PACSIN 1 forms tetramers via its N-terminal F-BAR domain Arndt Halbach 1 , Matthias Mo ¨ rgelin 2 , Maria Baumgarten 2 , Mark Milbrandt 1 , Mats Paulsson 1 and Markus Plomann 1 1 Center for Biochemistry and Center for Molecular Medicine (CMMC), Medical Faculty, University of Cologne, Germany 2 Department of Clinical Sciences, Section for Clinical and Experimental Infectious Medicine, University of Lund, Sweden In eukaryotic cells, complex regulatory mechanisms involving numerous proteins must operate to ensure the temporal and spatial specificity of intracellular membrane-trafficking pathways. We and others have identified three members of the protein kinase C and casein kinase 2 substrate in neurons (PACSIN) protein family, also named syndapin and focal adhesion pro- tein 52 (FAP52), which participate in rearrangements of actin networks during endocytosis [1–5]. In contrast to the neuron-specific PACSIN 1, other members of the PACSIN protein family show a broader tissue dis- tribution [2,4,6]. Via their C-terminal Src homology 3 (SH3) domains, PACSIN proteins bind to proline-rich domains of dynamin, synapsin and synaptojanin, three proteins also involved in vesicle endocytosis, as well as to neural Wiskott–Aldrich syndrome protein Keywords F-BAR domain; membrane; oligomerization; PACSIN 1; syndapin 1 Correspondence M. Plomann, Center for Biochemistry, Medical Faculty, University of Cologne, Joseph-Stelzmann-Str. 52, D-50931 Cologne, Germany Fax: +49 221 478 6977 Tel: +49 221 478 6944 E-mail: markus.plomann@uni-koeln.de (Received 13 October 2006, revised 23 November 2006, accepted 4 December 2006) doi:10.1111/j.1742-4658.2006.05622.x The ability of protein kinase C and casein kinase 2 substrate in neurons (PACSIN) ⁄ syndapin proteins to self-polymerize is crucial for the simulta- neous interactions with more than one Src homology 3 domain-binding partner or with lipid membranes. The assembly of this network has pro- found effects on the neural Wiskott–Aldrich syndrome protein-mediated attachment of the actin polymerization machinery to vesicle membranes as well as on the movement of the corresponding vesicles. Also, the sensing of vesicle membranes and ⁄ or the induction of membrane curvature are more easily facilitated in the presence of larger PACSIN complexes. The N-ter- minal Fes-CIP homology and Bin-Amphiphysin-Rvs (F-BAR) domains of several PACSIN-related proteins have been shown to mediate self-inter- actions, whereas studies using deletion mutants derived from closely related proteins led to the view that oligomerization depends on the formation of a trimeric complex via a coiled-coil region present in these molecules. To address whether the model of trimeric complex formation is applicable to PACSIN 1, the protein was recombinantly expressed and tested in four different assays for homologous interactions. The results showed that PACSIN 1 forms tetramers of about 240 kDa, with the self-interaction having a K D of 6.4 · 10 )8 m. Ultrastructural analysis of these oligomers after negative staining showed that laterally arranged PACSIN molecules bind to each other via a large globular domain and form a barrel-like structure. Together, these results demonstrate that the N-terminal F-BAR domain of PACSIN 1 forms the contact site for a tetrameric structure, which is able to simultanously interact with multiple Src homology 3 bind- ing partners. Abbreviations ADAM, a disintegrin and metalloprotease; BAR, Bin-Amphiphysin-Rvs; BS 3 , bis[sulfosuccinimidyl]suberate; FAP52, focal adhesion protein 52; F-BAR, Fes-CIP4 homology and Bin-Amphiphysin-Rvs; FCH, Fes-CIP4 homology; GST, glutathione S-transferase; NMDA, N-methyl- D- aspartate; NOSTRIN, eNOS-trafficking inducer; N-WASP, neural Wiskott–Aldrich syndrome protein; PACSIN, protein kinase C and casein kinase 2 substrate in neurons; PACSIN 1-CS, PACSIN 1 carrying a C-terminal Strep II tag; PCH, pombe CDC15 homology; PSTPIP, proline, serine, threonine phosphatase-interacting protein; SH3, Src homology 3; Sulfo-EGS, ethylene glycol bis[sulfosuccinimidylsuccinate]. FEBS Journal 274 (2007) 773–782 ª 2007 The Authors Journal compilation ª 2007 FEBS 773 (N-WASP), a stimulator of actin-related protein 2 ⁄ 3-induced actin polymerization [4,6]. Accordingly, the PACSIN proteins have been implicated not only in vesicle endocytosis at the plasma membrane but in a variety of membrane traffic events, most of which occur at membranes of intracellular sorting compart- ments [7,8]. PACSIN proteins also interact with specific trans- membrane proteins, such as ADAM (a disintegrin and metalloprotease) metalloproteinases [9,10], the CD95 ligand [11] and, in the case of PACSIN 1, the phos- phodiesterase 6c [12] and the N-methyl-d-aspartate (NMDA) receptor chain NR3A [13]. These interac- tions, most of which also involve the SH3 domains, indicate a role of PACSIN proteins in the regulation of the surface expression of some transmembrane molecules by endocytosis. For NR3A, we were able to demonstrate an activity-dependent mechanism by which PACSIN 1 regulates NMDA receptor expres- sion at synapses during development [13]. PACSIN proteins represent a subgroup within a lar- ger protein family, named pombe CDC15 homology (PCH), displaying a similar arrangement of domains, including at least one C-terminal SH3 domain and a conserved N-terminal region (Fig. 1A). The latter was originally defined as the CDC15-N-terminal (CDC15- NT) domain, spanning about 250 amino acids in PAC- SIN proteins [4]. Later, others distinguished between a region covering the N-terminal circa 100 amino acids, named the Fes-CIP4 homology (FCH) domain, and the adjacent a-helical stretch, which is believed to form a coiled-coil structure [14]. So far, no function has been reported for the FCH domain, whereas the a-helical region has been shown to be responsible for oligomerization of PCH proteins [15–17]. Recently, the whole N-terminal region, corresponding to the CDC15-NT domain and renamed the FCH and Bin- Amphiphysin-Rvs (BAR) (F-BAR) domain, has been characterized in more detail [18]. Like classical BAR domains, this related domain is able to bind to lipid bilayers. Moreover, the PCH proteins tested, including PACSIN ⁄ syndapin 1, bind to liposomes containing phosphatidylserine and phosphoinositides, and are alone sufficient to deform them into tubules [18]. In agreement with this, PACSIN 1 was recently identified as the key interaction partner of dynamin 1 in synaptic vesicle endocytosis [19]. In this process, PACSIN 1 is thought to bind to the synaptic plasma membrane, and induce a curvature in the membrane and neck for- mation prior to vesicle fission through the action of dynamin 1. Another proposed role for PACSIN proteins is in attachment of the actin polymerization machinery to vesicles after endocytosis. They are believed to act as linkers between the endocytic protein dynamin and N-WASP, thereby directing the actin propulsion machinery to the site of vesicle fission [4,6]. As several important interaction partners, including dynamin and N-WASP, interact with PACSIN’s single SH3 domain [4], only PACSIN oligomers would be able to act as linkers. Previously, we showed that, in vitro, all three PACSIN proteins are able to to bind to each other and might exist as homo-oligomers and ⁄ or hetero- oligomers [4]. Recently, the interconnecting function of PACSIN oligomers was shown to be essential for PACSIN-mediated cytoskeletal rearrangements and A BC Fig. 1. Domain structure and purification of recombinant PACSIN 1. (A) Domain struc- ture of PACSIN 1 with an N-terminal F-BAR domain and a C-terminal SH3 domain. The F-BAR domain was formerly known as the CDC15-NT domain, and includes a con- served region, also named the FCH domain, and an a-helical polypeptide stretch postula- ted to act as a coiled-coil domain. Addition- ally, PACSIN 1 contains two asparagine- proline-phenylalanine (NPF) motifs. (B) Fractions eluted from a StrepTactin Sepharose column were purified by an additional ultracentrifugation step and tested for purity by SDS ⁄ PAGE under reducing conditions. C, Coomassie Brilliant Blue; S, silver nitrate. The arrow marks the recombinant PACSIN 1-CS. (C) MALDI-TOF MS analysis of the same purified PACSIN 1 sample. PACSIN 1 oligomerization A. Halbach et al. 774 FEBS Journal 274 (2007) 773–782 ª 2007 The Authors Journal compilation ª 2007 FEBS endocytosis [20]. We report here that purified PAC- SIN 1 exists as a 230–240 kDa tetrameric complex in solution. Although it shares many structural and func- tional properties with other members of the PCH pro- tein family, PACSIN 1 is distinct in that it forms barrel-like homotetramers in vitro, held together by the N-terminal F-BAR domain. Results Purification and characterization of PACSIN 1 PACSIN 1, carrying a C-terminal Strep II tag (PAC- SIN 1-CS) and expressed in HEK293 cells, was puri- fied by a single affinity chromatography step. The homogeneity of the preparation was demonstrated by reducing SDS ⁄ PAGE followed either by staining with Coomassie Brilliant Blue and silver nitrate (Fig. 1B) or by immunoblotting (data not shown). MALDI- TOF MS analysis confirmed the purity (Fig. 1C). Secondary structure and thermal stability of PACSIN 1 In order to analyze the conformation and thermal sta- bility of PACSIN 1, the purified protein was studied by CD spectroscopy. Aliquots of PACSIN 1-CS (100 lgÆmL )1 )in5mm Tris ⁄ HCl (pH 7.4) were incu- bated at 20 °C, 37 °C, or 60 °C, and CD spectra were recorded. At the two lower temperatures, PACSIN 1 showed similar CD spectra with characteristic minima in the 208–223 nm region, indicative of a high a-helical content (Fig. 2A). When PACSIN 1 CD spectra were recorded at 60 °C, changes were observed, reflecting a loss of a-helical structure (Fig. 2A). This heat denatur- ation was irreversible, as a subsequent decrease in tem- perature was unable to restore the original structure (results not shown). Furthermore, calculation of the relative proportions of different secondary structure elements by three different algorithms revealed that, at 37 °C, PACSIN 1 contains about one-third a-helix, a portion of which is lost upon heating, accompanied by an increase in b-structure (supplementary Table S1). A melting point of 44 °C was determined from the mid- point of the transition at 221 nm (Fig. 2B). Oligomeric structure of PACSIN 1 complexes To determine the oligomerization state of PACSIN 1, purified PACSIN 1-CS was analyzed by gel filtration. In a KCl-containing buffer at cytosolic ion strength, PACSIN 1 eluted as a single peak between the marker proteins aldolase (158 kDa) and ferritin (440 kDa) (Fig. 3A). The presence of PACSIN 1 in individual fractions was confirmed by immunoblotting (results not shown). For the calculation of the molecular mass of the PACSIN protein complex, the K av values of the marker proteins used as a standard were plotted against log M r . With this method, the molecular mass of the native PACSIN complex was determined as being 234 kDa, corresponding to an oligomerization state of about 4.5 (Fig. 3B, supplementary Table S2A). Furthermore, performing the same experiment with a bis(sulfosuccinimidyl)suberate (BS 3 )-crosslinked sample led to comparable results (M r ¼ 240 kDa; oligomeriza- tion state 4.6), showing that crosslinking with this rea- gent captures PACSIN complexes in a native state (Fig. 3C, supplementary Table S2B). We next used protein crosslinking to study intermo- lecular interactions between PACSIN 1 subunits. Three homobifunctional imidoester reagents, disuccinimidyl suberate (DSS), BS 3 and ethylene glycol bis(sulfosuc- Fig. 2. CD spectra of PACSIN 1. (A) PACSIN 1-CS protein (100 lgÆ mL )1 in 5 mM Tris ⁄ HCl, pH 7.4) was measured at 20 °C (solid line), 37 °C (dotted line) and 60 °C (dashed line). Changes in the content of a-helix and b-structure were observed between 37 °C and 60 °C. (B) For the determination of the melting temperature, recordings were performed at 221 nm with a linear temperature gradient from 20 °Cto80°C. The midpoint of the conformational transition was at 44 °C. A. Halbach et al. PACSIN 1 oligomerization FEBS Journal 274 (2007) 773–782 ª 2007 The Authors Journal compilation ª 2007 FEBS 775 cinimidylsuccinate) (Sulfo-EGS) were tested for their ability to covalently link recombinantly expressed PACSIN 1-CS molecules. All reagents crosslinked PACSIN 1 to dimers (Fig. 4A, lanes 2 and 3 for BS 3 , and lane 4 for Sulfo-EGS) or, at higher crosslinker concentrations, to tetramers (Fig. 4A, lanes 4–7 for BS 3 , and lanes 5–7 for Sulfo-EGS). Sulfo-EGS differs from DSS and BS 3 by having a slightly longer spacer arm (16.1 A ˚ versus 11.4 A ˚ ), and was less efficient in crosslinking PACSIN 1 to tetramers. Adducts larger than tetramers were only occasionally observed at higher protein concentrations (50 lg versus 20 lg), corroborating the tetrameric structure indicated by the gel filtration experiments. MS analysis of crosslinked fractions confirmed the presence of dimers, but ionization was insufficient for the detection of higher oligomers (Fig. 4B,C). The mod- erate increase in mass of PACSIN 1 monomers resulted from bound crosslinker molecules. The actual presence of tetrameric complexes in a BS 3 -crosslinked sample was confirmed by size exclusion chromatography, which resulted in a single symmetrical peak comparable to the previously analyzed native sample (Fig. 3C). As oligomerization is a prerequisite for the proposed function of PACSIN as a linking protein, surface plas- mon resonance was used to further support the pres- ence of a self-interaction between PACSIN 1 subunits and to determine the strength of this binding (Fig. 5). High-affinity binding between PACSIN 1 monomers could indeed be detected (Fig. 5), and a k a of 1.44 · 10 5 m )1 Æs )1 ,ak D of 4.3 · 10 )3 and a K D of 6.4 · 10 )8 m were calculated (supplementary Table S3). Purified full-length PACSIN 1-CS (Fig. 1) was also submitted to electron microscopy after negative stain- ing with uranyl formate (Fig. 6). The protein particles were heterogeneous in size, and closer examination revealed that both monomers and, predominantly, higher aggregates were present in the sample (Fig. 6A). Most of the monomers formed elongated curved struc- tures, but fully extended 7–8 nm monomeric particles could occasionally be seen (Fig. 6B). Dimers showed a lateral alignment of PACSIN molecules joined at one end. Tetramers displayed a barrel-like structure, often with a more heavily stained hole in the middle and most mass at the periphery (Fig. 6C,D). The average diameter of the tetramers was 8 nm. Occasionally, par- ticles were oriented to give a top view (Fig. 6D, top panel). We never observed a waist-like structure, which would have indicated that a rod-like coiled-coil a-helix might assemble the higher-order structure. Electron microscopy of negatively stained recombinant F-BAR domains of PACSIN 1 again showed monomeric (Fig. 6E, upper row) and dimeric (Fig. 6E, lower row) Fig. 3. Analysis of PACSIN 1 oligomers by gel filtration. (A) Elution profile of PACSIN 1-CS from size exclusion chromatography using a Superdex 200 column (solid line). The elution profile of a mixture of thyroglobin (669 kDa), ferritin (440 kDa), aldolase (158 kDa), ovalbu- min (43 kDa) and ribonuclease A (13.7 kDa) is shown as a dotted line. Numbers at the peaks represent the molecular masses of the corresponding marker proteins. The Y-axes show the relative fluor- escence intensities at 280 nm, with the left axis corresponding to the marker proteins, and the right axis to PACSIN 1-CS. AU ¼ arbi- trary units. (B) Plot of K av values of marker proteins against log M r . Numbers at the open squares represent the molecular masses. The relative mass of the native PACSIN complex was determined as being about 234 kDa (closed square), corresponding to an olig- omerization state of 4.5. (C) Analysis of BS 3 -crosslinked PACSIN 1 gave comparable results, with a molecular mass of 240 kDa and an oligomerization state of 4.6 (closed square). Numbers at the open squares represent the molecular masses of the corresponding mar- ker proteins. PACSIN 1 oligomerization A. Halbach et al. 776 FEBS Journal 274 (2007) 773–782 ª 2007 The Authors Journal compilation ª 2007 FEBS particles, demonstrating that this domain is sufficient for oligomerization. Discussion The present characterization of the neurospecific repre- sentative of the PACSIN proteins, PACSIN 1 [1,3], supports the hypothesis that PACSINs act as linking molecules in vesicular trafficking. It was previously shown that PACSIN proteins bind to both dynamin and N-WASP, and that impairment of these inter- actions leads to changes in actin dynamics, block of endocytosis, and mislocalization of involved proteins [4,6,17,20–22]. As both binding partners are recognized by the single C-terminal SH3 domain of PACSINs, multiple simultaneous interactions are only possible if PACSINs form oligomers. Members of the PCH fam- ily contain at least one a-helical polypeptide stretch, which is assumed to form a coiled-coil and thereby enable oligomerization. Several studies on individual members of this protein family have shown their abil- ity to homo-oligomerize to dimers [proline, serine, threonine phosphatase-interacting protein (PSTPIP) and PSTPIP 2 [15]], or trimers [FAP52 [16] and (endothelial nitric oxide synthase) eNOS-trafficking inducer (NOSTRIN) [17]]. We previously observed that all PACSIN isoforms are able to interact with each other in two-hybrid assays [4], and recently another study confirmed the ability of PACSIN proteins to self-associate [20]. To determine the A B C Fig. 4. Crosslinking of PACSIN 1. (A) PACSIN 1 was incubated at 50 lgÆmL )1 with 0.5 lM–10 mM BS 3 (left panel) or at 20 lgÆmL )1 with 0–5 mM Sulfo-EGS cross- linker (right panel) and analyzed by reducing SDS ⁄ PAGE on 5–15% gels. The crosslinked products are labeled. (B, C) The products of PACSIN 1 crosslinked with 25 l M (*, B) and 2.5 m M Sulfo-EGS (#, C) were analyzed by MALDI-TOF MS. A. Halbach et al. PACSIN 1 oligomerization FEBS Journal 274 (2007) 773–782 ª 2007 The Authors Journal compilation ª 2007 FEBS 777 stoichiometry of PACSIN 1 oligomers, we expressed recombinant PACSIN 1 in eukaryotic cells and, by use of CD spectroscopy, confirmed that the protein was correctly folded. The content of a-helix in the full- length protein was found to be 33.7–37.9%, depending on the software used, which is significantly higher than the 25.8% calculated for the closely related PACSIN 2 ortholog FAP52 [16]. Heat treatment leads to an irre- versible loss of a high proportion of these helices and, interestingly, to an increase of b-structure (supplement- ary Table S1). When subjected to size exclusion chromatography, PACSIN 1 eluted as a complex of about 234 kDa, indicating that oligomers are formed in solution. This mass slightly exceeds that of a tetramer, which may result from the shape of the PACSIN 1 complexes when compared to the standard proteins. A similar oligomerization has been shown for a recombinantly produced GST–NOSTRIN fragment [17], but here gel filtration indicated a trimerization. In a recent publi- cation, it was suggested that PACSIN 1 predomin- antly forms dimers in vivo [20]. This was concluded from crosslinking studies in brain and cell extracts, but less well-resolved higher molecular weight com- plexes were also observed [20]. To address the appar- ent discrepancy with our gel filtration results, we used increasing concentrations of three different crossl- inkers with varying spacer arm lengths, and clearly detected the preferential formation of PACSIN 1 dimers and tetramers. In contrast to the other studies, we avoided the use of cell lysates in which potential exogenous interaction partners might be present and Fig. 5. Surface plasmon resonance binding curves obtained for the PACSIN 1 self-interaction. Antibodies to GST were coupled (15 000 RU) to a CM5 sensor chip and saturated with GST– PACSIN 1 or GST as a control. PACSIN 1-CS was injected at differ- ent concentrations. The binding curves shown have been corrected by subtracting the values obtained with GST alone. A B C D E Fig. 6. Electron microscopy of negatively stained recombinant PACSIN 1-CS and the PACSIN 1 F-BAR domain. The overview (A) shows aggregates as well as monomeric PACSIN 1 molecules; the lower panels show selected monomeric (B), dimeric (C) and tetra- meric (D) particles. The bottom panel (E) shows PACSIN 1 F-BAR monomers (upper row) and dimers (lower row). The picture at the lower right includes a monomer located next to the dimer. The bars correspond to 10 nm. PACSIN 1 oligomerization A. Halbach et al. 778 FEBS Journal 274 (2007) 773–782 ª 2007 The Authors Journal compilation ª 2007 FEBS in which complexes may include proteins other than PACSIN. The analysis of PACSIN 1 self-association by real-time surface plasmon resonance gave a K D of 6.4 · 10 )8 m, which is comparable to the K D of 4.7 · 10 )9 m calculated for FAP52 [16]. Although clo- sely related to PACSIN 2, FAP52 appears to partici- pate in different processes. In chicken embryo heart fibroblasts, it localizes to focal adhesion contacts [5], and this was never observed for any of the three PACSIN isoforms. To complement our biochemical analysis, we also employed electron microscopy of negatively stained PACSIN 1 complexes; this confirmed the formation of dimers and tetramers observed in solution. The PACSIN tetramers form a barrel-like structure in the absence of any lipids or other proteins. Taken together, our results suggest that tetramers are the highest oligo- mers formed by PACSIN 1, and that dimers may be intermediates in the assembly. The proportions of dimers and tetramers detected may depend on the ana- lytical method used. None of the particles seen with electron microscopy showed a waist-like structure, which would have been indicative of assembly via a coiled-coil a-helix. Instead, subunit contacts appear to be mediated by domains of globular shape. These may be the PACSIN 1 F-BAR domains, particularly as elec- tron microscopy of isolated PACSIN 1 F-BAR domains also shows oligomerization (Fig. 6E). The importance of this region for self-assembly has been confirmed for PACSIN 1 [20], and also been reported for other F-BAR domain-containing proteins [16,17]. The exact role of PACSINs is controversial, especi- ally with regard to when and where PACSIN mole- cules contribute to vesicle formation and removal. PACSIN proteins have been proposed to play a role in the regulation of transferrin endocytosis. However, these findings were based on overexpression of either isolated SH3 domains [23] or full-length proteins [4], and may reflect an impairment of proper dynamin localization, as overexpression of other dynamin-bind- ing SH3 domains has comparable effects [22–24]. A recent study demonstrated that F-BAR domains are able to bind to phospholipids, in particular to mem- branes containing phosphatidylserine, and that they are able to cause membranes to form tubules in vitro [18]. This suggests an involvement of PACSIN proteins early in vesicle formation at donor membranes, which has recently been confirmed for PACSIN 1 at nerve terminals [19]. Here, the phosphorylation-dependent interaction of PACSIN 1 with dynamin 1 is essential for synaptic vesicle endocytosis. The authors propose a new model in which PACSIN 1 induces membrane curvature and ⁄ or formation of a neck at endocytic sites before dynamin 1 facilitates vesicle fission inde- pendently of the actin cytoskeleton. PACSIN 1 F-BAR domain oligomers might be required for this function, as the related BAR domains need to dimerize to be active [25]. Also, PACSIN 1 tetramers localized around a vesicle neck could provide multiple docking sites for dynamin molecules. However, biochemical analysis revealed that PAC- SIN 1 is present in microsomal and cytosolic fractions from brain ([26]; unpublished results) and can only occasionally be detected at the plasma membrane [3,4,27]. Immunofluorescence microscopy clearly shows that most endogenous PACSIN 1 molecules are distri- buted throughout the neuron, including synapses, pro- cesses and cell bodies [3,4]. Also, proteomic studies of the composition of clathrin-coated vesicles [28], postsy- naptic densities [29] and brain plasma membranes [30] failed to identify PACSIN 1. Recently, we demonstra- ted a postsynaptic role for PACSIN 1 in regulating NR3A endocytosis [13], which may represent an exam- ple of PACSIN 1 acting as a linker molecule on moving in neurons. It has previously been shown that dynamin remains attached to vesicle membranes after scission, and serves as an anchoring site for actin tails [31,32]. The proline-rich region of dynamin is essential for the formation of actin comet tails, indicating that interac- tion partners that bind via their SH3 domains to this region, such as PACSIN proteins, may also be neces- sary. The role of PACSIN proteins in connecting the GTPase dynamin with N-WASP through oligomeriza- tion has recently been confirmed [20]. However, tetramerization of PACSIN proteins provides a more efficient mode of interconnection. The results presented here show that PACSIN forms tetramers via its F-BAR domain. Such tetramers may participate in synaptic vesicle endocytosis by deform- ing the corresponding membrane, and ⁄ or in the assem- bly of the vesicle docking site for actin-mediated propulsion [19,20]. The increasing number of trans- membrane molecules identified as PACSIN-binding partners may represent the cargo molecules transpor- ted by these vesicles, and may provide the specificity of the PACSIN–vesicle association. Experimental procedures Expression and purification of PACSIN 1 A full-length murine PACSIN 1 cDNA clone in pBlue- script [1] was used as template for PCR using AmpliTaq DNA Polymerase (Perkin Elmer, Wellesley, MA, USA) and specific primers (sense, 5¢-AAG CTT GCC ACC ATG TCT GGC TCC TAC GAT GAG GCC-3¢; antisense, A. Halbach et al. PACSIN 1 oligomerization FEBS Journal 274 (2007) 773–782 ª 2007 The Authors Journal compilation ª 2007 FEBS 779 5¢-GCG GCC GCT ATA GCC TCA ACG TAG TTG G-3¢). The a mplified DNA fragment was cloned into the pCR2.1 vector (Invitrogen, Karlsruhe, Germany), and after sequence confirmation was digested with HindIII and NotI and cloned into the HindIII–NotI-digested expression vector pCEP- puBM40-cStrep [33]. This produced a fusion protein in which a Strep II tag was placed in frame with the PACSIN 1 coding region. The plasmid was transfected into 293-EBNA cells by electroporation, and the cells were subsequently selected for puromycin resistance. Cell pellets were lysed in NaCl ⁄ P i (pH 7.5) containing 0.25 mm sucrose and 1 mm phenyl- methanesulfonyl fluoride by sonification, and centrifuged at 20 000 g for 15 min at 4 °C (Beckman ultracentrifuge L7-55, SW41 Ti rotor), and finally at 180 000 g for 2 h at 4 °C. Supernatants containing Strep II-tagged PACSIN 1 were loaded on a StrepTactin Sepharose column (IBA, Gottingen, Germany) at a flow rate of 0.5 mLÆmin )1 . After being washed with 10 column volumes of 100 mm Tris ⁄ HCl (pH 8.0) con- taining 1 mm EDTA and 1 mm phenylmethanesulfonyl fluor- ide, the proteins were eluted with the same buffer containing 2.5 mm desthiobiotin. The protein samples were resolved by SDS ⁄ PAGE, and analyzed either by Coomassie or silver staining of the gel, or transferred to a poly(vinylidene difluo- ride) membrane and detected with antibodies against PAC- SIN 1 [1]. Glutathione S-transferase (GST) fusion proteins of PACSIN 1 were produced by cloning cDNAs correspond- ing to either the complete coding region of PACSIN 1 or the F-BAR domain (amino acids 1–285) into the pGEX-6P vector (Amersham Pharmacia Biotech, Freiburg, Germany) and then expressing in Escherichia coli (BL21). The fusion proteins were purified by affinity chromatography on gluta- thione–Sepharose 4B, and GST was removed by cleavage with Precission protease (Amersham Pharmacia Biotech) for some applications. Gel filtration analysis For the size determination of purified PACSIN complexes, freshly purified recombinant PACSIN 1-CS was dialyzed against 10 mm Pipes ⁄ KOH (pH 7.4) containing 100 mm KCl, 3 mm NaCl and 3.5 mm MgCl 2 . The sample (50 lgÆ mL )1 ) was applied to a Pharmacia SMART Superdex 200 column and analyzed at a flow rate of 10 lLÆmin )1 . For size calculation, the standard proteins ribonuclease A (13.7 kDa), ovalbumin (43 kDa), aldolase (158 kDa), fer- ritin (440 kDa) and thyroglobin (669 kDa) were treated equally and analyzed. The eluted fractions were monitored at 280 nm by UV photometry. CD measurements CD spectra were recorded in a Jasco (Gross-Umstadt, Germany) J-715 spectropolarimeter. PACSIN 1-CS was dialyzed against 5 mm Tris ⁄ HCl (pH 7.5), at a concentra- tion of 100 lgÆmL )1 . Crosslinking assays Crosslinking assays were carried out using the three lysine side-chain-reactive crosslinkers (Pierce, Rockford, IL, USA), BS 3 , DSS and Sulfo-EGS. The PACSIN 1-CS was added at a concentration of 20 or 50 lgÆmL )1 . The reaction was carried out in a final volume of 40 lL in NaCl ⁄ P i (pH 7.4) for 1 h at 4 °C and was stopped by the addition of 10 lLof1m Tris ⁄ HCl (pH 8.0). MALDI-TOF MS For MALDI-TOF MS analysis, the samples were dissolved in 5 l L of 0.1% aqueous trifluoroacetic acid. MS was carried out in linear mode on a Bruker Reflex IV equipped with a video system, a nitrogen UV laser (O max ¼ 337 nm) and a HiMass detector (Bruker, Bremen, Germany). One microliter of the sample solution was placed on the target, and 1 lL of a freshly prepared saturated solution of sinapi- nic acid in acetonitrile ⁄ H 2 O (2 : 1) with 0.1% trifluoroace- tic acid was added. The spot was then recrystallized by addition of another 1 lL of acetonitrile ⁄ H 2 O (2 : 1), which resulted in a fine crystalline matrix. For recording of spec- tra, an acceleration voltage of 20 kV was used, and the detector voltage was adjusted to 1.9 kV. About 500 single laser shots were summed into an accumulated spectrum. Calibration was carried out using the single and double protonated ion signal of BSA for external calibration. Surface plasmon resonance binding assays Assays were performed using a Biacore 2000 (BIAcore AB). Coupling of antibodies to GST (BIAcore, Freiburg, Ger- many) to the CM5 chip was performed in 10 mm sodium acetate (pH 5.0), at a flow rate of 5 lLÆmin )1 . A 6 min pulse of 0.05 mm N-hydroxysuccinimide ⁄ 0.2 m N-ethyl-N¢- dimethylaminopropyl carbodiimide was used to activate the surface. The antibodies to GST (30 lgÆ mL )1 ) were injected for 7 min in 10 mm sodium acetate (pH 5.0), until the desired amount was coupled (15 000 RU), and excess reactive groups were deactivated by a 7 min pulse of 1 m ethanolamine hydrochloride (pH 8.5). The antibodies were saturated with GST–PACSIN 1 (100 lgÆmL )1 ), or, as a control, GST alone until saturation. Measurements were carried out in NaCl ⁄ P i (pH 7.4) containing 2.5 mm desthio- biotin at a flow rate of 30 lLÆmin )1 . The injection of 90 lL of the PACSIN 1-CS solution (0.1–2 lm) and the 180 s association was followed by a 180 s dissociation. Each ana- lysis was carried out a minimum of four times with two parallel samples. After subtraction of the data obtained for GST, they were analyzed with biaevaluation software 3.0, PACSIN 1 oligomerization A. Halbach et al. 780 FEBS Journal 274 (2007) 773–782 ª 2007 The Authors Journal compilation ª 2007 FEBS according to the Langmuir model for 1 to 1 binding. All binding curves could be fitted with an accuracy of v 2 < 0.5. Electron microscopy Purified PACSIN 1-CS (10 lgÆmL )1 ) or a purified recom- binant PACSIN 1 fragment containing the F-BAR domain was adsorbed onto a 400-mesh carbon-coated copper grid, which was rendered hydrophilic by glow discharge at low pressure in air. The grid was immediately washed with two drops of water, and stained with 0.75% uranyl for- mate for 15 s. Specimens were observed in a Jeol JEM 1230 transmission electron microscope (Jeol, Tokyo, Japan) operated at 60 kV accelerating voltage. The images were recorded with a Gatan Multiscan 791 CCD camera (Gatan, Munich, Germany). Evaluation of the data from electron micrographs was done as described previously [34]. Acknowledgements We would like to thank the Bioanalytical Laboratory of the Center for Molecular Medicine Cologne for the MS analysis. This work was supported by Deut- sche Forschungsgemeinschaft grant PL233 ⁄ 1-2 (to M. Plomann) and by a grant from the Ko ¨ ln Fortune program of the Medical Faculty of the University of Cologne. References 1 Plomann M, Lange R, Vopper G, Cremer H, Heinlein UAO, Scheff S, Baldwin SA, Leitges M, Cramer M, Paulsson M et al. (1998) PACSIN, a brain protein that is upregulated upon differentiation into neuronal cells. Eur J Biochem 256, 201–211. 2 Ritter B, Modregger J, Paulsson M & Plomann M (1999) PACSIN2, a novel member of cytoplasmic adap- ter proteins. FEBS Lett 454, 356–362. 3 Qualmann B, Roos J, DiGregorio PJ & Kelly RB (1999) Syndapin I, a synaptic dynamin-binding protein that associates with the neural Wiskott–Aldrich syn- drome protein. Mol Biol Cell 10, 501–513. 4 Modregger J, Ritter B, Witter B, Paulsson M & Plo- mann M (2000) All three PACSIN isoforms bind to endocytic proteins and inhibit endocytosis. J Cell Sci 113, 4511–4521. 5 Merila ¨ inen J, Lehto VP & Wasenius VM (1997) FAP52, a novel, SH3 domain-containing focal adhesion protein. J Biol Chem 272, 23278–23284. 6 Qualmann B & Kelly RB (2000) Syndapin isoforms par- ticipate in receptor-mediated endocytosis and actin organization. J Cell Biol 148, 1047–1062. 7 Braun A, Pinyol R, Dahlhaus R, Koch D, Fonarev P, Grant BD, Kessels MM & Qualmann B (2005) EHD proteins associate with syndapin I and II and such inter- actions play a crucial role in endosomal recycling. Mol Biol Cell 16, 3642–3658. 8 Kessels MM, Dong J, Leibig W, Westermann P & Qualmann B (2006) Complexes of syndapin II with dynamin II promote vesicle formation at the trans-Golgi network. J Cell Sci 119, 1504–1516. 9 Cousin H, Gaultier A, Bleux C, Darribere T & Alfan- dari D (2000) PACSIN 2 is a regulator of the metalloprotease ⁄ disintegrin ADAM13. Dev Biol 227, 197–210. 10 Mori S, Tanaka M, Nanba D, Nishiwaki E, Ishiguro H, Higashiyama S & Matsuura N (2003) PACSIN3 binds ADAM12 ⁄ meltrin alpha and up-regulates ect domain shedding of heparin-binding epidermal growth factor- like growth factor. J Biol Chem 278, 46029–46034. 11 Ghadimi MP, Sanzenbacher R, Thiede B, Wenzel J, Jing Q, Plomann M, Borkhardt A, Kabelitz D & Jans- sen O (2002) Identification of interaction partners of the cytosolic polyproline region of CD95 ligand (CD178). FEBS Lett 519, 50–58. 12 Houdart F, Girard-Nau N, Morin F, Voisin P & Vannier B (2006) The regulatory subunit of PDE6 interacts with PACSIN in photoreceptors. Mol Vis 11, 1061–1070 13 Pe ´ rez-Otan ˜ o I, Lujan R, Tavalin SJ, Plomann M, Mod- regger J, Liu XB, Jones EG, Heinemann SF, Lo DC & Ehlers MD (2006) Endocytosis and synaptic removal of NR3A-containing NMDA receptors by PACSIN1 ⁄ syndapin1. Nat Neurosci 9, 611–621. 14 Aspenstro ¨ m P (1997) A Cdc42 target protein with homology to the non-kinase domain of FER has a potential role in regulating the actin cytoskeleton. Curr Biol 7, 479–487. 15 Wu Y, Dowbenko D & Lasky LA (1998) PSTPIP 2, a second tyrosine phosphorylated, cytoskeletal-associated protein that binds a PEST-type protein-tyrosine phos- phatase. J Biol Chem 273, 30487–30496. 16 Nikki M, Merila ¨ inen J & Lehto VP (2002) Focal adhe- sion protein FAP52 self-associates through a sequence conserved among the members of the PCH family pro- teins. Biochemistry 41, 6320–6329. 17 Icking A, Matt S, Opitz N, Wiesenthal A, Mu ¨ ller-Esterl W & Schilling K (2005) NOSTRIN functions as a homotrimeric adaptor protein facilitating internalization of eNOS. J Cell Sci 118, 5059–5069. 18 Itoh T, Erdmann KS, Roux A, Habermann B, Werner H & De Camilli P (2005) Dynamin and the actin cytoskele- ton cooperatively regulate plasma membrane invagina- tion by BAR and F-BAR proteins. Dev Cell 9, 791–804. 19 Anggono V, Smillie KJ, Graham ME, Valova VA, Cousin MA & Robinson PJ (2006) Syndapin I is the A. Halbach et al. PACSIN 1 oligomerization FEBS Journal 274 (2007) 773–782 ª 2007 The Authors Journal compilation ª 2007 FEBS 781 phosphorylation-regulated dynamin I partner in synap- tic vesicle endocytosis. Nat Neurosci 9, 752–760. 20 Kessels MM & Qualmann B (2006) Syndapin oligomers interconnect the machineries for endocytic vesicle for- mation and actin polymerization. J Biol Chem 281, 13285–13299. 21 Kessels MM & Qualmann B (2002) Syndapins integrate N-WASP in receptor-mediated endocytosis. EMBO J 21, 6083–6094. 22 Shupliakov O, Lo ¨ w P, Grabs D, Gad H, Chen H, David C, Takei K, De Camilli P & Brodin L (1997) Synaptic vesicle endocytosis impaired by disruption of dynamin–SH3 domain interactions. Science 276, 259– 263. 23 Simpson F, Hussain NK, Qualmann B, Kelly RB, Kay BK, McPherson PS & Schmid SL (1999) SH3-domain- containing proteins function at distinct steps in clathrin- coated vesicle formation. Nat Cell Biol 1, 119–124. 24 Wigge P, Vallis Y & McMahon HT (1997) Inhibition of receptor-mediated endocytosis by the amphiphysin SH3 domain. Curr Biol 7, 554–560. 25 Peter BJ, Kent HM, Mills IG, Vallis Y, Butler PJ, Evans PR & McMahon HT (2004) BAR domains as sensors of membrane curvature: the amphiphysin BAR structure. Science 303, 495–499. 26 Wasiak S, Quinn CC, Ritter B, de Heuvel E, Baranes D, Plomann M & McPherson PS (2001) The Ras ⁄ Rac guanine nucleotide exchange factor mammalian Son-of- sevenless interacts with PACSIN 1 ⁄ syndapin I, a regula- tor of endocytosis and the actin cytoskeleton. J Biol Chem 276, 26622–26628. 27 Modregger J, DiProspero NA, Charles V, Tagle DA & Plomann M (2002) PACSIN 1 interacts with huntingtin and is absent from synaptic varicosities in presympto- matic Huntington’s disease brains. Hum Mol Genet 11, 2547–2558. 28 Blondeau F, Ritter B, Allaire PD, Wasiak S, Girard M, Hussain NK, Angers A, Legendre-Guillemin V, Roy L, Boismenu D et al. (2004) Tandem MS analysis of brain clathrin-coated vesicles reveals their critical involvement in synaptic vesicle recycling. Proc Natl Acad Sci USA 101, 3833–3838. 29 Li B, Otsu Y, Murphy TH & Raymond LA (2003) Developmental decrease in NMDA receptor desensitiza- tion associated with shift to synapse and interaction with postsynaptic density-95. J Neurosci 23, 11244–11254. 30 Nielsen PA, Olsen JV, Podtelejnikov AV, Andersen JR, Mann M & Wisniewski JR (2005) Proteomic mapping of brain plasma membrane proteins. Mol Cell Proteo- mics 4, 402–408. 31 Lee E & De Camilli P (2002) Dynamin at actin tails. Proc Natl Acad Sci USA 99, 161–166. 32 Orth JD, Krueger EW, Cao H & McNiven MA (2002) The large GTPase dynamin regulates actin comet for- mation and movement in living cells. Proc Natl Acad Sci USA 99, 167–172. 33 Sardy M, Odenthal U, Karpati S, Paulsson M & Smyth N (1999) Recombinant human tissue transglutaminase ELISA for the diagnosis of gluten-sensitive enteropathy. Clin Chem 45, 2142–2149. 34 Engel J & Furthmayr H (1987) Electron microscopy and other physical methods for the characterization of extracellular matrix components: laminin, fibronectin, collagen IV, collagen VI, and proteoglycans. Methods Enzymol 145, 3–78. Supplementary material The following supplementary material is available online: Table S1. PACSIN 1 secondary structure at different temperatures. Table S2. Calculation of K av values for native (A) and crosslinked (B) PACSIN 1 complexes. Table S3. Surface plasmon resonance analysis of PACSIN 1 self-interaction. This material is available as part of the online article from http://www.blackwell-synergy.com Please note: Blackwell Publishing is not responsible for the content or functionality of any supplementary materials supplied by the authors. Any queries (other than missing material) should be directed to the corres- ponding author for the article. PACSIN 1 oligomerization A. Halbach et al. 782 FEBS Journal 274 (2007) 773–782 ª 2007 The Authors Journal compilation ª 2007 FEBS . PACSIN 1 forms tetramers via its N-terminal F-BAR domain Arndt Halbach 1 , Matthias Mo ¨ rgelin 2 , Maria Baumgarten 2 , Mark Milbrandt 1 , Mats. purification of recombinant PACSIN 1. (A) Domain struc- ture of PACSIN 1 with an N-terminal F-BAR domain and a C-terminal SH3 domain. The F-BAR domain was formerly