support the idea that the GAPHD70 peptide and antibodies against GAPDH suppress the nuclear membrane fusion step in nuclear assembly.

Analysis of Proteins Participating in Nuclear Assembly by
Means of Autoimmune Sera-Autoimmune sera have contributed to the analyses of a proliferating cell nuclear antigen (41), MAN1 (42), and many other proteins. In this study we applied autoimmune sera to the analysis of the nuclear assembly mechanism. We previously examined the reactivities of PBC, RA, and Sjs sera with rat liver NE proteins by means of an immunoblotting method. It was shown that 100% of PBC, 70% of RA, and 97% of Sjs sera reacted with some proteins in a rat liver NE fraction. 2 We have reported that some PBC, RA, and Sjs sera react with several GlcNAc-bearing proteins in rat liver NE (27). Thus, we examined the reactivities of PBC, RA, and Sjs sera with Xenopus egg extract proteins and found that these sera react with many proteins in both Xenopus egg membrane and cytosol fractions and inhibit nuclear assembly in a Xenopus egg extract cell free system to various degrees. 2 An in vitro system comprising a Xenopus egg extract and sperm chromatin was chosen for the screening of proteins participating in nuclear assembly, because with this method it is easy to assay many samples, and sufficient amounts of protein for biochemical analyses can be obtained. K199 serum, found by means of this method, was used in this study.
We showed that K199 contains antibodies against RCC1. An antibody against RCC1 was identified first by Bischoff et al. in autoimmune serum from a Raynaud phenomenon patient (43). We suggested in this study that RCC1 participates in nuclear membrane assembly. RCC1 is a nucleus-specific GTP exchange factor for Ran. Localization of RCC1 on the surface of chromatin establishes the polarity of the Ran-GTP concentration gradient, which is essential for nuclear assembly (14,15). It is also known that RCC1 is essential for the fusion of vesicles on the surface of chromatin (14). Therefore, the suppression of nuclear membrane vesicle fusion by anti-RCC1 antibodies purified from K199 serum observed in this study is consistent with the known function of RCC1 in nuclear assembly. In other words, the finding of RCC1 implies that the method worked well for our purpose.
On the other hand, the anti-fructose-1,6-bisphosphate aldolase antibodies found in K199 are known to be present in 10% of sera from RA patients (44). Fructose-1,6-bisphosphate aldol-  Fig. 4 (12 units) was separated with a gel-permeation HPLC column, Superdex 200HR, equilibrated with 0.1 M ammonium acetate buffer (pH 8.3) at 0.5 ml/min, and the eluate was collected in 1.0-ml fractions. A 0.01 volume of each fraction was subjected to 8% SDS-PAGE followed by silver staining. A part of the stained gel, which contained 35-54-kDa proteins, is shown. B, the remaining proteins in the fractions were electrophoresed and then transferred to PVDF papers. Bands corresponding to 1-4 in A were cut out and incubated with K199 serum. After washing, the bound antibodies were eluted and concentrated. Suppression of nuclear assembly by these antibodies was examined using half of the thus-obtained antibodies, as in Fig. 3. PBS, Normal serum, and K199 in the figure are negative and positive controls for 1-4. PBS, normal serum, and K199 serum were added instead of purified antibodies. C and D, the 100 mM fraction in Fig. 4 was analyzed as in A and B. E and F, the 1 M fraction in Fig. 4 was also analyzed as in A and B. The results of the nuclear assembly assay are expressed as the averages of duplicate determinations. ase, suggested in this study to participate in nuclear assembly, is a glycolytic enzyme. How this enzyme participates in nuclear assembly and whether or not its enzymatic activity is necessary for nuclear assembly should be studied in the future. To our knowledge, GAPDH is not regarded as an autoimmune antigen. The GAPDH found in this study may be a novel autoimmune antigen. We suggest that GAPDH participates in nuclear assembly, as judged by means of a method involving the combination of an autoimmune serum and a Xenopus egg extract in this study. We also found another novel intrinsic membrane protein, which seems to participate in nuclear assembly, by the same method using another autoimmune serum (data not shown). Therefore, this method is very useful for searches for novel nuclear assembly factors.
Participation of GAPDH in Nuclear Membrane Fusion-We were first surprised by the suppression of nuclear assembly by the anti-GAPDH antibodies found in K199 serum because GAPDH was considered to be a classical glycolytic protein due to its pivotal role in energy production. However, there has been recent evidence that mammalian GAPDH exhibits a number of diverse activities unrelated to its glycolytic function (32). These activities include ones in microtube bundling (45), phosphotransferase activity (46), nuclear RNA export (47) 1 and 3, 30 g) and purified chicken GAPDH (lanes 2 and 4, 2 g) were electrophoresed and stained with Coomassie Brilliant Blue R-250 (lanes 1 and 2) or transferred to a PVDF paper and immunoblotted with the IgG fraction of anti-human GAPDH serum (lanes 3 and 4). B, the IgG fraction of anti-human GAPDH (circles) or a non-immune rabbit-IgG fraction (triangles) was added to a Xenopus egg crude extract and incubated at 4°C for 90 min. The remaining nuclear assembly activity was determined as in Fig. 3. C, the antibodies used in B were converted to Fab fragments by treatment with immobilized papain. The effects of the thus-generated Fab fragments were examined as in B with a concentration of Fab fragments corresponding to 3.75 mg/ml IgG. The vehicle (1) and Fab fragments of non-immune (2) and anti-GAPDH (3) IgGs were used.
repair (48), membrane fusion (28 -30, 49, 50), and others. It is known that a wide variety of genes (more than 200 in some species) contain sequences that can encode proteins exhibiting strong similarity to GAPDH (51,52). Rabbit brain cytosolic GAPDH was found to have least 16 isoforms when it was separated by two-dimensional gel electrophoresis (29). Therefore, some diverse functions of the enzyme can be allocated to specific isoforms. Indeed, a rabbit brain GAPDH isoform exhibiting fusogenic activity showed no glycolytic activity (29). In Xenopus embryos, multiple isoforms of GAPDH have been detected under some conditions (39). Therefore, a Xenopus protein participating in nuclear membrane fusion may be a GAPDH isoform for membrane fusion.
It is known that GAPDH can cause the fusion of reconstituted phospholipid vesicles in vitro (49). We recently reported that rabbit GAPDH binds to phosphatidylserine in lipid bilayers and stimulates membrane fusion of reconstituted phospholipid vesicles (28). There have also been several studies involving naturally occurring membranes. Han et al. (30) provide evidence suggesting that one GAPDH isoform purified from rabbit brain can reconstitute in vitro protein-catalyzed fusion between purified rat pancreatic ␤-cell plasma membranes and secretory granules involved in insulin exocytosis. Hessler et al. (49) suggest the participation of human neutrophil cytosolic GAPDH in granule-plasma membrane fusion based on the results of in vitro studies. Robbins et al. (50) characterize GAPDH in Chinese hamster ovary cell mutant FD1.3.25 in vivo. One isoform of GAPDH in FD1.3.25 cells was mutated at Ser-234 to Pro. Their studies suggested that GAPDH plays a specific role in the fusion of granules with the plasma membrane in vivo. On the other hand, it was suggested in this study with a Xenopus egg extract cell free system that GAPDH participates in nuclear membrane fusion. This is the first report of the participation of GAPDH in nuclear membrane fusion other than the case of granules and the plasma membrane.
Antibodies against the PS binding region of GAPDH suppress nuclear membrane fusion (Figs. 8 and 10). aPSD2, which is an antibody against the PS binding region of phosphatidylserine decarboxylase and reacts with the site comprising amino acid residues 70 -94 of rabbit GAPDH, suppressed the membrane fusion (Fig. 8). The GAPDH70 peptide itself, which constitutes the PS binding region of GAPDH, also suppressed membrane fusion (Figs. 9 and 10). The nuclear membrane contains PS, which comprises about 6% of the lipid. These results suggested that GAPDH binds to PS in the membrane and stimulates nuclear membrane fusion.
GAPDH-participating Step in Nuclear Membrane Fusion-Nuclear membrane fusion in a Xenopus egg extract cell free system comprises three steps, i.e. formation of a reticular network on chromatin, fusion of this network into a closed nuclear membrane, and subsequent expansion (23). In the first step, NE precursor vesicles bound to chromatin fuse with each other and form a reticular membrane network on the chromatin. This step involves Ran and is inhibited by GTP␥S and N-ethylmaleimide (8,14). When nuclear membrane fusion was inhibited with N-ethylmaleimide at this step, round vesicles bound to the chromatin surface were observed on microscopy (data not shown). In the second step, the reticular network fuses and forms a closed nuclear membrane (23). This step involves the p97⅐Ufd1⅐Npl4 complex (23). When nuclear envelope fusion is inhibited at this step, the reticular network is observed on the surface of the chromatin (23). In the third step, chromatin surrounded by a closed nuclear membrane becomes round and large with nuclear membrane expansion. This step involves the p97⅐p47 complex (23). When the reaction is inhibited at this step, small nuclei are observed (23). When nuclear membrane formation was suppressed by anti-GAPDH antibodies, a reticular network was observed on the majority of chromatin (Fig.  10). When the GAPDH70 peptide was added to the nuclear membrane formation system instead of antibodies, a very similar reticular network was observed (Fig. 10). These results strongly suggest that GAPDH is involved in the second step of nuclear membrane fusion, i.e. the fusion of the reticular network on chromatin into a closed nuclear membrane. However, the possible participation of GAPDH in another membrane fusion step(s) cannot be excluded.
As mentioned above, Hetzer et al. (23) show that AAA-ATPase p97 in a complex with adaptor proteins, Ufd1 and Npl4, functions in the second step of nuclear membrane fusion (23). Therefore, the p97⅐Ufd1⅐Npl4 complex and GAPDH seem to act in the same step. Although the mechanism underlying the p97⅐Ufd1⅐Npl4 action in the membrane fusion step is unclear, this activity may require a SNARE on the target membrane, like the p97⅐p47 action, and catalyze highly specific membrane fusion (23). Alternatively, the ternary complex may function in membrane assembly through ubiquitin-and proteasomedependent modification of a target protein needed for nuclear membrane assembly (53). On the other hand, GAPDH  1 and 3, 30 g) and purified chicken GAPDH (lanes 2 and 4, 3 g) were electrophoresed, transferred to a PVDF paper, and then incubated with anti-GAPDH70 (1 and 2) and aPSD2 (3 and 4) antibodies. Bound antibodies were detected by an enhanced chemiluminescence method. A part of the GAPDH band in each lane is shown. B, various amounts of non-immune IgG (triangles), anti-GAPDH70 (circles), and aPSD2 antibodies (square) were added to a Xenopus egg crude extract followed by incubation at 4°C for 90 min. The remaining nuclear assembly activity was determined as in Fig. 3.
FIG. 9. Suppression of nuclear assembly by the GAPDH70 peptide. To a Xenopus egg crude extract (8 l), 2 l of GAPDH70 (circles), control peptide ␥-endorphin (triangles), or parathyroid hormone (square) was added followed by incubation at 4°C for 90 min. The remaining nuclear assembly activity was determined as in Fig. 3. can catalyze very rapid vesicle fusion (29). It is known that GAPDH catalyzes the fusion of vesicles at a rate that satisfies the mathematical constraints imposed by the observed rate of fusion of synaptic vesicles with presynaptic membranes in vivo (29). GAPDH can catalyze one fusion event between two reconstituted phospholipid vesicles every millisecond (29), but one round of SNARE-mediated fusion of reconstituted liposomes takes 30 -40 min (54). Therefore, GAPDH may enhance the rate of nuclear membrane fusion that is tightly controlled by p97⅐Ufd1⅐Npl4 and SNARE on the target membrane. Alternatively, GAPDH may act through an unknown mechanism separate from p97⅐Ufd1⅐Npl4. In that case, involvement of the glycolytic activity of GAPDH in the nuclear envelope assembly cannot be excluded.