(C) PLOS One [1]. This unaltered content originally appeared in journals.plosone.org.
Licensed under Creative Commons Attribution (CC BY) license.
url:https://journals.plos.org/plosone/s/licenses-and-copyright

------------



High-affinity anti-Arc nanobodies provide tools for structural and functional studies

['Sigurbjörn Markússon', 'Department Of Biomedicine', 'University Of Bergen', 'Bergen', 'Erik I. Hallin', 'Helene J. Bustad', 'Arne Raasakka', 'Ju Xu', 'Gopinath Muruganandam', 'Vib-Vub Center For Structural Biology']

Date: 2022-06

Activity-regulated cytoskeleton-associated protein (Arc) is a multidomain protein of retroviral origin with a vital role in the regulation of synaptic plasticity and memory formation in mammals. However, the mechanistic and structural basis of Arc function is poorly understood. Arc has an N-terminal domain (NTD) involved in membrane binding and a C-terminal domain (CTD) that binds postsynaptic protein ligands. In addition, the NTD and CTD both function in Arc oligomerisation, including assembly of retrovirus-like capsids involved in intercellular signalling. To obtain new tools for studies on Arc structure and function, we produced and characterised six high-affinity anti-Arc nanobodies (Nb). The CTD of rat and human Arc were both crystallised in ternary complexes with two Nbs. One Nb bound deep into the stargazin-binding pocket of Arc CTD and suggested competitive binding with Arc ligand peptides. The crystallisation of the human Arc CTD in two different conformations, accompanied by SAXS data and molecular dynamics simulations, paints a dynamic picture of the mammalian Arc CTD. The collapsed conformation closely resembles Drosophila Arc in capsids, suggesting that we have trapped a capsid-like conformation of the human Arc CTD. Our data obtained with the help of anti-Arc Nbs suggest that structural dynamics of the CTD and dimerisation of the NTD may promote the formation of capsids. Taken together, the recombinant high-affinity anti-Arc Nbs are versatile tools that can be further developed for studying mammalian Arc structure and function, as well as mechanisms of Arc capsid formation, both in vitro and in vivo. For example, the Nbs could serve as a genetically encoded tools for inhibition of endogenous Arc interactions in the study of neuronal function and plasticity.

Funding: This work was supported by a Research Council of Norway ( www.forskningsradet.no ) TOPPFORSK grant (249951) to CRB, as well as the Research Council of Norway ( www.forskningsradet.no ) NORCRYST (grant number 245828) infrastructure consortium. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.

Copyright: © 2022 Markússon et al. This is an open access article distributed under the terms of the Creative Commons Attribution License , which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.

In order to develop new tools for Arc structural and functional studies, we produced and characterised six high-affinity anti-Arc nanobodies (Nbs). Two of the Nbs facilitated crystallisation of the rat and human Arc-CTD, in an extended and collapsed conformation, visualising conformational dynamics of the Arc-CTD. The complementarity-determining region (CDR) 3 of NbArc-H11 bound deep into the peptide binding site of the Arc-CTD and inhibited peptide binding in vitro, suggesting applicability for studying Arc function at the molecular level. Our data suggest a mechanism of Arc capsid assembly, in which structural dynamics of the CTD facilitate capsomer formation, and NTD dimerisation leads to capsomer linking and formation of the mature capsid. The Arc Nbs provide an excellent starting point for further development of tools for both structural biology, imaging, immunodetection, and functional studies.

High-resolution structures of the capsids formed by Drosophila melanogaster Arc isoforms 1 and 2 (dArc1 and dArc2, respectively) revealed assemblies of icosahedral symmetry with 240 Arc protomers [ 40 ]. There are, however, fundamental differences in mArc and dArc. dArc lacks the coiled-coil NTD found in mArc, and the two likely originated from two separate domestication events, suggesting evolutionary convergence [ 20 ]. Both dArc isoforms spontaneously assemble into capsids at ionic strengths mimicking RNA binding in vitro [ 40 ]. This is thought to occur in part due to the inherent tendency of the dArc CTD lobes to oligomerise [ 41 ]. In contrast, the mArc CTD is monomeric in solution [ 31 , 34 , 35 ], but dimeric forms of the CTD appear to be important in capsid assembly [ 42 ]. mArc is unable to associate with mRNA and form higher-order oligomers in the absence of the NTD, and it has been suggested that full-length Arc is needed for the formation of the capsids, with the long linker region connecting the two domains facilitating domain swapping [ 18 , 35 ]. Alanine scanning identified a motif within the NTD, 113 MHVWREV 119 , as a main facilitator of higher-order oligomerization and capsid formation in rat Arc (rArc). Furthermore, a crystal structure of short peptide harbouring the motif revealed a dimeric coiled coil. Upon replacement of the motif by poly-Ala, the full-length protein lost higher-order oligomerisation and is dimeric [ 18 ]. The structure of the mArc capsid and the mechanism of capsid formation are unknown.

Transposable elements (TEs), such as retrotransposons, constitute a major component of vertebrate genomes that can cause deletions and genomic instability. However, depending on the genomic context at their insertion site, they can sometimes be positively selected for in an exaptation, or domestication, event. Arc, likely introduced into the tetrapod genome via retroviral insertion, was initially identified in a search for Gag homology proteins descendant from Ty3/Gypsy retrotransposons [ 38 ]. Structural studies on the Arc-CTD show striking similarity with capsid domains of retroviruses [ 26 , 35 , 39 ]. Arc retains its ancestral ability to form viral-like capsids, as it forms large capsids that encapsulate RNA to transfer it between neurons, or, as shown in Drosophila, from neuron to muscle [ 17 , 20 ]. Arc preferably forms these assemblies upon binding of its own mRNA [ 18 ]. In this way, Arc encapsulates its own mRNA and transports it between cells in a viral infection-like manner.

The mArc NTD has low solubility, which has hindered detailed structural characterisation. It was suggested to form a coiled coil, and using small-angle X-ray scattering (SAXS) and Förster resonance energy transfer (FRET), it was shown to pack against the CTD, resulting in relatively compact full-length Arc [ 35 ]. The NTD is required for the association of Arc with endophilin [ 23 ], although direct binding has not been demonstrated. Full-length Arc, but not the CTD, associates with membranes rich in anionic lipids, suggesting a role for the NTD in membrane association [ 35 ]. Membrane association of Arc can introduce negative curvature in anionic membranes [ 36 ], and the NTD contains a cysteine cluster ( 94 CLCRC 98 ), which is S-palmitoylated in vivo, likely to facilitate membrane binding [ 37 ].

Mammalian Arc (mArc) consists of two folded domains, a basic N-terminal domain (NTD) and an acidic C-terminal domain (CTD), connected by a flexible linker [ 19 ]. The linker region and the N- and C-termini are disordered. The CTD of Arc consists of the N-lobe and the C-lobe [ 26 , 34 ]. The solution structure of the rat Arc CTD revealed a rigid bilobar structure, where the two lobes are connected by a short non-helical region [ 31 ]. The N-lobe of the CTD harbours a peptide binding groove, which mediates many protein-protein interactions of Arc, including binding of Stg, GluN2A, GKAP and other peptides [ 26 , 31 , 34 ]. The binding pocket recognises a PxF/Y sequence motif, where the proline takes part in conserved C-H…π interactions, and the aromatic residue binds into a π-cluster within the hydrophobic core [ 34 ]. The N-terminal end of the CTD N-lobe is extended from the domain in the unbound state, but upon peptide binding, it folds against the bound peptide to form a β-sheet [ 31 , 34 ]. The C-lobe does apparently not participate in peptide binding, despite its structural similarity to the N-lobe [ 34 ].

Arc is a fundamental regulator of synaptic plasticity, and high expression levels of Arc lead to the internalisation of AMPA-type glutamate receptors at the postsynaptic membrane [ 21 , 22 ]. Furthermore, Arc directly interacts with both endophilin, dynamin, and clathrin-adaptor protein 2 (AP2) [ 23 , 24 ]. Dynamin and endophilin both have essential late roles for clathrin-mediated synaptic receptor cycling, whereas AP2 has a critical role in initiating the process, as it coordinates cargo recruitment and selection by recruiting clathrin to the membrane [ 25 ]. A direct interaction of Arc with the AMPA receptor has been confirmed, as Arc binds the cytosolic tail of stargazin (Stg) [ 26 , 27 ]. Stg is an auxiliary subunit of AMPA receptors, associating with the receptor to modulate ion gating and receptor trafficking [ 28 ]. Long-term potentiation (LTP) is often induced via the activation of NMDA receptors, and Arc co-localises with NMDA receptors in postsynaptic complexes [ 7 , 29 , 30 ]. While Arc requires postsynaptic density protein 95 (PSD95) for localisation to the postsynaptic density, direct interactions with NMDA receptor subunits GluN2A and GluN2B have been identified [ 26 , 31 ]. GluN2A and GluN2B are vital for NMDA receptor-dependent LTP and long-term depression (LTD) [ 32 , 33 ]. Whether Arc regulates NMDA receptor function or serves as an integrator between LTP and LTD, is not yet fully understood.

Activity-regulated cytoskeleton-associated protein (Arc) is a highly conserved protein in vertebrates that has emerged as a key regulator of long-term synaptic plasticity, with roles in postnatal cortical development, memory, and cognitive flexibility [ 1 – 3 ]. Arc is an immediate early gene in glutamatergic neurons that is highly expressed in a transient manner upon synaptic activation and salient behavioural experience [ 4 – 6 ]. Following transcription, Arc mRNA is transported from soma to dendrites, and translation occurs locally in or near dendritic spines of activated glutamatergic synapses [ 7 ]. The expression of Arc is dynamic, and the protein is quickly degraded [ 8 – 11 ]. Although loss-of-function studies have established causal roles for Arc [ 12 – 15 ], the molecular basis and cellular mechanisms of Arc are not fully understood [ 11 , 16 ]. Arc serves as protein interaction hub, with several binding partners in the postsynaptic membrane and dendritic spines of excitatory synapses, as well as in the neuronal nucleus [ 16 ]. Arc can also self-assemble, forming oligomers and large retroviral-like-capsids implicated in intercellular communication [ 17 – 20 ].

As the NTD of Arc had previously been shown to be highly insoluble in the absence of its MBP fusion partner, solubilisation in L-arginine (L-Arg) was attempted. Purified MBP-2rNT fusion, at ~1 mg/mL, was dialysed against 1 L of 20 mM HEPES pH 7.5, 150 mM NaCl and 375 mM L-Arg pH 7.5. Following ~3 hr dialysis, the sample was supplemented with ~500 μg TEV protease and further dialysed against the same buffer overnight. For removal of the cleaved His-MBP affinity tag, the protein was subjected to negative NiNTA affinity purification and the unbound protein further purified on amylose resin to remove residual MBP. Unbound protein from the amylose resin was concentrated to 500 μL and injected on a Superdex 75 increase 10/300 GL column equilibrated in 20 mM HEPES pH 7.5, 150 mM NaCl and 375 mM L-Arg pH 7.5. Purity of the eluted NTD peak was assessed through SDS-PAGE. Although soluble NTD could be obtained through this method, the low yield did not allow for further biophysical characterisation.

Dynamic light scattering (DLS) of MBP-2rNT was measured on a Zetasizer Nano ZS DLS instrument (Malvern Panalytical) on 1 mg/mL sample in a 3x3 mm quartz cuvette (Hellma analytical) at 25°C. Prior to the measurement, samples were filtered through 0.22 μm centrifugal filters at 6,000 g and 4°C for 8 min. 30 μL of sample were loaded in the cuvette, allowed to equilibrate for 120 s and measured three times, with each run consisting of 10 accumulative measurements. Hydration radius (R H ) was determined as the maximum of the volume fraction size distribution profile.

Atomistic MD simulations were carried out on the hArc-CTD in GROMACS [ 64 ]. The hArc CTD crystal structure was stripped of Nbs, protons and ions, and assigned the predicted protonation state at pH 7.0 using PROPKA in PDB2PQR [ 65 ]. Protonated hArc-CTD was placed in a cubic box with a 10 Å extension around the protein. Solvation was done with the TIP3P water model in 0.15 M NaCl. The models were subjected to conjugate gradient energy minimisation with a steepest decent step every 50th step and a maximum of 5000 steps. Temperature (NVT) equilibration to 300 K and pressure (NPT) equilibration, via isotropic pressure coupling, were carried out using the Berendsen thermostat. MD simulations were carried out using the OPLS-AA/L force field [ 66 ] and leap-frog algorithm, while retaining constant temperature and pressure using a velocity-rescale thermostat and a Parinello-Rahman barostat, respectively. The MD trajectories were uploaded to Zenodo [ 67 – 70 ].

Diffraction data were processed in XDS [ 56 ]. Analysis of data quality was carried out in XTRIAGE [ 57 ] and phases solved using molecular replacement (MR) in PHASER [ 58 ]. Refinement was carried out in PHENIX.REFINE [ 57 ] and manual model building in Coot [ 59 ]. Anisotropy analysis and anisotropic scaling were carried out using the STARANISO and the UCLA-DOE lab—Diffraction anisotropy [ 60 ] web servers. Structure validation was performed using MolProbity [ 61 ]. Analysis of Nb-epitope interfaces, symmetry-based oligomers, and interacting residues was carried out using PDBsum [ 62 ] and PISA [ 63 ]. Figures of crystal structures were prepared in PyMOL (Schrödinger, LLC). Crystal diffraction and refinement statistics are shown in S1 Table . The refined coordinates and structure factors were deposited at the Protein Data Bank with entry codes 7R20 (Nb E5), 7R24 (rArc-CTD complex), 7R23 (extended hArc-CTD complex), and 7R1Z (collapsed hArc-CTD complex).

Protein crystallisation was carried out in TTP3 sitting drop 96-well crystallisation plates (SPT Labtech, Melbourn, UK) using a Mosquito LCP crystallisation robot (SPT Labtech, Melbourn, UK). For crystallisation of Nb complexes, Nb was mixed with the target protein in 1.2–1.5 fold molar excess and the complex isolated via SEC on a Superdex 75 Increase 10/300 GL or Superdex 200 Increase 10/300 GL (GE Healthcare) column in 20 mM Tris-HCl pH 7.5, 150 mM NaCl. Fractions corresponding to the complex were pooled and concentrated to 10–37 mg/mL in 10 kDa MWCO spin concentrators and centrifuged at 20,000 g and 4°C for 5–10 min before concentration determination via absorbance at 280 nm and set up of 96-well screening plates with 270–600 nL drops with varying protein:reservoir ratios (2:1, 1:1 and 1:2), sitting over a 70 μL reservoir of the precipitant solution. Plates were incubated at either 8°C or 20°C.

For negative-staining transmission electron microscopy, 300-mesh copper grids (#Cu-300, Electron Microscopy Sciences, Hatfield, PA, USA) with formvar (#15820, Electron Microscopy Sciences) and carbon coating (provided by MIC, Department of Biomedicine, University of Bergen) were glow discharged for 1 minute at 30 mA in a Dieno pico 100 vacuum chamber (Diener Electronics). 5 μL FLhArc from the Nb association assay were applied to grids and incubated for 2 min before excess sample was removed using Whatman ™ filter paper. Grids were then washed two times in 20 μL water droplets before they were stained with 2% uranyl acetate for 1 min, with excess solution removed with filter paper in between each wash and after staining. FLhArc without Nbs were used as control. Micrographs were obtained with a JEOL JEM-1230 electron microscope operated at 80kV. Images were recorded at 80,000 × and 150,000 × magnification.

Full-length hArc (FLhArc) was cloned into the pETZZ_1a vector as a His-ZZ fusion protein construct with a TEV protease cleavage site [ 19 ]. E. coli BL21 CodonPlus cells transformed with the pETZZ_1a-FLhArc construct were used to inoculate 50 mL LB broth supplemented with 34 μg/mL chloramphenicol and 50 μg/mL kanamycin and incubated overnight at 28°C and 200 rpm. The starter cultures were diluted into 1 L LB broth with antibiotics, and incubated at 37°C and 200 rpm, until an OD 600 of 0.6–0.8 was obtained. Expression was then induced by adding 0.5 mM IPTG and further incubated at 25°C and 200 rpm overnight. Cells were harvested at 6000 × g for 30 minutes and pellets stored at –20°C. Pellets were thawn on ice and resuspended in lysis buffer (5 mL/g pellet: 50 mM Na 2 HPO 4 /KH 2 PO 4 , 150 mM NaCl, 0.2% Tergitol ™ solution (Merck, Darmstadt, Germany), and 2 mM DTT, pH 8.0, containing 1 tablet cOmplete ™ with EDTA (Roche, Basel, Switzerland), 10 mM benzamidine and 0.2 mM PMSF) and homogenised using a Thomas pestle tissue grinder, before sonication on ice for 3 x 45 s at 20 W, with 45-s pauses. The soluble fraction was harvested by centrifugation at 14 000 × g at 4°C for 30 min. The soluble fraction was transferred to NiNTA agarose (Qiagen GmbH, Düsseldorf, Germany) equilibrated in 50 mM Na 2 HPO 4 /KH 2 PO 4 , 150 mM NaCl, and 2 mM DTT, pH 8.0, and incubated on rotation for 2 h at 4°C. The matrix was then washed with 50 mM Na 2 HPO 4 /KH 2 PO 4 , 150 mM NaCl, 0.2% Tergitol ™ (Merck, Darmstadt, Germany), and 2 mM DTT, pH 8.0, until A 280 = 0, before washing overnight with at least 30 bed volumes of 50 mM Na 2 HPO 4 /KH 2 PO 4 , 150 mM NaCl, and 2 mM DTT, pH 8.0. The matrix was then washed 10 bed volumes of 50 mM Na 2 HPO 4 /KH 2 PO 4 , 1 M NaCl, and 2 mM DTT, pH 8.0 followed by 10 bed volumes 50 mM Na 2 HPO 4 /KH 2 PO 4 , 150 mM NaCl, and 2 mM DTT, pH 8.0, before elution with the same buffer containing 300 mM imidazole. The protein was dialysed against 50 mM Na 2 HPO 4 /KH 2 PO 4 , 150 mM NaCl, and 0.5 mM TCEP, pH 7.4 to remove imidazole, before cleaving off the fusion protein with TEV protease. To remove the HisZZ-tag, the cutting reaction was added to Talon (Takara Bio Inc., Kusatsu, Shiga, Japan) and gently agitated for 1 h at 4°C, before elution and concentration at 2 000 × g at 4°C. Concentration was determined using the theoretical extinction coefficient (Abs 0.1%) 1.71. A 260/280 ratio >1 indicated nucleic acid content, and capsid formation was checked by negative-staining electron microscopy.

Frame selection and buffer subtraction were carried out in CHROMIXS [ 47 ], primary analysis in PRIMUS [ 48 ] and distance distribution function analysis in GNOM [ 49 ]. Ab initio models were created using DAMMIN [ 50 ] and GASBOR [ 51 ]. Oligomer models were built from individual subunits based on SAXS data in CORAL [ 52 ]. Theoretical scattering curves of coordinate files were calculated using CRYSOL [ 53 ] and scattering-based normal mode analysis of crystal structures carried out in SREFLEX [ 54 ].

SAXS data from MBP-2rNT, and FLrArc-7A in the presence and absence of nanobodies, were collected on the SWING beamline of Soleil synchrotron, Saint Aubin, France in HPLC mode [ 44 ]. SAXS data from hArc-CTD in the absence and presence of Nbs were collected on the BM29 beamline of European Synchrotron Radiation Facility, Grenoble, France in HPLC mode [ 45 , 46 ]. Please see technical details of SAXS data collection in S2 Protocol .

The thermodynamics and affinity of Nb or peptide binding to FLrArc-7A were measured on a MicroCal iTC200 instrument (Malvern Panalytical, Malvern, UK). Binding of NbArcs to FLrArc-7A was measured in 20 mM Tris-HCl pH 7.4, 150 mM NaCl with 3.5–5 μM FLrArc-7A in the cell and 30–50 μM anti-Arc Nb in the syringe at 20°C and a reference power of 10–12 μcal/s, with a stirring speed of 1000 rpm. In all cases, except for NbArc-E5, a single 0.5-μL priming injection was followed by 19 × 2-μL 4 s injections of Nb with a spacing of 120 s and a filter period of 5 s. For NbArc-E5 binding to FLrArc-7A, the priming injection was followed by 38 × 1-μL injections. In the case of stargazin (Stg) binding to FLrArc-7A, 2 mM of the Stg peptide (RIPSYRYR with N-terminal acetylation and C-terminal amidation) dissolved in the assay buffer was injected (1 × 0.5 and 19 × 2 μL) into 198 μM FLrArc-7A. For the Stg-Nb displacement assay, 150 μM of NbArc-H11 was injected (1 × 0.5 and 19 × 2 μL) into 10 μM FLrArc-7A and 120 μM Stg. Data processing (peak integration and dilution heat subtraction) was carried out in the Origin Lab software (v. 2021). Binding enthalpy (ΔH), association constant (K a ) and binding entropy (ΔS) were obtained by fitting the integrated thermograms with a 1:1 binding model.

For assessment of Nb binding and crude epitope mapping, protein pulldown assays were performed. 0.5 mg/mL of His-tagged hArc-CTD or His-MBP-tagged 2rNT were mixed with an equimolar amount of Nb and incubated on ice for 20–45 min. 200 μL of the solution were loaded on 100 μL of NiNTA agarose resin equilibrated in 20 mM Tris-HCl pH 7.4, 150 mM NaCl, 20 mM imidazole pH 7.5 and the mixture incubated at +4°C under gentle agitation for 1 h followed by centrifugation at 200 g and +4°C for 5 min. The supernatant (unbound protein) was removed, and the resin washed three times in the same buffer. To elute bound protein, the resin was incubated in the same buffer with 300 mM imidazole for 30 min before centrifugation as earlier. The fractions were analysed using SDS-PAGE.

2.2 nmol of FLrArc-7A were mixed with an equimolar amount of Nb, incubated on ice for around 30 minutes before the sample (100 μL) was injected on an Superdex 200 Increase 10/300 GL SEC column in 20 mM Tris, 150 mM NaCl (pH 7.5). 2.2 nmol of the unbound protein (apo) and Nbs alone were run as controls.

To assess the thermal stabilisation of Arc constructs upon Nb binding, DSF was utilised. Proteins were diluted to 0.5–2 mg/mL in the assay buffer (20 mM Tris-HCl pH 7.4 and 150 mM NaCl) and mixed with 100x SYPRO-orange (in 50% (v/v) DMSO/assay buffer) in 384 well PCR plates to a final concentration of 5x, making the final DMSO concentration in the assay 2.5% (v/v). The final reaction volume was 18 μL. Fluorescence emission at 610 nm, following excitation at 465 nm, was measured in a LightCycler 480 LC RT-PCR system (Roche, Basel, Switzerland) over the temperature range 20–95°C with a temperature ramp rate of 2.4°C/min. Thermal denaturation midpoints (T m ) were determined as the maximum of the first derivative of the melting curves.

For Nb production, BL21(DE3) competent E. coli cells were transformed with the expression clones. LB medium supplemented with 50 μg/mL kanamycin (10–20 mL) was inoculated with a single transformed colony, incubated overnight at 37°C and 170 rpm, diluted 100-fold into 0.5–2 L of the same medium and incubated at 37°C and 200 rpm. When OD600 reached 0.4–0.6, expression was induced via addition of 1 mM IPTG and maintained for 4 h at 30°C. Cells were harvested and pellets resuspended in 40 mM HEPES pH 7.5, 100 mM NaCl, 20 mM imidazole, 0.1 mg/mL hen egg white lysozyme. Cells were lysed via a single freeze-thaw cycle and sonication and soluble fraction harvested via centrifugation at 30000 g and 4°C for 30 min. NiNTA affinity purification and TEV proteolysis was carried out as above before a single negative NiNTA affinity purification step, concentration to 1–2 mL in 10 kDa MWCO spin concentrators and applied to either a HiLoad Superdex 75 pg 16/600 or a Superdex 75 Increase 10/300 GL SEC column (GE Healthcare, IL, USA) equilibrated in 20 mM Tris-HCl (pH 7.4) and 150 mM NaCl. Fractions deemed pure via SDS-PAGE were pooled, concentrated to 5–25 mg/mL, split into 50 μL aliquots, snap-frozen in liquid and stored at -80°C.

Anti-Arc Nbs were obtained commercially from NanoTag Technologies (Göttingen, Germany). Briefly, two alpacas were immunized a total of 6 times with the wild-type full length hArc and the FLrArc-7A mutant, gradually increasing the fraction of FLrArc-7A. After screening the obtained single-domain antibody (sdAb or nanobody) screening library using phage display and ELISA, sdAbs D4, C11, B5, H11, E5, and B12 were chosen as representatives of the four different sequence families and two unique binders and obtained as His 6 -TEV-Nb constructs in prokaryotic pNT1433 expression vectors.

FLrArc-7A is full-length rArc (residues 1–397) with residues 113–119 mutated to Ala to hinder capsid formation [ 18 ]. hArc-CTD contains residues 207–362. Not shown is the N-terminal 6xHis-tag and the linker containing the TEV site. 2rNT is the NTD of rArc (residues 24–137) with the poly-Ala mutation. The domain was insoluble in the absence of its fusion partner, maltose binding protein (MBP). FLhArc corresponds to FLrArc-7A, but with the human Arc sequence without any mutations.

Fig 1 outlines the Arc constructs. FLrArc-7A, full-length rArc containing a poly-Ala mutation (residues 113–119), was in the pHMGWA vector [ 18 ], resulting in an N-terminal His 6 -maltose binding protein (MBP) fusion [ 43 ]. E. coli BL21(DE3)-RIPL cells were used for FLrArc-7A expression. The human Arc CTD (hArc-CTD) (residues 206–361) was expressed and purified as described [ 35 ]. The protein was expressed with an N-terminal His 6 tag in E. coli BL21(DE3) cells using the pTH27 vector. The NTD of FLrArc-7A (2rNT) was expressed as an N-terminal His 6 -MBP fusion construct in pHMGWA using E. coli BL21(DE3)-RIPL cells. Details of recombinant Arc construct expression and purification are given in S1 Protocol .

Results

Anti-Arc nanobodies and their interaction with FLrArc-7A To aid in structural and functional studies on Arc, Nbs were raised against the dimeric poly-Ala rArc mutant (FLrArc-7A). The immunisation resulted in 92 positive Nb clones, and Nbs D4, H11, B5, C11, B12 and E5 were chosen for further characterisation, as they represented different sequence families within the collection of clones. The sequence alignment (Fig 2A) highlights the variable CDR loops and an extended CDR3 in NbArc-H11. DSF was used to estimate thermal stability of the Nbs (Fig 2B). No sequence-specific trend could be identified, but as the framework region (FR) is highly conserved, the varying thermal stability likely depended upon paratope loops, mainly CDR3. NbArc-H11 showed the highest T m despite having the longest CDR3, suggesting some folding of the CDR3 or its interaction with the FR surface. PPT PowerPoint slide

PNG larger image

TIFF original image Download: Fig 2. Anti-Arc nanobodies. A Sequence alignment reveals the location of the three CDRs, highlighting the long CDR3 of NbArc-H11. Sequences are coloured by conservation, and Cys24 and Cys92 (Kabat numbering) are highlighted in yellow. The alignment was produced using Clustal Omega [71] and JalView [72]. B Determination of NbArc T m using DSF (N = 3). The T m of NbArc-D4 could not be determined. C 1.42-Å crystal structure of NbArc-E5. Colouring of CDRs is the same as in panel A. D The conserved central disulphide bond, formed by oxidation Cys24 and Cys97, packs against Trp38 and is partially reduced. Electron density is shown as a blue mesh at 1.5σ. https://doi.org/10.1371/journal.pone.0269281.g002 To further characterise the Nbs in their free state, Nb E5 was crystallised, allowing for structure determination at 1.42-Å resolution (Fig 2C). The structure shows the typical immunoglobulin fold, an incomplete β-barrel of nine anti-parallel β-strands. The overall structure is compact, and the three CDR loops extend from the protein. Despite its elongated nature, the CDR3 loop is rigid in relation to the rest of the protein (S1A Fig). Hydrophobic residues of the CDR3 pack onto the exterior of the central β-barrel to shield non-polar side chains (S1B Fig), which might account for the high solubility of this Nb. Furthermore, the two tryptophans and the basic residues in the CDR3 loop suggest a negatively charged, buried epitope. The central disulphide (Cys24-Cys97) appeared mostly reduced, possibly due to radiation damage. Upon further inspection, it became clear that it was partially oxidised, and was refined to 20 and 80% occupancies for the oxidised and reduced rotamers of Cys97, respectively (Fig 2D).

SAXS analysis of FLrArc-7A Nb complexes SAXS was used to estimate the solution structure of the Nb complexes of FLrArc-7A, and to detect conformational changes in Arc upon Nb binding (Fig 5, Table 2). To measure scattering from only complexes and not unbound FLrArc-7A or Nbs, the complexes were separated from unbound FLrArc-7A using SEC and SAXS frames collected as the proteins eluted. The data indicate that in all cases, the Nb bound to the FLrArc-7A dimer in 2:2 stoichiometry, confirming the Arc:Nb stoichiometry of 1:1 from ITC. The Kratky plot (Fig 5C) indicated varying rigidity of the complexes, without significant rigidification upon binding. The distance distributions (Fig 5D) differed in all cases from that of unbound FLrArc-7A, showing an expansion in structure. R g increased in all cases, accompanied by small changes in D max (Table 2). The highest increase in R g , accompanied with an almost unchanged D max , was observed in complex with C11. This indicates binding further from the centre of mass of FLrArc-7A. An increase in D max was observed for E5 and H11, which indicated binding to the end of the long axis of the Arc dimer; the epitope for these Nbs must lie at the ends of the elongated dimer, further allowing to map domain location in the Arc dimer. PPT PowerPoint slide

PNG larger image

TIFF original image Download: Fig 5. SEC-SAXS analysis of FLrArc-7A Nb complexes. A SEC-SAXS profiles of the complexes and unbound (Apo) FLrArc-7A. The estimated R g of the frames used for processing is shown in open circles. All complexes were measured in 20 mM HEPES, 150 mM NaCl, 0.5 mM TCEP, pH 7.5. rArc+H11 was measured on a separate occasion in 20 mM Tris, 150 mM NaCl, pH 7.4, which caused differences in intensity and the elution volume of the main peak. B Scattering curves of rArc and rArc-Nb complexes, offset for easier visualisation. Data fits from GNOM are shown as grey lines. C Dimensionless Kratky plot. The maximum of an ideal rigid spherical particle ( , 1.104) is marked by X. D Distance distribution profiles. E Ab initio models of Nb-bound FLrArc-7A in various orientations, aligned with the model of the unbound protein (in transparent blue). Arrows indicate additional volumes in the complex models, which might correspond to bound Nb. Black arrows indicate two-fold symmetric binding, and grey arrows indicate additional volumes following no specific symmetry. Models were produced in DAMMIN with no forced symmetry, with χ2 values of 0.992 (apo), 1.063 (+H11), 1.090 (+C11), 1.106 (+E5), 1.287 (+D4), 1.090 (+B5) and 1.090 (+B12). https://doi.org/10.1371/journal.pone.0269281.g005 PPT PowerPoint slide

PNG larger image

TIFF original image Download: Table 2. Parameters derived from SAXS of FLrArc-7A Nb complexes. https://doi.org/10.1371/journal.pone.0269281.t002 Ab initio dummy atom models are compared in Fig 5E. Changes were observed in the calculated molecular envelope, and with NbArc-H11, -C11 and -E5 bound, additional volumes indicated the presence Nbs bound to each monomer. The changes in the overall shape of the protein may be related to reduced fluctuations in the disordered regions of FLrArc-7A upon Nb binding. The SAXS experiments demonstrated that all Nbs bound the FLrArc-7A dimer in 2:2 stoichiometry, but the low-resolution nature of SAXS and the low molar mass of the Nbs compared to dimeric Arc did not allow for accurate epitope mapping.

Anti-Arc nanobodies associate with high-molecular-weight hArc without promoting disassembly Considering the reported ability of some Nbs raised against viral proteins to associate with capsids and facilitate deactivation [74,75], we investigated whether the same applied for the anti-Arc Nbs. The association of the Nbs to hArc capsid-containing fractions was assessed using analytical SEC (Fig 6). Although a minor peak of a similar elution volume as Nb-bound, dimeric FLrArc-7A, appeared upon addition of the Nbs, most co-eluted with the high-molecular-weight forms of hArc in the void volume (Fig 6B). Furthermore, virus-like capsid structures, like those present in hArc without Nbs, were apparent in negative staining TEM images in the presence of Nb H11 (Fig 6C and 6D). The dimer peak, not present in the absence of Nbs, was most prominent with H11, suggesting some degree of Nb-induced capsid dissociation. The Nbs could be associated with the capsids, or other high-MW forms, but did not induce conformational changes that would directly lead to capsid disassembly. PPT PowerPoint slide

PNG larger image

TIFF original image Download: Fig 6. Rudimental analysis of Nb association with hArc capsid fraction. A Analytical SEC of capsid hArc-Nb complexes. 2 nmol of Arc capsids were mixed with a 1.5-fold molar excess of Nbs and run on a Superdex 200 Increase 10/300 GL in PBS buffer. Noted is the presumably dimeric peak present in some Nb complexes, with an elution volume corresponding to that of the FLrArc-7A complexes. NbArc-E5 was accidentally mixed with 4 nmol of hArc, resulting in the large void peak and lack of Nb excess peak. B SDS-PAGE analysis of SEC fractions, demonstrating that Nbs co-elute with the capsids. C Negative staining TEM images of capsid hArc at 150,000x magnification with NbArc-H11 (left) and without Nb (right). Arrows point at capsids, and a magnified view of selected capsids is shown. https://doi.org/10.1371/journal.pone.0269281.g006

Crystal structure of an rArc-CTD ternary nanobody complex We attempted the crystallisation of the FLrArc-7A mutant, using the six anti-Arc Nbs as crystallisation chaperones. However, despite extensive efforts, none of the complexes with one Nb grew crystals of sufficient quality for diffraction data collection. Therefore, crystallisation with more than one Nb was attempted. Analytical SEC showed the E5+C11, H11+C11, H11+B5 and E5+B5 Nb pairs to be compatible, whereas in other cases, binding of the second Nb caused dissociation of the first (S3 Fig). Crystallisation of FLrArc-7A was attempted in complex with both E5/C11 and H11/C11, as these individually led to the most thermal stabilisation (Fig 4, Table 1). FLrArc-7A in complex with H11 and C11 showed crystal growth after extended incubation. The FLrArc-7A+H11+C11 crystals diffracted to 2.7-Å resolution, but the asymmetric unit was too small to contain full-length Arc and both Nbs. Indeed, the crystals contained the rArc CTD in complex with both Nbs, suggesting to proteolytic cleavage of the linker region. Residues 211–356 of rArc, corresponding to the whole Arc-CTD, could be built, while the NTD was not detected in electron density. Nbs H11 and C11 could be fully built, apart from the first two N-terminal and the last C-terminal residues (Fig 7). The ternary complex contains the rArc CTD, i.e. the N-lobe, a 4-helix orthogonal bundle, and the C-lobe, a 5-helix orthogonal bundle, connected by a rigid central α-helix (Fig 7A). Nbs H11 and C11 bound to distinct, acidic epitopes on the N-lobe and the C-lobe, respectively. Accordingly, the electrostatic surface of the CDRs of both Nbs is basic (Fig 7B). A large majority of crystal contacts were formed exclusively by the Nbs (panel A in S4 Fig). PPT PowerPoint slide

PNG larger image

TIFF original image Download: Fig 7. The crystal structure of rArc-CTD in complex with NbArc-H11 and -C11. A The overall structure. The CTD of rArc is shown in blue to red from the N- to the C-terminus. NbArc-H11 is shown in orange and NbArc-C11 in purple. CDR 1, 2, and 3 are coloured blue, green, and red, respectively. B Both Nbs bind acidic surface patches of the CTD (above); the surface of the Nb CDRs is positive (below). Electrostatic potential was calculated using ABPS [76]. The epitopes correspond to the ligand peptide binding site of the N-lobe (H11) and the C-lobe oligomerisation surface (C11). C Pro103 and Tyr105 of H11 extend into the Arc N-lobe hydrophobic pocket. D C11 binding of the rArc-CTD. Polar contacts (yellow) and π-π interactions (red) are indicated with dashes. https://doi.org/10.1371/journal.pone.0269281.g007 The elongated CDR3 loop of H11 protrudes from the β-barrel fold and extends into the N-lobe of the CTD. The binding buries a large area on the CTD, 1044 Å2, and the first helix of the N-lobe is engulfed in a crevice between CDR loops 1 and 3 of H11 (Fig 7C). CDR3 accounts for most contacts, and Pro103 and Tyr105 of CDR3 show CH…π and π-π interactions with aromatic residues within the pocket of the N-lobe. This groove of the N-lobe is the ligand peptide binding site of Arc [26,31,34], indicating that Nb H11 could be used to modulate protein-protein interactions of the Arc N-lobe. Additionally, H11 showed interactions with the N-terminal tail of the N-lobe, suggesting that it might affect the conformation of the long linker connecting the NTD and CTD in full-length Arc. NbArc-C11 bound a solvent-exposed epitope in the C-lobe, burying 738 Å2 of Arc surface (Fig 7D). The interacting area is typical for antibody-antigen interactions [77]. The binding site overlaps with the suggested oligomerisation surface of the C-lobe [42], indicating that C11 could be used to modify Arc interactions and oligomeric state.

Crystallisation of human Arc CTD Limited structural information has been available on hArc, apart from the individual N- and C-lobes [34]. Taking advantage of NbArc-H11 and -C11, crystallisation of hArc-CTD was carried out. The structure of the hArc-Nb ternary complex was refined to 2.77-Å resolution and contained the CTD of hArc (residues 209–355) in complex with both Nbs (panel A in S5 Fig). The two lobe domains are almost identical to rArc (panel B in S5 Fig), with aligned Cα-RMSD of 0.65 Å for both lobes. However, the central helix connecting the two lobes breaks in hArc at residues 275–277, resulting in an altered relative orientation of the lobes. Compared to rArc, the hArc complex had altered crystal packing, and the Nbs play an even larger role in crystal contacts (panel B in S4 Fig). The binding mode of H11 to hArc was similar to that observed in the rArc structure, burying 1073 Å2 on the CTD. However, due to the bending of the central helix in hArc-CTD, the binding mode by C11 was slightly altered, and the buried surface area of hArc was increased to 848 Å2. This was achieved by additional interactions of C11 with both the central helix and the N-lobe, suggesting that C11 might prefer a slightly bent conformer of the CTD by interacting with both lobes.

Exploring the dynamic nature of the Arc-CTD central helix hinge region The crystal structures of the human and rat Arc CTD showed almost identical folds, except in the conformation of the central helix. In the rArc structure, the helix is straight (Fig 7A), whereas in hArc, it is kinked (panel A in S5 Fig), resulting in dislocation of the C-lobe in relation to the N-lobe. Upon comparison with the NMR model of the CTD from rArc [31], this conformational variation becomes even more apparent (Fig 9A). Breaking of the helix at Gly277 results in a conformational shift, corresponding to a 3.7 Å translation of the C-terminal end of the helix in the hArc crystal and 14.4 Å in the NMR model (Fig 9B). The dynamics of the central helix might have a role in the function of Arc, and dynamic movements of the CTD could differ between monomeric/dimeric Arc, with potential roles in synaptic plasticity, and the capsid form, which facilitates intercellular mRNA transport [11]. The corresponding helix in both dArc1 and dArc2 breaks, as the two lobes of CTD pack against each other in the capsid protomers [40]. To further explore this hypothesis, SAXS and MD simulations were carried out. PPT PowerPoint slide

PNG larger image

TIFF original image Download: Fig 9. The Arc-CTD is dynamic in structure. A Comparison of the Arc-CTD crystal structures with the NMR model of rArc-CTD (6GSE, [31]) reveals a hinge in the central helix. B Superimposition of the N-terminal end of the helix highlights a hinge at Gly277. C Analysis of the solution structure of the Arc-CTD using SAXS. Shown are the experimental scattering curve of hArc-CTD, theoretical curves for the crystal structure and the NMR model (6GSE, [31]), and the fit from volume fraction. D Central helix conformation and results of volume fraction analysis (left). Normal mode analysis (right) of the hArc-CTD crystal structure against the SAXS data gave an average solution conformer of the central helix. The N- and C-lobes are not shown in panels C and D for clarity. https://doi.org/10.1371/journal.pone.0269281.g009 To estimate the conformational flexibility of the CTD in solution, SAXS was used (Fig 9C and 9D). Fitting of theoretical scattering curves from the hArc-CTD crystal structure and the NMR model of the rArc-CTD against SAXS data from hArc-CTD [35] showed that neither seemed fully representative of the solution structure. Simultaneous fitting of both conformers gave fractions of around 80% and 20% for the crystal and NMR conformations, respectively. Hence, the mean solution conformer was close to the crystal structure, but not identical. Normal mode analysis of the hArc-CTD crystal structure against the SAXS data gave an average conformer that showed a slightly kinked helix. Thus, considerable conformational flexibility is present in the dynamic CTD in solution. To obtain a more detailed insight into conformational plasticity of the CTD, the hArc-CTD crystal structure, stripped of bound Nbs, was subjected to a 640-ns atomistic MD simulation (S7 Fig). In the simulation, the breaking of the central helix at Gly277 facilitates extension of the central hinge, and after rotation, the two lobes pack onto each other, resulting in collapse after only 50 ns. This collapsed conformation seemed more stable than the extended conformer, as little conformational change was observed for the remainder of the simulation. Therefore, the simulation suggested the CTD to be dynamic in structure and implied that hArc-CTD can fold into conformers similar to the capsid protomers of dArc1 and dArc2. Thr278 of the Arc-CTD undergoes phosphorylation by TNIK kinase in Neuro2A cells, and a phosphomimicking mutation reduces the ability of recombinant mouse Arc to form capsids [78]. Furthermore, the G277D mutation was suggested to reduce the ability of Arc to oligomerise [42]. As both mutations are located in the dynamic hinge region of the CTD central helix, MD simulations were run for G277D and T278E, to examine if altered dynamics would be observed (S7 Fig). In G277D, the helix did break but full compaction was restricted by the negative charge on Asp277. In the T278E phosphomimic, Glu278 formed a salt-bridge with the Arg281 to restrict extension of the hinge region, inhibiting the collapse of the domain. As these mutations reduce Arc to oligomerisation, the simulations suggest a role for the CTD structural dynamics in Arc capsid assembly.

Crystal structure of a collapsed Arc-CTD Considering the dynamic nature of the Arc-CTD, crystallisation of the protein in different conformers was attempted. Optimisation of previously neglected crystallisation conditions for hArc-CTD in complex with NbArc-H11 and -C11 and matrix microseeding produced crystals for structure determination, allowing for structure determination at 1.94-Å resolution (Fig 10 and S8 Fig). PPT PowerPoint slide

PNG larger image

TIFF original image Download: Fig 10. Structure of the collapsed CTD of hArc. A The overall structure of the dimeric hArc-CTD+H11+C11 complex, viewed down the dimer symmetry axis. The two CTDs are shown in rainbow, H11 in orange, and C11 in purple. CDR1, 2, and 3 are coloured blue, green, and red, respectively. B A monomer, highlighting the collapse of the CTD and lack of the last two helices of the C-lobe. Shown are the sidechains with interactions on the interface. Gly277, at the central helix hinge, is shown in magenta. C Dimerisation results in rearrangement of charges, resulting in one highly acidic surface. D The CTD dimer interface (stereo view). The residues that form the interface are shown, including the inter-monomer salt bridge between Glu319 and Arg306. https://doi.org/10.1371/journal.pone.0269281.g010 The complex crystallised in a different unit cell, when compared to the other structures of the rat and human Arc CTD, and the structure of the complex was different (Fig 10A). In this conformer, breaking of the central helix at Gly277 leads to a complete collapse of the CTD and packing of the C-lobe against the N-lobe (Fig 10B). The interface between the lobes remained partially hydrated and was mainly characterised by indirect hydrogen bonding or salt bridges. Moreover, the CTD appeared dimeric. Using PISA, the calculated free energy gain upon dimerisation (ΔGint) of -54.4 kJ mol-1 indicated a stable dimeric assembly. The putative dimer interface (Fig 10D), which buries 600 Å2 of surface area from each monomer, formed between the C-lobes and involves 10 residues of each subunit. Residues 212–330 of the CTD could be built into the electron density; the last two helices were missing. Therefore, the dimer interface formed between incomplete C-lobes, fusing the exposed hydrophobic cores of the two protomers. Accordingly, the dimer interface is dominated by non-polar interactions. What led to the absence of the C-terminal end of the CTD, and dimerisation, is unresolved. Either the missing C-terminal portion was lost due to degradation, or it was not observed due to flexibility. Large empty spaces were observed in the crystal lattice in proximity of the C-termini (panel C in S8 Fig). The observed dimerisation led to a rearrangement of the electrostatic potential of the CTD, resulting in one side of the dimer being highly acidic (Fig 10C). This might facilitate interactions with cationic components, such as the Arc NTD. The unique CTD conformer observed in the crystal resembles that observed in MD simulations and could be representative of the hArc capsid protomer, as the conformation closely resembles a variety of capsid protomers, including dArc1 and dArc2 [40]. With the ability of Nbs to select for unique protein conformers, it was surprising that both NbArc-H11 and -C11 bound to both the extended and collapsed conformer (Fig 10A). In particular, a portion of the C11 epitope, present on the last two helices of the C-lobe, was absent in the structure, and the surface area C11 buries from the C-lobe was reduced to 658 Å2. This was compensated by interactions of the FR of C11 with the N-lobe of the CTD, not observed in the other crystal structures, with an additional interacting surface of 271 Å2. Whether these additional interactions are representative of binding interactions in solution is uncertain. In addition, the binding of H11 did not restrict the conformational shift. Upon the collapse of the CTD, the second loop of the C-lobe (the ⍺5-⍺6 loop of the CTD) comes into proximity of the CDR2 loop of H11, showing modest interactions but no steric hindrance. Otherwise, the overall binding mode of H11 was nearly identical to what was observed in the other crystal structures.

[END]

[1] Url: https://journals.plos.org/plosone/article?id=10.1371/journal.pone.0269281

(C) Plos One. "Accelerating the publication of peer-reviewed science."
Licensed under Creative Commons Attribution (CC BY 4.0)
URL: https://creativecommons.org/licenses/by/4.0/


via Magical.Fish Gopher News Feeds:
gopher://magical.fish/1/feeds/news/plosone/

You are viewing proxied material from magical.fish. The copyright of proxied material belongs to its original authors. Any comments or complaints in relation to proxied material should be directed to the original authors of the content concerned. Please see the disclaimer for more details.