Search tips
Search criteria 


Logo of nihpaAbout Author manuscriptsSubmit a manuscriptHHS Public Access; Author Manuscript; Accepted for publication in peer reviewed journal;
Cell Host Microbe. Author manuscript; available in PMC 2010 July 23.
Published in final edited form as:
PMCID: PMC2745191

Responses to amyloids of microbial and host origin are mediated through Toll-like receptor 2


Amyloid contributes to incapacitating disorders, such as Alzheimer’s disease, by eliciting local inflammation in the brain. However, little is known about how amyloid deposition elicits inflammation. Salmonella enterica serotype Typhimurium produces amyloid-like curli fibrils composed of the protein CsgA. We show that curli fibrils contributed to Nos2 expression in a mouse sepsis model by stimulating Toll-like receptor (TLR) 2. The TLR2 agonist activity of CsgA was markedly reduced by an amino acid substitution (N122A) that also lowered its amyloidogenicity. Synthetic peptides corresponding to residues 111-151 of CsgA or β-amyloid 1-42 from plaques of Alzheimer’s disease stimulated Nos2 production in macrophages and microglia cells through a TLR2-dependent mechanism. This activity was abrogated when a N122A substitution was introduced into the synthetic CsgA peptide. The induction of TLR2-mediated responses by bacterial and eukaryotic amyloids may provide a novel linchpin of pathogenesis, with implications for innate immunity and the immunopathogenesis of Alzheimer’s disease.


Amyloids, named after their amylose-like ability to stain with iodine, are fibrils composed of orderly repeats of protein with a β sheet structure, in which the β sheets are oriented perpendicular to the fibrillar axis (Otzen and Nielsen, 2007). The formation of amyloids by human proteins characterizes several illnesses, including Alzheimer’s disease, Huntington’s disease, type 2 diabetes, secondary amyloidosis and prion diseases. The pathogenesis of these conditions involves marked inflammation at sites of amyloid deposition, resulting in tissue injury (Cherny et al., 2005). However, the mechanisms by which amyloid deposits elicit inflammatory responses remain poorly characterized.

The csg gene cluster of Escherichia coli encodes the production of curli fibrils (Olsen et al., 1989), proteinaceous surface structures mediating biofilm formation and binding of host proteins. The biogenesis of curli on the bacterial cell surface involves formation of amyloid fibrils composed of the major subunit CsgA. Unlike formation of eukaryotic amyloids, curli formation requires a specific assembly machinery, including nucleator proteins (CsgB and CsgF) that initiate the polymerization of secreted CsgA subunits into fibrils (Chapman et al., 2002).

Studies on the pathogenesis of E. coli and S. enterica infections have implicated curli fibrils in the induction of inflammatory responses. Curli fibrils elicit IL-8 expression in macrophages and contribute to the development of hypotension and increased plasma nitrite/nitrate levels in a mouse model of E. coli sepsis (Bian et al., 2000; Bian et al., 2001). Deletion of csgBA in the closely related pathogen S. enterica serotype Typhimurium reduces its ability to elicit intestinal inflammation in a calf model (Tukel et al., 2005). Studies on the CsgA curlin subunit suggest that the generation of these host responses involves the stimulation of Toll-like receptor (TLR) 2 signaling by curli fibrils (Tukel et al., 2005). The identification of CsgA as a pathogen associated molecular pattern (PAMP) that signals through TLR2 is somewhat surprising, because the csg gene cluster is only present in a small number of closely related genera belonging to the family Enterobacteriaceae, including Salmonella, Shigella, Escherichia, Citrobacter, Enterobacter and Klebsiella (Zogaj et al., 2003). In contrast, other PAMPs of gram-negative bacteria, such as lipopolysaccharide (LPS) or flagellin, are widely conserved among members of the phylum Proteobacteria. To better understand which properties make curli fibrils a logical target for recognition of conserved molecular patterns by the innate immune system, we studied the mechanism by which this bacterial product is recognized by TLR2.


The S. Typhimurium csg operon contributes to Nos2 expression in vivo

The E. coli csg operon has recently been implicated in contributing to an elevated expression of the inducible nitric oxide synthase gene (Nos2) in a mouse shock model (Bian et al., 2001). We investigated whether TLR2 stimulation by curli fibrils is the mechanism responsible this increased Nos2 expression. S. Typhimurium bacteremia was modeled using intraperitoneal infection of C57BL/6 mice to synchronize the arrival of bacteria at systemic sites of infection. To establish the relevance of curli fibrils in inducing inflammatory responses in this model, we investigated host responses in C57BL/6 mice infected intraperitoneally with the curli producing S. Typhimurium wild type (IR715) or a strain deficient for curli production (csgBA mutant, SF15). As a control, mice were either mock-infected or inoculated with a S. Typhimurium msbB mutant (RPW3). A mutation in msbB reduces the acylation of lipid A, which abrogates the ability of LPS to function as a TLR4 agonist (Somerville et al., 1996). Compared to mock-infected animals, expression of Nos2 was induced more than 4,000 fold in the liver of mice 8 hours after S. Typhimurium infection (Figure 1). In contrast, Nos2 mRNA levels were markedly blunted in wild type mice infected with the msbB mutant, confirming the importance of LPS for inducing host responses (Khan et al., 1998). Nos2 levels were also markedly reduced in TLR2 deficient mice infected with the S. Typhimurium wild type, suggesting that both TLR2 and TLR4 are necessary for fully inducing this response in vivo. Importantly, the level of Nos2 mRNA was significantly (P = 0.006) lower in mice infected with a strain deficient for curli production (csgBA mutant) than in mice infected with the curli producing S. Typhimurium wild type. In contrast, the curli producing S. Typhimurium wild type and a strain deficient for curli production (csgBA mutant) elicited similar Nos2 mRNA levels when the experiment was repeated with TLR2 deficient mice (Figure 1A). These data provided genetic evidence for a contribution of curli fibrils to the induction Nos2 expression in vivo. Deletion of csgBA did not reduce Nos2 expression to the same degree as TLR2 deficiency, suggesting that curli cooperates with other TLR2 ligands (e.g. lipoprotein) to fully induce nitric oxide synthase (iNOS) production in vivo.

Figure 1
Contribution of curli fibrils to iNOS production during S. Typhimurium sepsis. (A) Nos2 mRNA levels observed in the liver of mice 8 hours after intraperitoneal infection of wild type mice or TLR2 deficient mice with 108 colony forming units (CFU) of ...

TLR2 stimulation by curli fibrils contributes to iNOS production

The induction of host responses during bacteremia results from the initial interactions of bacteria with tissue macrophages and/or blood monocytes through bacterial specific TLRs (Cristofaro and Opal, 2003; Weighardt and Holzmann, 2007). To study the mechanism by which curli fibrils contribute to host responses during bacteremia, we investigated whether these surface structures were able to stimulate bone marrow-derived macrophages (BMDM) isolated from C57BL/6 mice to produce iNOS. The compound produced by iNOS is nitric oxide, a product that is labile in the presence of oxygen. Nitrite, the major byproduct of nitric oxide upon reaction with oxygen, was detected after stimulation of BMDM using the Griess assay. Native curli fibrils were gel purified from the S. Typhimurium msbB mutant to avoid contamination with LPS. Stimulation of BMDM with increasing concentrations of curli fibrils did not result in increased nitrite production (Figure 1B). However, when the same assay was performed in the presence of interferon (IFN) γ, a cytokine present in blood during sepsis, stimulation of BMDM with increasing concentrations of curli fibrils resulted in increased nitrate production. These data suggested that iNOS production in BMDM was induced synergistically by IFN-γ and curli. We hypothesized that TLR2 stimulation may be a potential mechanism by which curli fibrils may contribute to iNOS production in BMDM. To test this idea, nitrite production induced by curli fibrils in IFN-γ stimulated BMDM from control mice (C57BL/6) was compared to that in IFN-γ stimulated BMDM from TLR2 deficient (TLR2−/−) mice. TLR2 deficiency abrogated the ability of curli to induce iNOS production in IFN-γ stimulated BMDM. Similarly, synthetic lipopeptide (Pam3CSK4), a well-characterized TLR2 agonist (Aliprantis et al., 1999), induced nitrite production in IFN-γ stimulated BMDM from wild type mice but not from TLR2−/− mice (Figure 1C). Collectively, these data suggested that native curli fibrils contributed to iNOS production during bacteremia by stimulating macrophages through TLR2.

A N122A substitution reduces the TLR2 agonist activity of S. Typhimurium CsgA

To characterize the interaction of curli fibrils with TLR2 in more mechanistic detail, we initially reduced the complexity of our experimental model. To this end, we investigated stimulation of human kidney (HEK293) cells stably transfected with human TLR2 and the CD14 adaptor protein (HEK293-TLR2CD14 cells) with purified recombinant CsgA, the major subunit of curli fibrils. HEK293-TLR2CD14 cells were stimulated with increasing concentrations of purified Gluthathion S transferase (GST) protein (negative control), GST-CsgA fusion protein (Humphries et al., 2003), or Pam3CSK4 (positive control) (Figure 2). Measurements of IL-8 production were used to follow stimulation, because TLR2-dependent production of this cytokine does not require IFN-γ. Pam3CSK4 and GST-CsgA elicited IL-8 secretion by HEK293-TLR2CD14 cells, while no stimulation was observed with GST protein (Figure 2A). Neither ligand elicited IL-8 production in HEK293 cells (data not shown). Our data suggested that the GST-CsgA fusion protein was a potent TLR2 agonist, although its activity appeared to be lower than that of Pam3CSK4. To identify regions within CsgA contributing to TLR2 stimulation, we constructed GST fusion proteins containing the 22 residue N-terminal domain of mature CsgA (residues 21-42 in the CsgA primary sequence, GST-CsgA-N), or the amyloidogenic C-terminal domain, which contains 5 conserved repeat regions (residues 43-151, GST-CsgA-R1-5) (Collinson et al., 1999). Only a fusion protein containing the C-terminal domain elicited IL-8 production in HEK293-TLR2CD14 cells at levels similar to GST-CsgA (Figure 2B). Next we constructed GST fusions containing the R1 repeat (GST-CsgA-R1), the R1-R3 repeats (GST-CsgA-R1-3) or the R4-R5 repeats (GST-CsgA-R4-5). Only GST-CsgA-R4-5 elicited IL-8 production in HEK293-TLR2CD14 cells, although at reduced levels compared to GST-CsgA (Figure 2B). These data suggested that the R4-R5 repeat region is necessary and sufficient for eliciting IL-8 production through TLR2.

Figure 2
Characterization of the TLR2 agonist activity of CsgA. Production of IL-8 was determined 24 hours after stimulation of HEK293-TLR2CD14 cells with different affinity purified fusion proteins or synthetic peptides (Pam3CSK4) using ELISA. Data are shown ...

To determine whether residues conserved among different bacterial pathogens are important for TLR2 stimulation, we constructed GST fusions to CsgA proteins from both E. coli and Shigella sonnei (Figure 3). Purified fusion proteins containing CsgA from E. coli or S. sonnei elicited similar levels of IL-8 production in HEK293-TLR2CD14 cells as GST-CsgA from S. Typhimurium (Figure 3A). These data suggested that residues important for TLR2 recognition are conserved between CsgA proteins from different genera. Conserved amino acids residues within the R4-R5 repeat region (Figure 3B) were replaced in GST-CsgA-R1-5 by alanine residues using site-directed mutagenesis. Of 14 fusion proteins containing alanine substitutions of conserved amino acid residues, 6 did not form stable proteins and could not be purified (data not shown). The remaining 8 fusion proteins were purified and tested for their ability to elicit IL-8 production in HEK293-TLR2CD14 cells. Only one fusion protein, containing amino acid substitution N122A (GST-CsgA-R1-5N122A), exhibited a greatly attenuated ability to stimulate IL-8 secretion compared to GST-CsgA-R1-5 (Figure 3C). Interestingly, the asparagine residue 122 is located in a region of the protein, whose corresponding region in E. coli CsgA forms amyloid fibrils when produced as a synthetic hexapeptide (Cherny et al., 2005).

Figure 3
Identification of CsgA residues required for its TLR2 agonist activity. Production of IL-8 was determined 24 hours after stimulation of HEK293-TLR2CD14 cells with different affinity purified fusion proteins using ELISA. Data are shown as averages ± ...

The ability of CsgA to form amyloids is required for signaling through TLR2

In the absence of the CsgB and CsgF nucleator proteins, CsgA adopts a soluble form that upon prolonged incubation assembles into curli amyloid fibrils (Chapman et al., 2002). To investigate whether the N122A substitution affected amyloid formation of S. Typhimurium CsgA, increasing concentrations of purified fusion proteins were incubated with the amyloid specific dye Thioflavin T (ThT) (LeVine, 1993; Wang et al., 2007) (Figure 4). GST-CsgA-R1-5 exhibited significantly greater ThT fluorescence than GST-CsgA-R1-5N122A, indicating that the N122A substitution reduced amyloid formation (Figure 4A). Analysis of the protein secondary structure with circular dichroism (CD) spectroscopy revealed that GST-CsgA-R1-5 produced a peak at approximately 197nm and a trough at approximately 220nm, which is characteristic of proteins containing predominantly β-sheet structure (CD spectroscopy predicted 95% β-sheet structure) In contrast, GST-CsgA-R1-5N122A failed to show these features in the CD spectrum (Figure 4B), suggesting that this protein had much less secondary structure (CD spectroscopy predicted 25% β-sheet structure). These biochemical features were consistent with the idea that the N122A substitution reduced amyloidogenicity. The reduced ThT binding and altered CD spectrum correlated well with a reduced ability of GST-Csg-A-R1-5N122A to elicit IL-8 secretion in HEK293-TLR2CD14 cells (Figure 4C). Collectively, these data suggested that the TLR2 agonist activity of CsgA was related to its ability to form amyloids.

Figure 4
Effect of an N122A substitution on the ability of affinity purified GST-CsgA-R1-5 to form amyloids and serve as TLR2 agonist. (A) Increasing concentrations of GST-CsgA-R1-5 and GST-CsgA-R1-5N122A fusion proteins were stained with the amyloid specific ...

Glutamine/asparagine-rich domains of proteins have a high propensity to assemble amyloid fibrils (Michelitsch and Weissman, 2000) by forming β-sheets linked together by hydrogen bonds (Perutz et al., 2002). The C-terminal amyloid domain of CsgA consists of five imperfect direct repeats of 19–23 amino acids (R1-R5), each containing a hexapeptide with conserved glutamine and asparagine residues (Barnhart and Chapman, 2006) and a chemical composition similar to prion protein repeats (Cherny et al., 2005). A role of E. coli CsgA hexapeptide sequences in amyloid formation is suggested by the finding that the corresponding synthetic peptides form amyloid fibrils (Cherny et al., 2005; Wang et al., 2007). We further investigated the effect of a N122A substitution on amyloid formation by electron microscopic examination of synthetic S. Typhimurium CsgA hexapeptides (Figure 5). E. coli CsgA contains the amyloidogenic hexapeptide QFGGGN (residues 117-122) which can form amyloid fibrils (Cherny et al., 2005). The N122A substitution was located in the corresponding hexapeptide QYGGNN (residues 117-122) of the S. Typhimurium CsgA protein (Figure 3B). A positive control (β-amyloid 1-40 peptide) and the synthetic peptides QYGGNN and QYGGNA (corresponding to an N122A substitution) were incubated at 37°C for 7 days to promote amyloid fibril formation. Synthetic β-amyloid 1-40 and the QYGGNN peptide formed fibrils, while the QYGGNA peptide did not (Figure 5).

Figure 5
Formation of amyloid fibrils the CsgA QYGGNN hexapeptide. Formation of amyloid fibrils by synthetic QYGGNN hexapeptide (A), synthetic QYGGNA hexapeptide (B) and synthetic β-amyloid 1-40 peptide from amyloid plaques of Alzheimer’s disease ...

Responses to amyloids of host and microbial origin are mediated through TLR2

To investigate whether responses to amyloids of both host and microbial origin are mediated through TLR2, human macrophage-like (THP-1) cells were stimulated with purified curli fibrils or β-amyloid 1-40 and induction of IL-8 mRNA was followed by quantitative real-time PCR analysis (Figure 6). Stimulation of THP-1 cells with curli fibrils or β-amyloid 1-40 resulted in marked induction of IL-8 mRNA expression, which could be partially inhibited by pre-incubation with a blocking anti-TLR2 monoclonal antibody (Figure 6A). These data supported the idea that both microbial (curli) and host (β-amyloid 1-40) amyloids contribute to inflammatory cytokine production by stimulating human TLR2.

Figure 6
Pro-inflammatory responses elicited by amyloids of bacterial and host origin in macrophages. (A) Stimulation (4 hours) of human THP-1 cells with purified curli fibrils or β-amyloid 1-40 in the presence (open bars) or absence (gray bars) of inhibitory ...

IFN-γ and host amyloids are thought to contribute to neurotoxicity during Alzheimer’s disease by eliciting the production of reactive nitrogen species (Meda et al., 1995). We therefore determined whether, similar to curli fibrils (Figure 1), host amyloids would elicit Nos2 expression in IFN-γ stimulated murine BMDM. Purified curli fibrils, synthetic β-amyloid 1-40 and synthetic β-amyloid 1-42 induced significantly (P < 0.05) higher levels of Nos2 mRNA in IFN-γ stimulated BMDM than IFN-γ alone. In contrast, stimulation with a peptide consisting of a scrambled amino acid sequence of β-amyloid 1-42 (scrambled β-amyloid 1-42) did not induce Nos2 expression above background levels. Expression of Nos2 elicited in BMDM macrophages upon stimulation with amyloids was TLR2-dependent, since this response was abrogated in BMDM from TLR2 deficient mice.

To stringently test whether curli fibrils stimulate host responses because CsgA has an intrinsic TLR2-agonist activity, it was necessary to rule out the possibility that preparations used for stimulation of host cells were contaminated with other bacterial PAMPs. To this end we synthesized synthetic peptides corresponding to the R4-R5 repeats of CsgA (CsgA-R4-5) and CsgAN122A (CsgA-R4-5N122A). Synthetic peptides were incubated with ThT to monitor amyloid formation (Figure 7). As expected, the positive control (β-amyloid 1-42) exhibited greater ThT fluorescence than the negative control peptide (scrambled β-amyloid 1-42). CsgA-R4-5 exhibited significantly greater ThT fluorescence than CsgA-R4-5N122A, indicating that the N122A substitution reduced amyloid formation (Figure 7A). The biological activity of synthetic peptides was investigated using BV2 murine microglial cells, a cell type relevant for the generation of host responses to β-amyloid 1-42 in the brain. Expression of TLR2 on the surface of microglia cells was verified using flow cytometry (Figure 7B). Microglia cells were treated with IFN-γ and stimulated with synthetic peptides. Synthetic lipopeptide (Pam3CSK4) induced Nos2 expression in microglia cells at levels similar to synthetic β-amyloid 1-42 and synthetic CsgA-R4-5. In contrast, scrambled β-amyloid 1-42 and CsgA-R4-5N122A did not induce Nos2 expression above background levels. Induction of Nos2 expression by Pam3CSK4, β-amyloid 1-42 or CsgA-R4-5 could be inhibited by pre-incubation of microglia cells with a blocking anti-murine TLR2 monoclonal antibody (Figure 7C). Collectively, these data demonstrated that responses to both host (β-amyloid 1-42) and microbial amyloids (CsgA) are mediated through TLR2.

Figure 7
Amyloidogenicity and TLR2 agonist activity of synthetic peptides. (A) Synthetic peptides were stained with the amyloid specific dye Thioflavin T (ThT) and fluorescence was measured (438nm excitation, 495nm emission) to assess their amyloidogenicity. Data ...


Amyloid formation in the host is thought to result from pathologic conditions leading to protein misfolding. In contrast, the csgBA and csgDEFG operons of E. coli and S. Typhimurium encode a nucleation-precipitation machinery specifically dedicated to the assembly of curli amyloid fibrils (Chapman et al., 2002). Thus, formation of curli fibrils is the desired outcome rather than the result of a misguided protein-folding pathway, a difference relevant for in vivo assembly. While specific pathologic conditions resulting in host amyloid deposition in vivo are a critical factor in the pathogenesis of amyloidosis, the nucleation-precipitation machinery assembles curli fibrils independently of specific environmental factors (Hammer et al., 2007). Conversely, host proteins involved in the pathogenesis of amyloidosis are soluble under physiologic condition, but release of soluble CsgA only occurs in bacteria with defective nucleation-precipitation machinery (e.g. a csgB mutant) (Chapman et al., 2002). Prolonged incubation of synthetic CsgA peptides resulted in formation of amyloid fibrils, thereby eliciting host responses through TLR2. The use of synthetic peptides excluded contamination with other TLR ligands as a possible explanation for the observed host responses. However, assembly of CsgA peptides into amyloid fibrils is not a good model for curli biogenesis in vivo, which requires the nucleator protein CsgB. The presence of an intact curli nucleation-precipitation machinery was associated with significantly increased Nos2 mRNA levels during S. Typhimurium bacteremia, thereby providing genetic evidence that curli biogenesis contributed to host responses in vivo.

Our finding that the CsgA protein stimulates TLR2 signaling because it forms amyloid fibrils has several important implications. First, although the csg operon is only present in a small number of genera within the family Enterobacteriaceae, extracellular amyloid adhesins are widely distributed among several bacterial phyla, including the Proteobacteria (Alpha-, Beta-, Gamma- and Deltaproteobacteria), Bacteriodetes, Chloroflexi and Actinobacteria (Larsen et al., 2007). Thus, sensing the presence of bacterial amyloids through TLR2 may provide the host’s innate immune system the capacity to detect a large group of microbes by conserved pattern recognition. Second, amyloid formation underlies several common human chronic diseases and TLR2-mediated recognition may provide a novel linchpin of pathogenesis.

For example, amyloidosis is the hallmark of Alzheimer’s disease, Huntington’s disease, type 2 diabetes, secondary amyloidosis and prion diseases, conditions marked by inflammation at sites of amyloid deposition (Cherny et al., 2005). Alzheimer’s disease is characterized by the presence in the brain of extracellular plaques consisting of compacted β-amyloid fibril deposits. By chronically activating microglia, the main resident macrophage-like cell type in the central nervous system, these β-amyloid deposits elicit foci of local inflammatory reactions that contribute to neuronal injury (Akiyama et al., 2000; Town et al., 2005). Recently, innate immune receptors, including TLR2 and CD14 have been implicated in neurodegeneration during Alzheimer’s disease (Fassbender et al., 2004; Jana et al., 2008; Letiembre et al., 2007; Richard et al., 2008; Udan et al., 2007, 2008) and in responses to acute-phase serum protein A (Cheng et al., 2008; He et al., 2009). CD14 directly binds β-amyloid and increases its uptake by microglial cells (Liu et al., 2005) while TLR2 triggers an inflammatory response to aggregated β-amyloid in THP-1 macrophage-like cells (Udan et al., 2007, 2008). Similarly, TLR2 contributed to cytokine production elicited by curli fibrils and β-amyloid in THP-1 macrophage-like cells. CsgA elicits IL-8 production in HEK293 cells transfected with TLR2, but cytokine production is increased further when cells are transfected with TLR2 and CD14 (Tukel et al., 2005). Here we show that cytokine production elicited by CsgA was dependent on the ability of this protein to form amyloid fibrils. Collectively, these data suggest that pattern recognition of amyloid fibrils of both bacterial and host origin underlies critical triggering of the innate immune system.

Activation of microglia cells is suggested to contribute to progressive neuronal injury by the release of neurotoxic products, such as nitric oxide (Hu et al., 1997; Meda et al., 1995). The induction of iNOS in microglial cultures stimulated with β-amyloid requires the presence of augmenting factors, such as IFN-γ or LPS (Goodwin et al., 1995; Meda et al., 1995; Nakamura et al., 1999). Similarly, we found that Nos2 expression in murine BMDM and microglia cells was induced synergistically by β-amyloid and IFN-γ. Importantly, our data provided evidence that the mechanism for inducing Nos2 expression involved recognition of amyloids by TLR2. This observation has implications for the pathogenesis of bacterial infections and of amyloid-associated diseases. For instance, our data suggest that curli fibrils may contribute to sepsis, because TLR2 stimulation contributes to the production of nitric oxide, a potent vasodilator (Kilbourn et al., 1990; Petros et al., 1991). Consistent with this idea, curli fibrils of E. coli have been shown to contribute to the development of hypotension and to increased plasma nitrite/nitrate levels in a mouse sepsis model (Bian et al., 2000; Bian et al., 2001). Furthermore, our data point to amyloid stimulation of TLR2 as a potential mechanism contributing to neuronal injury during amyloidosis. Hence, this research identifies potential targets for anti-inflammatory therapeutics.


Bacterial strains

S. Typhimurium strain IR715 is a fully virulent, spontaneous nalidixic acid-resistant derivative of strain ATCC14028 (Stojiljkovic et al., 1995). RPW3 and SF15 are derivatives of strain IR715 carrying mutations in msbB and csgBA, respectively (Raffatellu et al., 2005; Weening et al., 2005).

GST fusion proteins and synthetic peptides

A plasmid pSW5-50 encoding GST-CsgA has been described previously (Humphries et al., 2003). E. coli and S. sonnei csgA and truncated forms of S. Typhimurium csgA were amplified using the primer pairs listed in TableS1 and cloned into PCR2.1 (Invitrogen). GST fusion proteins were constructed in pGEX-4T-2 and affinity purified as described previously (Humphries et al., 2003). The protein concentration in each sample was determined by Bradford assay (Ausubel et al., 1994).

Synthetic β-amyloid-1-40-HCl and β-amyloid-1-42 and scrambled β-amyloid-1-42 were purchased from Anaspec (San Jose, CA). Synthetic lipopeptide (Pam3CSK4) was purchased from Invivogen. Synthetic peptides CsgAR4-5, CsgA-R4-5N122A, QYGGNN and QYGGNA were chemically synthesized by Biosynthesis (Lewisville, TX).

Site-directed mutagenesis

Mutations resulting in single amino acid substitutions were introduced into csgA using the primer pairs listed in TableS1 using the Quickchange Site-Directed Mutagenesis Kit (Stratagene, La Jolla).

Tissue culture

To isolate bone marrow-derived macrophages (BMDM) from C57BL/6 mice or TLR2-deficient (TLR2−/−) mice (Takeuchi et al., 1999) (Jackson Laboratory), femurs were removed and flushed with complete medium (Dulbecco’s modified Eagle’s medium, 10% fetal bovine serum, 2 mM glutamine, 100 μg ml1 penicillin and streptomycin). The bone marrow was then cultured for three days in complete medium supplemented with L-cell conditioned medium, which was prepared as described (Rolan and Tsolis, 2007). On day 3, the media was replaced with fresh complete medium supplemented with L-cell conditioned medium. On day 7, adherent cells were treated with Trypsin-EDTA (Gibco) and the resulting cell suspension was centrifuged at 1,000 rpm for 10 min. The cells were seeded in 24 well plates at a density of 5×105 cells per well. Cells were then incubated overnight in complete media without L-cell conditioned media or antibiotics. Curli fibrils were purified from the S. Typhimurium msbB mutant as described previously (Tukel et al., 2005). A blocking anti-human TLR2 monoclonal antibody (TL2.1) was purchased from eBiosciences and was added to cells at 10μg/well in 500μl one hour prior to stimulation. Recombinant mouse IFN-γ was purchased from eBioscience and added to BMDM (100units/well in 500μl). Supernatants from stimulated BMDM were collected 24 hours after stimulation and analyzed for nitrite production by incubating them with an equal volume of Griess reagents (Sigma). Absorbance was measured at 550 nm to determine nitrite concentration. For real-time PCR analysis, BMDM were collected 4 hours after stimulation and RNA was isolated with TriReagent (Molecular Research Center) according to the instruction of the manufacturer. Real-time PCR was performed as described below (see animal experiments).

HEK293 cells stably transfected with human TLR2/CD14 (Invivogen) were cultured according to manufacturer’s instructions. Appropriate ligands were added to HEK293-TLR2CD14 cells and supernatants were harvested at 24 hours after to determine IL-8 production by enzyme linked immunosorbent assay (ELISA) (Biolegend, San Diego, USA).

THP-1 cells were seeded at approximately 5×105cells/well in 24 well plates containing RPMI1640+10% FCS were differentiated with 50ng/ml phorbol 12-myristate 13-acetate (PMA) for 48 hours followed by growth without PMA for 4 days. THP-1 cells were incubated for 1 hour with mouse anti-human TLR2 antibody (E-bioscience, 10μg/well in 500μl) or with medium (control), followed by incubation with β-amyloid 1-40 (20μM) or curli fibrils (1μg/ml) for 4 hours. RNA was extracted to determine IL-8 mRNA levels by real-time PCR as described previously (Tukel et al., 2005).

BV-2 microglia cells were a kind gift of Dr. Lee-Way Jin (MIND Institute, UC Davis). BV-2 microglia were seeded at approximately 5×105cells/well in 24 well plates containing DMEM+10% FCS and incubated overnight. Microglia were treated with anti mouse/human TLR2 antibody (clone T2.5, E-bioscience, 10μg/well in 500μl) or with medium (control) for 1 hour. CsgA-R4-5 (10μM), CsgA-R4-5N122A (10μM), β-amyloid 1-42 (10μM), scrambled β-amyloid 1-42 (10μM), Pam3CSK4 (10μM), and recombinant mouse IFNγ (100units/well in 500μl) were added to the wells for 4 hours. RNA was extracted to determine iNOS mRNA levels by real-time PCR as described above.

Thioflavin T (ThT) Assay

CsgA samples were incubated for 3 days at 37°C and then mixed with 10mM ThioflavinT (Sigma) 1:1 at a final volume of 200μl. Fluorescence was measured at 438nm excitation, 495nm emission and 475nm cutoff settings using a Spectramax M2 plate reader (Molecular Devices, Sunnydale. CA).

CD Spectroscopy

Samples of purified GST-CsgA-R1-5 GST-CsgA-R1-5N122A fusion protein (0.7mg/ml) were analyzed in a Jasco J-715 (Tokyo, Japan) spectropolarimeter from 190 to 260 nm in a quartz cell at room temperature.

Peptide preparation and electron microscopy

A volume of 10μl of each sample was loaded on a Formvar-coated copper grid and incubated for 2 min and negatively stained with 2% uranyl acetate for 20 seconds.

Flow Cytometry

Microglia cell suspensions containing 5×105cells were washed once with FACS buffer and treated with rat anti-Mouse CD16/CD32 (Fc block, BD). Cells were then stained for 30 min in the dark at 4°C with optimized concentrations of Alexa Fluor® 488 conjugated anti-mouse/human Toll-like receptor 2 or IgG1 K isotype control antibodies (eBioscience). Cells were then washed once and resuspended in FACS buffer and analyzed using an LSR II flow cytometer (Becton Dickinson, San Jose, CA). The data were analyzed by using FlowJo software (Treestar, Inc., Ashland, OR).

Animal Experiments

For mouse experiments, 4 to 6 week old female C57BL/6 mice (Jackson Laboratory) were used. Groups of 4 mice were intraperitoneally infected with 1×108 CFU in PBS or sterile PBS (mock infection). At 8 hours after infection, mice where sacrificed a sample of the liver was collected from each mouse, immediately snap-frozen in liquid nitrogen and stored at −80°C. RNA was extracted from snap-frozen tissue with TriReagent (Molecular Research Center) according to the instruction of the manufacturer. 1000 ng of RNA from each sample was reverse transcribed in 0.05 ml volume (Taqman reverse transcription reagent, Applied Biosystems). 0.005 ml of cDNA was used for each Real-Time reaction. Real-time PCR was performed using SYBR Green (Applied Biosystems) and the 7900HT Fast Real-Time PCR System. The data were analyzed using the comparative Ct method (Applied Biosystems). Fold-increases in cytokine expression in infected mice were calculated relative to the average level of the respective cytokine in four mock-infected mice. The primers for Gapdh and Nos2 have been described previously (Roux et al., 2007; Wilson et al., 2008).

Statistical analysis

For statistical analysis of data, fold-changes in mRNA levels measured by real-time PCR underwent logarithmic transformation. Statistical analysis of data was performed using a Student’s t test.


We would like to thank Charles L. Bevins and Renée M. Tsolis for helpful comments on the manuscript. Work in AJB’s laboratory is supported by Public Health Service grants AI040124, AI044170, AI076246 and AI079173. CT is supported by Scientist Development Grant 0835248N from the American Heart Association.


Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.


  • Akiyama H, Barger S, Barnum S, Bradt B, Bauer J, Cole GM, Cooper NR, Eikelenboom P, Emmerling M, Fiebich BL, et al. Inflammation and Alzheimer’s disease. Neurobiol Aging. 2000;21:383–421. [PMC free article] [PubMed]
  • Aliprantis AO, Yang RB, Mark MR, Suggett S, Devaux B, Radolf JD, Klimpel GR, Godowski P, Zychlinsky A. Cell activation and apoptosis by bacterial lipoproteins through toll-like receptor-2. Science. 1999;285:736–739. [PubMed]
  • Ausubel FM, Brent R, Kingston RE, Moore DD, Seidman JG, Smith JA, Struhl K, editors. Current protocols in molecular biology. J. Wiley & Sons; 1994.
  • Barnhart MM, Chapman MR. Curli biogenesis and function. Annu Rev Microbiol. 2006;60:131–147. [PMC free article] [PubMed]
  • Bian Z, Brauner A, Li Y, Normark S. Expression of and cytokine activation by Escherichia coli curli fibers in human sepsis. J Infect Dis. 2000;181:602–612. [PubMed]
  • Bian Z, Yan ZQ, Hansson GK, Thoren P, Normark S. Activation of inducible nitric oxide synthase/nitric oxide by curli fibers leads to a fall in blood pressure during systemic Escherichia coli infection in mice. J Infect Dis. 2001;183:612–619. [PubMed]
  • Chapman MR, Robinson LS, Pinkner JS, Roth R, Heuser J, Hammar M, Normark S, Hultgren SJ. Role of Escherichia coli curli operons in directing amyloid fiber formation. Science. 2002;295:851–855. [PMC free article] [PubMed]
  • Cheng N, He R, Tian J, Ye PP, Ye RD. Cutting edge: TLR2 is a functional receptor for acute-phase serum amyloid A. J Immunol. 2008;181:22–26. [PMC free article] [PubMed]
  • Cherny I, Rockah L, Levy-Nissenbaum O, Gophna U, Ron EZ, Gazit E. The formation of Escherichia coli curli amyloid fibrils is mediated by prion-like peptide repeats. J Mol Biol. 2005;352:245–252. [PubMed]
  • Collinson SK, Parker JM, Hodges RS, Kay WW. Structural predictions of AgfA, the insoluble fimbrial subunit of Salmonella thin aggregative fimbriae. J Mol Biol. 1999;290:741–756. [PubMed]
  • Cristofaro P, Opal SM. The Toll-like receptors and their role in septic shock. Expert Opin Ther Targets. 2003;7:603–612. [PubMed]
  • Fassbender K, Walter S, Kuhl S, Landmann R, Ishii K, Bertsch T, Stalder AK, Muehlhauser F, Liu Y, Ulmer AJ, et al. The LPS receptor (CD14) links innate immunity with Alzheimer’s disease. Faseb J. 2004;18:203–205. [PubMed]
  • Goodwin JL, Uemura E, Cunnick JE. Microglial release of nitric oxide by the synergistic action of beta-amyloid and IFN-gamma. Brain research. 1995;692:207–214. [PubMed]
  • Hammer ND, Schmidt JC, Chapman MR. The curli nucleator protein, CsgB, contains an amyloidogenic domain that directs CsgA polymerization. Proc Natl Acad Sci U S A. 2007;104:12494–12499. [PubMed]
  • He RL, Zhou J, Hanson CZ, Chen J, Cheng N, Ye RD. Serum amyloid A induces G-CSF expression and neutrophilia via Toll-like receptor 2. Blood. 2009;113:429–437. [PubMed]
  • Hu J, Ferreira A, Van Eldik LJ. S100beta induces neuronal cell death through nitric oxide release from astrocytes. Journal of neurochemistry. 1997;69:2294–2301. [PubMed]
  • Humphries AD, Raffatellu M, Winter S, Weening EH, Kingsley RA, Droleskey R, Zhang S, Figueiredo J, Khare S, Nunes J, et al. The use of flow cytometry to detect expression of subunits encoded by 11 Salmonella enterica serotype Typhimurium fimbrial operons. Mol Microbiol. 2003;48:1357–1376. [PubMed]
  • Jana M, Palencia CA, Pahan K. Fibrillar amyloid-beta peptides activate microglia via TLR2: implications for Alzheimer’s disease. J Immunol. 2008;181:7254–7262. [PMC free article] [PubMed]
  • Khan SA, Everest P, Servos S, Foxwell N, Zahringer U, Brade H, Rietschel ET, Dougan G, Charles IG, Maskell DJ. A lethal role for lipid A in Salmonella infections. Mol Microbiol. 1998;29:571–579. [PubMed]
  • Kilbourn RG, Jubran A, Gross SS, Griffith OW, Levi R, Adams J, Lodato RF. Reversal of endotoxin-mediated shock by NG-methyl-L-arginine, an inhibitor of nitric oxide synthesis. Biochemical and biophysical research communications. 1990;172:1132–1138. [PubMed]
  • Larsen P, Nielsen JL, Dueholm MS, Wetzel R, Otzen D, Nielsen PH. Amyloid adhesins are abundant in natural biofilms. Environ Microbiol. 2007;9:3077–3090. [PubMed]
  • Letiembre M, Liu Y, Walter S, Hao W, Pfander T, Wrede A, Schulz-Schaeffer W, Fassbender K. Screening of innate immune receptors in neurodegenerative diseases: A similar pattern. Neurobiol Aging 2007 [PubMed]
  • LeVine H., 3rd Thioflavine T interaction with synthetic Alzheimer’s disease beta-amyloid peptides: detection of amyloid aggregation in solution. Protein Sci. 1993;2:404–410. [PubMed]
  • Liu Y, Walter S, Stagi M, Cherny D, Letiembre M, Schulz-Schaeffer W, Heine H, Penke B, Neumann H, Fassbender K. LPS receptor (CD14): a receptor for phagocytosis of Alzheimer’s amyloid peptide. Brain. 2005;128:1778–1789. [PubMed]
  • Meda L, Cassatella MA, Szendrei GI, Otvos L, Jr, Baron P, Villalba M, Ferrari D, Rossi F. Activation of microglial cells by beta-amyloid protein and interferon-gamma. Nature. 1995;374:647–650. [PubMed]
  • Michelitsch MD, Weissman JS. A census of glutamine/asparagine-rich regions: implications for their conserved function and the prediction of novel prions. Proc Natl Acad Sci U S A. 2000;97:11910–11915. [PubMed]
  • Nakamura Y, Si QS, Kataoka K. Lipopolysaccharide-induced microglial activation in culture: temporal profiles of morphological change and release of cytokines and nitric oxide. Neuroscience research. 1999;35:95–100. [PubMed]
  • Olsen A, Jonsson A, Normark S. Fibronectin binding mediated by a novel class of surface organelles on Escherichia coli. Nature. 1989;338:652–655. [PubMed]
  • Otzen D, Nielsen PH. We find them here, we find them there: Functional bacterial amyloid. Cell Mol Life Sci 2007 [PubMed]
  • Perutz MF, Pope BJ, Owen D, Wanker EE, Scherzinger E. Aggregation of proteins with expanded glutamine and alanine repeats of the glutamine-rich and asparagine-rich domains of Sup35 and of the amyloid beta-peptide of amyloid plaques. Proc Natl Acad Sci U S A. 2002;99:5596–5600. [PubMed]
  • Petros A, Bennett D, Vallance P. Effect of nitric oxide synthase inhibitors on hypotension in patients with septic shock. Lancet. 1991;338:1557–1558. [PubMed]
  • Raffatellu M, Chessa D, Wilson RP, Dusold R, Rubino S, Bäumler AJ. The Vi capsular antigen of Salmonella enterica serotype Typhi reduces Toll-like receptor-dependent interleukin-8 expression in the intestinal mucosa. Infect Immun. 2005;73:3367–3374. [PMC free article] [PubMed]
  • Richard KL, Filali M, Prefontaine P, Rivest S. Toll-like receptor 2 acts as a natural innate immune receptor to clear amyloid beta 1-42 and delay the cognitive decline in a mouse model of Alzheimer’s disease. J Neurosci. 2008;28:5784–5793. [PubMed]
  • Rolan HG, Tsolis RM. Mice lacking components of adaptive immunity show increased Brucella abortus virB mutant colonization. Infect Immun. 2007;75:2965–2973. [PMC free article] [PubMed]
  • Roux CM, Rolan HG, Santos RL, Beremand PD, Thomas TL, Adams LG, Tsolis RM. Brucella requires a functional Type IV secretion system to elicit innate immune responses in mice. Cell Microbiol. 2007;9:1851–1869. [PubMed]
  • Somerville JE, Jr, Cassiano L, Bainbridge B, Cunningham MD, Darveau RP. A novel Escherichia coli lipid A mutant that produces an antiinflammatory lipopolysaccharide. J Clin Invest. 1996;97:359–365. [PMC free article] [PubMed]
  • Stojiljkovic I, Bäumler AJ, Heffron F. Ethanolamine utilization in Salmonella typhimurium: nucleotide sequence, protein expression and mutational analysis of the cchA cchB eutE eutJ eutG eutH gene cluster. J Bacteriol. 1995;177:1357–1366. [PMC free article] [PubMed]
  • Takeuchi O, Hoshino K, Kawai T, Sanjo H, Takada H, Ogawa T, Takeda K, Akira S. Differential roles of TLR2 and TLR4 in recognition of gram-negative and gram-positive bacterial cell wall components. Immunity. 1999;11:443–451. [PubMed]
  • Town T, Nikolic V, Tan J. The microglial “activation” continuum: from innate to adaptive responses. J Neuroinflammation. 2005;2:24. [PMC free article] [PubMed]
  • Tükel C, Raffatellu M, Humphries AD, Wilson RP, Andrews-Polymenis HL, Gull T, Figueiredo JF, Wong MH, Michelsen KS, Akcelik M, et al. CsgA is a pathogen-associated molecular pattern of Salmonella enterica serotype Typhimurium that is recognized by Toll-like receptor 2. Mol Microbiol. 2005;58:289–304. [PubMed]
  • Udan ML, Ajit D, Crouse NR, Nichols MR. Toll-like receptors 2 and 4 mediate Abeta(1-42) activation of the innate immune response in a human monocytic cell line. Journal of neurochemistry 2007 [PubMed]
  • Udan ML, Ajit D, Crouse NR, Nichols MR. Toll-like receptors 2 and 4 mediate Abeta(1-42) activation of the innate immune response in a human monocytic cell line. Journal of neurochemistry. 2008;104:524–533. [PubMed]
  • Wang X, Smith DR, Jones JW, Chapman MR. In vitro polymerization of a functional Escherichia coli amyloid protein. The Journal of biological chemistry. 2007;282:3713–3719. [PMC free article] [PubMed]
  • Weening EH, Barker JD, Laarakker MC, Humphries AD, Tsolis RM, Bäumler AJ. The Salmonella enterica serotype Typhimurium lpf, bcf, stb, stc, std, and sth fimbrial operons are required for intestinal persistence in mice. Infect Immun. 2005;73:3358–3366. [PMC free article] [PubMed]
  • Weighardt H, Holzmann B. Role of Toll-like receptor responses for sepsis pathogenesis. Immunobiology. 2007;212:715–722. [PubMed]
  • Wilson RP, Raffatellu M, Chessa D, Winter SE, Tukel C, Baumler AJ. The Vi-capsule prevents Toll-like receptor 4 recognition of Salmonella. Cell Microbiol. 2008;10:876–890. [PubMed]
  • Zogaj X, Bokranz W, Nimtz M, Romling U. Production of cellulose and curli fimbriae by members of the family Enterobacteriaceae isolated from the human gastrointestinal tract. Infect Immun. 2003;71:4151–4158. [PMC free article] [PubMed]