Journal of Neurochemistry, 2001, 77, 730±740
A unique spacer domain of synaptotagmin IV is essential for Golgi localization Mitsunori Fukuda,* Keiji Ibata*,² and Katsuhiko Mikoshiba*,² *Laboratory for Developmental Neurobiology, Brain Science Institute, RIKEN, Hirosawa, Wako, Saitama, Japan ²Division of Molecular Neurobiology, Department of Basic Medical Science, The Institute of Medical Science, The University of Tokyo, Shirokanedai, Minato-ku, Tokyo, Japan
Abstract Synaptotagmin (Syt) family members consist of six separate domains: a short amino terminus, a single transmembrane domain, a spacer domain, a C2A domain, a C2B domain and a short carboxyl (C) terminus. Despite sharing the same domain structures, several synaptotagmin isoforms show distinct subcellular localization. Syt IV is mainly localized at the Golgi, while Syt I, a possible Ca21-sensor for secretory vesicles, is localized at dense-core vesicles and synaptic-like microvesicles in PC12 cells. In this study, we sought to identify the region responsible for the Golgi localization of Syt IV by immunocytochemical and biochemical analyses as a means of de®ning the distinct subcellular localization of the synaptotagmin family. We found that the unique C-terminus of the
spacer domain (amino acid residues 73±144) between the transmembrane domain and the C2A domain is essential for the Golgi localization of Syt IV. In addition, the short C-terminus is probably involved in proper folding of the protein, especially the C2B domain. Without the C-terminus, Syt IVDC proteins are not targeted to the Golgi and seem to colocalize with an endoplasmic reticulum (ER) marker (i.e. induce crystalloid ER-like structures). On the basis of these results, we propose that the divergent spacer domain among synaptotagmin isoforms may contain certain signals that determine the ®nal destination of each isoform. Keywords: C2 domain, exocytosis, glycosylation, Golgi, immediate early genes, synaptotagmin. J. Neurochem. (2001) 77, 730±740.
Synaptotagmins (Syts) constitute a family of at least 13 proteins in the rat or mouse and are thought to be involved in membrane traf®c (reviewed in SuÈdhof and Rizo 1996; Fukuda and Mikoshiba 1997, 2001; Schiavo et al. 1998; MarqueÁze et al. 2000). All members are composed of a short extracellular (or luminal) amino (N) terminus, a single transmembrane domain, and two C2 domains (known as the C2A and C2B domains) in the large cytoplasmic domain, but an alternatively spliced variant that lacks the transmembrane domain (e.g. Syt VIDTM) has been described in some of them (Fukuda and Mikoshiba 1999; Craxton and Goedert 1999). All synaptotagmins show type I membrane topology (Nextracellular (ext)/Ccytoplasmic (cyt)) and lack a signal peptide sequence (Fukuda et al. 1999; Fukuda and Mikoshiba 2000b). Recent studies have indicated that at least several isoforms show distinct subcellular localization in the brain: Syt I, an evolutionarily conserved and wellcharacterized isoform, has been shown to be essential for synaptic vesicle exocytosis and endocytosis in neurons and for secretory vesicle exocytosis in some endocrine cells (Fukuda et al. 1995b, 2000; Mikoshiba et al. 1995; Lang
et al. 1997; Mochida et al. 1997; Ohara-Imaizumi et al. 1997; Thomas and Elferink 1998; reviewed in MarqueÁze et al. 2000). Syt I (or II) is also present at growth cone vesicles and has been shown to be essential for neurite outgrowth of chick dorsal root ganglion neurons and PC12 cells (Kabayama et al. 1999; Fukuda and Mikoshiba 2000c). Syt III is mainly localized at synaptic plasma membrane rather than synaptic vesicles (Butz et al. 1999), and it can heterodimerize with Syts V, VI, and X through the N-terminal cysteine motif by disul®de bonding (Fukuda et al. 1999). Syt VIDTM, a major alternatively spliced
730
Received November 10, 2000; revised manuscript received January 9, 2001; accepted January 12, 2001. Address correspondence and reprint requests to M. Fukuda, Laboratory for Developmental Neurobiology, Brain Science Institute, RIKEN, 2±1 Hirosawa, Wako, Saitama 351±0198, Japan. E-mail:
[email protected] Abbreviations used: C, carboxyl; cyt, cytoplasmic; ER, endoplasmic reticulum; ext, extracellular; HRP, horse radish peroxidase; N, amino; PAGE, polyacrylamide gel electrophoresis; PBS, phosphate-buffered saline; PCR, polymerase chain reaction; Syt(s), synaptotagmin(s).
q 2001 International Society for Neurochemistry, Journal of Neurochemistry, 77, 730±740
A regulatory region of the Syt IV transport 731
variant that lacks the transmembrane domain, is attached to various membrane fractions and is also present in the cytosol (Butz et al. 1999; Fukuda and Mikoshiba 1999). Although Syt IV was ®rst proposed to be present at synaptic vesicles and to be involved in the regulation of neurotransmitter release in concert with Syt I (Littleton et al. 1999; Thomas et al. 1999), our recent immunocytochemical and immunoelectron microscopic analyses (K. Ibata et al., manuscript in preparation) using a highly speci®c anti-Syt IV antibody has indicated that Syt IV and Syt I have different subcellular localizations, with Syt IV being mainly localized at the Golgi and distal portion of the neurites in nerve growth factor-differentiated PC12 cells and hippocampal primary neurons (Berton et al. 2000; Ibata et al. 2000). In addition, the biochemical characters of the C2 domains (even of the C2A and C2B domains) of synaptotagmin isoforms differ somewhat in terms of phospholipid, inositol polyphosphate, or syntaxin binding (Fukuda et al. 1995a, 1996, 1997; Li et al. 1995; Mehrotra et al. 1997; Ibata et al. 1998; Fukuda and Mikoshiba 2000a). All these observations suggest that different isoforms may have different functions and be involved in different membrane traf®c processes. If each Syt isoform participates in speci®c membrane traf®c, how does it reach the destination where it functions, and what determines the membrane targeting? This information is quite important to understanding the molecular basis of membrane traf®c, but the membrane targeting signals of synaptotagmin isoforms have never been elucidated. In the present study, we investigated the transport of Syt IV protein to the Golgi by performing deletion and mutational analyses and identi®ed a regulatory region, and in this paper we discuss the membrane targeting signals of the synaptotagmin family based on our ®ndings. Materials and methods Materials Restriction enzymes and recombinant Taq DNA polymerase were obtained from Toyobo Biochemicals (Osaka, Japan). Polyclonal antibody against FLAG peptide was obtained from Zymed Laboratories, Inc. (San Francisco, CA, USA). Horseradish peroxidase (HRP)-conjugated anti-T7 tag antibody and anti-T7 tag antibody-conjugated agarose were from Novagen (Madison, WI, USA). Anti-TGN38 antibody was from Transduction Laboratories (Lexington, KY, USA). Anti-BiP (Grp78) and anti-GOS28 antibodies were from StressGen Biotechnologies Corp. (Victoria, BC, Canada). All other chemicals were commercial products of reagent grade. Solutions were prepared in deionized water. Expression constructs pEF-FLAG-Syt IV, and -T7-N-Gly-Syt IV (MASMTGGQQMGRNGS; T7-tag underlined, N-glycosylation (N-Gly) site in bold letters), were prepared as described previously (Fukuda et al. 1994, 1999; Fukuda and Mikoshiba 2000b). pEF-FLAG(or T7-N-Gly)-Syt
IVDCyto, -Syt IVDC2AB, -Syt IVDC2B, -Syt IVDC, -Syt IVDTM, pEF-FLAG-Syt IVD79-425, -T7-N2Syt IV, -N2Syt IVDC, and -N2Syt IVDC2AB were constructed similarly by polymerase chain reaction (PCR) using the following sets of primers with appropriate restriction enzyme sites (underlined) and/or termination codons (bold letters): 5 0 -GCACTAGTTCAACAGCAGATCCAGGCAAA-3 0 (DCyto primer; antisense; amino acid residues 33±38), 5 0 -GCACTAGTTCATTTCTCCTCTGACGTGAG-3 0 (DC2AB primer; antisense; amino acid residues 145±150), 5 0 -GCACTAGTTCATCTCTTGATGATCTCTCT-3 0 (DC2B primer; antisense; amino acid residues 275±280), 5 0 -GCACTAGTTCAGATCTCCTTCCAGTGCCC-3 0 (DC primer; antisense; amino acid residues 403±408), 5 0 -GCGGATCCCAGAGAAGATCAGCCAAATC-3 0 (DTM primer; sense; amino acid residues 39±45), 5 0 -GCACTAGTTCACTTGTCATCTCCTCCAAA-3 0 (D79 primer; antisense; amino acid residues 73±78), and 5 0 -GCGGATCCAACTCTACAGAGGCTCCTATCACCACCAGCCG-3 0 (N2Syt IV primer; sense; amino acid residues 2±7). pEF-T7-N2Syt IVD42-146, -N2Syt IVD42±146DC2B, -N2Syt IVD44-68, -N2Syt IVD73-144, -N2Syt IVD141-149, and -N2Syt IV(A8) were essentially produced by means of two-step PCR techniques as described previously (Fukuda et al. 1995a) using the following mutagenic oligonucleotides: D42 primer 5 0 -GCAGATCTTCTCTGACAGCAGATCC-3 0 and D146 primer 5 0 -GCAGATCTGAGGAGAAACAAGAGAA-3 0 ; D44 primer 5 0 -TCAGCTAGCTGATCTTCTCTGACA-3 0 and D68 primer 5 0 -TCAGCTAGCAAAAAGAAGTTTGGA-3 0 ; D73 primer 5 0 -TCAAAGCTTCTTTTTGCTACTTAG-3 0 and D144 primer 5 0 -CGGAAGCTTACGTCAGAGGAGAAA-3 0 ; D141 primer 5 0 -TCCCAAGCTTCTCTTGTTTCTTCAAGCTCTCAGGGGAA-3 0 and D149 primer 5 0 -CAAGAGAAGCTTGGGACA-3 0 ; and A8 antisense primer 5 0 -AAAAAGCTTTGGGGTGACATTTTCCAGAGCAGAAGAGGCGCCAGCTTTGGCGTTGGCTTTGGGGAAGTTGCCATTGGCGTCTCGCTTCGCTAGGTCAAGATGCGCGGA-3 0 and A8 sense primer 5 0 -CCAAAGCTTTTTACGGAGACAGCAAAGGAGGCCAATGCCCCT-3 0 . All constructs were veri®ed by DNA sequencing using the Hitachi SQ-5500 DNA sequencer. Cell culture and transfections Glass-bottom dishes (35 mm-dish, MatTek Corp., MA, USA) were coated with collagen type IV (Becton Dickinson, Bedford, MA, USA). PC12 cells (0.8±1 105 or 4 106 cells, the day before transfection) were cultured on these dishes or 10-cm collagen type I-coated dishes (Becton Dickinson) in Dulbecco's modi®ed Eagle's medium containing 10% horse serum and 10% fetal bovine serum at 378C, 5% CO2. Transfection was achieved by using the LipofectAmine Plus reagent according to the manufacturer's notes (Life Technologies, Rockville, MD, USA). Transfection of various pEF-T7-N-Gly-Syt constructs into COS-7 cells (5 105 cells, the day before transfection/10-cm plate) was carried out by the DEAE-dextran method as described elsewhere (Fukuda et al. 1999). N-glycosidase F, endoglycosidase H, and neuraminidase digestion of T7-(N-Gly)-Syts Immunoprecipitation of T7-Syts by anti-T7 tag antibody-conjugated agarose was performed as described previously (Fukuda et al. 1999). The beads were resuspended in 30 mL of the buffer [20 mm HEPES-KOH, pH 7.2, 0.1% sodium dodecyl sulfate (SDS), 0.5%
q 2001 International Society for Neurochemistry, Journal of Neurochemistry, 77, 730±740
732 M. Fukuda et al.
IGEPAL CA-630 (Sigma), 20 mm EDTA, 0.1 mm phenylmethylsulfonyl ¯uoride, 10 mm leupeptin, and 10 mm pepstatin A for N-glycosidase F; and 50 mm sodium acetate, pH 5.3, 0.1% SDS, 0.5% IGEPAL CA-630, and protease inhibitors for neuraminidase or endoglycosidase H], and then divided into two microtubes. After denaturation by boiling for 3 min and cooling to 378C, 1 U of N-glycosidase F, 5 mU of endoglycosidase H (Roche Diagnostics, Mannheim, Germany), or 10 U of neuraminidase (Sigma) was added to one tube, and the mixtures were incubated for 1 h at 378C. Reactions were stopped by adding SDS sample buffer and boiling for 3 min. Proteins were subjected to 7.5%, 12.5%, or 10±18.3% gradient SDS±polyacrylamide gel electrophoresis (PAGE) and immunoblotting as described previously (Fukuda et al. 1999). Immunocytochemistry Three days after transfection, PC12 cells (35 mm-dish) were washed twice with phosphate-buffered saline (PBS), ®xed in 4% paraformaldehyde in 0.1 m sodium phosphate buffer for 20 min at room temperature (238C), and washed with 0.1 m glycine. The ®xed cells were then permeabilized with 0.3% Triton X-100 in PBS for 2 min and immediately washed with blocking solution (1% BSA, 0.1% Triton X-100 in PBS) three times for 5 min. Next, the cells were incubated in blocking solution for 1 h at room temperature and then with anti-TGN38 (1/500 dilution), anti-BiP (1/200 dilution), anti-GOS28 (1/100 dilution), anti-T7 mouse monoclonal antibody (Novagen; 1/10,000), or anti-FLAG (0.5 mg/mL) rabbit polyclonal antibody for 1 h at room temperature. Primary antibodies were washed out with blocking solution three times for 5 min and the cells were then incubated with anti-rabbit Alexa 488 and/or anti-mouse Alexa 568 antibodies (Molecular Probe, Eugene,
OR, USA) for 1 h at room temperature. After washing out the secondary antibodies with blocking solution ®ve times for 5 min, immunoreactivity was analyzed with a ¯uorescence microscope (TE300, Nikon, Tokyo, Japan) attached to a laser confocal scanner unit CSU 10 (Yokogawa Electric Corp., Tokyo, Japan) and HiSCA CCD camera (C6790, Hamamatsu Photonics, Hamamatsu, Japan). Images were pseudo-colored and superimposed with Adobe Photoshop software (Ver. 4.0).
Results Determination of the minimum domain required for the Golgi localization of synaptotagmin IV The synaptotagmin family consists of at least six distinct domains: a short extracellular (or luminal) N-terminal domain, a transmembrane domain, a spacer domain between the transmembrane and C2 domains, a C2A and a C2B domain, which are highly homologous among the synaptotagmin family, and a carboxyl (C)-terminal tail of varying length. To identify the domain involved in the Golgi localization of Syt IV, we systematically produced deletion mutants of Syt IV (Fig. 1) and examined their subcellular localization by using PC12 cells, in which Syt IV is endogenously expressed (Hilbush and Morgan 1994; Vician et al. 1995; Thomas et al. 1999; Ibata et al. 2000). The Golgi localization of Syt IV deletion mutants was determined by the following three methods: (i) comparison with immunostaining of anti-TGN38 antibody (Fig. 2), which has been Fig. 1 Schematic representation of the various synaptotagmin IV deletion mutants. The synaptotagmin family consists of six separate domains: a short extracellular (or luminal) N-terminus, a transmembrane domain (TM; open boxes), a spacer domain, a C2A and C2B domain (hatched boxes), and a short C-terminus. FLAG-tags are represented by black boxes. Deletions were made systematically from the N- or C-terminus. The Golgi localization of each mutant is indicated after its name (1 or ±, respectively) and was determined on the basis of the immunocytochemical ®ndings (Fig. 2). Amino acid numbers, except for the FLAG-tag sequence, are given on both sides.
Fig. 2 Subcellular localization of FLAG-Syt IV deletion mutants. FLAG-tagged Syt IV deletion mutants (see Fig. 1) were expressed in PC12 cells as described under `Materials and Methods'. PC12 cells were ®xed, permeabilized and stained with anti-FLAG and antiTGN38 (or anti-BiP) antibodies. FLAG-Syt IV is localized at the Golgi-like compartment in PC12 cells (a, FLAG-Syt IV; d, light-®eld micrograph; and e, overlay). (b and c) FLAG-Syt IV was co-localized with TGN-38 (yellow in part c). (f and g) Differential distribution of FLAG-Syt IV or FLAG-Syt IVDC2AB (green) and TGN38 (red),
respectively, after brefeldin A treatment. (h±j) FLAG-Syt IVDC2AB (green in part h) was co-localized well with TGN38 (red in part i, and yellow in part j). FLAG-Syt IVDCyto (k), -Syt IVD79-425 (l), -Syt IVDTM (m), or -Syt IVDC (n) were not localized at the Golgi-like perinuclear regions. Note that more than 40% of cells expressing FLAGSyt IVDC proteins are contained in granular structures of various sizes (arrowheads in part n), which overlapped well with BiP, a marker for the ER (red in part o, and yellow in part p). Scale bars indicate 20 mm.
q 2001 International Society for Neurochemistry, Journal of Neurochemistry, 77, 730±740
A regulatory region of the Syt IV transport 733
q 2001 International Society for Neurochemistry, Journal of Neurochemistry, 77, 730±740
734 M. Fukuda et al.
shown to overlap with endogenous Syt IV immunocytochemically (Ibata et al. 2000), or anti-GOS28 antibody, a Golgi marker (data not shown); (ii) treatment of cells with brefeldin A, which causes the Golgi markers to be distributed to the endoplasmic reticulum (ER); and (iii) analysis of Golgi-speci®c glycosylation events (see below for details). To discriminate recombinant from endogenous proteins, we ®rst added FLAG-tag to the N-terminus of Syt IV (FLAG-Syt IV). Recombinant FLAG-Syt IV proteins were predominantly expressed at the perinuclear region (Fig. 2e, arrow) that overlaps with TGN38 (Fig. 2a, Syt IV in green; Fig. 2b, TGN38 in red; and Fig. 2c, overlay), the same as endogenous Syt IV as described previously (Ibata et al. 2000). FLAG-Syt IVDC2AB and -DC2B proteins (deletion of two C2 domains) were still predominantly localized at the perinuclear regions (Golgi), which overlapped well with TGN-38 (Fig. 2h, Syt IVDC2AB in green; Fig. 2i, TGN38 in red; Fig. 2j, overlay; and data not shown), and these signals were sensitive to brefeldin A treatment (Figs 2f or g, Syt IV or Syt IVDC2AB in green and TGN38 in red, respectively). By contrast, deletion of the whole cytoplasmic domain (FLAG-Syt IVDCyto; Fig. 2k), the C-terminus of the spacer domain (FLAG-Syt IVD79-425; Fig. 2l), the transmembrane domain (FLAG-Syt IVDTM; Fig. 2m), or the short C-terminus (FLAG-Syt IVDC; Fig. 2n) completely prevented the Golgi localization. In addition, more than 40% of cells expressing FLAG-Syt IVDC proteins contain some granular structures of various sizes (Fig. 2n, arrowheads), which overlapped well with BiP, a marker for the ER (Fig. 2o, BiP in red; Fig. 2p, overlay). These granules resembled the crystalloid ER previously described (Pathak et al. 1986; Yamamoto et al. 1996). FLAG-Syt IVDTM was concentrated at the edge of the cells (plasma membrane), the same as Syt VIDTM, as described previously (Fig. 2m) (Fukuda and Mikoshiba 1999; Fukuda et al. 2000). Therefore, the Syt IV protein contains at least two regulatory regions for Golgi localization (a region from the transmembrane domain before the C2A domain and the C-terminal tail), although the presence of the transmembrane domain is a prerequisite for the targeting of Syt IV protein to the Golgi. To rule out the possibility that some deletions affect the membrane topology of Syt IV (Next/Ccyt), which might also cause mistargeting of Syt IV proteins, we examined the membrane topology of Syt IV deletion mutants by introducing of an arti®cial N-glycosylation site into the N-terminus (T7-N-Gly-tag; see Materials and methods) (Fukuda and Mikoshiba 2000b). As shown in Fig. 3, with the exception of T7-N-Gly-Syt IVDTM, deletion mutants and wild-type proteins undergo N-glycosylation. T7-N-GlySyt IVDTM, on the other hand, lacks the transmembrane domain and therefore was not N-glycosylated (Fig. 3, lanes 11 and 12). This indicates that the mistargeting of Syt IVDCyto and Syt IVDC proteins is not attributable to
Fig. 3 Determination of the membrane topology of Syt IV deletion mutants by N-glycosidase F digestion. pEF-T7-N-Gly-Syt IV mutants or a control vector were transfected into COS-7 cells, and recombinant proteins were immunoprecipitated by anti-T7 tag antibodyconjugated agarose as described previously (Fukuda et al. 1999). After treatment with and without N-glycosidase F (N-Gly F), proteins were subjected to 10±18.3% gradient SDS±PAGE followed by immunoblotting with HRP-conjugated anti-T7 tag antibody. Note that, except for T7-N-Gly-Syt IVDTM, the apparent molecular weight of the T7-N-Gly-Syts shifted to lower bands after N-glycosidase digestion (broken boxes). Molecular weight markers ( 1023) are shown on the left.
membrane topology and that the region around the transmembrane domain is a main determinant in formation of the Next/Ccyt orientation of Syt IV. To examine further the Golgi localization of Syt IV deletion mutants biochemically, we focused on Golgispeci®c glycosylation events. Since Syt IV itself does not contain an N-glycosylation site, we produced a chimera protein between Syt II and Syt IV (addition of the N-glycosylation sequence of Syt II (NSTE) to the N-terminus of Syt IV; N2Syt IV in Fig. 4(a). Vertebrate Syts I and II contain a single N-glycosylation site at the N-terminus (# in Fig. 4a), and they undergo addition of sialic acid (complex form of oligosaccharides) at the Golgi in PC12 cells, because the apparent molecular weight of T7-Syts I and II was slightly lower after neuraminidase treatment (Fig. 4b, lanes 1, 2, 5 and 6). Since endoglycosidase H can not attack N-linked sugar when it converts to complex form of oligosaccharides at the Golgi, T7-Syts I and II were resistant to endoglycosidase H treatment, but sensitive to N-glycosidase F treatment which could cleave all N-linked sugar (Fig. 4b, lanes 3, 4, 7 and 8). Thus, we used endoglycosidase H to determine whether the proteins have reached the Golgi or not. By contrast, T7-Syt IV was resistant to all three enzymes (Fig. 4b, lanes 9±12). Since addition of four amino acids to Syt IV did not affect its subcellular localization immunocytochemically (data not
q 2001 International Society for Neurochemistry, Journal of Neurochemistry, 77, 730±740
A regulatory region of the Syt IV transport 735
Fig. 4 Analysis of N2Syt IV, a chimera between the Syt II N-terminus and Syt IV, by glycosidase treatment. (a) Sequence comparison of Syts I, II, IV, and N2Syt IV N-terminal domain. # indicates the N-glycosylation site. The transmembrane domains (TM) are shown by box. N2Syt IV has an additional four amino acids of Syt II (NSTE; underlined) at the N-terminal domain of Syt IV. (b) Treatment of T7-Syt I (lanes 1±4), -Syt II (lanes 5±8), and -Syt IV (lanes 9±12) with neuraminidase (N ), endoglycosidase H (H ), and N-glycosidase F (F ), and the control (±). T7-Syts I, II, and IV were expressed in PC12 cells and immunoprecipitated by anti-T7 tag antibody-conjugated agarose as described previously (Fukuda et al. 1999). Immunoprecipitants were exposed to or not exposed to glycosidase, subjected to 7.5% SDS±PAGE, and then analyzed by immunoblotting with HRP-conjugated anti-T7 tag antibody. Note that Syts I and II were sensitive to neuraminidase and N-glycosidase F, but not to endoglycosidase H (lanes 2±4 and 6±8). (c) Exposure of T7-N2Syt
IV, -N2Syt IVDC, and -N2Syt IVDC2AB to endoglycosidase H. T7-N2Syts were expressed in PC12 cells and immunoprecipitated by anti-T7 tag antibody-conjugated agarose. Immunoprecipitants were treated with (H ) and without (±) endoglycosidase H, subjected to 12.5% SDS±PAGE and then analyzed by immunoblotting with HRP-conjugated anti-T7 tag antibody. The broken boxes indicate the endoglycosidase H-resistant bands. Arrowhead points to the light chain of the anti-T7 tag antibody. (d) Proportion of the endoglycosidase H (endo H )-resistant band of T7-N2Syt IV, -N2Syt IVDC, and -N2Syt IVDC2AB. Immunoreactive bands in (c) were captured by Gel Print 2000i/VGA (Bio Image) and quanti®ed by Basic Quanti®er Software (version 1.0) (Bio Image). Columns indicate the proportion of endoglycosidase H-resistant bands (i.e. the proportion of proteins that were modi®ed at the Golgi). Note that all of the T7-N2Syt IVDC proteins were sensitive to endoglycosidase H treatment. The results shown are representative of three independent experiments.
q 2001 International Society for Neurochemistry, Journal of Neurochemistry, 77, 730±740
736 M. Fukuda et al.
shown), we examined the effect of exposure to endoglycosidase H as a means of evaluating the Golgi-speci®c modi®cations of N2Syt IV proteins. Interestingly, T7-N2Syt IV and -N2Syt IVDC2AB yielded two bands, but T7-Syt IVDC yielded only a single band (Fig. 4c). The major higher broad bands of T7-N2Syt IV and -N2Syt IVDC2AB were resistant to endoglycosidase H treatment (Fig. 4c, boxed), but were sensitive to both neuraminidase and N-glycosidase F treatment (data not shown), indicating that these protein bands were indeed modi®ed at the Golgi. By contrast, the minor lower bands of T7-N2Syt IV and -N2Syt IVDC2AB were sensitive to endoglycosidase H treatment. It is noteworthy that all of the Syt IVDC proteins were sensitive to endoglycosidase H, indicating that they were not targeted to the Golgi and were not modi®ed at the Golgi (Fig. 4d). Based on the above ®ndings together with the results of the immunocytochemical studies and drug treatment (Fig. 2), we concluded that the Syt IV and Syt IVDC2AB proteins are indeed present at the Golgi. Since about 30% of the N2Syt IV and N2Syt IVDC2AB were sensitive to endoglycosidase H (Fig. 4d) and these proteins were localized at the Golgi-like perinuclear regions immunocytochemically, both may be present at the early Golgi as well as the late Golgi, where complex modi®cations occur. The unique C-terminal spacer domain is essential for Golgi localization of synaptotagmin IV In the next set of experiments, we examined the effect of internal deletion of the spacer domain on Golgi localization
of Syt IV to determine the relationship between the possible regulatory regions determined above, i.e. the spacer domain and the short C-terminus (Fig. 5a). The Golgi localization of each mutant was evaluated both immunocytochemically (Fig. 5e) and by Golgi-speci®c modi®cations (Figs 5c and d), as described above. As shown in Fig. 5(e-1), deletion of the whole spacer domain (a, D42-146) completely prevented Golgi localization, the same as it prevented Syt IVDCyto and Syt IVD79-425 localization, as described above (Figs 2k and l). In addition, T7-N2Syt IVD42-146 and D42±146DC2B proteins hardly underwent any Golgispeci®c modi®cations (i.e. they were sensitive to endoglycosidase H treatment) (Fig. 5c, lanes 1±4, and Fig. 5d). These ®ndings indicate that only the spacer domain of Syt IV functions as a Golgi localization signal and that the short C-terminus itself is independent of the Golgi localization of Syt IV, although its deletion also prevented the Golgi localization (Fig. 2n). Since crystalloid ER is known to be induced by overexpression of malfolded cytochrome P-450 (Ishihara et al. 1995), deletion of the C-terminus of Syt IV probably impairs correct folding, and that may induce the crystalloid ER. To further specify the essential region for the Golgi localization of Syt IV, we produced three internal deletion mutants (Fig. 5a; D44-68, D73-144, and D141-149) by comparison with Syt XI, a closely related isoform of Syt IV. In general, the spacer domain between the transmembrane domain and C2 domains is not conserved among the synaptotagmin family, and only a subclass of synaptotagmins (Syts IV and XI) show signi®cant homology within
Fig. 5 Effect of internal deletion of the spacer domain of Syt IV on Golgi localization. (a) Schematic representation of Syt IV internal deletion mutants and comparison between mouse Syt IV and Syt XI. Amino acid identities of each domain of Syt IV and Syt XI are indicated by percentages. T7-N2-tag, the transmembrane domain of Syt IV (and XI), and two C2 domains are represented by black boxes, open boxes, and hatched boxes, respectively. The Golgi localization of each mutant is indicated after its name, which was determined from the results of the immunocytochemical study (e) as well as the Golgi-speci®c modi®cations (c,d). ` ^' indicates that only a small proportion of Syt IVD73-144 localized at the Golgi. Amino acid numbers, except for the T7-N2-tag sequence, are given on both sides. (b) Amino acid sequence comparison between mouse Syt IV and Syt XI spacer domain. Double dots and single dots indicate the identical and similar amino acids, respectively, in Syt IV and Syt XI. Note that the N-terminal half of the spacer domain of Syts IV and XI is highly conserved (58.3% identity, see part a), but the C-terminal region did not show signi®cant homology (16.9% identity). # indicates the identical amino acids in Syts IV and XI (within the underlined region), all of which were replaced by Ala residues in T7-N2Syt IV(A8) mutant. (c) Exposure of T7-N2Syt IVD42-146, -N2Syt IVD42±146DC2B -N2Syt IVD44-68, and -N2Syt IVD73-144 to endoglycosidase H. T7-N2Syt IV deletion mutants were expressed in PC12 cells and
immunoprecipitated by anti-T7 tag antibody-conjugated agarose. Immunoprecipitants were treated with (H) and without (±) endoglycosidase H, subjected to 10% SDS±PAGE and then analyzed by immunoblotting with HRP-conjugated anti-T7 tag antibody. The broken boxes indicate the endoglycosidase H-resistant bands. Note that most of the N2Syt IVD42-146, N2Syt IVD42±146DC2B, and N2Syt IVD73-144 proteins were sensitive to endoglycosidase H treatment. (d) Proportion of endoglycosidase H (endo H )-resistant band of T7-N2Syt IVD42-146, -N2Syt IVD42±146DC2B, -N2Syt IVD44-68, -N2Syt IVD73-144, -N2Syt IVD141-149, and -N2Syt IV(A8). Immunoreactive bands in (c) were captured by Gel Print 2000i/VGA and quanti®ed by Basic Quanti®er Software. Columns indicate the proportion of endoglycosidase H-resistant bands (i.e. the proportion of proteins that were modi®ed at the Golgi). Note that the deletion of the unique C-terminal spacer domain (underlined in part b) was most effective on the Golgi-speci®c modi®cations. The results shown are representative of three independent experiments. (e) Subcellular localization of T7-N2Syt IV deletion mutants. T7-N2Syt IV deletion mutants were expressed in PC12 cells, as described under `Materials and Methods'. PC12 cells were ®xed, permeabilized, and stained with anti-T7 tag antibody. (e-1) T7-N2Syt IVD42-146 (e-2) -N2Syt IVD44-68 (e-3) -N2Syt IVD73-144, and (e-4) -N2Syt IVD141-149. Scale bar indicates 20 mm.
q 2001 International Society for Neurochemistry, Journal of Neurochemistry, 77, 730±740
A regulatory region of the Syt IV transport 737
this region (Figs 5a and b). The overall amino acid identity of Syt IV and Syt XI within the spacer domain between the transmembrane and C2 domains is 30.1%, but only the N-terminal half is signi®cantly conserved (Figs 5a and b; the
N-terminus, 58.3% and the C-terminus, 16.9% identities). Since we found that the Syt XI was mainly localized at the Golgi when it was expressed in PC12 cells (Fukuda and Mikoshiba 2001), we ®rst expected that the N-terminal
q 2001 International Society for Neurochemistry, Journal of Neurochemistry, 77, 730±740
738 M. Fukuda et al.
conserved region of the spacer domain is involved in Golgi localization. Unexpectedly, however, deletion of the N-terminal (D44±68) or short C-terminal region (D141±149) could not prevent the Golgi localization or the Golgi-speci®c modi®cations (Figs 5c±e). By contrast, deletion of the Syt IV unique region (D73±144) dramatically reduced the Golgi-speci®c modi®cations (Fig. 5d), and Syt IV D73±144 almost completely prevented the Golgi localization (Fig. 5e-3). To con®rm that the different Golgi localization signals were used for Syts IV and XI, we mutated the amino acids that are identical in Syts IV and XI within the C-terminal spacer domain by Ala-based sitedirected mutagenesis (Leu-92, Glu-97, Leu-101, Pro-110, Ser-114, Asp-117, Glu-128, and Ser-135 marked by # in Fig. 5b were replaced by Ala residues; named N2Syt IV(A8)). However, these mutant proteins were found to be mainly localized at the Golgi immunocytochemically (data not shown) and to have undergone the Golgi-speci®c modi®cations, like the wild-type proteins (Fig. 5d). Therefore, the localization signals of the two isoforms are probably different, even though both proteins are present at the Golgi. Discussion The synaptotagmins are a family of membrane proteins that share six separate domains. Their two C2 domains are believed to be the functional domain of the family, because the Syt I (or II) C2 domains are essential for Ca21-activated secretory vesicle exocytosis and endocytosis (reviewed in SuÈdhof and Rizo 1996; Fukuda and Mikoshiba 1997; Schiavo et al. 1998; MarqueÁze et al. 2000). However, the functions of the other domains, especially the spacer domain and C-terminus, remain obscure. In this study, we used deletion and mutation analyses in combination with Golgispeci®c modi®cations to show that the unique C-terminus of the spacer domain of Syt IV is required for Golgi localization (Figs 2 and 5), and the short C-terminus seemed to be required for proper folding of the proteins. Interestingly, the N-terminal half of the C-terminus is conserved among vertebrate Syts I-XI as well as during evolution (from nematoda to humans), suggesting that this region has some important role, and the C-terminus in Syt IV is involved in proper folding of the protein (especially the C2B domain), because deletion of the C-terminus prevented Golgi localization and induced crystalloid ER-like structures (Fig. 2n), while deletion of the whole C2B domain attenuated subcellular localization (Fig. 2h). Moreover, Syts I-XI proteins without the short C-terminus (Syts I±XIDC) showed completely different subcellular localizations from those of wild-type proteins when transiently expressed in PC12 cells, and in many cases we observed granules (crystalloid ER-like structures), the same as in Syt IVDC-expressing cells (M. Fukuda, unpublished
data). Syts IDC and IIDC proteins were not transported to the Golgi (mainly retained at the ER), and consequently they were not present at dense-core vesicles derived from TGN (trans-Golgi network). Consistent with this, Krasnov and Enikolopov recently reported that when the short C-terminus of Syt II was mutated or deleted, it was not transported to the tips of neurites (where dense-core vesicles are abundant) in nerve growth factor-differentiated PC12 cells (Krasnov and Enikolopov 2000). The origin and mechanism of formation of the crystalloid ER-like structures remain obscure, but they may contain malfolded Syt IV proteins, as in the case of expression of malfolded P-450 (Ishihara et al. 1995). Recent demonstrations of the three-dimensional structure of the Syt III C2A 1 C2B domains or the rabphilin-3 A C2B domain have indicated that the short C-terminus is part of the C2B domain rather than an independent domain (Sutton et al. 1999; Ubach et al. 1999). Therefore, it is likely that deletion of the short C-terminus destabilizes the C2B domain structure, and that this may induce the crystalloid ER structures. On the basis of these results, we speculate that the short C-terminus of the synaptotagmin family is essential for proper folding of the C2B domain. Our deletion mutant analysis showed that the C-terminal spacer domain of Syt IV is involved in Golgi localization, but we were unable to identify the speci®c amino acid residues responsible for the Golgi localization, as in the case of the YXXF motif of TGN38 (where X is any amino acid, and F is a strong hydrophobic amino acid). Interestingly, the C-terminal spacer domain of Syt IV (amino acids 73±144) is not conserved in other synaptotagmin family members, including Syt XI, a closely related isoform of Syt IV. Given that the spacer domain of Syt family is the highly diversi®ed region, one attractive explanation for the distinct subcellular localization of Syts I, III, IV, and VI in the brain (Butz et al. 1999; Fukuda and Mikoshiba 1999; Berton et al. 2000; Ibata et al. 2000) is that it results from the different sequences of their spacer domain. Moreover, there are some potential phosphorylation sites for certain kinases within the spacer domain of Syt IV. Previously, we and colleagues showed that protein kinase A is involved in the induction of the Syt IV mRNA (Vician et al. 1995) and that it seems to promote Syt IV protein transport from the Golgi to the cell periphery (Ibata et al. 2000). Further work is required to elucidate the relationship between Syt IV protein sorting and protein phosphorylation. In summary, we identi®ed the regulatory region for the Golgi localization of Syt IV, the C-terminus of the spacer domain. Since the spacer domain is one of the diversi®ed domains among synaptotagmin isoforms, it is tempting to speculate that it might determine the ®nal destination of each isoform. Nevertheless, because the spacer domain of Syt I has not been well conserved during evolution, the role of the spacer domain of Syt IV described here may not apply to other isoforms.
q 2001 International Society for Neurochemistry, Journal of Neurochemistry, 77, 730±740
A regulatory region of the Syt IV transport 739
Acknowledgements We thank Dr Shigekazu Nagata for the expression vector, Eiko Kanno for technical assistance, and Drs Masao Sakaguchi, and Akihiro Mizutani for critical comments on the manuscript. This work was supported in part by grants from the Science and Technology Agency to Japan (to K.M.) and Grants 11780571 and 12053274 from the Ministry of Education, Science, and Culture of Japan (to M.F.).
References Berton F., Cornet V., Iborra C., Garrido J., Dargent B., Fukuda M., Seagar M. and MarqueÁze B. (2000) Synaptotagmin I and IV de®ne distinct populations of neuronal transport vesicles. Eur. J. Neurosci. 12, 1294±1302. Butz S., Fernandez-Chacon R., Schmitz F., Jahn R. and SuÈdhof T. C. (1999) The subcellular localizations of atypical synaptotagmins III and VI. J. Biol. Chem. 274, 18290±18296. Craxton M. and Goedert M. (1999) Alternative splicing of synaptotagmins involving transmembrane exon skipping. FEBS Lett. 460, 417±422. Fukuda M. and Mikoshiba K. (1997) The function of inositol high polyphosphate binding proteins. Bioessays 19, 593±603. Fukuda M. and Mikoshiba K. (1999) A novel alternatively spliced variant of synaptotagmin VI lacking a transmembrane domain: implications for distinct functions of the two isoforms. J. Biol. Chem. 274, 31428±31434. Fukuda M. and Mikoshiba K. (2000a) Distinct self-oligomerization activities of synaptotagmin family: unique calcium-dependent oligomerization properties of synaptotagmin VII. J. Biol. Chem. 275, 28180±28185. Fukuda M. and Mikoshiba K. (2000b) Calcium-dependent and -independent hetero-oligomerization in the synaptotagmin family. J. Biochem. 128, 627±635. Fukuda M. and Mikoshiba K. (2000c) Expression of synaptotagmin I or II promotes neurite outgrowth in PC12 cells. Neurosci. Lett. 295, 33±36. Fukuda M. and Mikoshiba K. (2001) Characterization of KIAA1427 protein as an atypical synaptotagmin (Syt XIII). Biochem. J. 354, 249±257. Fukuda M., Aruga J., Niinobe M., Aimoto S. and Mikoshiba K. (1994) Inositol-1,3,4,5-tetrakisphosphate binding to C2B domain of IP4bp/synaptotagmin II. J. Biol. Chem. 269, 29206±29211. Fukuda M., Kojima T., Aruga J., Niinobe M. and Mikoshiba K. (1995a) Functional diversity of C2 domains of synaptotagmin family: mutational analysis of inositol high polyphosphate binding domain. J. Biol. Chem. 270, 26523±26527. Fukuda M., Moreira J. E., Lewis F. M. T., Sugimori M., Niinobe M., Mikoshiba K. and LlinaÂs R. (1995b) Role of the C2B domain of synaptotagmin in vesicular release and recycling as determined by speci®c antibody injection into the squid giant synapse preterminal. Proc. Natl Acad. Sci. USA 92, 10708±10712. Fukuda M., Kojima T. and Mikoshiba K. (1996) Phospholipid composition dependence of Ca21-dependent phospholipid binding to the C2A domain of synaptotagmin IV. J. Biol. Chem. 271, 8430±8434. Fukuda M., Kojima T. and Mikoshiba K. (1997) Regulation by bivalent cations of phospholipid binding to the C2A domain of synaptotagmin III. Biochem. J. 323, 421±425. Fukuda M., Kanno E. and Mikoshiba K. (1999) Conserved N-terminal cysteine motif is essential for homo- and heterodimer formation
of synaptotagmins III, V, VI, and X. J. Biol. Chem. 274, 31421±31427. Fukuda M., Moreira J. E., Liu V., Sugimori M., Mikoshiba K. and LlinaÂs R. R. (2000) Role of the conserved WHXL motif in the C terminus of synaptotagmin in synaptic vesicle docking. Proc. Natl Acad. Sci. USA 97, 14715±14719. Hilbush B. S. and Morgan J. I. (1994) A third synaptotagmin gene, Syt3, in the mouse. Proc. Natl Acad. Sci. USA 91, 8195±8199. Ibata K., Fukuda M. and Mikoshiba K. (1998) Inositol 1,3,4,5tetrakisphosphate binding activities of neuronal and non-neuronal synaptotagmins: identi®cation of conserved amino acid substitutions that abolish inositol 1,3,4,5-tetrakisphosphate binding to synaptotagmins III, V, and X. J. Biol. Chem. 273, 12267±12273. Ibata K., Fukuda M., Hamada T., Kabayama H. and Mikoshiba K. (2000) Synaptotagmin IV is present at the Golgi and distal parts of neurites. J. Neurochem. 74, 518±526. Ishihara N., Yamashina S., Sakaguchi M., Mihara K. and Omura T. (1995) Malfolded cytochrome P-450 (M1) localized in unusual membrane structures of the endoplasmic reticulum in cultured animal cells. J. Biochem. 118, 397±404. Kabayama H., Takei K., Fukuda M., Ibata K. and Mikoshiba K. (1999) Functional involvement of synaptotagmin I/II C2A domain in neurite outgrowth of chick dorsal root ganglion neuron. Neuroscience 88, 999±1003. Krasnov P. A. and Enikolopov G. (2000) Targeting of synaptotagmin to neurite terminals in neuronally differentiated PC12 cells. J. Cell Sci. 113, 1389±1404. Lang J., Fukuda M., Zhang H., Mikoshiba K. and Wollheim C. B. (1997) The ®rst C2 domain of synaptotagmin is required for exocytosis of insulin from pancreatic b-cells: action of synaptotagmin at low micromolar calcium. EMBO J. 16, 5837±5846. Li C., Ullrich B., Zhang J. Z., Anderson R. G. W., Brose N. and SuÈdhof T. C. (1995) Ca21-dependent and -independent activities of neural and non-neural synaptotagmins. Nature 375, 594±599. Littleton J. T., Serano T. L., Rubin G. M., Ganetzky B. and Chapman E. R. (1999) Synaptic function modulated by changes in the ratio of synaptotagmin I and IV. Nature 400, 757±760. MarqueÁze B., Berton F. and Seagar M. (2000) Synaptotagmins in membrane traf®c: which vesicles do the tagmins tag? Biochimie 82, 409±420. Mehrotra B., Elliott J. T., Chen J., Olszewski J. D., Pro®t A. A., Chaudhary A., Fukuda M., Mikoshiba K. and Prestwich G. D. (1997) Selective photoaf®nity labeling of the inositol polyphosphate binding C2B domains of synaptotagmins. J. Biol. Chem. 272, 4237±4244. Mikoshiba K., Fukuda M., Moreira J. E., Lewis F. M. T., Sugimori M., Niinobe M. and LlinaÂs R. (1995) Role of the C2A domain of synaptotagmin in transmitter release as determined by speci®c antibody injection into the squid giant synapse preterminal. Proc. Natl Acad. Sci. USA 92, 10703±10707. Mochida S., Fukuda M., Niinobe M., Kobayashi H. and Mikoshiba K. (1997) Roles of synaptotagmin C2 domains in neurotransmitter secretion and inositol high-polyphosphate binding at mammalian cholinergic synapses. Neuroscience 77, 937±943. Ohara-Imaizumi M., Fukuda M., Niinobe M., Misonou H., Ikeda K., Murakami T., Kawasaki M., Mikoshiba K. and Kumakura K. (1997) Distinct roles of C2A and C2B domains of synaptotagmin in the regulation of exocytosis in adrenal chromaf®n cells. Proc. Natl Acad. Sci. USA 94, 287±291. Pathak R. K., Luskey K. L. and Anderson R. G. W. (1986) Biogenesis of the crystalloid endoplasmic reticulum in UT-1 cells: evidence that newly formed endoplasmic reticulum emerges from the nuclear envelope. J. Cell Biol. 102, 2158±2168.
q 2001 International Society for Neurochemistry, Journal of Neurochemistry, 77, 730±740
740 M. Fukuda et al.
Schiavo G., Osborne S. L. and Sgouros J. G. (1998) Synaptotagmins: more isoforms than functions? Biochem. Biophys. Res. Commun. 248, 1±8. SuÈdhof T. C. and Rizo J. (1996) Synaptotagmin: C2-domain proteins that regulate membrane traf®c. Neuron 17, 379±388. Sutton R. B., Ernst J. A. and Brunger A. T. (1999) Crystal structure of the cytosolic C2A-C2B domains of synaptotagmin III: implications for Ca21-independent snare complex interaction. J. Cell Biol. 147, 589±598. Thomas D. M. and Elferink L. A. (1998) Functional analysis of the C2A domain of synaptotagmin 1: implications for calcium-regulated secretion. J. Neurosci. 18, 3511±3520. Thomas D. M., Ferguson G. D., Herschman H. R. and Elferink L. A.
(1999) Functional and biochemical analysis of the C2 domains of synaptotagmin IV. Mol. Biol. Cell. 10, 2285±2295. Ubach J., Garcia J., Nittler M. P., SuÈdhof T. C. and Rizo J. (1999) Structure of the Janus-faced C2B domain of rabphilin. Nat. Cell Biol. 1, 106±112. Vician L., Lim I. K., Ferguson G., Tocco G., Baudry M. and Herschman H. R. (1995) Synaptotagmin IV is an immediate early gene induced by depolarization in PC12 cells and in brain. Proc. Natl. Acad. Sci. USA 92, 2164±2168. Yamamoto A., Masaki R. and Tashiro Y. (1996) Formation of crystalloid endoplasmic reticulum in COS cells upon overexpression of microsomal aldehyde dehydrogenase by cDNA transfection. J. Cell Sci. 109, 1727±1738.
q 2001 International Society for Neurochemistry, Journal of Neurochemistry, 77, 730±740
