Differential Interactome and Innate Immune Response Activation of Two Structurally Distinct Misfolded Protein Oligomers

Depósito de investigación de la Universidad de Sevilla https://idus.us.es/ Esta es la versión aceptada del artículo publicado en: This is an accepted manuscript of a paper published in: ACS Chemical Neuroscience: 2019 DOI: 10.1021/acschemneuro.9b00088. Copyright: © 2019 American Chemical Society El acceso a la versión publicada del artículo puede requerir la suscripción de la revista. Access to the published version may require subscription. “This article has been published in a revised form in [ACS Chemical Neuroscience] [ 10.1021/acschemneuro.9b00088. ]. This version is free to download for private research and study only. Not for redistribution, re-sale or use in derivative Works”. 1 Differential interactome and innate immune response activation of two structurally distinct misfolded protein oligomers. Benedetta Mannini1, Giulia Vecchi1, Adahir Labrador-Garrido2,3, Bertrand Fabre4, Kathryn Lilley4, David Pozo2,3*, Michele Vendruscolo1, Fabrizio Chiti5, Christopher M. Dobson1*, Cintia Roodveldt2,3* 1Centre for Misfolding Diseases, Department of Chemistry, University of Cambridge, Cambridge, UK 2Andalusian Center for Molecular Biology and Regenerative Medicine (CABIMER), Universidad de Sevilla-Consejo Superior de Investigaciones Científicas (CSIC) Universidad Pablo de Olavide, Seville, Spain 3Department of Medical Biochemistry, Molecular Biology and Immunology, Universidad de Sevilla, Seville, Spain 4Cambridge Centre for Proteomics, Systems Biology Centre, Department of Biochemistry, University of Cambridge, Cambridge, UK 5Department of Biomedical Experimental and Clinical Sciences, Section of Biochemistry, University of Florence, Florence, Italy ABSTRACT: The formation of misfolded protein oligomers during early stages of amyloid aggregation and the activation of the inflammatory response are two characteristic events associated with the onset and progression of a wide range of neurodegenerative diseases. Although it has been established that misfolded oligomers are involved in the neuroinflammatory process, the links between their structural and physicochemical features and their functional effects on the immune response remain largely unknown. To explore such links, we have taken advantage of two structurally distinct and well-characterized types of soluble oligomers of the protein HypF-N, denoted type A and B, which have been shown to have different surface-exposed hydrophobicity and cytotoxicity. We analysed the microglial inflammatory response triggered by these oligomers and found that, even though they both stimulate the release of pro-inflammatory cytokines, type B oligomers induced a substantially greater response than type A oligomers. By using a combination of confocal microscopy, protein pull-down assays and high-throughput mass spectrometric-based proteomic analysis, we found that the two types of oligomers interact differentially with microglial cells. In particular, despite the fact that both types of oligomers were found to bind to a common pool of microglial proteins, type B relative to type A oligomers showed enhanced protein binding, a finding that correlated with the observed inflammatory response. In addition, the receptors and signalling molecules in the interactome associated with inflammatory-mediated neurodegeneration revealed a number of previously unidentified proteins likely to be involved in the oligomer-elicited innate immune response. INTRODUCTION The misfolding and aggregation of proteins into amyloid fibrils is the hallmark of a wide range of human neurodegenerative disorders, including Alzheimer's disease (AD), Parkinson's disease (PD), Hungtington’s disease (HD) and amyotrophic lateral sclerosis (ALS) 1. Increasing evidence suggests that the soluble oligomeric species preceding fibril formation, or released by mature fibrils, are believed to be responsible for generating cellular dysfunction through a variety of mechanisms 2 3 4 5 6 7 8. Thus, a comprehensive understanding of such mechanisms is vital for designing effective approaches for diagnostic and therapeutic interventions 9 10 11. The chronic neuroinflammatory state resulting from the massive microglial activation that develops in neurodegenerative diseases including PD 12, AD 13, HD 14 and ALS (Roodveldt et al., Trends Mol. Med. in press), is known to be a central feature in the development of these pathologies 15. It has become established that neurodegeneration can progress in a non-neuronal autonomous manner in which glial cells actively contribute to neuronal death by generating and sustaining neuroinflammation 16 15. Oligomers formed during the aggregation of a series of different proteins have generally been found to share a range of physicochemical characteristics, including a significant content of β-sheet structure, a small size (compared to amyloid fibrils), a high exposure of hydrophobic surfaces to the solvent and a strong reactivity to conformation-specific antibodies 17. Such physicochemical properties have also been found to be determinants of the cellular toxicity of the oligomers 18 19 20 2 21 22. However, the impact of the physicochemical features of such oligomers on microglial responses in the context of misfolding diseases is still unknown. In this context, it has been suggested that the main trigger of innate immunity is the aberrant exposure of hydrophobic surfaces in certain biological molecules and assemblies 23 24, and such exposed regions are likely to be characteristic of virtually all misfolded protein aggregates 25 19 2 21 22. It remains to be established, however, how the structural, physicochemical and biological features of these potentially pathogenic species are correlated with the stimulation of an innate immune response. The aggregates formed by misfolded and amyloidogenic proteins have been recognized as potential triggers of inflammation in the context of neurodegenerative disorders 14 16. A number of disease-linked protein oligomers present in the extracellular medium have been shown to elicit strong inflammatory responses in microglia, which are the principal 2 immunocompetent cells in the central nervous system (CNS), by acting as potent danger-associated molecular patterns (DAMPs) and strongly activating innate immunity 26 27 28 29. In the present study we investigated the inflammatory response triggered by misfolded oligomers of the N-terminal domain of the HypF protein from E. coli (HypF-N). This protein is a convenient model system for exploring the link between neurodegeneration and protein aggregation for a number of reasons. First, under conditions that destabilize its native structure it forms oligomers and amyloid fibrils in vitro that are closely similar to those observed for proteins associated with human diseases 2 30 31 32 33. Second, HypF-N oligomers formed in vitro impair cellular viability in a similar manner to protein oligomers associated with many amyloid diseases, both when added to the extracellular medium of cultured cells 34 35 36 37 38 21 39 and when administered directly into the brain in experiments with animal models 40 39. Third, HypF-N oligomers co-localise with protein dense regions in the postsynaptic membrane of cultured neurons, inhibit long-term potentiation in hippocampal slices and induce cognitive impairment upon administration into rat brains, all of which are recognised biochemical, electrophysiological and biological effects of the Ab42 oligomers associated with AD 41. Fourth, HypF-N oligomers are stable and maintain their morphological and structural properties under a wide range of experimental conditions, allowing detailed structural, biological and functional studies to be carried out 2 37 21 41 39. Finally, and of particular importance in the context of the present study, the nature and cytotoxicity of the different types of HypF-N oligomers is clearly different: indeed, those designated as type A are cytotoxic toward neuroblastoma cells and primary neurons, and in animal models, while type B do not elicit such effects 2 39. Despite their different cytotoxic properties, these two types of oligomers share a similar size and morphology when observed by atomic force microscopy 2. In fact, both oligomer types are roughly spherical, with diameters of 2-6 nm, and have similar thioflavin T (ThT) binding properties that are indicative of prefibrillar, amyloidogenic structures 2. Characterization of these species at a molecular level, however, revealed that the toxic type A oligomers have a higher degree of surface hydrophobicity than the benign type B oligomers, indicating a potential causative link between such structural characteristics and cytotoxicity 2 21 22. In the present study, by examining the effects of the these well-characterized and robust HypF-N oligomers on microglial cells, we report that the hydrophobicity and degree of interaction of the oligomers with the microglial proteome are highly correlated with the elicited immune response. Moreover, we identify the immune interactome for both oligomer types, which enables the identification of key proteins responsible for aggregate-elicited microglial neuroinflammation. RESULTS Low concentrations of oligomers do not affect cell viability but trigger inflammation. We first tested the ability of type A and type B oligomers to stimulate N13 microglial cells in culture. For this purpose, we searched for conditions in which neither type of oligomers affected cellular viability, so that we could enable a direct comparison of the inflammatory responses triggered by the two types of aggregates. We carried out a MTT reduction test, a standard assay based on mitochondrial metabolic integrity as a readout of cell viability, on N13 cells treated for 24 h with different concentrations of type A or type B HypF- N oligomers (Figure 1). As previously shown with a variety of cell lines, including human neuroblastoma (SH-SY5Y), mouse endothelial (Hend), Chinese hamster ovary (CHO), and human embryonic kidney 293 (HEK293) cells 2 21 38 42 43, at a concentration of 12 µM (monomer equivalents), the two types of oligomers showed a marked difference in cytotoxicity, with type A aggregates causing a 30% decrease in cell viability and type B oligomers appearing essentially innocuous (Figure 1). In N13 cells the cytotoxic effect of type A oligomers decreased in a dose-dependent manner and at concentrations below 4.3 µM did not detectably alter cell viability, whereas type B oligomers were found to be innocuous at all concentrations tested (Figure 1). Figure 1. Viability of N13 microglia cells after treatment with type A and type B oligomers. The degree of MTT reduction for murine N13 microglial cells treated for 24 h with different concentrations (in monomer equivalents) of type A (red) and type B (blue) oligomers added to the extracellular medium. The error bars correspond to standard errors of the means (SEM) of four independent experiments (n=4). Single, double, and triple asterisks (*) refer to p<0.05, p<0.01, and p<0.001, respectively, with respect to untreated cells. We next determined by ELISA assays the levels of a set of key disease-linked immune mediators, namely the proinflammatory IL-1β, IL-6, and TNFα cytokines, and the antiinflammatory IL-10 cytokine, after incubation of N13 cells with type A or B oligomers for 6 h or 24 h (Figure 2). As a positive control, we used bacterial lipopolysaccharide (LPS) as a conventional stimulus to assess the innate immune response by N13 cells where all four cytokines tested have been shown to be increased upon LPS engagement 44. Compared to untreated cells, the release of IL-1β elicited by the type B oligomers was found to be significantly higher at all concentrations tested, the lowest being 0.005 µM (monomer equivalents), whereas in the case of type A aggregates it was significantly higher only when present at the highest concentrations tested in this study, above 8.6 µM (monomer equivalents). Both types of oligomers induced the release of IL-6 at concentrations above 0.05 µM and 0.02 µM (monomer equivalents), for type A and B, respectively. The amount of TNFα detected for cells treated with type A and type B oligomers was significantly higher than that for untreated cells, with effects still detectable at concentrations of 0.1 µM and of 0.005 µM (monomer equivalents), respectively. Stimulation of cells with native non-aggregated HypF-N protein at 0.125 µM, 0.25 µM, 2 µM and 8 µM resulted in the secretion of two cytokines tested in this study, IL-6 and TNFα at levels indistinguishable from untreated cells (Supplementary Figure 1). We found, in addition, that treatment with either type of oligomer at concentrations below 12 µM did not stimulate the release of the anti-inflammatory cytokine IL-10; indeed, the levels of IL-10 detected in the presence of both type A and B 3 oligomers were lower than those found in the untreated cells (Figure 2). Only at oligomer concentrations of 12 µM (monomer equivalents), the highest tested here, were the IL-10 levels significantly higher than those of untreated cells. Three conclusions can be drawn from these data. First, the levels of the released pro-inflammatory mediators IL-1β, IL-6 and TNFα, and the lack of stimulation of IL-10 secretion, except at very high concentrations, indicate that both types of oligomers elicit a significant pro-inflammatory response but not an immunoregulatory one. Second, the concentration of HypF- N oligomers able to trigger an inflammatory response is about three orders of magnitude lower than that needed to cause cytotoxicity, suggesting that the inflammatory response could be particularly important during the early events of the protein aggregation process, when the concentration of misfolded protein oligomers is still very low. Third, the two types of oligomers have a different ability to induce the release of cytokines, as we will discuss in detail below. Figure 2. Cytokine release profile of N13 microglia cells after treatment with type A and type B oligomers. The release of IL- 1β, IL-6, TNFα and IL-10 was measured by ELISA assays in the supernatants of cultured N13 cells treated with different concentrations of type A and type B oligomers (in monomer equivalents). The culture medium with and without addition of lipopolysaccharide (LPS) was used as a negative and positive control, respectively. The error bars correspond to standard errors of the mean (SEM) of 4 independent experiments (n=4). The shaded area in yellow indicates the range between 0.005 and 3.6 µM, where the oligomers are unable to cause cell death as assessed by the MTT reduction viability test. Single, double, and triple asterisks (*) refer to p<0.05, p<0.01, and p<0.001, respectively, with respect to cells treated with the other type of oligomers at the same concentration. Single, double, and triple points ( • ) refer to p<0.05, p<0.01, and p<0.001, respectively, relative to untreated cells. The two types of oligomers display different abilities to elicit a pro-inflammatory response. The comparison between the release of cytokines mediated by the two types of oligomers showed that type B are more potent than type A in triggering the release at low concentrations of all three pro-inflammatory mediators tested. The levels of IL-1β and TNFα were significantly higher when microglia cells were treated with type B than with type A oligomers, from a concentration of 0.005 µM, the lowest tested here, up to 2.9 µM and 0.2 µM, respectively; the release of IL-6 was significantly higher for type B oligomers at almost all the concentrations where an effect was observed, i.e. at concentrations above 0.05 µM. For clarity, the ranges in which type B oligomers were found to be more efficient respect to type A oligomers have been highlighted in Figure 2 with a dark yellow shaded area. At the protein concentrations in which the viability of cells was observed to be affected by the type A oligomers, the levels of the two cytokines, IL-1β and TNFα, are similar for both types of oligomers, although at the highest protein concentrations, type A oligomers were more potent. In conclusion, these sets of data show that at low concentrations of oligomers, at which the cell viability is not detectably compromised (light yellow shaded area in Figure 2), type B oligomers are stronger inducers of the inflammatory response than type A oligomers. Figure 3. Interaction of type A and B oligomers with the surfaces of N13 cells. (A-G) Representative confocal microscopy images showing the interaction of the oligomers with the cultured N13 cells. Cells were incubated for 24 h with type A (A, B) or B (C, D) oligomers and then washed. The samples were then permeabilized by using 0.5% (v/v) Triton X-100 (A, C) or not permeabilized (B, D). The red fluorescence arises from the immunolabelling of the oligomers and the blue indicates the nuclei stained with DAPI. In (E) and (F) enlarged images from (A) and (C), respectively, are shown. (G) Untreated cells. (H) Quantification of the fluorescence emission arising from the oligomers normalized to the number of cells labelled with Alexa Fluor 633-conju- gated secondary antibodies. The error bars correspond to standard deviations of the mean of three different fields. Triple asterisks refer to p<0.001 relative to permeabilized cells. 4 Figure 4. Interaction of type A and B oligomers with the protein components of N13 cells. (A) Pull-down assay using type A and type B oligomers incubated for 1 h in isolation or in combination with microglial N13 cell lysates. (B) Log10 abundances (iBAQ values) of proteins binding to type B oligomers compared to type A oligomers as obtained by proteomic analysis. Each blue point is a protein detected and quantified. Points within the dashed line have been detected in pull-downs with only one of the two oligomers, and have been assigned with a default value of 1 in order to allow their graphical representation. The points on the left side of the plot are proteins detected to bind only to the type B oligomers, while points on the lower edge of the plot are proteins detected to bind only to the type A oligomers. Type A and B oligomers exhibit differential penetration profiles across the cell membrane. To shed light on the molecular mechanisms responsible for the differential immunogenic activity of the oligomers, we investigated their ability to interact with cell surfaces and their propensity to become internalized within cells. First, we evaluated the ability of the oligomers to penetrate the membranes of N13 cells by means of immunofluorescence (IF) and confocal microscopy analysis (Figure 3A-G). Cultured cells were incubated in the presence of either 0.7 µM (monomer equivalents) of type A or type B oligomers added to the extracellular medium for 24 h, washed thoroughly, and then either permeabilized or left non-permeabilized before labelling with anti-HypF-N primary antibodies and subsequently with Alexa Fluor 633-conjugated secondary antibodies (red fluorescence). Cell nuclei were stained with 4’,6-dia- midino-2-phenylindole (DAPI, blue fluorescence). In non-per- meabilized cells, the antibodies are unable to enter the cells and hence cannot detect any oligomers located inside them; any red fluorescence signal must therefore arise from HypF-N aggregates outside the cells or attached to cell membranes. By contrast, in the permeabilized samples the antibody can penetrate the membrane and the red fluorescence can result from the oligomers inside or outside the cells 45. Numerous red fluorescent puncta are visible within the permeabilized cells treated with type A oligomers (Figure 3A and at higher magnification in Figure 3E), whereas such puncta are hardly detectable in the non-permeabilized cells (Figure 3B). These results suggest that the majority of the type A oligomers added to the cell media were localized within the cells. In the images obtained from cells incubated with type B oligomers, the numbers of red puncta visible in the permeabilized cells (Figure 3C and magnification in Figure 3F) and non-permeabilized cells (Figure 3D) are similar, indicating that the number of oligomers that had penetrated the cell membrane was very low and that the majority of oligomers were attached to the cell surfaces. These data are summarized in Figure 3H showing the quantification of the fluorescence signals derived from the oligomers. Type A and B oligomers bind to a common pool of membrane proteins in microglial cells but with different affinities. As a second part of this study, we assessed the ability of the two different types of oligomers to bind to the proteins present within N13 cells. To this aim, we performed pull-down experiments by incubating type A and type B oligomers for 1 h in isolation or in combination with proteins extracted from N13 cell lysates. During the protein extraction step, the membrane protein fraction that would otherwise be under-represented was enriched by using an appropriate protein lysis protocol to ensure that these proteins remained in their native conformations despite the absence of the lipid membrane (see Methods section for details). After incubation with the extracts, the oligomers were pulled down by centrifugation and the pelleted fractions were analyzed by SDS-PAGE (Figure 4A). The results indicate that type B oligomers have a generally higher ability than type A oligomers to bind to proteins, indicating a higher preference of type B oligomers for the protein component of the membrane. As a control, no proteins from N13 cells were found in the pellet fraction in lysates incubated in the absence of either type of oligomer (Figure 4A). We then subjected the protein fractions pulled-down with type A and type B oligomers to liquid chromatographytandem mass spectrometry (GeLC-MS/MS) followed by highthroughput proteomic analysis in order to assess and compare the N13 cell interactome of the two oligomers, i.e. the set of proteins bound in each case. The result of a representative experiment is shown in Figure 4B and indicates the abundance of the proteins bound to type B oligomers compared to type A oligomers. Analysis of the data supports the conclusion from the pull-down experiments, as the large majority (>90%) of the common binders are above the bisector line (Figure 4B), indicating a higher protein binding capability of type B compared to type A oligomers. It is also evident that only a small number of proteins have a single value of abundance (Figure 4B, horizontal and vertical points within the dashed blue lines), indicating that they were detected to bind to only one type of oligomer in that replica. In order to identify the proteins that could be key players in the microglial immune response to the oligomers, we selected for additional analyses only those proteins known to be located on cell membranes, by filtering the N13 cell interactomes of both types of oligomers on the basis of their subcellular locations reported in the UniProt database 46. Figure 5A shows the abundances of N13 membrane proteins found in the pull-down samples of type B compared to type A oligomers. Proteins detected to bind to only one type of oligomer (type A or type B) in a given replica are reported on the left and lower edge of each graph, within the dashed blue lines (horizontal or vertical points, respectively). The data reveal that the results for the membrane proteins alone are very similar to those obtained for all proteins (Figure 4B), as the majority of the shared points (Figure 5A) (>95% in total) are again above the bisector line. The results, therefore, reveal that type B oligomers have a higher ability than type A oligomers to bind to the set of membrane proteins of microglial cells (Figure 5A). They also show that type B oligomers have a higher ability than type A oligomers to bind to native cytosolic proteins, indicating that they have a higher general propensity to bind to both types of proteins (Supplementary Figure 2). 5 To shed light on the identity of the membrane proteins of N13 cells present in the interactomes obtained for the two types of oligomer, we combined the results of all the independent pull-down GeLC-MS/MS experiments (biological replicas) in each case. We then defined as ‘common binders’ those proteins detected in at least two biological replicas for each oligomer type, and as ‘specific binders’ those proteins detected at least twice for a single type of oligomer, but never for the other type. We found that 518 membrane proteins out of 573 (90.4% of all the membrane proteins detected in this study) are common binders of both types of oligomer and that the relative proportion of each protein is essentially the same in the pull-down assays for the two types of oligomer (Figure 5B). These results show, therefore, that the large majority of the membrane proteins interact with both types of oligomer (i.e. are common binders) and their relative abundances are independent of the type of oligomer to which they bind. Only a very small number of proteins (14) are found to be specific binders; of these, 12 are for type B oligomers (~2.1%) and 2 are for type A oligomers (~0.3%) (Figure 5B, left and bottom shaded area respectively). Nevertheless, even though type A and B oligomers bind to essentially the same pool of proteins (Figure 4B), albeit with significantly different affinities (Figure 5A), there are 14 proteins that appear to be specific binders. Figure 5. Interaction of type A and B oligomers with membrane proteins from N13 cells. (A) Replica-specific scatterplot of protein abundances (log10 iBAQ values) of N13 microglia cell membrane proteins bound to type B oligomers (y axis) vs those bound to type A oligomers (x axis) from the four biological replicas analysed in this study. The black points in each subplot represent membrane proteins detected and quantified in each replica. Proteins shown vertically on the left side and horizontally on the lower edge of each scatterplot within the dashed blue lines are membrane proteins binding to type B and type A oligomers, respectively. (B-C) Scatterplot of the normalised protein abundances comparing type A and type B oligomers. The abundance of each membrane protein detected in at least two replicas has been averaged and normalised over the biological replicas (see Materials and Methods). The points shown vertically and horizontally within the dashed blue lines in the left side and lower edge of the plot are proteins detected to interact with one type of oligomer but not the other. The red square points are membrane proteins known to be associated with the immune response. The errors are median absolute deviations corrected with the propagation law of errors for the z-score normalisation. (B) Scatterplot of the normalised abundances of membrane proteins binding to type B vs type A oligomers. Proteins on the left side of the plot within the dashed blue lines are type B ‘specific binders’, while proteins on the lower edge are type A ‘specific binders.’ (C) Scatterplot of the normalised abundances of the total set of membrane proteins in microglia, obtained from available MS data 47 vs those found to bind to type B oligomers. The interactome of HypF-N oligomers suggests key proteins in microglia immune signaling. Of the 518 membrane proteins found to bind to both types of oligomer, 77 proteins were identified from an extensive literature search to be related to the inflammatory response (red squares points in Figure 5B), hence defining the proteins that could potentially be responsible for the inflammatory response observed in N13 microglia cells. The list of these proteins, which we define as the immune interactome, is shown in Table 1. In order to identify the proteins involved in the immune response elicited in microglia, we compared the abundances of the membrane proteins found to interact with both types of oligomer in the whole oligomer interactome (see Materials and Methods for details), with the corresponding protein levels at which they are naturally present in microglial cells obtained from a recent MS proteomic study 47. In Figure 5C, we report the abundances of membrane proteins bound to type B oligomers, compared to the abundances of the proteins in microglia, normalised as z-scores 47 and filtered for membrane localisation. Data from the type B oligomer binders were used as representative of abundance rankings of the proteins in the oligomer interactome since abundance rankings are conserved between oligomer types (Figure 4B) and the abundance values are more accurate from the analysis of the type B oligomer samples, because of the higher intensities of the MS signal. We then sorted the proteins listed in Table 1 according to specific criteria to reveal a ‘top ranked binders’ selection: (i) proteins that are most abundant in the oligomer interactome, i.e. are in the top 10% of the abundance distribution; (ii) proteins binding to the oligomers that are found in the top 10% of the abundance distribution in the microglia and in the top 25% of the abundance distribution in the oligomer interactome, (iii) proteins that are observed to bind strongly to the oligomers despite their low abundance, i.e. those in the bottom 10% of the abundance protein distribution in microglia and at least average abundance in the oligomer interactome; (iv) proteins with the highest ratio of binding to type B relative to type A oligomers, 6 Table 1. The immune interactome. List of the 77 membrane proteins found to bind to both type A and type B oligomers and known to be related to the immune response. Protein_IDs Leading_ID Entry name Gene names Protein names Q922R8 Q922R8 PDIA6_MOUSE Pdia6 Txndc7 Protein disulfide-isomerase A6 (EC 5.3.4.1) P06800-3;P06800-2;P06800 P06800-3 PTPRC_MOUSE Ptprc Ly-5 Receptor-type tyrosine-protein phosphatase C (EC 3.1.3.48) O35658 O35658 C1QBP_MOUSE C1qbp Gc1qbp Complement component 1 Q subcomponent-binding protein Q61830 Q61830 MRC1_MOUSE Mrc1 Macrophage mannose receptor 1 (MMR) (CD antigen CD206) Q9JJX6-4;Q9JJX6-2;Q9JJX6-3;Q9JJX6 Q9JJX6-4 P2RX4_MOUSE P2rx4 P2x4 P2X purinoceptor 4 (P2X4) Q91ZX7 Q91ZX7 LRP1_MOUSE Lrp1 A2mr Prolow-density lipoprotein receptor-related protein 1 (LRP-1) (Alpha-2-macroglobulin receptor) (A2MR) (CD antigen CD91) Q61549 Q61549 AGRE1_MOUSE Adgre1 Emr1 Gpf480 Adhesion G protein-coupled receptor E1 (Cell surface glycoprotein F4/80) P17710-3;P17710-4;P17710-2;P17710 P17710-3 HXK1_MOUSE Hk1 Hexokinase-1 (EC 2.7.1.1) Q9Z0M6-2;Q9Z0M6-3;Q9Z0M6 Q9Z0M6-2 CD97_MOUSE Cd97 CD97 P07901 P07901 HS90A_MOUSE Hsp90aa1 Hsp86 Hsp86-1 Hspca Heat shock protein HSP 90-alpha (Heat shock 86 kDa) (HSP 86) (HSP86) O70172 O70172 PI42A_MOUSE Pip4k2a Pip5k2a Phosphatidylinositol 5-phosphate 4-kinase type-2 alpha (EC 2.7.1.149 P41251 P41251 NRAM1_MOUSE Slc11a1 Bcg Ity Lsh Nramp1 Natural resistance-associated macrophage protein 1 (NRAMP 1) Q9WTR1 Q9WTR1 TRPV2_MOUSE Trpv2 Grc Transient receptor potential cation channel subfamily V member 2 (TrpV2) P01899;P01897;P14430;P14429;P01896 ;P14431;P01898;P01902 P01899 HA11_MOUSE H2-D1 H-2 class I histocompatibility antigen D-B alpha chain (H-2D(B)) B2RXS4 B2RXS4 PLXB2_MOUSE Plxnb2 Plexin-B2 P21279;P21278 P21279 GNAQ_MOUSE Gnaq Guanine nucleotide-binding protein G(q) subunit alpha Q8R422 Q8R422 CD109_MOUSE Cd109 CD109 antigen (GPI-anchored alpha-2 macroglobulinrelated protein Q9QUN7 Q9QUN7 TLR2_MOUSE Tlr2 Toll-like receptor 2 (CD antigen CD282) Q922S4 Q922S4 PDE2A_MOUSE Pde2a cGMP-dependent 3',5'-cyclic phosphodiesterase (EC 3.1.4.17) P63017;P17156 P63017 HSP7C_MOUSE Hspa8 Hsc70 Hsc73 Heat shock cognate 71 kDa protein (Heat shock 70 kDa protein 8) Q61093;Q8CIZ9-2;Q8CIZ9;Q8CIZ9-3 Q61093 CY24B_MOUSE Cybb Cgd Cytochrome b-245 heavy chain P09581;O35942;Q6P5G0;Q62406;Q624 06-2;Q61532;Q78DX7 P09581 CSF1R_MOUSE Csf1r Csfmr Fms Macrophage colony-stimulating factor 1 receptor (CSF-1 receptor) Q6A0D4 Q6A0D4 RFTN1_MOUSE Rftn1 Kiaa0084 Raftlin (Raft-linking protein) Q99JB2 Q99JB2 STML2_MOUSE Stoml2 Slp2 Stomatin-like protein 2 Q61462;Q61462-2 Q61462 CY24A_MOUSE Cyba Cytochrome b-245 light chain P20491 P20491 FCERG_MOUSE Fcer1g Fce1g High affinity immunoglobulin epsilon receptor subunit gamma (Fc receptor gamma-chain) Q9D8Y0;Q9D4J1 Q9D8Y0 EFHD2_MOUSE Efhd2 Sws1 EF-hand domain-containing protein D2 (Swiprosin-1) P04441-2;P04441 P04441-2 HG2A_MOUSE Cd74 Ii H-2 class II histocompatibility antigen gamma chain (Ia antigen-associated invariant chain) (CD74) Q6P9J9;Q6P9J9-2 Q6P9J9 ANO6_MOUSE Ano6 Tmem16f Anoctamin-6 (Small-conductance calcium-activated nonselective cation channel) P28867;P28867-2;REV__Q6ZPL9 P28867 KPCD_MOUSE Prkcd Pkcd Protein kinase C delta type (EC 2.7.11.13) Q80WJ7 Q80WJ7 LYRIC_MOUSE Mtdh Lyric Protein LYRIC (3D3/LYRIC) Q8VDD8 Q8VDD8 WASH1_MOUSE Wash1 Orf19 Wash WAS protein family homolog 1 Q9D1D4;Q9D1D4-2 Q9D1D4 TMEDA_MOUS E Tmed10 Tmp21 Transmembrane emp24 domain-containing protein 10 P62821 P62821 RAB1A_MOUSE Rab1A Rab1 Ras-related protein Rab-1A O70472 O70472 TM131_MOUSE Tmem131 D1Bwg0491e Kiaa0257 Rw1 Transmembrane protein 131 P41233 P41233 ABCA1_MOUSE Abca1 Abc1 ATP-binding cassette sub-family A member 1 P41241;P41242-3;P41242;P41242-2 P41241 CSK_MOUSE Csk Tyrosine-protein kinase CSK (EC 2.7.10.2) P35991;Q60629;Q8CBF3- 2;Q03145;P29319;Q8CBF3;P54761;Q61 772-2;P54754;Q61772- 3;P54763;P54763- 2;Q61772;O09127;Q62413 P35991 BTK_MOUSE Btk Bpk Tyrosine-protein kinase BTK (EC 2.7.10.2) Q9CQW9 Q9CQW9 IFM3_MOUSE Ifitm3 Interferon-induced transmembrane protein 3 (Dispanin subfamily A member 2b) Q9JKF1;Q3UQ44 Q9JKF1 IQGA1_MOUSE Iqgap1 Ras GTPase-activating-like protein IQGAP1 Q9D3P8 Q9D3P8 PLRKT_MOUSE Plgrkt Plasminogen receptor (KT) (Plg-R(KT)) 7 Q60875-5;Q60875-4;Q60875;Q60875- 3;Q60875-2 Q60875-5 ARHG2_MOUSE Arhgef2 Kiaa0651 Lbcl1 Lfc Rho guanine nucleotide exchange factor 2 P70658-2;P70658 P70658-2 CXCR4_MOUSE Cxcr4 Cmkar4 Lestr Sdf1r C-X-C chemokine receptor type 4 (CXC-R4) P04223-2;P04223;P06339;P14432 P04223-2 HA1K_MOUSE H2-K1 H2-K H-2 class I histocompatibility antigen Q3TBT3;Q3TBT3-3;Q3TBT3-2 Q3TBT3 STING_MOUSE Tmem173 Eris Mita Mpys Sting Stimulator of interferon genes protein (mSTING) P03991;P14428 P03991 HA1W_MOUSE H2-K1 H2-K H-2 class I histocompatibility antigen Q6R5N8 Q6R5N8 TLR13_MOUSE Tlr13 Toll-like receptor 13 P25911-2 P25911-2 LYN_MOUSE Lyn Tyrosine-protein kinase Lyn (EC 2.7.10.2) P43406 P43406 ITAV_MOUSE Itgav Integrin alpha-V (Vitronectin receptor subunit alpha) Q80TH2-1;Q80TH2;Q80TH2- 2;Q80TE7 Q80TH2-1 LAP2_MOUSE Erbb2ip Erbin Kiaa1225 Lap2 Protein LAP2 (Densin-180-like protein) Q9QVP9 Q9QVP9 FAK2_MOUSE Ptk2b Fak2 Pyk2 Raftk Protein-tyrosine kinase 2-beta (EC 2.7.10.2) P68181-4;P68181-2;P68181-3;P68181 P68181-4 KAPCB_MOUSE Prkacb Pkacb cAMP-dependent protein kinase catalytic subunit beta (PKA C-beta) (EC 2.7.11.11) Q9JJ28 Q9JJ28 FLII_MOUSE Flii Fli1 Fliih Protein flightless-1 homolog Q18PI6;Q18PI6-2;Q18PI6-3 Q18PI6 SLAF5_MOUSE Cd84 Slamf5 SLAM family member 5 (Leukocyte differentiation antigen CD84) Q64343 Q64343 ABCG1_MOUSE Abcg1 Abc8 Wht1 ATP-binding cassette sub-family G member P62482;P63143 P62482 KCAB2_MOUSE Kcnab2 Ckbeta2 I2rf5 Kcnb3 Voltage-gated potassium channel subunit beta-2 (EC 1.1.1.-) Q9WVK4;Q8BH64 Q9WVK4 EHD1_MOUSE Ehd1 Past1 EH domain-containing protein 1 (PAST homolog 1) (mPAST1) Q8BGU5-2;Q8BGU5 Q8BGU5-2 CCNY_MOUSE Ccny Cfp1 Cyclin-Y (Cyclin fold protein 1) P48025 P48025 KSYK_MOUSE Syk ptk72 Sykb Tyrosine-protein kinase SYK (EC 2.7.10.2) Q8K4Q8 Q8K4Q8 COL12_MOUSE Colec12 Clp1 Srcl Collectin-12 P63158 P63158 HMGB1_MOUSE Hmgb1 Hmg-1 Hmg1 High mobility group protein B1 P31996-2;P31996 P31996-2 CD68_MOUSE Cd68 Macrosialin (CD68) P25911;P16277 P25911 LYN_MOUSE Lyn Tyrosine-protein kinase Lyn (EC 2.7.10.2) P49300 P49300 CLC10_MOUSE Clec10a Mgl Mgl1 C-type lectin domain family 10 member A (MMGL) P49769;Q61144-2;Q61144 P49769 PSN1_MOUSE Psen1 Ad3h Psnl1 Presenilin-1 (PS-1) (EC 3.4.23.-) Q61072 Q61072 ADAM9_MOUS E Adam9 Kiaa0021 Mdc9 Mltng Disintegrin and metalloproteinase domain-containing protein 9 (ADAM 9) (EC 3.4.24.-) Q60766-2;Q60766 Q60766-2 IRGM1_MOUSE Irgm1 Ifi1 Iigp3 Irgm Immunity-related GTPase family M protein 1 (EC 3.6.5.-) O54885 O54885 TYOBP_MOUSE Tyrobp Dap12 Karap TYRO protein tyrosine kinase-binding protein (DNAX-activation protein 12) P05555-2;P05555 P05555-2 ITAM_MOUSE Itgam Integrin alpha-M (CD11 antigen-like family member B) O88351 O88351 IKKB_MOUSE Ikbkb Ikkb Inhibitor of nuclear factor kappa-B kinase subunit beta (I-kappa-B-kinase beta) Q8C863;Q8BZZ3;Q8C863-2 Q8C863 ITCH_MOUSE Itch E3 ubiquitin-protein ligase Itchy (EC 6.3.2.-) Q99NH8;Q99NH8-2 Q99NH8 TREM2_MOUSE Trem2 Trem2a Trem2b Trem2c Triggering receptor expressed on myeloid cells 2 (TREM-2) Q3U1F9 Q3U1F9 PHAG1_MOUSE Pag1 Cbp Pag Phosphoprotein associated with glycosphingolipid-en- riched microdomains 1 (Csk-binding protein) (Transmembrane phosphoprotein Cbp) Q08857 Q08857 CD36_MOUSE Cd36 Platelet glycoprotein 4 (Glycoprotein IIIb) Q3TD49-2;Q3TD49 Q3TD49-2 SPP2B_MOUSE Sppl2b Signal peptide peptidase-like 2B (SPP-like 2B) (SPPL2b) (EC 3.4.23.-) Q9JJR8 Q9JJR8 TMM9B_MOUS E Tmem9b Transmembrane protein 9B O88597 O88597 BECN1_MOUSE Becn1 Beclin-1 (Coiled-coil myosin-like BCL2-interacting protein) as the former are more effective in inducing cytokine release. These criteria allowed us to rank and screen the list of immune-related type B oligomer binders, and the resulting selection of the top ranked proteins is shown in Table 2. The list discloses proteins involved in cell signalling of particular relevance to the mechanisms of innate immunity. DISCUSSION Studies of the mechanisms of neurodegeneration associated with protein misfolding and aggregation has led to the conclusion that multiple lines of investigation are required to understand the multifactorial nature of such neuropathologies 9 48. The inflammatory process associated with protein aggregation and deposition has been recognized to be a major contribution to the onset and progression of neurodegenerative misfolding diseases 15 16. However, for a number of reasons, including the challenges involved in the structural characterisation of disease-linked protein oligomers, their low stability under experimental conditions, and the lack of comparative studies performed in parallel, there is still a lack of information related to correlations of oligomer structure and function. Such difficulties have limited the elucidation of crucial and common molecular mechanisms underlying oligomerelicited immune responses in neurodegenerative diseases. As we have reported in this study, the HypF-N oligomer model enables one to overcome many of these constraints. Our results from experiments involving the stimulation of microglial cells using type A and B oligomers of HypF-N showed that both types of oligomer stimulate the release of pro-inflammatory cytokines, validating the use of the 8 Table 2. Highest ranked immune-related candidates. Selection of the highest ranked immune-related proteins identified in the type B oligomer interactome according to the selection criteria i-iv discussed in the text. The proteins highlighted in the list are those matching two or more selection criteria. The proteins ID and name, gene names and experimental values of binding ratio for type B oligomers vs type A oligomers (log2(B/A)), the proportion of a given protein in the type B oligomer sample (log10(% in B)), proportion of the protein in the type A oligomer sample (log10(% in A)), and in microglia (log10(% in microglia)) are indicated. CRITERIAa Leading ID Gene names log2(B/A ) log10(% in B) log10(% in A) log10(% in microglia) Protein names (i) P63017 Hspa8 Hsc70 Hsc73 2.6 -0.48 -0.75 -0.276 Heat shock cognate 71 kDa protein (Heat shock 70 kDa protein 8) P62821 Rab1A Rab1 1.8 -0.534 -0.601 -0.855 Ras-related protein Rab-1A (YPT1-related protein) P20491 Fcer1g Fce1g 2.37 -0.89 -1.1 - High affinity immunoglobulin epsilon receptor subunit gamma (Fc receptor gamma-chain) Q9JKF1 Iqgap1 1.34 -0.976 -0.94 -0.326 Ras GTPase-activating-like protein IQGAP1 Q922R8 Pdia6 Txndc7 6.66 -1.07 -2.11 -0.699 Protein disulfide-isomerase A6 (EC 5.3.4.1) Q61462 Cyba 2.4 -1.21 -1.38 -2.474 Cytochrome b-245 light chain P07901 Hsp90aa1 Hsp86 Hsp86- 1 Hspca 3.79 -1.25 -1.9 -0.959 Heat shock protein HSP 90-alpha (Heat shock 86 kDa) Q99JB2 Stoml2 Slp2 2.42 -1.26 -1.46 -2.046 Stomatin-like protein 2 Q80WJ7 Mtdh Lyric 2.11 -1.29 -1.47 -1.687 Protein LYRIC (3D3/LYRIC) O35658 C1qbp Gc1qbp 4.97 -1.29 -2.36 -1.266 Complement component 1 Q subcomponent-binding protein Q61093 Cybb Cgd 2.57 -1.38 -1.7 -1.589 Cytochrome b-245 heavy chain Q9CQW9 Ifitm3 1.42 -1.47 -1.44 -1.75 Interferon-induced transmembrane protein 3 Q9D3P8 Plgrkt 1.21 -1.54 -1.46 -2.09 Plasminogen receptor (KT) (Plg-R(KT)) P01899 H2-D1 3.15 -1.65 -2.06 -2.74 H-2 class I histocompatibility antigen, D-B alpha chain (H-2D(B)) P41251 Slc11a1 Bcg Ity Lsh Nramp1 3.43 -1.67 -2.02 -2.57 Natural resistance-associated macrophage protein 1 (NRAMP 1) Q6A0D4 Rftn1 Kiaa0084 2.48 -1.72 -1.91 -2.6 Raftlin (Raft-linking protein) Q9D8Y0 Efhd2 Sws1 2.22 -1.78 -1.96 -1.18 EF-hand domain-containing protein D2 (Swiprosin-1) (ii) Q8R422 Cd109 3.08 -2.34 -2.76 -4.558 CD109 antigen (GPI-anchored alpha-2 macroglobulin-related protein) P25911-2 Lyn 2.21 -2.5 -2.55 -3.301 Tyrosine-protein kinase Lyn (EC 2.7.10.2) P62482 Kcnab2 Ckbeta2 I2rf5 Kcnb3 1.33 -1.74 -1.66 -3.482 Voltage-gated potassium channel subunit beta-2 (EC 1.1.1.-) (iii) Q922R8 Pdia6 Txndc7 6.66 -1.07 -2.11 -0.699 Protein disulfide-isomerase A6 (EC 5.3.4.1) P06800-3 Ptprc Ly-5 5.42 -2.18 -3.05 -0.969 Receptor-type tyrosine-protein phosphatase C (EC 3.1.3.48) O35658 C1qbp Gc1qbp 4.97 -1.29 -2.36 -1.266 Complement component 1 Q subcomponent-bind- ing protein Q61830 Mrc1 4.93 -2.57 -3.04 -0.673 macrophage mannose receptor 1 (MMR) Q9JJX6-4 P2rx4 P2x4 4.58 -2.26 -2.93 -1.921 P2X purinoceptor 4 (P2X4) Q61549 Adgre1 Emr1 Gpf480 4.12 -2.28 -2.95 -2.28 Adhesion G protein-coupled receptor E1 (Cell surface glycoprotein F4/80) P17710-3 Hk1 4.11 -3.11 -3.67 -1.555 Hexokinase-1 (EC 2.7.1.1) Q9Z0M6-2 Cd97 3.99 -2.48 -2.88 -3.302 CD97 P07901 Hsp90aa1 Hsp86 Hsp86- 1 Hspca 3.79 -1.25 -1.9 -0.959 Heat shock protein HSP 90-alpha (Heat shock 86 kDa) O70172 Pip4k2a Pip5k2a 3.76 -2.26 -2.8 -1.92 Phosphatidylinositol 5-phosphate 4-kinase type-2 alpha (EC 2.7.1.149) P41251 Slc11a1 Bcg Ity Lsh Nramp1 3.43 -1.67 -2.02 -2.57 Natural resistance-associated macrophage protein 1 (NRAMP 1) Q9WTR1 Trpv2 Grc 3.37 -2.06 -2.44 -1.982 Transient receptor potential cation channel subfamily V member 2 (TrpV2) (iv) P04441-2 Cd74 Ii 2.19 -2.06 -2.21 -3.019 H-2 class II histocompatibility antigen gamma chain (Ia antigen-associated invariant chain) Q9QUN7 Tlr2 2.95 -2.6 -2.85 -1.41 Toll-like receptor 2 (CD antigen CD282) a Criteria as defined in the text HypF-N oligomers as a model of neuroinflammation induced by protein misfolding 14 29. We note that the stimulation of cytokine release was observed following exposure to very low concentrations of oligomers, as many as three orders of magnitude lower than those needed to cause detectable cytotoxicity, suggesting the importance of this phenomenon at early stages of pathology, when the concentration of misfolded protein oligomers is still very low. Interestingly, when the ability of type A and type B oligomers to stimulate the immune response was evaluated within the lower concentration range, type B oligomers were found to be stronger inducers of proinflammatory cytokine release than type A oligomers. As the Supplementary Figure 2. Interactions of type A and type B oligomers with the cytosolic proteins from N13 cells. Replica-specific scatterplot of protein abundances (log10 iBAQ values) of N13 microglia cell proteins bound to type B oligomers (y axis) vs those bound to type A oligomers (x axis) in the four biological replicas analysed. The green points in each subplot represent cytosolic proteins detected and quantified in each replica. Proteins shown vertically on the left and horizontally on the lower edge of each scatterplot are proteins binding only to type B and type A oligomers, respectively.