Journal of Structural Biology 184 (2013) 43–51 Contents lists available at SciVerse ScienceDirect Journal of Structural Biology journal homepage: www.elsevier.com/locate/yjsbi Clathrin-coated vesicles from brain have small payloads: A cryo-electron tomographic study J. Bernard Heymann a,⇑, Dennis C. Winkler a, Yang-In Yim b, Evan Eisenberg b, Lois E. Greene b, Alasdair C. Steven a a b Laboratory of Structural Biology Research, National Institute of Arthritis, Musculoskeletal and Skin Diseases, Bethesda, MD 20892, United States Laboratory of Cell Biology, National Heart Lung and Blood Institute, National Institutes of Health, Bethesda, MD 20892, United States a r t i c l e i n f o Article history: Available online 18 May 2013 Keywords: Clathrin-mediated endocytosis Cryo-electron microscopy Three-dimensional image reconstruction Adaptor proteins Fullerenes a b s t r a c t Clathrin coats, which stabilize membrane curvature during endocytosis and vesicular trafficking, form highly polymorphic fullerene lattices. We used cryo-electron tomography to visualize coated particles in isolates from bovine brain. The particles range from 66 to 134 nm in diameter, and only 20% of them (all P80 nm) contain vesicles. The remaining 80% are clathrin ‘‘baskets’’, presumably artifactual assembly products. Polyhedral models were built for 54 distinct coat geometries. In true coated vesicles (CVs), most vesicles are offset to one side, leaving a crescent of interstitial space between the coat and the membrane for adaptor proteins and other components. The latter densities are fewer on the membraneproximal side, which may represent the last part of the vesicle to bud off. A small number of densities – presumably cargo proteins – are associated with the interior surface of the vesicles. The clathrin coat, adaptor proteins, and vesicle membrane contribute almost all of the mass of a CV, with most cargoes accounting for only a few percent. The assembly of a CV therefore represents a massive biosynthetic effort to internalize a relatively diminutive payload. Such a high investment may be needed to overcome the resistance of membranes to high curvature. Published by Elsevier Inc. 1. Introduction Clathrin-mediated endocytosis is responsible for cellular uptake in the context of receptor recycling (LDL) (Ehrlich et al., 2004), synaptic vesicle recycling (Augustine et al., 2006), virus infection (e.g. (Ehrlich et al., 2004; Matlin et al., 1981; Rust et al., 2004)), and import of the prion protein (Taylor et al., 2005), among other processes. Clathrin-coated vesicles are also involved in protein sorting at the trans-Golgi network (Traub, 2005), and the assembly of the Golgi apparatus itself requires clathrin (Radulescu et al., 2007). This functional diversity requires assembling polymorphic scaffolds that are able to accommodate large variations in the size, shape, and molecular nature of the cargoes. The building-block – the clathrin triskelion – is a remarkable structure with three hinged 52 nm-long legs connected at a trimeric hub (Brodsky, 2012; Kocsis et al., 1991; Ungewickell and Branton, 1981). It is able to assemble into many different forms, including flat lattices (Heuser, 1989), clathrin baskets (CBs) which Abbreviations: CB, clathrin basket; CV, coated vesicle; AP, adaptor protein. ⇑ Corresponding author. Address: Bldg. 50, Room 1515, 50 South Drive MSC 8025, N.I.H., Bethesda, MD 20892-8025, United States. E-mail address: Bernard_Heymann@nih.gov (J.B. Heymann). 1047-8477/$ - see front matter Published by Elsevier Inc. http://dx.doi.org/10.1016/j.jsb.2013.05.006 are proteinaceous particles devoid of lipid (Crowther and Pearse, 1981; Crowther et al., 1976; Pearse and Robinson, 1984; Vigers et al., 1986b), and clathrin-coated vesicles (CVs) (Crowther et al., 1976). The coats of CBs and CVs adopt a wide range of polyhedral shapes and sizes. In CV assembly, the main role of clathrin is to impose curvature of the membrane or to stabilize curvature otherwise accomplished (Hinrichsen et al., 2006). This calls for an energetically unfavorable distortion of the lipid bilayer. The endocytic process starts with the formation of a coated pit, followed by deepening the invagination until it is pinched off from the membrane of origin as a CV through the action of the GTPase dynamin (Hinshaw, 2000). Once the CV detaches from the plasma membrane, it is rapidly uncoated by the ATPase, Hsc70, and the freed clathrin triskelions recycle back to the membrane. The lifetime of a CV is only a few seconds before it is uncoated (Taylor et al., 2011). In view of these kinetics, it is likely that biochemical isolates contain, in addition to bona fide CVs recently budded off, also CBs assembled in the homogenate and CVs completed from coated pits during the isolation procedure. In CVs, the clathrin network is coupled to the vesicle membrane through various proteins, the major ones being the adaptor proteins (APs), AP-1 and AP-2, which also function in cargo selection 44 J.B. Heymann et al. / Journal of Structural Biology 184 (2013) 43–51 (Edeling et al., 2006). Each AP is composed of a heterotetrameric complex, forming a four-chain core, with two chains extending from the core to form two appendages or ‘‘ears’’. Various parts of the APs have been put forward as the clathrin-binding parts, such as the AP-2 core (Matsui and Kirchhausen, 1990; Peeler et al., 1993), the a-appendage (Goodman and Keen, 1995), and the linker or hinge of the ß2-appendage (Shih et al., 1995) at a motif called the clathrin box (Dell’Angelica et al., 1998; ter Haar et al., 2000). It remains unsettled which potential interactions are important in vivo. The polymorphism of clathrin-coated particles complicates analyses of their three-dimensional structures by ‘‘single particle’’ reconstructions from cryo-electron microscopy. Classification of images into homogeneous subsets represents one approach to overcoming this obstacle and most studies to date have focused on the D6, 36-triskelion, coat which is relatively abundant in preparations from brain (Fotin et al., 2004; Heymann et al., 2005; Smith et al., 1998, 2004; Xing et al., 2010). However, this structure is a CB lacking a cargo and per se it casts little direct light on the wide range of clathrin lattices that form nor on the interactions of the clathrin coat with other components. The more recently introduced technique of cryo-electron tomography (cryo-ET) (Baumeister and Steven, 2000; McEwen and Frank, 2001) has the advantage of rendering three-dimensional structures for individual particles and its potentiality for investigating clathrin-coated particles has been demonstrated (Cheng et al., 2007). Here we have followed a generally similar approach, working with a larger data set and focusing to a greater extent on coat polymorphisms, their complements of APs, the presence of vesicles, and, in particular, the quantitation of the mass contributions of the respective constituents. 2. Materials and methods 2.1. Preparation of clathrin-coated particles Material was isolated from fresh bovine brains essentially following the procedure of Nandi et al. (1982) and using a 12% sucrose-D2O ultracentrifugation step (SW28 rotor, 100,000g for 3 h) for purification. The preparation was stored at a protein concentration of about 5 mg/ml at 4 °C in the homogenization buffer (0.1 M MES, 0.1 mM EGTA, 0.5 mM MgCl2, and 3 mM NaN3, pH 6.5). Multiple preparations were done and each was used within a week (longer storage resulted in precipitation and freezing with cryo-protectants also failed). 2.2. Data acquisition The specimen was mixed with an equal volume of 10 nm colloidal gold particles to serve as fiducial markers (BBInternational, Ltd), giving 3  1012 particles/ml, applied to a glow-discharged lacy carbon grid and plunge-frozen (Reichert Jung KF80 Cryofixation System). The SerialEM package (Mastronarde, 2005) was used to record tilt series on a Tecnai T12 electron microscope (FEI) operating at 120 kV, using a post-column energy filter (in zero-loss mode with a 20 eV energy slit width) and a CCD camera of 2048  2048 pixels (GIF2002, Gatan, Inc.). Tilt series consisted of up to 141 micrographs taken from 70° to 70° in 1° steps. The total dose for a series was approximately 60 e /nm2, and the target defocus was 4 lm under focus, putting the first zero of the CTF at (3.7 nm) 1. They were recorded at magnifications of 54,000 (3 tomograms), 38,500 (11 tomograms) and 26,000 (9 tomograms), giving pixel sizes of 0.55 nm, 0.78 nm and 1.14 nm, respectively. The intermediate magnification yielded the best tomograms (obtained from two sample preparations), providing a sufficient number of fiducial markers and particles within the field-of-view. The lower magnification appeared to decrease the quality of the clathrin spars in the tomograms. All tomograms further discussed are at the intermediate magnification. 2.3. Tomogram reconstruction Tilt series were aligned and tomograms reconstructed using the package Bsoft (Heymann and Belnap, 2007; Heymann et al., 2008a). The tomograms were denoised by a non-linear anisotropic diffusion algorithm (Frangakis and Hegerl, 2001). Individual clathrin particles were selected and extracted. The resolution of each individual micrograph was estimated by Fourier ring correlation using a cutoff of 0.3 (FSC0.3) (Cardone et al., 2005; Heymann et al., 2008a). The zero-tilt micrographs showed resolutions of 5.4–6.5 nm. 2.4. Modeling the clathrin lattice Polyhedral models of the clathrin network were built into the tomographic subvolumes using the program Chimera (Pettersen et al., 2004). Symmetry was determined by visual inspection. The coordinates of the polyhedral vertices were regularized (Heymann et al., 2008b). The reference inter-spar angles were set to the canonical angle for the associated polygon (90° for a tetragon, 108° for a pentagon, 120° for a hexagon and 128.6° for a heptagon). 2.5. Modeling the vesicle membrane The vesicles in CVs are approximately spherical. Accordingly, for each vesicle, a set of closely spaced points was generated, evenly distributed at a radius approximately equal to that of the vesicle. The model was generated and positioned relative to the tomogram by cross-correlation. The positions of points were refined by crosscorrelation to a reference membrane patch. The offset of the vesicle within each CV was calculated as the difference between the geometric centers of the coat polyhedron and the vesicle model. To represent the volume of the membrane, a shell mask with a thickness of 5 nm was generated (in a eukaryotic bilayer the P–P distance is 4 nm (Mitra et al., 2004) and typically phospholipid bilayers are 5 nm (Woodka et al., 2012)). 2.6. Segmenting the coated vesicle For a given CV, a synthetic clathrin coat map was constructed using the polyhedron and averaged spar density from a single particle reconstruction of the 36-vertex D6 barrel (Heymann et al., 2005). The use of this map avoids including spurious densities that might be present in a tomogram. This map was binarized by thresholding at a level that gave a good representation of the N-termini. This mask and a mask covering the region outside the polyhedral model were used to isolate the interior of the CV, excluding the coat. The membrane model (see above) was refined within this interior mask, and a membrane mask produced as a shell with a thickness of 5 nm. The membrane mask and a mask outside the membrane model were used to isolate the vesicle lumen. The membrane mask and a mask of the vesicle lumen were used to produce a mask of the interstitial volume. With these four masks (coat, interstitial, membrane and lumen), each region was isolated as a separate map. 2.7. Counting adaptor protein (AP) densities The AP core-sized densities in the interstitial space of each CV were counted manually. In addition, potential APs in the interior of CBs and interstitial spaces of CVs were located as follows: The J.B. Heymann et al. / Journal of Structural Biology 184 (2013) 43–51 segmented map was cross-correlated with a soft-edged spherical reference density of 8 nm in diameter. Peaks in the cross-correlation map higher than 30% of the maximum correlation were identified and when two or more were closer than 8 nm, the one with the higher value was retained. Most of the densities corresponding to these peaks (>70%) were consistent with the manually selected densities, with only small differences in assigned location (from 0 to 6 nm). Because of the greater objectivity of the automated method, it was used to give the AP numbers used in calculating their contributions to the particle masses of CBs and CVs. 2.8. Masses of components Masses of the individual proteins were obtained from UniProtKB/Swiss-Prot (http://www.uniprot.org), or the RCSB/PDB (http://www.rcsb.org). Clathrin bovine heavy chain: P49951, 191.6 kDa; bovine light chain A: P04973, 26.7 kDa; bovine light chain B: P04975, 25.1 kDa. The average of the light chains (25.9 kDa) was used in the calculation of the triskelion mass: 3 (191.6 kDa + 25.9 kDa) = 652.5 kDa. AP-2 mass: Q0VCK5, alpha-2 103.8 kDa, P63009, beta 104.6 kDa, Q3ZC13, mu 49.7 kDa, Q17QC5, sigma 17.0 kDa. Total mass for AP-2: 275.1 kDa. The AP-2 core was estimated at 219 kDa from the residue masses in the crystal structure 1GW5 (PDB). The appendages were estimated from the crystal structures: alpha (1QTS) 26 kDa and beta (1E42) 24 kDa. Membrane mass: Typical membrane lipids have densities of 1 g/ml (Greenwood et al., 2006). However, the CV vesicles also contain an unspecified amount of protein. Assuming that the membrane contains about 30% protein at 1.35 g/ml, the density is estimated at 1.1 g/ml or 660 Da/nm3 (comparable to that estimated for synaptic vesicles – (Takamori et al., 2006)). This density was then used to estimate the vesicle membrane mass from the volume of the membrane mask (see above). 45 overexpressed (Zhao et al., 2001) or depleted (Hirst et al., 2008), disrupting the normal ratios of CV components. CBs assemble readily in vitro without the participation of membranes (Keen, 1990). Since our (and other similar) preparations are made at pH 6.5, conditions that promote the polymerization of clathrin, the CBs are probably formed during the isolation procedure. 3.2. Modeling clathrin lattices The clathrin lattices in the tomograms are well defined and were modeled as polyhedral fullerene shells. To model a given coat, a vertex site was placed at each point identified as a triskelion hub. Links were then generated between adjacent vertices, in most cases closing the polyhedron (e.g. Fig. 1D, E). Subsequently, the models were refined by an automated regularization procedure (see Section 2). Each model was examined to determine its symmetry and the number and types of its polygonal facets. Many different lattices (N = 54) were identified and cataloged (Table 2). Despite this overall diversity, a subset with 28–38 vertices accounted for the majority of these coats (196/246; 80%; mostly CBs); the other geometries observed were relatively rare. Even the 54 forms observed here represent only a small fraction of the vast number of possibilities: there are 32 possible forms with 28–36 vertices (Katsura, 1983), and 5767 with 28–60 vertices (Brinkmann and Dress, 1997). The large majority of facets in these coats are either pentagons (60.5%) or hexagons (39.0%), while some of the larger lattices also include a few tetragons (0.07%) and heptagons (0.4%) (e.g., Fig. 1E). The heptagonal faces have been reported before in CVs (Cheng et al., 2007) and in flat sheets (Heuser, 1980). To our knowledge, the rarer tetragonal (square) faces have not been seen before. As there is a one-to-one correspondence between vertices and triskelia, the number of vertices in a given model affords a measurement of the coat mass, taking the mass of an individual triskelion to be 653 kDa (see Section 2). The closed clathrin lattices in our data set range in mass from 18 to 56 MDa. 3. Results and discussion 3.3. Vesicle size limits and size determination 3.1. Clathrin baskets and coated vesicles The tomograms show many coated particles, readily identifiable as such by their distinctive clathrin lattices (Fig. 1A). Similar particles have been identified as coated vesicles (CVs) in cellular tomograms (Ladinsky et al., 1999; Zampighi et al., 2005). Our detailed analysis focused on 5 tomograms of ice layers that were relatively thin (80–160 nm, as measured directly from the tomograms) but nevertheless thick enough for the coated particles to be completely embedded. They range from 66 nm to 134 nm in diameter (average 74 nm, SD = 10 nm), as assessed from central sections. These dimensions are in agreement with earlier observations of unstained freeze-dried specimens by dark-field STEM (Steer et al., 1988) and measurements in situ (Hirst et al., 2008) and extend the size range covered in a previous study by cryo-ET (Cheng et al., 2007). The smaller particles lack an internal membrane and we take them to be clathrin baskets (CBs) (Fig. 1B), similar to those studied in ‘‘single particle’’ cryo-EM analyses of the abundant D6 basket (see Introduction). However, many of the larger ones (i.e. P80 nm in diameter) clearly contain vesicles, which are 30– 68 nm in diameter (Fig. 1C). These, the true CVs, make up 20% of the total population (Table 1). Similar estimates in the 25–30% range have been reported (Merisko et al., 1982; Steer et al., 1988). It follows that any bulk analysis of such preparations refers primarily to CBs. In vivo baskets are rare, most likely because the free triskelions are chaperoned (Jiang et al., 2000). Increased numbers of baskets have been observed only in cases where auxilin is 35 Bona fide CVs with closed polyhedral coats were present in our data set (Table 2). Their vesicles are generally small (minimum diameter, 30 nm) and, at least approximately, spherical. The smallest CVs that we found have coats with 38 vertices and a diameter of 80 nm, values similar to those reported by Vigers et al. (1986a), and slightly larger than the 36-vertex CV described by Cheng et al. (2007). These vesicles are similar in size to the smallest pure lipid vesicles (Haque et al., 2001; Lapinski et al., 2007; Tenchov et al., 1985). Other vesicles and the surrounding coats are larger, for instance the average coat diameter for CVs is 89 nm (SD = 10 nm) and the largest one observed in this preparation, 116 nm. To assess the shape of a given vesicle, a spherical sheet of the appropriate radius made up of evenly spaced points was generated. This model was then used to generate a mask 5 nm thick that represents the vesicle membrane (see Section 2 for discussion of the assigned mean thickness). Such a model is shown in Fig. 1E. The mask volume was then converted into a mass using a density of 660 Da/nm3 (see Section 2). Thus calculated, these vesicles range from 10 to 32 MDa in mass (excluding a large outlier at 45 MDa). Vesicles are confined within their surrounding coats, but their sizes do not strictly follow the maximum accommodatable size (i.e., the average coat diameter minus 28 nm) and is often smaller. Moreover, in agreement with Cheng et al. (2007), we find the vesicles to be offset relative to the coats. The average offset is 6 nm (SD = 3 nm) and the range covered is 2–13 nm. Smaller vesicles show larger offsets than bigger ones (Fig. 2). Using the 46 J.B. Heymann et al. / Journal of Structural Biology 184 (2013) 43–51 Fig.1. (A) A field of coated particles in a 2.3 nm-thick slice through a denoised tomogram. The white arrow marks a vesicle-containing CV. The black arrow marks a CB (no vesicle). White arrowheads mark pentagonal and hexagonal facets sampled in this slice. (B and C): Serial, 3.1 nm-thick, slices through a CB (B) and a CV (C). (D and E) Ball-andstick models of the clathrin coats of the particles in (B) and (C), also depicting the vesicle model in (C). The white arrows in panels in (C) and (E) point to the same vertex, part of a heptagonal face. The black arrow in (E) marks a tetragonal face. Table 1 Percentages of complete and/or vesicle-containing coated particles extracted from five tomograms (300 particles). No vesicle With vesicle Total Partial coat Complete coat Total 11 7 18 70 12 82 81 19 100 polyhedral models of the coats and the spherical shell models of the vesicles, we determined the closest approach between the membrane and a triskelion hub for each CV. These measurements range from 8 to 18 nm, with an average of 14 nm (SD = 3 nm). Appraisal of the models led to the conclusion that the closest a membrane can approach the hub without overlapping the N-termini of the clathrin heavy chains is about 14 nm. It therefore appears that the vesicles are close to or in contact with the clathrin N-termini on one side, leaving a crescent of interstitial space occupied by APs on the other side (Fig. 3). This arrangement may reflect the geometry of the coated pit, where the APs are required to couple the coat to the membrane as invagination proceeds, while the side where the vesicle closes involves other proteins, e.g., dynamin and endophilin (Sundborger et al., 2011). We did observe some incomplete coats (Table 1). It is therefore possible that the coat is not necessarily closed completely after scission of the vesicle. Once the support for a curved membrane is no longer required, the coat can be recycled for subsequent endocytosis. J.B. Heymann et al. / Journal of Structural Biology 184 (2013) 43–51 47 Table 2 Size and symmetry of complete clathrin baskets and coated vesicles from five tomograms. Number of vertices Possible fullerenes Observed symmetries Observed forms ( ) Particle counts Coated vesicles 20 24 26 28 30 32 34 36 38 40 42 1 1 1 2 3 6 6 15 17 40 45 0 0 0 1 1 1 4 5 3 5 5 (2) (1) (1) (1) (1) – – – 61 17 25 23 54 16 11 6 – – – 0 0 0 0 0 3 6 3 44 46 48 50 52 54 56 58 60 62 64 86 Total 89 116 199 271 437 580 924 1205 1812 – – – T C2 D3 C1, C2, C1, C1, C1, D7 C1, C1, C1, C2 C1, C1 C2 C1 – C1 C1 C1 5 (1) 4 (2) 5 (1) 1 5 (2) 1 (1) 1 2 (1) 0 2 (2) 2 (1) 1 54 (17) 8 4 5 2 5 1 1 2 – 2 2 1 246 4 3 2 2 3 1 1 2 – 2 2 1 35 C2 D2, D6 C2 C2, D2 C2, D3, C2, D2 C2 C2, D2 D2 Forms with tetragons or heptagons in parentheses (i.e., non-fullerenes). Symmetries: C<n>, n-fold cyclic; D<n>, n-fold dihedral; T, tetrahedral. Fig.3. Cutaway renderings of CVs, including two typical ones (A: 48 triskelions, B: 58 triskelions), the smallest one (C: 38 triskelions), and the largest one (D: 86 triskelions). Each is segmented to show the clathrin coat (pink), the vesicle membrane (blue), interstitial densities (green), and cargo densities (red). In most CVs the vesicles are offset so that the lipid bilayer is close to or in contact with the clathrin N-termini (yellow arrows in A and D), while on the other side the interstitial densities, mainly APs, form a cap (crescent-shaped in cross-section) between the coat and the vesicle (white arrows in A and B). 3.4. Counting adaptor proteins (APs) The majority of densities in the interstitial spaces of the CVs are of the right size (8 nm diameter) and shape for the AP core (220 kDa), and APs have been shown to be (by far) the most abundant protein component other than clathrin (Ahle and Ungewickell, 1989; Blondeau et al., 2004; Girard et al., 2005; Keen et al., 1991; Pearse and Robinson, 1984). We marked these densities with points in each CV (the green densities in Fig. 3). This procedure was performed both manually and automatically. The manual method yielded on average 50 APs/CV (SD = 21), while the automated method with an appropriate threshold for selecting densities yielded 45 APs/CV (SD = 15). The automated approach therefore yielded a basically consistent result. We also searched for AP densities inside the CB interiors with the automated method, finding 34 APs/CB (SD = 9). All further references to AP counts are from the automated procedure. The total AP mass in a given CB or CV was taken to be the number of detected points times the mass of an AP complex (275 kDa, including the appendages). The resulting numbers range from 2 to 18 MDa for CBs and from 6 to 27 MDa for CVs. The number of APs was found to correlate with the triskelion count, giving a ratio of 0.92 AP/triskelion (SD = 0.19, n = 35, R = 0.79) for CVs, and 1.02 AP/triskelion (SD = 0.19, n = 214, R = 0.73) for CBs. This result is similar to the molar ratios determined by quantitative SDS–PAGE for bulk preparations of coated vesicles (Ahle and Ungewickell, 1989; Keen et al., 1991; Pearse and Robinson, 1984) and somewhat lower than the value of 1.5 reported from mass spectrometry (Blondeau et al., 2004; Girard et al., 2005). 3.5. Total masses of CVs and CBs Fig.2. The offsets of vesicles relative to their coat centers are related to their size; the larger ones tend to be less eccentric and more closely follow the clathrin coat. The cartoon at top right maps the centers of a given coat (large white disk) and its enclosed vesicle (smaller white disk), connected by a double-headed arrow to indicate the offset. So far, we have used the cryo-tomograms and a priori information to estimate the mass contributions of the main parts of each CB and CV in our data set: the clathrin coat; the vesicle (only in the CVs); and the APs. The sums of these parts can be compared to distributions of particle masses (Steven et al., 1983) and compositional studies (Blondeau et al., 2004) of similar preparations. The calculated masses were obtained by summing the two major parts in CBs (Fig. 4A) and the three major parts in CVs (Fig. 4B). In CBs, which are systematically smaller than CVs (an average diameter of 71 nm vs. 89 nm), clathrin accounts on average for 70% of the total mass. In CVs, clathrin contributes about half of the mass, although this fraction falls with increasing particle size: the remaining mass divides nearly equally between the vesicle and APs. 48 J.B. Heymann et al. / Journal of Structural Biology 184 (2013) 43–51 Fig.4. Distributions of partial masses for (A) CBs and (B) CVs. The clathrin masses (triangles) were calculated from the numbers of polyhedral vertices at 653 kDa per vertex triskelion. The AP masses (diamonds) were calculated from automated counts in the interior of each CB and the interstitial space of each CV at 275 kDa per AP. The membrane mass (gray disks) was calculated from the membrane mask volume (660 Da/nm3). The line fits and bar insets show the component fractions: (A) clathrin, 0.696 ± 0.039; AP, 0.304 ± 0.039; (B) clathrin, 0.514 ± 0.050; AP, 0.203 ± 0.031; membrane, 0.284 ± 0.052 (standard deviation). The masses for the CBs and CVs when combined give an expected distribution for the bulk preparation. Table 3 shows that the large proportion of CBs skews the composition such that most of the overall mass is attributed to clathrin and APs, with vesicles making only a minor overall contribution. We cannot directly measure the total masses of individual CBs and CVs in our tomograms. In particular, the payload proteins are not accounted for. However, the distribution of two- and three-part mass sums can be compared with the distribution of particle masses determined earlier by STEM for the same type of preparation (Steven et al., 1983). Fig. 5 shows the calculated distributions of masses for the CBs (gray bars) and the CVs (black bars), with a transition in the 40–45 MDa range. The combined distribution follows the pattern of the STEM data quite closely, reproducing the two peaks at 27 and 33 MDa associated with abundant CBs (Fig. 5A, gray curve). The decrease in numbers of CVs with size is also reflected in both data sets (Fig. 5B). The average mass per particle from the STEM data is 35.2 MDa, while that for the combined CB and CV populations is 35.1 MDa. The agreement between these two data sets indicates that little mass remains to be accounted for, including other accessory proteins and the CV cargoes. Separating the CBs and CVs, the average mass per particle is 31.0 MDa and 60.4 MDa, respectively. If it is assumed that the difference between the STEM and tomography data reflects only additional mass in CVs, it amounts to 0.7 MDa per CV. Given the uncertainties in our analysis, this can vary to some extent, but not beyond a few percent of the CV mass. 3.6. Coated vesicle cargo CV cargoes are diverse (see e.g., (McMahon and Boucrot, 2011; Pizarro-Cerda et al., 2010)). The red densities in Fig. 3 illustrate the variety of shapes and levels of occupancy inside the vesicles in our tomograms. Some of the green densities in contact with the vesicle membranes may be membrane-associated or -embedded proteins. Preparations from bovine brain are likely to contain CVs with proteins typical of synaptic vesicles (Blondeau et al., 2004). The narrow size range of vesicles (30–50 nm, average 16 nm, SD = 8 nm) and their masses (average 16.4 MDa, SD = 7.7 MDa) in our CVs (blue shells in Fig. 3A–C) correspond to the values obtained for synaptic vesicles (Takamori et al., 2006). We therefore expect that some of the densities seen in our tomograms to represent proteins from synaptic vesicles. It is not yet possible to identify specific cargo molecules on morphological grounds. Densities are visible protruding from the outer surface of a vesicle (formerly, the cytoplasmic surface of the cell membrane – Fig. 6). These could be adaptor proteins involved in cargo selection or linking the coat to the vesicle, or parts of receptors and other membrane-embedded proteins. In many of our CVs, elongated densities extend from the outer surface of the membrane, reminiscent of the abundant synaptobrevin or the bulky V-ATPase (Fig. 6B, C, D, E, K, L). Inside the vesicle lumen (formerly on the outer surface of the cell membrane), densities clearly extend beyond the membrane surface, but always with some indication of connections. The vesicle membrane appears to vary in thickness, sometimes showing distortions (Fig. 6M) and large protrusions (Fig. 6N, O). 3.7. Preferred curvatures of clathrin coats and vesicles CBs are smaller, averaging only 71 nm in diameter (SD = 6 nm). It is likely that this size reflects the preferred (intrinsic) curvature for clathrin assembly under these conditions. This suggests that the larger coats formed around vesicles are assembled under conditions in which the growing lattice is under some stress, pushed outwards from the shape it would naturally assume and it, in turn, exerts inwards pressure on the invaginating membrane pit (soon to be vesicle). Even though the individual interactions between the triskelion legs may be weak, the coordinated effect may be sufficient for this process (Wakeham et al., 2003). In this way, the growing clathrin coat may contribute actively to the endocytic process. Table 3 Total mass contributions for CVs and CBs (MDa). Clathrin baskets Coated vesicles Total Clathrin APs Membrane Total 4640 1112 5752 (65.8%) 1985 427 2412 (27.6%) 0 575 575 (6.6%) 6625 (75.8%) 2114 (24.2%) 8739 J.B. Heymann et al. / Journal of Structural Biology 184 (2013) 43–51 49 Fig.5. (A) Comparison of the summed masses of the clathrin, APs, and CV membranes (black bars) and of the first two components for CBs (gray bars) as calculated from the tomograms, with the distribution of total particle masses for the same kind of preparation as determined by STEM measurements (gray curve) (Steven et al., 1983). (B) An enlargement of (A) in the range of masses covered by CVs. Each bin covers 3 MDa. CVs larger than 90 MDa not shown: 2 from STEM analysis (104 and 108 MDa) and one from a tomogram (128 MDa). Fig.6. Gallery of central slices through 15 CVs, illustrating the wide variation in vesicle morphology and contents. The vesicle membranes exhibit local variations in thickness, resulting from embedded proteins. Some vesicles depart significantly from a spherical shape (arrows in M). Densities in vesicle lumens (arrowheads) are typically connected to the membrane, as expected for cargo proteins. Long thin densities protruding from the vesicle surface are common (arrows in B–E, G). Some vesicles have large structures attached to the membrane (arrows in N, O). Scale bar: 50 nm. However, other factors may also affect coat size. In the case of virus entry, the size must be dictated by the cargo, assuming the virus to be essentially incompressible. The bovine brain preparation studied here is likely to contain precursors to synaptic vesicles (Blondeau et al., 2004) and indeed the vesicles observed fall within the range reported for synaptic vesicles (30–50 nm, 10–30 MDa (Qu et al., 2009; Takamori et al., 2006)). As precursors to synaptic vesicles, the size range may be set in the CVs, as suggested in a recent report where reduction in the expression of the coat accessory proteins, AP180 and CALM, led to larger synaptic vesicles (Petralia et al., 2013). 3.8. Brain CVs have relatively small payloads CVs are large particles, being at least 80 nm in outer diameter, with coats made up of at least 38 triskelions (114 clathrin heavy 50 J.B. Heymann et al. / Journal of Structural Biology 184 (2013) 43–51 chains) and total masses of 40 MDa and more. It is striking that, as illustrated in Fig. 5, almost all of the mass of a CV can be accounted for in terms of its clathrin coat, complement of APs, and the vesicle; in contrast, the cargo mass is small. We detect only 5–10 densities associated with the inner surface of a 30 nm vesicle. Most of them are visibly connected to the surface of the vesicle and it is likely that the others are also connected via linkers too fine to be visualized at the current resolution. Moreover, these densities are quite variable in shape, implying that a given CV takes up a variety of different cargo molecules. It may be that there are other, lower-profile, cargo proteins embedded in the lipid bilayer but still the payload cannot amount to a few percent of the particle mass. This is much less than the corresponding numbers typical of another class of carrier particle bounded by a protein coat - viral capsids. In bacteriophage HK97, for example, (Duda et al., 2009), the payload (DNA genome) accounts for 60% of the nucleocapsid mass and the protein coat, 40%. In the polyomavirus SV40 (Liddington et al., 1991), the payload of the chromatinized genome accounts for about 28% of the virion mass and the protein shell for the remaining 72%. In larger CVs, for instance, those that engulf entire virions in cell entry, the payload will represent a larger fraction of total particle mass. Nevertheless, the apparent inefficiency of the trafficking processes represented by the relatively small CVs typical of brain preparations is striking. Acknowledgments Molecular graphics images were produced using the UCSF Chimera package from the Resource for Biocomputing, Visualization and Informatics at the University of California, San Francisco (supported by NIHP41RR-01081). This work was supported by the Intramural Research Programs of NIAMS and NHLBI, NIH. References Ahle, S., Ungewickell, E., 1989. Identification of a clathrin binding subunit in the HA2 adaptor protein complex. J. Biol. Chem. 264, 20089–20093. Augustine, G.J., Morgan, J.R., Villalba-Galea, C.A., Jin, S., Prasad, K., et al., 2006. Clathrin and synaptic vesicle endocytosis: studies at the squid giant synapse. Biochem. Soc. Trans. 34, 68–72. Baumeister, W., Steven, A.C., 2000. Macromolecular electron microscopy in the era of structural genomics. Trends Biochem. Sci. 25, 624–631. Blondeau, F., Ritter, B., Allaire, P.D., Wasiak, S., Girard, M., 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. Brinkmann, G., Dress, A.W.M., 1997. A constructive enumeration of fullerenes. J. Algorithm 23, 345–358. Brodsky, F.M., 2012. Diversity of clathrin function: new tricks for an old protein. Annu. Rev. Cell Dev. Biol. 28, 309–336. Cardone, G., Grunewald, K., Steven, A.C., 2005. A resolution criterion for electron tomography based on cross-validation. J. Struct. Biol. 151, 117–129. Cheng, Y., Boll, W., Kirchhausen, T., Harrison, S.C., Walz, T., 2007. Cryo-electron tomography of clathrin-coated vesicles: structural implications for coat assembly. J. Mol. Biol. 365, 892–899. Crowther, R.A., Pearse, B.M., 1981. Assembly and packing of clathrin into coats. J. Cell Biol. 91, 790–797. Crowther, R.A., Finch, J.T., Pearse, B.M., 1976. On the structure of coated vesicles. J. Mol. Biol. 103, 785–798. Dell’Angelica, E.C., Klumperman, J., Stoorvogel, W., Bonifacino, J.S., 1998. Association of the AP-3 adaptor complex with clathrin. Science 280, 431–434. Duda, R.L., Ross, P.D., Cheng, N., Firek, B.A., Hendrix, R.W., et al., 2009. Structure and energetics of encapsidated DNA in bacteriophage HK97 studied by scanning calorimetry and cryo-electron microscopy. J. Mol. Biol. 391, 471–483. Edeling, M.A., Mishra, S.K., Keyel, P.A., Steinhauser, A.L., Collins, B.M., et al., 2006. Molecular switches involving the AP-2 beta2 appendage regulate endocytic cargo selection and clathrin coat assembly. Dev. Cell 10, 329–342. Ehrlich, M., Boll, W., Van Oijen, A., Hariharan, R., Chandran, K., et al., 2004. Endocytosis by random initiation and stabilization of clathrin-coated pits. Cell 118, 591–605. Fotin, A., Cheng, Y., Sliz, P., Grigorieff, N., Harrison, S.C., et al., 2004. Molecular model for a complete clathrin lattice from electron cryomicroscopy. Nature 432, 573– 579. Frangakis, A.S., Hegerl, R., 2001. Noise reduction in electron tomographic reconstructions using nonlinear anisotropic diffusion. J. Struct. Biol. 135, 239– 250. Girard, M., Allaire, P.D., McPherson, P.S., Blondeau, F., 2005. Non-stoichiometric relationship between clathrin heavy and light chains revealed by quantitative comparative proteomics of clathrin-coated vesicles from brain and liver. Mol. Cell. Proteomics 4, 1145–1154. Goodman Jr., O.B., Keen, J.H., 1995. The alpha chain of the AP-2 adaptor is a clathrin binding subunit. J. Biol. Chem. 270, 23768–23773. Greenwood, A.I., Tristram-Nagle, S., Nagle, J.F., 2006. Partial molecular volumes of lipids and cholesterol. Chem. Phys. Lipids 143, 1–10. Haque, M.E., McIntosh, T.J., Lentz, B.R., 2001. Influence of lipid composition on physical properties and peg-mediated fusion of curved and uncurved model membrane vesicles: ‘‘nature’s own’’ fusogenic lipid bilayer. Biochemistry 40, 4340–4348. Heuser, J., 1980. Three-dimensional visualization of coated vesicle formation in fibroblasts. J. Cell Biol. 84, 560–583. Heuser, J., 1989. Effects of cytoplasmic acidification on clathrin lattice morphology. J. Cell Biol. 108, 401–411. Heymann, J.B., Belnap, D.M., 2007. Bsoft: image processing and molecular modeling for electron microscopy. J. Struct. Biol. 157, 3–18. Heymann, J.B., Cardone, G., Winkler, D.C., Steven, A.C., 2008a. Computational resources for cryo-electron tomography in Bsoft. J. Struct. Biol. 161, 232–242. Heymann, J.B., Butan, C., Winkler, D.C., Craven, R.C., Steven, A.C., 2008b. Irregular and semi-regular polyhedral models for Rous sarcoma virus cores. Comput. Math. Methods Med. 9, 197–210. Heymann, J.B., Iwasaki, K., Yim, Y.I., Cheng, N., Belnap, D.M., et al., 2005. Visualization of the binding of Hsc70 ATPase to clathrin baskets: implications for an uncoating mechanism. J. Biol. Chem. 280, 7156–7161. Hinrichsen, L., Meyerholz, A., Groos, S., Ungewickell, E.J., 2006. Bending a membrane: how clathrin affects budding. Proc. Natl. Acad. Sci. USA 103, 8715–8720. Hinshaw, J.E., 2000. Dynamin and its role in membrane fission. Annu. Rev. Cell Dev. Biol. 16, 483–519. Hirst, J., Sahlender, D.A., Li, S., Lubben, N.B., Borner, G.H., et al., 2008. Auxilin depletion causes self-assembly of clathrin into membraneless cages in vivo. Traffic 9, 1354–1371. Jiang, R., Gao, B., Prasad, K., Greene, L.E., Eisenberg, E., 2000. Hsc70 chaperones clathrin and primes it to interact with vesicle membranes. J. Biol. Chem. 275, 8439–8447. Katsura, I., 1983. Theory on the structure and stability of coated vesicles. J. Theor. Biol. 103, 63–75. Keen, J.H., 1990. Clathrin and associated assembly and disassembly proteins. Annu. Rev. Biochem. 59, 415–438. Keen, J.H., Beck, K.A., Kirchhausen, T., Jarrett, T., 1991. Clathrin domains involved in recognition by assembly protein AP-2. J. Biol. Chem. 266, 7950–7956. Kocsis, E., Trus, B.L., Steer, C.J., Bisher, M.E., Steven, A.C., 1991. Image averaging of flexible fibrous macromolecules: the clathrin triskelion has an elastic proximal segment. J. Struct. Biol. 107, 6–14. Ladinsky, M.S., Mastronarde, D.N., McIntosh, J.R., Howell, K.E., Staehelin, L.A., 1999. Golgi structure in three dimensions: functional insights from the normal rat kidney cell. J. Cell Biol. 144, 1135–1149. Lapinski, M.M., Castro-Forero, A., Greiner, A.J., Ofoli, R.Y., Blanchard, G.J., 2007. Comparison of liposomes formed by sonication and extrusion: rotational and translational diffusion of an embedded chromophore. Langmuir 23, 11677– 11683. Liddington, R.C., Yan, Y., Moulai, J., Sahli, R., Benjamin, T.L., et al., 1991. Structure of simian virus 40 at 3.8-A resolution. Nature 354, 278–284. Mastronarde, D.N., 2005. Automated electron microscope tomography using robust prediction of specimen movements. J. Struct. Biol. 152, 36–51. Matlin, K.S., Reggio, H., Helenius, A., Simons, K., 1981. Infectious entry pathway of influenza virus in a canine kidney cell line. J. Cell Biol. 91, 601–613. Matsui, W., Kirchhausen, T., 1990. Stabilization of clathrin coats by the core of the clathrin-associated protein complex AP-2. Biochemistry 29, 10791–10798. McEwen, B.F., Frank, J., 2001. Electron tomographic and other approaches for imaging molecular machines. Curr. Opin. Neurobiol. 11, 594–600. McMahon, H.T., Boucrot, E., 2011. Molecular mechanism and physiological functions of clathrin-mediated endocytosis. Nat. Rev. Mol. Cell Biol. 12, 517– 533. Merisko, E.M., Farquhar, M.G., Palade, G.E., 1982. Coated vesicle isolation by immunoadsorption on Staphylococcus aureus cells. J. Cell Biol. 92, 846–857. Mitra, K., Ubarretxena-Belandia, I., Taguchi, T., Warren, G., Engelman, D.M., 2004. Modulation of the bilayer thickness of exocytic pathway membranes by membrane proteins rather than cholesterol. Proc. Natl. Acad. Sci. USA 101, 4083–4088. Nandi, P.K., Irace, G., Van Jaarsveld, P.P., Lippoldt, R.E., Edelhoch, H., 1982. Instability of coated vesicles in concentrated sucrose solutions. Proc. Natl. Acad. Sci. USA 79, 5881–5885. Pearse, B.M., Robinson, M.S., 1984. Purification and properties of 100-kd proteins from coated vesicles and their reconstitution with clathrin. EMBO J. 3, 1951– 1957. Peeler, J.S., Donzell, W.C., Anderson, R.G., 1993. The appendage domain of the AP-2 subunit is not required for assembly or invagination of clathrin-coated pits. J. Cell Biol. 120, 47–54. Petralia, R.S., Wang, Y.X., Indig, F.E., Bushlin, I., Wu, F., et al., 2013. Reduction of AP180 and CALM produces defects in synaptic vesicle size and density. Neuromolecular Med. 15, 49–60. J.B. Heymann et al. / Journal of Structural Biology 184 (2013) 43–51 Pettersen, E.F., Goddard, T.D., Huang, C.C., Couch, G.S., Greenblatt, D.M., et al., 2004. UCSF Chimera – a visualization system for exploratory research and analysis. J. Comput. Chem. 25, 1605–1612. Pizarro-Cerda, J., Bonazzi, M., Cossart, P., 2010. Clathrin-mediated endocytosis: what works for small, also works for big. BioEssays 32, 496–504. Qu, L., Akbergenova, Y., Hu, Y., Schikorski, T., 2009. Synapse-to-synapse variation in mean synaptic vesicle size and its relationship with synaptic morphology and function. J. Comp. Neurol. 514, 343–352. Radulescu, A.E., Siddhanta, A., Shields, D., 2007. A role for clathrin in reassembly of the Golgi apparatus. Mol. Biol. Cell 18, 94–105. Rust, M.J., Lakadamyali, M., Zhang, F., Zhuang, X., 2004. Assembly of endocytic machinery around individual influenza viruses during viral entry. Nat. Struct. Mol. Biol. 11, 567–573. Shih, W., Gallusser, A., Kirchhausen, T., 1995. A clathrin-binding site in the hinge of the beta 2 chain of mammalian AP-2 complexes. J. Biol. Chem. 270, 31083–31090. Smith, C.J., Grigorieff, N., Pearse, B.M., 1998. Clathrin coats at 21 A resolution: a cellular assembly designed to recycle multiple membrane receptors. EMBO J. 17, 4943–4953. Smith, C.J., Dafforn, T.R., Kent, H., Sims, C.A., Khubchandani-Aswani, K., et al., 2004. Location of auxilin within a clathrin cage. J. Mol. Biol. 336, 461–471. Steer, C.J., Bisher, M.E., Trus, B.L., Hainfeld, J.F., Wall, J.S., et al., 1988. Membrane contents of distinct subpopulations of coated vesicles determined by scanning transmission electron microscopy. Biochim. Biophys. Acta 938, 167–180. Steven, A.C., Hainfeld, J.F., Wall, J.S., Steer, C.J., 1983. Mass distributions of coated vesicles isolated from liver and brain: analysis by scanning transmission electron microscopy. J. Cell Biol. 97, 1714–1723. Sundborger, A., Soderblom, C., Vorontsova, O., Evergren, E., Hinshaw, J.E., et al., 2011. An endophilin-dynamin complex promotes budding of clathrin-coated vesicles during synaptic vesicle recycling. J. Cell Sci. 124, 133–143. Takamori, S., Holt, M., Stenius, K., Lemke, E.A., Gronborg, M., et al., 2006. Molecular anatomy of a trafficking organelle. Cell 127, 831–846. Taylor, D.R., Watt, N.T., Perera, W.S., Hooper, N.M., 2005. Assigning functions to distinct regions of the N-terminus of the prion protein that are involved in its copper-stimulated, clathrin-dependent endocytosis. J. Cell Sci. 118, 5141–5153. 51 Taylor, M.J., Perrais, D., Merrifield, C.J., 2011. A high precision survey of the molecular dynamics of mammalian clathrin-mediated endocytosis. PLoS Biol. 9, e1000604. Tenchov, B.G., Yanev, T.K., Tihova, M.G., Koynova, R.D., 1985. A probability concept about size distributions of sonicated lipid vesicles. Biochim. Biophys. Acta 816, 122–130. ter Haar, E., Harrison, S.C., Kirchhausen, T., 2000. Peptide-in-groove interactions link target proteins to the beta-propeller of clathrin. Proc. Natl. Acad. Sci. USA 97, 1096–1100. Traub, L.M., 2005. Common principles in clathrin-mediated sorting at the Golgi and the plasma membrane. Biochim. Biophys. Acta 1744, 415–437. Ungewickell, E., Branton, D., 1981. Assembly units of clathrin coats. Nature 289, 420–422. Vigers, G.P., Crowther, R.A., Pearse, B.M., 1986a. Location of the 100 kd-50 kd accessory proteins in clathrin coats. EMBO J. 5, 2079–2085. Vigers, G.P., Crowther, R.A., Pearse, B.M., 1986b. Three-dimensional structure of clathrin cages in ice. EMBO J. 5, 529–534. Wakeham, D.E., Chen, C.Y., Greene, B., Hwang, P.K., Brodsky, F.M., 2003. Clathrin self-assembly involves coordinated weak interactions favorable for cellular regulation. EMBO J. 22, 4980–4990. Woodka, A.C., Butler, P.D., Porcar, L., Farago, B., Nagao, M., 2012. Lipid bilayers and membrane dynamics: insight into thickness fluctuations. Phys. Rev. Lett. 109, 058102. Xing, Y., Bocking, T., Wolf, M., Grigorieff, N., Kirchhausen, T., et al., 2010. Structure of clathrin coat with bound Hsc70 and auxilin: mechanism of Hsc70-facilitated disassembly. EMBO J. 29, 655–665. Zampighi, G.A., Zampighi, L., Fain, N., Wright, E.M., Cantele, F., et al., 2005. Conical tomography II: a method for the study of cellular organelles in thin sections. J. Struct. Biol. 151, 263–274. Zhao, X., Greener, T., Al-Hasani, H., Cushman, S.W., Eisenberg, E., et al., 2001. Expression of auxilin or AP180 inhibits endocytosis by mislocalizing clathrin: evidence for formation of nascent pits containing AP1 or AP2 but not clathrin. J. Cell Sci. 114, 353–365.