Structural characterization of the neurabin sterile alpha motif domain



Neurabin1 (SWISS-PROT/TrEMBL O35867) and its isoform spinophilin2, 3 (SWISS-PROT/TrEMBL O35274) are neuronal scaffolding proteins that are critical for synaptic transmission and synaptic plasticity.4, 5 They are highly enriched in dendritic spines, the site of excitatory neurotransmission, because of their ability to interact with F-actin in this subcellular compartment.6 Neurabin is expressed exclusively in neuronal cells, whereas spinophilin is expressed ubiquitously, although it is also highly enriched in neurons. Spinophilin is sometimes referred to as neurabinII.3 Neurabin consists of 1095 residues (MW 122,730 Da), whereas spinophilin is smaller consisting of only 817 residues (MW 89,640 Da).

A recent report by Allen et al.7 showed, using neurabin and spinophilin knockout mice, that both proteins are necessary for monoamineric signal transduction. Thus, despite their similarity in domain structure (see Fig. 1), neurabin and spinophilin have distinct signaling and regulatory roles. Therefore, it is essential to understand the structural differences upon which this functional difference is based. As is typical for scaffolding proteins, both proteins contain multiple common protein interaction domains (see Fig. 1). Both neurabin and spinophilin contain an F-actin binding, a functionally critical Protein Phosphatase 1 (PP1)-binding, a PDZ,8, 9 and a C-terminal coiled-coil domain; sequence similarity within these functional domains can be as high as 86% (PDZ domains). The F-actin binding domain is important for targeting these proteins to dendritic spines.6, 10 The PP1-binding domain of both isoforms targets PP1 towards its substrates, one of which is Ser845 of the GluR1 AMPA receptor subunit. The PDZ domains of both proteins have also been reported to bind the C-termini of p70S6 kinase11 and kalirin-7.12 Finally, the coiled-coil domains have been implicated in playing a central role in the homo- and hetero-multimerization of these two proteins with themselves or with coiled-coil domains from other proteins.13, 14 The most significant difference in the domain structure between these two proteins is that neurabin, but not spinophilin, contains a sterile alpha motif (SAM) domain at its C-terminus (Pfam PF07647). Only in invertebrates spinophilin can have a SAM domain.15 Currently, no functional role has been identified for the neurabin SAM domain.

Figure 1.

A, Domain structure of spinophilin (top) and neurabin (bottom). Sequence identity between the domains is annotated in the figure. B, Primary sequence alignment of the neurabin SAM and shank3 SAM domains. The zinc binding residues in shank3 and the corresponding residues in neurabin are highlighted by boxes (see text). The shank3 (M1803E) mutation that is necessary to achieve high enough protein solubility and reduced aggregation is underlined. Secondary structure for the neurabin SAM domain is included above the sequence alignment.

In 1995, Ponting16 identified 14 proteins that each contained a conserved domain of ∼70 residues. This domain was named SAM because of its presence in yeast proteins that are essential for sexual differentiation. SAM domains have since been found in more than 1300 proteins,17 comparing well with ∼1600 proteins that have SH2 domains (Src Homology 2), underlying its importance as a functional protein motif. But unlike SH2 domains, which nearly always bind to phosphorylated tyrosine residues, there has so far been no common functional theme identified for SAM domains. Instead SAM domains appear to have a wide diversity of functions. SAM domains can either form homodimers with themselves or heterodimers with other SAM domains.18 Recently, reports have shown that SAM domains are not only protein:protein interaction sites but also protein:RNA recognition motifs.19, 20

Baron et al.21 demonstrated that the SAM domain of the neuronal scaffolding protein shank3 also forms homo-polymeric α-helical sheet-like structures and suggested that the shank3 SAM domain may be essential for the organization of the post synaptic density (PSD). This polymerization and organization-like behavior of the shank3 SAM domain is promoted in vitro by the addition of zinc, which is present in the PSD (4.1 nmol/mg of protein).22 In the present study, we have determined the structure of the neurabin SAM domain by solution state nuclear magnetic resonance (NMR) spectroscopy. Notably, our results show that the shank3 SAM domain is the closest structural neighbor of the neurabin SAM domain. We have also examined the polymerization state of the neurabin SAM domain and find, in contrast to the shank3 SAM domain, that it is a monomer.


Expression and purification of the neurabin SAM domain

After optimal construct screening,23 the neurabin SAM domain (residues 986–1056) was cloned into a his-tagged bacterial expression vector (Thio6-His6-TEV-). This construct was transformed into the Escherichia coli strain BL21-CodonPlus (DE3)-RIL (Stratagene) and a single colony was used to inoculate a 100-mL culture of LB containing kanamycin (50 μg/mL) and chloramphenicol (34 μg/mL). The culture was grown overnight at 37°C with shaking at 250 rpm The next morning the cells were diluted 1:50 into fresh LB medium with appropriate antibiotics. Cells were grown at 37°C until they reached an OD600 of 0.6–0.9. The cultures were then placed at 4°C and the shaker temperatures adjusted to 18°C. Cultures were induced with 1 mM isopropyl β-D-1-thiogalactopyranoside and allowed to express overnight (∼18 h) at 18°C under vigorous shaking (250 rpm). Cultures were harvested by centrifugation and the pellets stored at −80°C until purification. The expression of uniformly 13C/15N-labeled and 15N-labeled protein was carried out by growing freshly transformed cells in M9 minimal medium containing 4 g/L [13C]-D-glucose and/or 1 g/L 15NH4Cl as the sole carbon and nitrogen sources, respectively.

For purification, the pellets were resuspended in lysis buffer (50 mM Tris pH 8.0, 5 mM imidazole, 500 mM NaCl, and 0.1% Triton-X, Complete tabs-EDTA free (Roche)). The cells were lysed by three passes through a C3 Emulsiflex cell cracker (Avestin) and the cell debris removed by centrifugation (40,000g/30 min/4°C). The clarified lysate was filtered through a 0.22-μm membrane (Millipore) and loaded onto a HisTrap HP column (GE Healthcare) equilibrated with 50 mM Tris pH 8.0, 5 mM imidazole, and 500 mM NaCl. The protein was eluted with a 5–500-mM imidazole gradient. The fractions containing the target protein were identified by SDS-PAGE gel electrophoresis and were pooled and dialyzed against 50 mM Tris pH 7.5 and 50 mM NaCl. Tobacco etch virus (TEV) NIa protease was used to cleave the purification tag. Cleavage, verified by SDS-PAGE gel electrophoresis, was achieved overnight at room temperature under steady rocking. The cleaved sample was then dialyzed against 50 mM Tris pH 7.5 and 250 mM NaCl for 5 h and further purified by a second his-tag purification step. The cleaved, now untagged SAM domain was collected in the flow through. The purity of the final sample was checked by SDS-PAGE gel analysis. The sample was concentrated and finally exchanged into a buffer containing 20 mM sodium phosphate buffer pH 6.5 and 50 mM NaCl using an Amicon Ultra centrifugation filter device with a 5000 MWCO (Millipore). Ten percent D2O and 0.02% NaN3 were added into the sample that reached a final concentration of ∼2 mM.

All NMR measurements were performed at 298 K on a Bruker AvanceII 500 MHz spectrometer using a TCI HCN-z cryoprobe. Proton chemical shifts were referenced to internal 3-(trimethyl-silyl)-1-propanesulfonic acid, sodium salt (DSS). Using the absolute frequency ratios, the 13C and 15N chemical shifts were referenced indirectly to DSS.

Chemical shift assignment and structure calculation

The following spectra were used to achieve the sequence-specific backbone and sidechain assignments of all aliphatic residues: 2D [1H,15N]-HSQC, 2D [1H,13C]-HSQC, 3D HNCACB, 3D CBCA(CO)NH, 3D CC(CO)NH, 3D HNCO, 3D HNCA, 3D HBHA(CO)NH, 3D 15N-resolved [1H,1H]-TOCSY, 3D HC(C)H-TOCSY.24 The 2D [1H,1H]-NOESY, 2D [1H,1H]-TOCSY, and 2D [1H,1H]-COSY spectra of the neurabin986–1056 samples in D2O solution after complete H/D exchange of the labile protons were used for the assignment of the aromatic side chains. The NMR spectra were processed with Topspin1.3 (Bruker, Billerica, MA) and analyzed with the CARA software package (

The following spectra were used for structure calculation: 3D 15N-resolved [1H,1H]-NOESY (mixing time of 85 ms), 3D 13C-resolved [1H,1H]-NOESY (mixing time of 85 ms), and 2D [1H,1H]-NOESY (mixing time 85 ms, D2O solution). The amino acid sequence, the chemical shift assignment, and the NOESY spectra were input for the automated NOESY peak picking and NOE assignment method of ATNOS/CANDID/CYANA.25–27 The results of ATNOS/CANDID/CYANA were refined by manual peak adjustment and additional calculations in CYANA. The 20 conformers from the final CYANA cycle with the lowest residual CYANA target function values (of 100 calculated) were energy-minimized in a water shell with the program CNS28 using the RECOORD29 script package. The quality of the structures was assessed by the programs WHATCHECK,30 AQUA,31 NMR-PROCHECK,31 and MOLMOL.32

Chemical shift assignments of neurabin986–1056 were deposited in the BMRB under accession number 7118 and coordinates were submitted to the PDB under accession code 2GLE.

Zinc titration of the neurabin SAM domain

Purified neurabin986–1056 was exchanged from the NMR based phosphate buffer into the following buffer, 10 mM Tris pH 7.5 and 50 mM NaCl, using a HiTrap desalting column (GE Healthcare). This extra step is necessary to prevent precipitation of Zn3(PO4)2. A 2D [1H,15N] HSQC spectrum was used to monitor perturbations in 1H and 15N chemical shifts upon zinc titration. A 50-μM neurabin986–1056 sample was used to obtain the reference spectrum and 100 μM of Zn2+ was subsequently added.

Neurabin coiled-coil domain (neurabin627–822) expression and purification

DNA encompassing the neurabin coiled-coil domain (neurabin627–822) was subcloned into a modified pETM41 (His6-MBP-TEV-, EMBL) vector. The plasmid was transformed into the E. coli strain BL21-CodonPlus (DE3)-RIL (Stratagene). Cultures were grown under vigorous shaking until they reached an OD of ∼0.7. Expression was induced by 1 mM IPTG and cultures were incubated overnight (∼18 h) at 18°C under vigorous shaking. Cells were collected by centrifugation and stored at −80°C until purification. For purification, cells were resuspended in lysis buffer (50 mM Tris HCl pH 8.0, 5 mM Imidazole, 500 mM NaCl, and 0.1% Triton X-100) containing one Complete EDTA-free Protease Inhibitor Cocktail Tablet (Roche) and lysed by three passes through a C3 Emulsiflex cell cracker (Avestin). Cellular debris was pelleted by centrifugation at 45,000g/35 min/4°C. The filtered soluble fraction was loaded onto a Histrap HP column (GE Healthcare) pre-equilibrated with 50 mM Tris-HCl pH 8.0, 5 mM imidazole, and 500 mM NaCl. Protein was eluted using a gradient of 5–500 mM imidazole in 36 column volumes. Fractions containing MBP-neurabin627–822 were pooled, TEV NIa protease was added and the sample was dialyzed overnight (∼18 h, 4°C) against 25 mM Tris pH 8.0 and 500 mM NaCl. Successfully cleaved protein was subsequently loaded onto a second his-tag column that had been pre-equilibrated with 50 mM Tris-HCl pH 8.0, 5 mM imidazole, and 500 mM NaCl and incubated at 4°C with shaking for 1.5 h. All fractions containing untagged neurabin627–822 were collected and pooled. The sample was concentrated to ∼400 μM and exchanged into the following buffer: 20 mM sodium phosphate buffer pH 6.5 and 50 mM NaCl.

Biophysical characterization of neurabin627–822

CD spectroscopy was used to determine the folded state of neurabin627–822 (Jasco J-810). The data were recorded using the following parameters: 20 mM sodium phosphate buffer, pH 6.5, 500 mM NaCl, and 0.02% NaN3, room temperature, continuous scanning mode, scanning speed of 20 μm/min, 3 scans/experiment, and a cell length of 1 cm. The CD spectrum of neurabin627–822 confirmed its expected α-helical state. To test the multimeric state of neurabin627–822, dynamic light scattering (DLS, Viscotec 802), native gel electrophoresis, and size exclusion chromatography (Superdex-200, GE Healthcare) were performed.


The neurabin SAM domain has a typical SAM domain fold

The 3D structure of the neurabin986–1056 has been solved by solution state NMR spectroscopy. The structure resembles the typical SAM domain fold consisting of five helices (α-helix1: 9–19; α-helix2: 21–29; α-helix3: 34–40; α-helix4: 42–48; and α-helix5: 53–74). In other SAM domains, such as TEL or ETS-1, the loops that link the helical core are highly flexible.33, 34 However, this is not the case for the neurabin SAM domain where all loops are highly structured. Helices 1–4 are relatively short (2–3 helical turns) and surround the considerably longer (5 turns) C-terminal Helix 5 [Fig. 2(C)]. The most striking characteristic of the neurabin986–1056 SAM domain is that all five helices form a compact bundle. The four aromatic residues in the neurabin SAM domain (2 Trp, 1 Tyr, and 1 Phe) are buried in the hydrophobic core of the structure [Fig. 2(A), orange]. The surface of the neurabin SAM domain is predominately hydrophilic, with a positively charged patch along the extended Helix 5 and a more hydrophobic region interspersed with negative charge on the opposite side, composed of Helices 1 and 2 [Fig. 2(B)]. This mixed charged and hydrophobic surface is surprising, because most known SAM domain interactions are mediated by hydrophobic surface residues. In addition, it excludes clearly the possibility that the neurabin SAM domain is an RNA interacting SAM domain, such as the Smaug and VTS1 SAM domains,19, 35–37 which have large positively charged surfaces critical for RNA interaction.

Figure 2.

A, Stereo view of the bundle of 20 energy-minimized CYANA conformers representing the structure of neurabin986–1056 (PDB ID 2GLE). The superposition is the best fit of the backbone atoms N, Cα, C′ of residues 4–70. Side chains with low local RMSD (<0.7) within the bundle are shown in cyan. Side chains of residues with high local RMSD (>0.7) are shown in green. Backbones are shown in dark blue. Aromatic heavy side chains are colored in orange. The local RMSD of Trp7, Trp15, Tyr24, and Phe28 are 0.31, 0.11, 0.87, and 0.62, respectively. B, Electrostatic surface of the neurabin SAM domain (left) and 180° rotation around x-axes (right). Red represents negative charges and blue positive charges. C, Ribbon presentation of the closest conformer to the mean coordinates of the bundle of 20 conformers used to represent the NMR structure of the neurabin SAM (left) and the shank3 SAM domain (right). The regular secondary structure elements are identified.

Tyr24 of the neurabin SAM domain is a proposed phosphorylation site that is conserved in numerous SAM domains. The specific tyrosine kinase for the neurabin SAM domain is currently unknown. Tyr24 seems to be placed in a protected pocket formed by Helices 2, 3, and 4 and not accessible for direct phosphorylation. Therefore, the SAM domain needs to be structurally rearranged to become phosphorylated. This phosphorylation might be necessary for neurabin SAM domain to interact with its potential protein interaction partner.

A total of 2121 NOESY derived distance constraints (∼29 NOE constraints per residue) were used for the structure calculation of neurabin986–1056 (Table I). The neurabin986–1056 model has excellent stereochemistry, as tested with NMR-PROCHECK,38, 39 with 83.8% of the residues in the most favored region, 16.2% in the additionally allowed region of the Ramachandran diagram, and no residues in the generously allowed region and the disallowed region (Table I). Figure 2(A) shows the local RMSD of the side chain heavy atoms. Flexible residues (green) are located on the surface of the structure whereas the hydrophobic core residues (cyan) are well defined and show very low mobility.

Table I. Structural and CNS Refinement Statistics for the Neurabin SAM Domain
 Neurabin SAM
Number of restrains
 Unambiguous distance restrains (all)2121
  Short range |ij| ≤ 1982
  Medium range 1 < |ij| < 5612
  Long range |ij| ≥ 5527
 Unambiguous distance restrains per residue28.7
Deviations from idealized covalent geometry
 Bonds (Å)0.0108 ± 0.0002
 Angles (°)1.27 ± 0.035
 Impropers (°)1.43 ± 0.07
 No. of violation NOE (>0.5 Å)0 ± 0
 No. of violation dihedral (>5°)0 ± 0
Structural quality
 Ramachandran plot (NMR-PROCHECK30) 
  Most favored region (%)83.8
  Additionally allowed region (%)16.2
  Generously allowed region (%)0
  Disallowed region (%)0
 RMSD within the bundle (Å) 
  Backbone (N, Cα, C, and O) (4–70)0.31 ± 0.05
  All heavy atoms (4–70)0.63 ± 0.05

The neurabin SAM domain is a monomer in solution

SAM domains are versatile protein:protein interaction domains. They are known to self-associate to form homo- and hetero-oligomers,40 long helical homopolymers,34 or in some cases even multihelical sheets, as recently reported for the shank3 SAM domain in the PSD.21 To determine the role of the neurabin SAM domain in the multimerization of neurabin, it is necessary to first determine the oligomeric state of isolated neurabin SAM domain in solution. Three techniques were employed. First, measurements of the transverse relaxation times (T2) of the amide protons were carried out using a 1D 1H one–one spin echo experiments.41, 42 At different lengths of the echo delay (100 and 2.9 ms), the intensity of the amide proton signals was differentially modulated so as to enable a determination of the transverse relaxation time. The intensities of ∼10 separated amide resonances in the 1D 1H NMR spectrum of the neurabin SAM domain (MW 8.29 kDa) between 8.0 and 10 ppm were used to extract an average amide proton T2 relaxation time of 30.6 ± 4.2 ms. For comparison, lysozyme, which is known to be a monomer in solution with a molecular weight of 14.4 kDa, was dissolved in the identical buffer used for the neurabin SAM measurements and its averaged amide proton T2 relaxation time was 18.2 ± 1.7 ms. Therefore, the significantly longer average amide proton transverse relaxation time of the neurabin SAM domain when compared with lysozyme confirms directly that the neurabin SAM is monomeric in solution. Secondly, this result is also supported by analytical size exclusion chromatography data (Superdex 200, GE Healthcare) and DLS (Viscotec 802) data, which yielded a molecular weight of about 8 kDa for the neurabin SAM domain (data not shown). This observation demonstrates that the neurabin SAM domain is not a polymerization domain, rather it must function via protein:protein interaction with other proteins.

The coiled-coil domain, and not the SAM domain, is the solely responsible multimerization domain of neurabin

To understand the function of the neurabin and how it differs from spinophilin, it is important to understand the function of each individual neurabin domain and their intramolecular interactions. In neurabin, domains fulfill joint functions, that is, the PP1-binding and PDZ domain are critical for targeting and anchoring of PP1 close to post-synaptic substrates. Since it has been shown that in vivo neurabin can form homo- or hetero-multimers1, 14 and deletion of the coiled-coil and SAM domains in neurabin abolishes its dimerization,6 it was of interest to see if these domains function as one entity or are functionally distinct. Although the neurabin SAM domain is clearly a monomer in solution, the SAM domain could play an indirect dimerization role through influencing the interaction of the coiled-coil domain. Recombinantly expressed neurabin coiled-coil domain is clearly a multimer, as tested using DLS, SEC, and native gel analysis (see Methods). Although it is possible to investigate its multimerization state, it is more difficult to distinguish between its exact multimerization status, because all techniques are calibrated for globular proteins and not anisotropic proteins such as the more extended neurabin coiled-coil domain. However, all of our data suggest that the neurabin coiled-coil domain is at least a trimer in solution. To test for neurabin coiled-coil:SAM domain interaction, purified 15N labeled neurabin SAM domain was titrated with 1 equiv. of purified, unlabeled neurabin coiled-coil domain. No changes in the 2D [1H,15N] HSQC spectrum of the neurabin SAM domain were observed (extreme line broadening beyond detect-ability was expected in case of binding, because of the high molecular weight of the neurabin coiled-coil multimer) upon the addition of the coiled-coil domain (data not shown). Thus, the neurabin SAM domain does not interact intra or intermolecularly with the neurabin coiled-coil domain.

The shank3 SAM domain is the closest structural neighbor of the neurabin SAM domain

The 3D structure of the neurabin SAM domain was used to further guide the identification of its functional role in the PSD. As expected for SAM domains, the structural similarity between the neurabin SAM domain and other SAM domains is high, even though the sequence identity is low (between 12 and 28%). A pair-wise comparison between the neurabin SAM domain structure and other SAM domain structures was performed using DALI43 and Superpose,44 respectively. Both 3D structure-based methods identified that the closest structural neighbor of the neurabin SAM domain is the shank3 SAM domain (z-score 10.4; PDBid shank3 SAM domain PDBid: 2F3N).

The shank3 SAM domain was recently hypothesized to play a critical organizational scaffolding role in the PSD. Namely, in vitro, it can form 2D multihelical sheets, which have been proposed to form the underlying architectural framework for the organization of the PSD. In contrast to the shank3 SAM domain, which forms homopolymeric sheets, the neurabin SAM domain is a monomer is solution. However, it was experimentally shown that Zn2+, which is present in vivo in the PSD, can bind to the shank3 SAM domain and thereby dramatically foster this multihelical sheet formation. The Zn2+ binding pocket of the shank3 SAM domain (E1768, H1769, H1801) is located in inter and intrapolymer interfaces,45 which enhances the organization of the sheet structure. Given the structural similarity of the shank3 and neurabin SAM domains, we examined whether Zn2+ might bind to the neurabin SAM domain. Specifically, a 2D [1H,15N] HSQC spectrum of 15N-labeled neurabin SAM was used to monitor chemical shift perturbations upon titration with Zn2+. It was possible to identify a potential Zn2+ binding pocket in the neurabin SAM domain formed by His2, His987, Glu988, and Glu1054 (changes in chemical shifts monitored by 2D [1H,15N] HSQC measurements). Importantly, His2 is a cloning artifact following TEV HIS-tag cleavage. In the native neurabin SAM sequence this residue is an arginine (Arg984). Moreover, this zinc binding pocket is not located in the corresponding shank3 SAM domain structure. Finally, the neurabin SAM domain remains monomeric in solution upon zinc binding. Therefore, it is apparent that the native neurabin SAM domain does not have a preformed Zn2+ binding site, other than a cloning artifact based one, and does not interact with Zn2+ ions. Thus, despite their similarity in structure, the scaffolding behavior of the neurabin SAM domain is clearly distinct from that of its closest structural neighbor, the shank3 SAM domain.


We thank R. Page for careful reading of the manuscript. The DLS instrument was purchased using Brown University Seed Funds. ITC measurements were performed in the NSF EPSCoR supported proteomics facility at Brown University. CD measurements were performed in the RI-INBRE Research Core Facility. We thank Dr. John Marshall (Brown University) for the shank3 plasmid and Dr. James Bowie (UCLA) for discussion and the shank3 (L47E) protein.