www.nature.com/npjbiofilms ARTICLE OPEN The biofilm-associated surface protein Esp of Enterococcus faecalis forms amyloid-like fibers Agustina Taglialegna 1,5, Leticia Matilla-Cuenca 1,2,5 , Pedro Dorado-Morales 2 , Susanna Navarro 3 , Salvador Ventura 3 , 1✉ James A. Garnett4, Iñigo Lasa 2 and Jaione Valle Functional amyloids are considered as common building block structures of the biofilm matrix in different bacteria. In previous work, we have shown that the staphylococcal surface protein Bap, a member of the Biofilm-Associated Proteins (BAP) family, is processed and the fragments containing the N-terminal region become aggregation-prone and self-assemble into amyloid-like structures. Here, we report that Esp, a Bap-orthologous protein produced by Enterococcus faecalis, displays a similar amyloidogenic behavior. We demonstrate that at acidic pH the N-terminal region of Esp forms aggregates with an amyloid-like conformation, as evidenced by biophysical analysis and the binding of protein aggregates to amyloid-indicative dyes. Expression of a chimeric protein, with its Esp N-terminal domain anchored to the cell wall through the R domain of clumping factor A, showed that the Esp N-terminal region is sufficient to confer multicellular behavior through the formation of an extracellular amyloid-like material. These results suggest that the mechanism of amyloid-like aggregation to build the biofilm matrix might be widespread among BAP-like proteins. This amyloid-based mechanism may not only have strong relevance for bacteria lifestyle but could also contribute to the 1234567890():,; amyloid burden to which the human physiology is potentially exposed. npj Biofilms and Microbiomes (2020)6:15 ; https://doi.org/10.1038/s41522-020-0125-2 INTRODUCTION machinery that controls the expression, secretion and polymer- In the last two decades, an increasing amount of studies have ization of protein subunits. A more simplistic type of amyloid- shown that amyloid structures are not restricted to protein like material is the one formed by the biofilm-associated protein aggregation disorders but also, they can be found as functional (Bap) of S. aureus14. This surface protein undergoes partial amyloids for a number of relevant biological processes in proteolytic cleavage that releases fragments containing the N- eukaryotic cells, fungi, and bacteria1,2. In bacteria, one of the terminal region of Bap only under specific environmental most predominant processes in which functional amyloids have conditions14. This region has a molten globule-like conformation a biological role is in the context of multicellular behavior and that, when the pH becomes acidic, switches to a β-sheet-rich biofilm formation. Biofilms are the most common bacterial conformation and polymerizes to form amyloid-like fibers. The growth in natural environments, medical, and industrial N-terminal region of Bap contains two EF-hand calcium-binding settings3. In biofilms, bacteria are embedded in an extracellular motifs. Binding of calcium to the EF-hand motifs stabilizes the matrix that provides structural integrity, protects bacteria from molten globule conformation. This prevents β-sheet transition the action of the immune system and antimicrobial treatments, and the subsequent self-assembly into amyloid fibers, hindering favors nutrient adsorption and keeps extracellular enzymes biofilm formation. Therefore, inactivation of the EF-hand close to the cells4. Amyloids are especially well suited for calcium-binding domains by site-directed mutagenesis did not assembling the biofilm matrix in the extracellular environment affect biofilm formation in the presence of calcium14,15. because they provide a stable chassis for its inherent resistance The enterococcal surface protein Esp is a Bap-orthologous against protease digestion and denaturation5. Amyloids can also protein16,17. Esp is a high molecular weight protein of 1873 amino modify the viscoelastic properties of the extracellular matrices acids. It contains well-differentiated modules including the N- helping biofilms to adapt and respond to environmental terminal domain, sharing 26% sequence identity with Bap; the fluctuations6. Therefore, proteins with amyloid or amyloid-like central core region with two domains (A and B) and a series of conformation have been proven widespread in biofilms of many tandem repeats sharing 23% sequence identity with C-repeats of unrelated bacteria7. The most studied and characterized Bap; and the C-terminal domain that includes the LPXTG-like motif bacterial amyloid system is the curli fimbriae present in found in surface-associated proteins. Esp contributes to biofilm Escherichia coli and in other enterobacterial species, which is formation by E. faecalis. It was demonstrated that the minimal composed of the major structural subunit CsgA8,9. Other region contributing to Esp-mediated biofilm enhancement was examples of biofilm-related amyloids are FapC in Pseudomo- confined to the N-terminal domain17. However, the molecular nas10, TasA in Bacillus subtilis11, chaplins in Streptomyces mechanisms by which the N-terminal region of Esp performs this coelicolor12 or the phenol-soluble modulins (PSMs) in the function is unknown. Based in our previous results obtained with pathogenic bacteria Staphylococcus aureus13. In all these Bap of S. aureus, we hypothesize that these proteins could examples, functional amyloid assembly involves specialized mediate intercellular connection through a common molecular 1 Instituto de Agrobiotecnología (IDAB), CSIC-UPNA-Gobierno de Navarra, Avenida Pamplona 123, Mutilva-31192, Navarra, Spain. 2Navarrabiomed-Universidad Pública de Navarra- Departamento de Salud, IDISNA, Pamplona-31008, Navarra, Spain. 3Institut de Biotecnologia i de Biomedicina and Departament de Bioquimica i Biologia Molecular, Universitat Autonoma de Barcelona, Bellaterra, Spain. 4Centre for Host Microbiome Interactions, Dental institute, King’s College London, London, UK. 5These authors contributed equally: Agustina Taglialegna, Leticia Matilla-Cuenca. ✉email: jaione.valle@csic.es Published in partnership with Nanyang Technological University A. Taglialegna et al. 2 a Bap A B C-repeats COOH 1 361 819 2148 2276 26% identity 23% identity 20% Esp N-term A B C-repeats COOH 1 743 995 1064 1648 1873 b * * * * * * * 1 * * * * * Normalized aggregation propensity * 0.75 TANGO Aggrescan 0.5 WALTZ FoldAmyl 0.25 0 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 Amino acid (x100) Fig. 1 Prediction of the amyloidogenic potential of Esp. a Structural similarity between Bap and Esp. The N-terminal domain of Esp (residues 49–743) and region B of Bap (residues 361–819) share 26% sequence identity. The C-repeats of Bap (amino acids 820–2147) and C-repeats of Esp (amino acids 1064–1648) share 23% sequence identity. b Comparison of the predicted aggregation propensities. Bars representing the 1234567890():,; aggregation propensities obtained by four bioinformatic algorithms TANGO (violet), Aggrescan (orange), WALTZ (black), and FoldAmyloid (Blue). Values have been normalized for each predictor. Asterisks denote the amino acid stretches predicted to be amyloidogenic by at least three of the algorithms used. mechanism of amyloid-like conformation. By using biophysical Table 1. Amino acid sequences corresponding to the predicted in vitro assays, cell-based and microscopic approaches and dye- amyloid peptides (marked with the asterisks in Fig. 1). binding analyses, we demonstrated that the N-terminal domain of Esp formed aggregates with amyloid-like properties when the pH Amino acids Sequence of the media became acidic. We further showed that the N- terminal domain of Esp is functional as it was able to induce 21–49 SVGVASVLVGVGLVFGATGIVN multicellular behavior in bacteria lacking the esp gene through the 158–174 TNIWKVNYIRANDGLFA formation of amyloid-like material, both when it was exogenously 211–221 RLMYRIYLVSH added to the culture and when it was expressed from a plasmid. 231–239 IESTGYFLT Overall, our results suggest that the mechanism of amyloid-like 323–332 YVSGIQMHMV aggregation might be a widespread mechanism of the BAP-like 440–448 TYGTGYYYLQ proteins to build the biofilm matrix. 667–686 NDAFSVLDVY 698–705 NTGIVTFT 1050–1058 KGTVVVTYSD RESULTS 1685–1694 GTMYFWKEKV Computational algorithms predict Esp aggregation tendency 1701–1711 NKKATVVVIYP The enterococcal surface protein Esp possesses high structural 1718–1726 VEVVISVVD similarities to the staphylococcal Bap. The N-terminal domain 1843–1865 VDSNIYTIAGGLILLGTLGLLGY (Esp_N) of Esp shows 26% identity with the N-terminal domain of Bap (Fig. 1a). Bearing in mind that Bap adopts an amyloid conformation to mediate biofilm formation, we have used four existing computational algorithms commonly used to predict the Esp_N expressed with the curli-dependent amyloid generator (C- amyloidogenic propensity of a protein. The selected predictors, DAG) system forms amyloid fibrils including WALTZ18, AGGRESCAN19, TANGO20, and FoldAmyloid21, To explore the amyloid-forming propensity of Esp_N, we used the rely on different principles (β-sheet formation, aggregation C-DAG system22. Amino acids 67–511 from Esp were cloned into propensity, hydrogen-bonding interactions, and expected packa- the pExport plasmid that allows the expression of the epitope of ging density). The aggregation properties predicted by these interest fused to the first 42 residues of the signal sequence of algorithms for Esp were normalized and they were compared in CsgA (ssCsgA) and the 6xHistidine tag (Fig. 2a). We also cloned Fig. 1b. All the algorithms identified a high number of amyloid- into the pExport plasmid a C-repeat of Esp (amino acids prone segments at the Esp_N region (Fig. 1b and Table 1 and 1148–1228), which showed no amyloidogenic-prone region Supplementary Fig. 1A). The alignment of peptides predicted to prediction on it, and one A-repeat and part of the B-domain be amyloidogenic by at least 3 of the 4 algorithms showed no (amino acids 913–1060) that possess only one aggregation-prone clear conserved domains among them (Supplementary Fig. 1B), peptide (Fig. 1a). Besides, plasmids carrying the S. aureus Bap_B or suggesting that the presence of aggregation-prone regions rather Bap_A regions fused to ssCsgA were included as positive and than amino acid motifs is conserved at the N-terminus of Esp. negative controls, respectively (Fig. 1a). The resulting plasmids npj Biofilms and Microbiomes (2020) 15 Published in partnership with Nanyang Technological University A. Taglialegna et al. 3 a b c Esp_N Esp_C Esp_AB d 1.5 ** ** OM * PS CsgG CsgG Bap_B Bap_A OD500 1 CM is is 0.5 PBAD 6xH PBAD 6xH amyloid non-amyloid sscsgA sscsgA 0 p_ N p_ C AB p_ B p_ A Es Es p_ Ba Ba Es e f g Esp_N Esp_N Esp_N h i j k Esp_AB Bap_B Bap_A Esp_C Fig. 2 Expression of Esp_N using the curli-dependent amyloid generator (C-DAG) system. a Representative scheme of C-DAG. Potential amyloid domains are expressed under the control of the inducible PBAD promoter, which allows the expression of the epitope fused to the first 42 residues of the signal sequence of CsgA. CsgG is expressed under the control of an IPTG-inducible promoter. Outer membrane (OM), cytoplasmic membrane (CM), periplasmic space (PS). b Variation in colony color phenotype of E. coli cells exporting the N-terminal domain of Esp (Esp_N), a C-repeat (Esp_C), and AB-repeats (Esp_AB) on agar supplemented with CR. E. coli expressing Bap_B and Bap_A from S. aureus was used as positive and negative controls respectively. c CR solubilized using ethanol–acetone from E. coli cells expressing Esp_N, Esp_C, Esp_AB, Bap_B, and Bap_A. d Quantification of CR. Absorbance was measured at 500 nm. Bars represent the standard deviations of the results of six independent experiments (n = 6). Statistically significant differences were determined using Mann–Whitney test *p < 0.05, **p < 0.01. Transmission electron micrographs of negatively stained fiber-like structures formed by E. coli cells that express Esp_N e, f and Bap_B h. E. coli cells that express Bap_A i, Esp_C j and Esp_AB k did not show the presence of fibers as shown by transmission electron microscopy. Immunogold labeled of fiber-like structures of samples from E. coli cells that express Esp_N using anti-His antibodies g. Scale bar of panels e, j, k represents 500 nm. Scale bar of panels f, h, i represents 200 nm. Scale bar of panel g represents 100 nm. were introduced in the carrier E. coli VS39 strain, which is deficient 2c). Results showed that E. coli strain that expressed Esp_N was in the curli system but expresses csgG gene that is required for the able to bind significantly higher levels of CR in comparison to E. protein export. In the presence of arabinose, chimeric proteins are coli strain that express Esp_C and Esp_AB (Fig. 2d). Interestingly, expressed and subsequently guided by the ssCsgA epitope at the electron microscopy analysis of the E. coli strain exporting the extracellular space through a pore-like structure that is formed by Esp_N domain revealed the presence of extracellular dense the CsgG protein in the outer membrane (Fig. 2a). The presence of aggregates surrounding the cells, that are most likely fibrous extracellular amyloid-like aggregates was detected by analyzing (Fig. 2e, f), and that specifically reacted with gold-labeled anti-His the capacity of the strains to bind Congo Red dye (CR). As shown antibody (Fig. 2g and supplementary Fig. 2). Nonspecific binding in Fig. 2b, E. coli cells expressing Esp_N domain were able to bind was determined in the presence of the secondary gold-labeled CR as the strain producing Bap_B of S. aureus, while no binding to antibody (Supplementary Fig. 2). Similar structures were also CR was observed in E. coli strains expressing Bap_A, Esp_C, and Esp_AB chimeras. We also quantified the CR attached to the cells observed in E. coli cells exporting Bap_B (Fig. 2h) but not in cells using a quantitative CR-binding assay. After induction of the that expressed Bap_A, Esp_AB, and Esp_C chimeras (Fig. 2i, j, k). proteins using the C-DAG system, E. coli cells were incubated in a Taken together these results showed that the N-terminal domain solution containing CR. The colorant bound to the amyloids-like of Esp expressed in a cell-based aggregation system formed material was solubilized from the cells using ethanol–acetone (Fig. amyloid-like aggregates. Published in partnership with Nanyang Technological University npj Biofilms and Microbiomes (2020) 15 A. Taglialegna et al. 4 a pH: 1610–1630 cm−1 range of the β-strand region of the amide I 3.6 3.8 4.0 spectrum, which is attributed to amyloid-like β-strand interactions b pH (Fig. 4e). Deconvolution of the spectra showed that the 4.2 4.2 majoritarian band of rEsp_N aggregates was detected at 1620–1630 cm−1, which is typically assigned to the presence of inter-molecular β-sheets that contributes 38.5% of the area (Fig. pH: 4.2 4.4 4.6 4f). The dominated structure of the soluble Esp_N (neutral pH) spectra is detected at 1640 cm−1 assigned to the presence of intra-molecular β-sheets covering 37.91% of the area. The second majoritarian structures appear at 1640 cm−1, covering 33.9% of 4.2 7.0 the area and at 1620–1630 cm−1 that contributes 29.01% of the area for the Esp_N aggregates and the soluble protein, pH: 4.8 5.0 5.4 respectively. A representation of the second derivative values, which are necessary for the deconvolution method of the bands, is shown in Supplementary Fig. 4. Overall, our data are consistent with the assembly of Esp into supramolecular β-sheet-enriched aggregates. pH: 6.0 6.4 6.8 c To confirm that the detected β-sheet-enriched aggregates display amyloid-like properties, we analyzed if rEsp_N aggregates could bind amyloid-indicative dyes, such as CR, Thioflavin-T (Th-T), and Proteostat. CR binding to rEsp_N aggregates was determined by analyzing the absorbance recorded from 380 to 680 nm. As it is pH: 7.0 7.5 8.0 characteristic for amyloid proteins, incubation of CR in the presence of rEsp_N aggregates formed at pH 4.2 resulted in a red shift of its spectrum compared to the protein incubated at pH 7 and the CR buffer. Moreover, the reversion of the aggregates at pH 7 displayed no red shift in the spectrum (Fig. 5a). Th-T fluorescence emission was enhanced in the presence of rEsp_N aggregates at the intensity of Th-T fluorescence spectral Fig. 3 Recombinant rEsp_N protein forms reversible aggregates maximum of 488 nm. In contrast, no change in Th-T fluorescence in acidic phosphate–citrate buffer. a 2 μM of purified rEsp_N protein was incubated in phosphate buffer solutions at different pH was observed when the protein was incubated at pH 7.0 (Fig. 5b). values. Aggregates started to be visible in buffers with pH 4 and 4.2 Finally, staining with Proteostat showed that the dye was (indicated by a red arrow). b Reversion assay showed complete significantly concentrated at the rEsp_N aggregates (Fig. 5c). disassembly after phosphate–citrate buffer exchange from pH 4.2 to The fact that rEsp_N aggregates bind all these amyloid-indicative pH 7. c Scanning electron microscopy of rEsp_N aggregates. Scale dyes confirmed the amyloid-like features of Esp. bar represents 1 µm. In addition, we tested the aggregation kinetics of rEsp_N at pH 4.2 and pH 7 by measuring Th-T fluorescence signal over time. We observed a very high Th-T signal from the very beginning of the Recombinant Esp_N assembles into β-sheet-rich structures with reaction when the protein was incubated in acidic buffer. The amyloidogenic properties at acidic pH rapid aggregation of rEsp_N gave a kinetics missing the lag-phase To study the biophysical properties of the N-terminal domain of that is typical of the sigmoidal aggregation kinetics of the amyloid Esp we purified a recombinant Esp_N domain (rEsp_N) (Supple- proteins (Fig. 5d). Taking together all these results showed that mentary Fig. 3). rEsp_N protein was incubated in the N-terminal domain of Esp self-assembled in structures with phosphate–citrate buffer solutions with different pH. As it was amyloidogenic properties at acidic pH. previously observed for Bap_B14, rEsp_N formed aggregates at acidic pH (pH 4.0–4.2) (Fig. 3a). At pH 4.2, rEsp_N formed a subtle but visible ring of protein adhered to the walls of the tube. Heterologous expression of Esp_N induces multicellular behavior Interestingly, the process was reversible and rEsp_N aggregates through the formation of fibrillar structures dissociated completely when the pH raised to 7 (Fig. 3b). Analysis We generated a chimeric protein comprising the Esp_N domain of the structure of the rEsp_N aggregates by scanning electron (amino acids 50–743) tagged with the 3xFLAG amino acid microscopy showed the formation of a dense mesh-like network sequence and linked to the R domain of the clumping factor A (Fig. 3c). (R_ClfA) that contains the LPXTG motif for anchoring to the We determined the relative size of rEsp_N aggregates by static light peptidoglycan. The chimeric protein was expressed from the scattering (Fig. 4a), showing the presence of significantly big pCN51 vector under the control of the Pcad-cadC promoter. This aggregates of rEsp_N at pH 4.2 that completely disassembled when plasmid was transformed into the biofilm negative S. aureus Δbap buffer was changed to pH 7. We also showed that these self- strain23. As a control, we expressed the entire Esp protein from assembled rEsp_N aggregates bound strongly to bis-ANS dye pCN51 plasmid. The localization of both Esp_N-R_ClfA chimeric evidencing their exposure of hydrophobic patches to solvent (Fig. 4b). protein and full Esp on the bacterial cell wall was confirmed by In order to test whether the aggregates formed by rEsp_N have immunofluorescence using anti-esp antibodies (Fig. 6a). As shown amyloid-like properties, we analyzed the secondary structure in Fig. 6b, bacterial cells expressing the Esp chimeric protein content by far-UV circular dichroism (CD) of the rEsp_N at pH 4.2 induced cell aggregation (or cell clumping) and biofilm formation and pH 7. Results showed that rEsp_N domain showed a similarly to those cells expressing the whole Esp protein. This preponderance of β-sheet content at acidic pH (Fig. 4c, d). We result confirmed that the N-terminal domain of Esp is sufficient to also analyzed the amide I region of the attenuated total encourage cell aggregation, similar to the full Esp. reflectance-Fourier transform infrared spectroscopy (ATR-FTIR) According to the in vitro assays with the rEsp_N protein, Esp- spectrum in the range 1700–1600 cm−1 of the rEsp_N aggregates mediated multicellular behavior was only observed under acidic (acidic pH) and its soluble state (neutral pH). The overall analysis of conditions, when bacteria were grown in media supplemented both spectra showed a left shift absorbance in the with glucose, TSB-gluc (Fig. 6b, c), or LB-gluc (Supplementary npj Biofilms and Microbiomes (2020) 15 Published in partnership with Nanyang Technological University A. Taglialegna et al. 5 a 600 b2000 c pH 7.0 pH 7.0 1 pH 4.2 pH 4.2 500 pH 7.0 pH 7.0 pH 7.0 pH 7.0 0 Fluorescence (a.u) Ligth scattering pH 4.2 pH 7.0 pH 4.2 pH 7.0 1500 buffer -1 400 [O]mdeg pH 7 -2 300 1000 pH 4.2 -3 200 -4 500 100 -5 -6 0 0 190 200 210 220 230 240 250 260 320 325 330 335 340 400 450 500 550 600 Wavelength (nm) Wavelength (nm) Wavelength (nm) d pH 4.2 pH 7 f 0.015 0.016 pH 4.2 0.014 pH 7 α helix (%) 8 6 Absorbance 0.012 β strand (%) 43 36 Absorbance Turn (%) 22 14 0.01 0.01 Coil (%) 25 35 0.008 NRMSD 0.008 0.022 0.006 0.005 0.004 e 0.016 0.002 pH 7.0 0.014 0 pH 4.2 0 0.012 1580 1620 1660 1700 1580 1620 1660 1700 1740 Absorbance 0.01 Wavenumber cm-1 Wavenumber cm-1 0.008 0.006 Bands (cm-1) % area Structure Bands (cm-1) % area Structure 0.004 1622.14 38.58 β strand (inter) 1622.96 29.014 β strand (inter) 0.002 1640.39 33.90 β strand (intra) 1639.61 37.917 β strand (intra) 0 1657.96 20.20 helix/loop 1656.75 23.644 helix/loop 1675.83 7.32 β turn 1673.58 9.423 β turn 1550 1600 1650 1700 1750 1800 Wavenumber cm-1 Fig. 4 Amyloid features of rEsp_N. a Synchronous light scattering of 0.1 mg/ml rEsp_N samples in phosphate–citrate buffer either at pH 7.0, 4.2, and pH-reverted samples to pH 7.0. b Bis-ANS fluorescence of 0.1 mg/ml rEsp_N at pH 7.0, pH 4.2, and pH-reverted samples to pH 7.0. Free bis-ANS is represented in black. c Far-UV CD spectra of 0.05 mg/ml rEsp_N at pH 7 and pH 4.2. d Secondary structure composition of rEsp aggregates. Data were analyzed with Dichroweb implementing the CDSSTR algorithm. e FTIR spectra in the amide I region of rEsp_N aggregates formed after 24 h in acidic buffer (blue line) and rEsp_N in its soluble state (black line). f Deconvolution of the FTIR spectra of the rEsp_N at pH 4.2 and at pH 7. Component bands after gaussian deconvolution are shown. Percentage of each secondary structure is detailed in the tables. Fig. 5), but not in LB without glucose (pH 7) (Supplementary Specifically, we extracellularly added rEsp_N to a bacterial culture Fig. 5). Finally, we determined by transmission electron micro- of the E. faecalis 23 strain that lack the esp gene and therefore is scopy (TEM) the presence of extracellular material in the cells unable to induce bacterial aggregation. As a control, we also used complemented with the chimeric Esp_N-R_ClfA protein. S. aureus the S. aureus Δbap strain. As shown in Fig. 7a, the addition of Δbap strain expressing the Esp_N-R_ClfA chimera formed a rEsp_N promoted bacterial aggregation in both strains (Fig. 7a). By fibrous-like structures that connected cells together (Fig. 6e) and using immunofluorescence labeling with anti-His antibodies we that were not observed in the Δbap cells (Fig. 6d). Moreover, were able to detect rEsp_N protein in the clusters formed by both extracellular material positively reacted with gold-labeled anti-Esp of E. faecalis and the S. aureus Δbap strains (Fig. 7b). Finally, we antibodies confirming that amyloids fibers are formed by Esp_N- analyzed the cell aggregates by SEM and TEM. We detected R_ClfA protein (Fig. 6f). Similarly, extracellular fibrous-like struc- extracellular tangles of fiber-like material on the cell surface of E. tures that reacted with gold-labeled anti-Esp antibodies were faecalis 23 connecting and attaching cells together (Fig. 7c, d). All detected in the E. faecalis 11279 strain that expresses Esp protein these results showed that rEsp_N is biologically active and induce (Fig. 6h, i) and not in its derivative Esp mutant (Fig. 6g). Taken cell-to-cell aggregation in E. faecalis. together all these results showed that the N-terminal domain of Esp is able to mediate multicellular behavior through the formation of extracellular fibrous-like structures at acidic pH. DISCUSSION Amyloid fibers have been recently discovered as the main group Exogenous addition of rEsp_N to esp-negative bacteria induces of proteins composing the extracellular matrix of many unrelated cell-to-cell aggregation bacteria5. Two different types of biofilm-related amyloids machi- Purified recombinant amyloid-like proteins usually have biological neries are described so far7. The first group consists in intrinsic activity, since they yield multicellular behavior phenotypes bacterial amyloids, where the amyloid state is the primary undistinguishable from those formed by the wild-type strain functional form of the proteins. Examples of amyloids of this when they are exogenously added to biofilm-deficient groups are: curli (csgACB-csgDEFG) in enterobacteria25,26, Fap strains11,14,24. To validate the functionality of rEsp_N, we (fapA-F) in Pseudomonas10 and chaplins in S. coelicolor24. The performed an exogenous complementation assay that consisted second type are facultative bacterial amyloid-like proteins where in determining the capacity of rEsp_N to induce bacterial the protein adopts a functional globular folded structure and clumping when exogenously added to a bacterial culture. depending on proteolytic processing and environmental triggers, Published in partnership with Nanyang Technological University npj Biofilms and Microbiomes (2020) 15 A. Taglialegna et al. 6 a 1.2 b 120 pH 7.0 pH 7.0 Th-T fluorescence (a.u) pH 4.2 pH 4.2 1 pH 4.2 pH 7 100 pH 7.0 pH 7.0 CR buffer pH 4.2 Absorbance pH 7.0 0.8 80 buffer 0.6 60 0.4 40 0.2 20 0 0 400 440 480 520 560 600 640 680 460 480 500 520 540 560 580 600 nm nm d c 400 Th-T fluorescence (a.u.) Fluorescence 350 pH 7.0 300 pH 4.2 Trans 250 buffer 200 150 100 50 0 0 300 600 900 1200 1500 1800 Time (s) Fig. 5 Binding of rEsp_N to amyloid-indicative dyes. a Shift in CR absorbance spectrum upon binding of CR dye to 0.1 mg/ml rEsp_N at pH 7.0, pH 4.2, and pH-reverted samples to pH 7.0. Free CR absorbance spectrum is represented in black. b Increase in Th-T emission fluorescence upon binding to 0.1 mg/ml of rEsp_N at pH 7.0, pH 4.2, and pH-reverted samples to pH 7.0. c Fluorescence of rEsp_N aggregates stained with Proteostat dye. Samples were visualized using a ×100 oil immersion lens. Scale bar represents 10 µm. d Aggregation kinetics of rEsp_N at pH 4.2 (red) and at pH 7 (green), as monitored by changes in relative Th-T fluorescence signal. Free Th-T fluorescence is shown in black. it can also switch to an amyloid conformation. It seems that the indicator dyes are not sufficient to affirm that Esp adopts a cross-β facultative amyloids play a dual role in biofilm formation acting as arrangement typical of amyloids. adhesins in their native conformation and as matrix scaffolds Another coincidence between Esp and Bap is the fact that acidic when they polymerize into amyloid-like fibrillar structures. An pH conditions favor amyloid polymerization of both proteins. As example of this type of proteins is Bap of S. aureus. Bap is secreted occurs with Bap, the pH at which the N-terminal domain of Esp and covalently anchored to the cell wall. At its native conforma- shows aggregation activity (pH 4.2) is close to the isoelectric point of tion, Bap is involved in primary adhesion to an abiotic surface and the full protein (pI 4.5) and the isoelectric point of the Esp N-terminal host cell ligands27. Then, Bap is processed releasing to the media domain (pI 4.68) and, therefore, lack of net charge at this pH favors fragments containing the N-terminal region to the media, whilst the conversion of the protein from its native state to the amyloid- part of the C-terminal repeat region remains anchored to the like state. Previous work stablished that high glucose concentration membrane. When the pH becomes acid the N-terminal domain of (0.5% (w/v) or higher) in the growth media promoted Esp-mediated Bap switches from its partially ordered native state to an biofilm formation without affecting esp expression16,29. Our results aggregation-prone conformation that facilitates polymerization provide a possible explanation to this situation: an acidification of into amyloid-like fibrillar structures. the growth media as a consequence of glucose metabolism would The results presented here, propose Esp as a novel facultative favor Esp amyloid folding and aggregation, thus inducing biofilm amyloid-like protein. The prediction of potentially amyloidogenic formation without relevant changes in esp gene expression. Of regions using several computational algorithms converged to course, we cannot exclude the possibility that other undetermined indicate regions capable of adopting an amyloid conformation in factors might participate in this process and interact with Esp. the N-terminal domain of Esp (Esp_N). Esp undergoes a limited Another major point to be considered in the amyloidogenic proteolytic or spontaneous cleavage at certain conditions, as capacities of Esp and Bap, is the fact that Esp does not contain shown by western-blot using antibodies against the N-terminal any EF-hand motif in its N-terminal domain30. For Bap, binding of region of Esp (Supplementary Fig. 6)14,28. This processing would calcium to the EF-hand motifs is a key process that regulates the enable the release of the N-terminal domain of the protein that formation of amyloid-like fibers. In the presence of calcium, the self-assembles into amyloid-like fibrillar structures. However, we protein undergoes tertiary rearrangements that increase its stability are unable to determine which part of the N-terminal region of the in solution and thus decrease its aggregation behavior. In case of protein are releasing the processed peptides that form part of the Esp, this process might be triggered by other metal ions or factors amyloid-like fibers. It is important to notice that sortase-mediated that require further investigation. anchoring of Esp is not strictly required for multicellular Bap and Esp belongs to the biofilm-associated protein (BAP) aggregation, since extracellular addition of rEsp_N-purified family. Members of this family have been identified in many protein lacking the LPXTG motif is biologically active and induce unrelated bacteria including BapA of Salmonella enterica ser. cell-to-cell aggregation. Although were able to detect extracellular typhimurium31, Bap of Acinetobacter baumannii32, LapA of fibrous dense aggregates in E. coli cells exporting the Esp_N Pseudomonas putida33, BapL of Listeria monocytogenes34, Bap of domain by electron microscopy analysis, we did not prove in a Burkholderia cepacia35, and BpfA of Shewanella oneidensis36. All the truthful way that they are canonical amyloid fibers. Secondary BAP proteins share several structural and functional features structure data (CD spectra and FTIR) and binding to amyloid (Supplementary Fig. 7). They are high-molecular-weight proteins npj Biofilms and Microbiomes (2020) 15 Published in partnership with Nanyang Technological University A. Taglialegna et al. 7 a Fluorescence DAPI DIC Merge b O Esp_N Esp Esp Aggregation Esp_N c Biofilm d e f g h i Fig. 6 Expression of Esp N-terminal region is sufficient to mediate biofilm phenotype in a heterologous strain. a Inmunofluorescence showing surface localization of Esp. Bacteria were fixed and labeled with anti-Esp antibodies and DAPI. Scale bars represent 3 µm. b S. aureus Δbap expressing chimeric N-terminal domain of Esp (Esp_N) or the entire protein (Esp) forms bacterial clumps (arrows) in overnight cultures grown in TSB-gluc under shaken conditions (200 rpm) at 37 °C. Ø refers to cells complemented with pCN51 empty plasmid. Images of the bottom part of the tubes are shown (bottom panel). c Biofilm formation of S. aureus Δbap cells that express the N-terminal domain of Esp (Esp_N) or the full protein (Esp). For biofilm formation, bacteria were cultured overnight at 37 °C in microtiter plates under static conditions. Electron micrographs of negatively stained S. aureus Δbap d and S. aureus Δbap expressing Esp_N e. f Immunogold-labeled samples of fibers formed by amyloid Esp_N. h, i Immunogold-labeled electron micrographs of E. faecalis 11279 (Esp positive) and its derivative Esp mutant g. Scale bar of panels d, f, g represents 500 nm. Scale bar of panels e, h, i represents 200 nm. that contain a core domain of tandem repeats. They play a amyloid domain has been proposed to mediate protein–protein relevant role in bacterial infectious processes and confer upon interactions and the tertiary structure of Bap in S. epidermidis. bacteria the capacity to form a biofilm37. Based on the functional Although the peptide is repeated 17 times in the protein Bap of similarities found between Esp and Bap, it is tempting to speculate the S. epidermidis strain C533, the peptide is present only one time that other members of the BAP family will also possess amyloid- in the Bap sequence of S. aureus42. Our analysis confirmed the like properties. In fact, the prediction of potentially amyloidogenic absence of amyloid-like domains in the central repeat regions of regions using WALTZ algorithm indicated that other BAP Bap of S. aureus and Esp of E. faecalis. Therefore, the presence of members might be capable of forming amyloid-like structures several aggregation-prone regions in a domain or the amyloido- from the N-terminal domain, in the case of BapL of L. genic peptides derived of the repeating units of a protein would monocytogenes and from the repeat units, in the case of Bap of be sufficient to promote an amyloid-like conformation of the BAP A. baumannii and BapA from S. enteritidis (Supplementary Fig. 7). It is known that repeating units may contain peptides with family of proteins. However, the putative amyloidogenic behavior amyloidogenic behavior38–40. In a previous work Lembré et al.41, of all these domains needs to be validated and will be the object demonstrated that a peptide derived from the repeats of the Bap of further research that will dig in the mechanistic insight to the homolog protein of S. epidermidis formed amyloid fibers. This likely aggregation mechanism of BAP family proteins. Published in partnership with Nanyang Technological University npj Biofilms and Microbiomes (2020) 15 A. Taglialegna et al. 8 a E. faecalis 23 S. aureus bap b Fluorescence DAPI DIC Merge rEsp_N O rEsp_N O E. faecalis 23 O E. faecalis 23 rEsp_N E. faecalis 23 c rEsp_N S. aureus bap d S. aureus bap rEsp_N Fig. 7 Exogenous complementation of Esp-negative strains with purified rEsp_N. a Bacterial clumping of overnight cultures of esp-negative E. faecalis 23 and S. aureus Δbap grown under shaken conditions (200 rpm) at 37 °C in the presence of rEsp_N. Ø indicates no addition of rEsp_N to bacterial culture. b Inmunofluorescence showing surface localization of rEsp_N. Bacteria were fixed and labeled with anti-His antibodies and DAPI. Scale bars represent 3 µm. c Scanning electron microscopy and d transmission electron micrographs of negatively stained E. faecalis 23 incubated with (right panel) and without rEsp_N (left panel). Scale bar represents 500 nm. Table 2. Strains and plasmids. Strains and plasmids Reference 16 E. faecalis 23 E. faecalis esp−, non-biofilm forming 16 E. faecalis 11279 E. faecalis esp+, medium biofilm forming 22 E. coli VS39 E. coli transformed with pVS76 (cat PlacUV5 csgG, pACYC184 ori; produces CsgG) E. coli VS39 Esp_N VS39 complemented with pEXPORT:esp_N This study E. coli VS39 Esp_C VS39 complemented with pEXPORT:esp_C This study E. coli VS39 Esp_AB VS39 complemented with pEXPORT:esp_AB This study 5 E. coli VS39 Bap_B VS39 complemented with pEXPORT:bap_B 5 E. coli VS39 Bap_A VS39 complemented with pEXPORT:bap_A S. aureus V329 Δbap S. aureus bap− 23 S. aureus Δbap Esp_N S. aureus Δbap containing pCN51:esp_N This study S. aureus Δbap Esp S. aureus Δbap containing pCN51:esp This study 5 S. carnosus TM400 S. carnosus TM400 Esp_N S. carnosus TM400 containing pCN51:esp_N This study S. carnosus TM400 Esp S. carnosus TM400 containing pCN51:esp This study pET46-eK-LIC_Esp_N Plasmid for expression of Esp N-terminal This study pCN51:esp_N Plasmid for expression of Esp N-terminal domain fused to the R-domain of ClfA, This study under the control of the Pcad promoter pCN51:esp Plasmid for expression of Esp under the control of the Pcad promoter This study pEXPORT:esp_N Expression vector for C-DAG system including the N-terminal domain of Esp This study METHODS Pronadisa), supplemented with glucose 0.5% (w/v) in E. faecalis or glucose Oligonucleotides, plasmids, bacterial strains, and culture 0.25% (w/v) in S. aureus, brain heart infusion (BHI) broth and B2 broth conditions media (casein hydrolysate 10 g/l, yeast extract 25 g/l, NaCl 25 g/l, K2HPO4 1 g/l, glucose 5 g/l). When required for selection, medium was supple- Bacterial strains, plasmids, and primers used in this study are listed in mented with appropriate antibiotics at the following concentrations: Tables 2 and 3. E. coli strains were grown in LB broth (Conda-Pronadisa). E. erythromycin (Em), 1.5 and 10 µg/ml; ampicillin (Amp), 100 µg/ml; faecalis and S. aureus strains were grown in trypticase soy broth (Conda- chloramphenicol (Cl), 20 µg/ml. npj Biofilms and Microbiomes (2020) 15 Published in partnership with Nanyang Technological University A. Taglialegna et al. 9 5′-GGTACCGACTACAAAGACCATGACGGTGATTATAAAGATCATGACATCGACTACAAGGATGACGATGACAAGGAAATTGAACCAATTCCAGAGGAT-3’ DNA manipulations and bacterial transformation General DNA manipulations were performed using standard procedures. Plasmids were purified using the NucleoSpin Plasmid miniprep kit (Macherey-Nagel) according to the manufacturer's protocol. FastDigest restriction enzymes and Rapid DNA ligation kit (Thermo Scientific) were used according to the manufacturer's instructions. Plasmids were transformed into E. coli XL1-Blue strain (Stratagene), E. faecalis, and S. aureus by electroporation. Briefly, S. aureus cells were cultured in B2 broth until the OD at 650 nm reaches 0.5. Cells were washed three times with glycerol 10% (v/v). Cells were resuspended in 15 ml of glycerol 10% (v/v) and incubated at 20 °C for 15 min. After incubation, cells were resuspended in 10% glycerol at the final concentration of 1 × 1010 CFU/ml. For staphylococcal electroporation, 50 μl of the electrocompetent cells were mixed with 2 μg of plasmid and the mix was added to an ice-cold 0.1-cm gap electrode Gene Pulser Cuvette (Bio-Rad). One pulse was given at 100 Ω, 1.25 V, and 25 F capacity. Immediately after the pulse, 1 ml of SMMP media was added to the cuvette. After incubation at 37 °C for 2 h, the cell mixture was plated onto TSA plates with selective antibiotics. For E. faecalis, cells were cultured in BHI media overnight at 37 °C. Cells were washed three times with glycerol 10% (v/v) and were resuspended at the final concentration of 1 × 1010 CFU/ml. For enterococcal electroporation, 50 μl of the electrocompetent cells were mixed with 3 μg of plasmid. The 5′-CTCGAGTTAGTGATGATGGTGATGGTGTGCTAATACAGTACCTTCTTTATCAAC-3′ mix was added to an ice-cold 0.1-cm gap electrode cuvette. One pulse was given at 400 Ω, 1.25 V, and 25 F capacity. Immediately after the pulse, 1 ml of BHI media with subinhibitory concentration of chloramphenicol (0.2 μg/ ml) was added to the cuvette. After incubation at 37 °C for 2 h, the cell 5′-CTGAATTCGAGCTCGGTACGGAAGAGACTTCTTCCTCTTTTCAGA-3′ mixture was plated onto TSA plates with the selective antibiotic. 5′-CGACTCTAGAGGATCCAATAAGGAAGGGAGCAAATTTAGTTTG-3′ Generation of chimeric proteins The N-terminal domain of the esp gene (Esp_N) was amplified from E. faecalis 11279 using primers Esp_Ori_Bam and Esp_B_Kpn (Table 3). To allow anchoring of amplified Esp_N domain to the bacterial cell wall, the R 5′-GGTACCATTTTTACTTACAGTTACTGCTAAATCG-3′ 5′-GAGGAGAAGCCCGGTTAACTGTAAGTAAAAAT-3′ region of clumping factor A (clfA) gene containing an LPXTG motif was 5′-GCGGCCGCACAAGAAAATCCGGGATATACG-3′ 5′-GGATCCAGGAAGGGAGCAAATTTAGTTTGT-3′ 5′-GACGACGACAAGATGCAAATGGGTGAAG-3′ amplified from S. aureus Newman strain using primer K-3xF-ClfA, contain- ing the 3xFlag sequence and a recognition sequence for KpnI, and primer ClfA-7cE with a recognition sequence for EcoRI. The KpnI/EcoRI-restricted R-clfA was cloned into pCN51 vector (pCN51:esp_N-clfA). The Esp_N-ClfA chimeric protein was expressed under the control of the Pcad-cadC 5′-TTACACCCTATTTTTTCGCC-3′ promoter, a cadmium inducible promoter. Esp gene was amplified using primers EF1-InPhu-pCN51-BamH and EF2-InPhu-pCN51-Kpn and the PCR product was cloned in pCN51 (pCN51:esp). Construction of plasmids for Esp expression in C-DAG system To obtain E. coli strains for C-DAG system, we PCR amplified the N-terminal region of Esp (esp_N) with primers EspB-NotI-5 and EspB-XhoI-3. Nucleotides from 2737 to 3192 for a A-repeat and the B-domain (esp_AB), Sequence and nucleotides 3442 to 3685 for a C-repeat of the Esp (esp_C) were gene synthetized by GeneArt (Invitrogen, ThermoFisher Scientific). Esp_N, esp_AB, and esp_C were ligated with NotI/XhoI-restricted pEXPORTXhoI plasmid. The final plasmids pEXPORT-Esp_N, -Esp_C and -Esp_AB were transformed in E. coli VS39 strain. Induction of protein production and presence of amyloid-like material was assessed on solid medium contain- ing 10 μg/ml CR by evaluating colony color phenotype. To quantify the amount of CR bound to fibers generated from arabinose-induced C-DAG strains, we resuspended and incubated E. coli VS39 cells in 1 ml of PBS containing 1.5 μl of CR 0.8% (w/v). The dye was solubilized from the cells using ethanol:acetone (80:20 v/v). Absorbance was measured at 500 nm using a Biospectrophotometer (Eppendorf). Independent experiments were performed six times. Mann–Whitney tests was used to assess significant differences (p < 0.05) of absorbance among tested strains. EF1-InPhu-pCN51-BamH Esp2-InPhu-pCN51-KpnI Multicellular behavior phenotypes For the biofilm formation assay strains were grown overnight at 37 °C and Primers. then diluted 1:40 in the corresponding media. Cell suspension was used to inoculate sterile 96-well polystyrene microtiter plates (Thermo Scientific). Esp_Ori_Bam EspB-XhoI-3 After 24 h of incubation at 37 °C wells were gently rinsed two times with EspB-NotI-5 Esp_B_Kpn Esp-Lic-Fw K-3xF-ClfA Esp-Lic-Rv water, dried, and stained with crystal violet for a few minutes. To quantify ClfA-7cE Table 3. Primer biofilm formation capacity, crystal violet adhered at the bottom of the wells was resuspended with 200 μl of a solution of ethanol:acetone (80:20 v/v) and the absorbance was quantified at 595 nm. Published in partnership with Nanyang Technological University npj Biofilms and Microbiomes (2020) 15 A. Taglialegna et al. 10 Protein expression and production of anti-Esp antibodies an excitation setting of 550 nm and an emission setting of 600 nm. Images N-terminus of Esp (amino acids 50–743) was PCR amplified from E. faecalis were acquired and analyzed with EZ-C1 software (Nikon). 11279 using high fidelity Phusion DNA Polymerase (Thermoscientific) and primers Esp-LIC-Fw and Esp-LIC-Rv (Table 3) designed for use in the LIC Far-UV CD cloning system. The resulting fragment was cloned in pET46-Ek/LIC vector Far-UV CD spectra were measured on a Chirascan (Applied Photophysics) (Novagen). Overnight cultures of E. coli BL21 DE3 containing the spectropolarimeter thermostated at 25 °C. rEsp_N (10 mg/ml) in 20 mM expression plasmid were diluted 1:100 and grown to an OD600nm of 0.6. Tris–HCl pH 8, 50 mM NaCl was diluted 200-fold to 0.05 mg/ml in 10 mM Isopropyl B-D-thiogalactopyranoside (IPTG) was added to a final concen- MOPS at either pH 4.2 or 7.0. Spectra were recorded from 260 to 190 nm, at tration of 0.4 mM and the cultures were shaken (150 rpm) at 18 °C 0.2 nm intervals, 1 nm bandwidth, and a scan speed of 50 nm/min. Ten overnight. After centrifugation, pellets were resuspended, sonicated, and accumulations were averaged for each spectrum. Deconvolution of data centrifuged. Supernatants were filtered (0.45 μm) and rEsp_N protein was performed using the Dichroweb server implementing the CDSSTR purified by Ni affinity chromatography using HisTrapTM FF columns (GE algorithm. Healthcare). To achieve the highest purity, size exclusion chromatography was applied with a HiLoad 16/600 Superdex 200 pg column (GE Healthcare). The concentration of the purified protein was determined Static light scattering by the bicinchoninic acid (BCA) Protein Assay (Pierce, Thermo Scientific) Fluorescence spectra were analyzed for rEsp_N samples diluted at 0.1 mg/ using BSA as standard. Chicken polyclonal antibodies raised against ml in phosphate–citrate buffer either at pH 7.0 or 4.2, and with samples purified rEsp_N protein were supplied by Davids Biotechnologie company. reverted from pH 4.2 to pH 7.0. Scattering was measured exciting at Antibodies were subsequently purified by chromatography. 330 nm and recording in the 320–340 nm range, using excitation and emission bandwidths of 5 nm at low sensitivity, in a Jasco FP-8200 Immunoblot analysis fluorescence spectrophotometer (Jasco Corporation, Japan). One milliliter of an overnight culture was harvested, washed, and finally resuspended in 100 μl of PBS buffer containing 30% (w/v) raffinose Bis-ANS binding (Sigma), 5 μl of lysostaphin 1 mg/ml (Sigma), and 2 μl of DNase 1 mg/ml Bis-ANS-binding assay was performed incubating 0.1 mg/ml rEsp_N (stock (Sigma). After 2 h of incubation at 37 °C, cells were centrifuged. The 0.2 mg/ml) in phosphate–citrate buffer at pH 4.2 and 7.0 containing 10 µM bis- supernatants from surface protein extracts were recovered and analyzed ANS. Samples were measured in a Jasco FP-8200 spectrophotometer (Jasco by SDS–PAGE (7.5% (w/v) separation gel; 5% (w/v) stacking gel). For Corporation, Japan) at 25 °C, in the 400–600 nm range with an excitation Western blot analysis, protein extracts were blotted onto Hybond-ECL wavelength of 370 nm using a slit width of 5 nm for excitation and emission. nitrocellulose membranes (Amersham Biosciences). Anti-Esp-purified anti- body was diluted 1:10,000 with PBS-Tween 5% (w/v) skim-milk. Horse- radish peroxidase-conjugated goat anti-chicken (Abcam Ab97135) diluted ATR-FTIR spectroscopy 1:20,000 in PBS-Tween 5% (w/v) skim-milk was used as a secondary ATR FTIR spectroscopy analyses of Esp_N samples in phosphate–citrate antibody for Esp detection and the subsequent chemiluminescence buffer pH 4.2 and in buffer pH 7 were performed with a Bruker Tensor 27 reaction was recorded. FTIR Spectrometer (Bruker Optics Inc.) supplied with a Golden Gate MKII ATR accessory. Each spectrum consists of 32 acquisitions measured at a resolution of 1 cm−1. Spectra were acquired, background subtracted, Formation of rEsp_N aggregates and reversion assay baseline corrected, and normalized, with the Min/Max normalization To determine the effect of the pH in protein aggregation, rEsp_N method, using the OPUS MIR Tensor 27 software. For the analysis of the IR recombinant protein at a final concentration of 2 μM was incubated in spectra, the second derivative was calculated and plotted to define the phosphate-citrate buffer at pH ranging from 3.6 to 8. Macroscopy protein number and position of band components under the absorbance spectra, aggregates were visualized as a ring of protein adhered to the glass wall at according to the local minima. Next, spectra were smoothed using the the air–liquid interface after 24 h of incubation at 37 °C under shaking non-parametric smoother Loess at a 5% level and automatically fitted with conditions (200 rpm). For reversion assay of rEsp_N aggregates formed the “Peaks I residuals” method using PeakFit package v4.12 (Systat after 24 h of incubation at 37 °C, the phosphate–citrate buffer at pH 4.2 was Software), placing initially as many peaks as local minima appeared in the removed and exchanged for phosphate–citrate buffer at pH 7. After an secondary derivative of the absorbance spectra. The resulting Gaussian overnight incubation at 37 °C and shaking conditions, dissolution of curves were plotted using the amplitude and center values of the fitted rEsp_N aggregates was macroscopically determined. bands. In order to assign secondary structures to band components, frequencies in the range 1611–1630 cm−1 were assigned to intermolecular β-sheet structures, 1630–1640 cm−1 to intramolecular β-sheet structures, Th-T binding 1648–1657 cm−1 to β-helical structures and 1662–1687 cm−1 to β-turns, as rEsp_N protein at 0.1 mg/ml was incubated at pH 7.0, pH 4.2, or pH- described in the literature. reverted from pH 4.2 to pH 7.0 and Th-T added at 25 µM final concentration before measuring. Fluorescence emission spectra were recorded in the range of 460–600 nm with an excitation wavelength of Immunofluorescence 445 nm using a 5 nm slit width for excitation and emission in a Jasco FP- Cells were grown overnight in the corresponding tested conditions and 8200 spectrophotometer (Jasco Corporation, Japan) at 25 °C. Aggregation fixed during 10 min with 3% (v/v) paraformaldehyde (SIGMA). 200 μl of kinetics of 0.05 mg/ml rEsp_N protein in phosphate–citrate buffer at pH 4.2 fixed bacteria were placed on coverslips and incubated for 30 min. After and pH 7.0 were recorded for 1800 s under agitation (800 rpm) at 25 °C, in several washes with phosphate-buffered saline (PBS), cells were saturated the presence of 25 μM Th-T. The kinetic traces were measured exciting at with PBS–0.5% (w/v) BSA, and finally stained with anti-Esp or monoclonal 440 nm and emission was recorded at 475 nm; 5 nm slit widths were used anti-polyHistidine antibody (Sigma Aldrich H1029) antibodies diluted for excitation and emission in a Jasco FP-8200 spectrophotometer. 1:1000. Alexa Fluor 488-conjugated goat anti-chicken (Abcam ab150169) or Alexa Fluor 488-conjugated goat anti-mouse (Invitrogen A-11029) diluted 1:200 were used as a secondary antibody and DAPI diluted 1:200 CR and ProteoStat binding was used to label bacteria. Binding of CR to 0.2 mg/ml rEsp_N protein was evaluated at pH 7.0, pH 4.2, and with pH-reverted samples from pH 4.2 to pH 7.0. rEsp_N samples were diluted to a final concentration of 0.1 mg/ml in buffer at pH 7.0 and Microscopy analysis incubated in the presence of 10 µM CR dye. After 5 min of equilibration, For TEM, cells were grown overnight in the corresponding tested the absorbance spectra were recorded from 375 to 700 nm in a conditions, washed twice with PBS and then fixed with 2% (v/v) Specord200 Plus spectrophotometer (Analytik Jena, Germany). paraformaldehyde (SIGMA) for 1 h at room temperature. Formvar/ For ProteoStat staining, 2 μM aggregated rEsp_N at pH 4.5 were carbon-coated nickel grids were deposited on a drop of fixed sample incubated for 30 min, at room temperature and in darkness with during 5 min and rinsed three times with PBS. Negative staining was ProteoStat Mix buffer (1X Assay Buffer, 1 μl ProteoStat). Preparations were performed using 2% (v/v) uranyl acetate (Agar Scientific, Stansted, UK). observed with an Axioskop 2 plus epifluorescence microscope (Zeiss) using Observations were made with a JEOL 1011 transmission electron npj Biofilms and Microbiomes (2020) 15 Published in partnership with Nanyang Technological University A. Taglialegna et al. 11 microscope. For Esp immunogold labeling, grids coated with the sample 17. Tendolkar, P. M., Baghdayan, A. S. & Shankar, N. The N-terminal domain of were washed and incubated for 45 min on a drop of PBS containing 1:10 enterococcal surface protein, Esp, is sufficient for Esp-mediated biofilm antibody against Esp. After washing with PBS, grids were incubated for enhancement in Enterococcus faecalis. J. Bacteriol. 187, 6213–6222 (2005). 45 min with gold-conjugated (6 nm) goat-anti-chicken secondary antibody 18. Oliveberg, M. Waltz, an exciting new move in amyloid prediction. Nat. Methods 7, (Abcam Ab3960). For 6xHis immunogold labeling, samples were incubated 187–188 (2010). with monoclonal anti-polyHistidine antibody (Sigma Aldrich H1029) (1:10). 19. Conchillo-Solé, O. et al. AGGRESCAN: a server for the prediction and evaluation of After washing with PBS, grids were incubated with gold-conjugated ‘hot spots’ of aggregation in polypeptides. BMC Bioinforma. 8, 65 (2007). (10 nm) goat anti-mouse IgG (H&L) secondary (AURION 810.022). Grids 20. Fernandez-Escamilla, A. M., Rousseau, F., Schymkowitz, J., Serrano, L. Prediction of were stained with 2% (v/v) uranyl acetate as described above. For scanning sequence-dependent and mutational effects on the aggregation of peptides and electron microscopy, cells on thermanox coverslips were fixed for 1 h in proteins. Nat. Biotechnol. 10, 1302–1306 (2004). 0.07 M sodium cacodylate buffer (pH 7.3) containing 1.2% glutaraldehyde 21. Garbuzynskiy, S. O., Lobanov, M. Y. & Galzitskaya, O. V. FoldAmyloid: a method of and 0.05% Ruthenium red (RR). The samples were then washed in the prediction of amyloidogenic regions from protein sequence. Bioinformatics 26, same buffer containing 0.05% RR and post‐fixed in 1% osmium tetroxide in 326–332 (2010). cacodylate buffer. Coverslips were treated with the critical point drying 22. Sivanathan, V. & Hochschild, A. A bacterial export system for generating extra- method and observed on a Zeiss JSM 982 scanning electron microscope at cellular amyloid aggregates. Nat. Protoc. 8, 1381–1390 (2013). the Laboratoire de Biologie Cellulaire et Microscopie Electronique, UFR 23. Valle, J. et al. Bap, a biofilm matrix protein of Staphylococcus aureus prevents Medecine (Tours, France). cellular internalization through binding to GP96 host receptor. PLoS Pathog. 8, e1002843 (2012). 24. Sawyer, E. B., Claessen, D., Haas, M., Hurgobin, B. & Gras, S. L. The assembly of Reporting summary individual chaplin peptides from Streptomyces coelicolor into functional amyloid Further information on experimental design is available in the Nature fibrils. PLoS One 6, e18839 (2011). Research Reporting Summary linked to this paper. 25. Robinson, L. S., Ashman, E. M., Hultgren, S. J. & Chapman, M. R. Secretion of curli fibre subunits is mediated by the outer membrane-localized CsgG protein. Mol. Microbiol. 59, 870–881 (2006). DATA AVAILABILITY 26. Hammer, N. D. et al. The C-terminal repeating units of CsgB direct bacterial The data that support the findings of this study are included in the article, its functional amyloid nucleation. J. Mol. Biol. 422, 376–389 (2012). supplementary information files, or are available from the corresponding author 27. Cucarella, C. et al. Bap, a Staphylococcus aureus surface protein involved in biofilm upon reasonable request. formation. J. Bacteriol. 183, 2888–2896 (2001). 28. Geoghegan, J. A. et al. Role of surface protein SasG in biofilm formation by Staphylococcus aureus. J. Bacteriol. 192, 5663–5673 (2010). Received: 3 October 2019; Accepted: 6 March 2020; 29. Tendolkar, P. M., Baghdayan, A. S., Gilmore, M. S. & Shankar, N. Enterococcal surface protein, Esp, enhances biofilm formation by Enterococcus faecalis. Infect. Immun. 72, 6032–6039 (2004). 30. Latasa, C., Solano, C., Penadés, J. R. & Lasa, I. Biofilm-associated proteins. C.R.— Biol. 329, 849–857 (2006). REFERENCES 31. Latasa, C. et al. BapA, a large secreted protein required for biofilm formation and 1. Blanco, L. P., Evans, M. L., Smith, D. R., Badtke, M. P. & Chapman, M. R. Diversity, host colonization of Salmonella enterica serovar Enteritidis. Mol. Microbiol. 58, biogenesis and function of microbial amyloids. Trends Microbiol. 20, 66–73 (2012). 1322–1339 (2005). 2. Fowler, D. M., Koulov, A. V., Balch, W. E. & Kelly, J. W. Functional amyloid-from 32. Loehfelm, T. W., Luke, N. R. & Campagnari, A. A. Identification and characteriza- bacteria to humans. Trends Biochem. Sci. 32, 217–224 (2007). tion of an Acinetobacter baumannii biofilm-associated protein. J. Bacteriol. 190, 3. O’Toole, G., Kaplan, H. B. & Kolter, R. Biofilm formation as microbial development. 1036–1044 (2008). Annu. Rev. Microbiol. 54, 49–79 (2000). 33. Hinsa, S. M., Espinosa-Urgel, M., Ramos, J. L. & O’Toole, G. A. Transition from 4. Flemming, H.-C. & Wingender, J. The biofilm matrix. Nat. Rev. Microbiol. 8, reversible to irreversible attachment during biofilm formation by Pseudomonas 623–633 (2010). fluorescens WCS365 requires an ABC transporter and a large secreted protein. 5. Taglialegna, A., Lasa, I. & Valle, J. Amyloid structures as biofilm matrix scaffolds. J. Mol. Microbiol. 49, 905–918 (2003). Bacteriol. 198, 2579–2588 (2016). 34. Jordan, S. J. et al. Listeria monocytogenes biofilm-associated protein (BapL) may 6. Lembré, P., Di Martino, P. & Vendrely, C. Amyloid peptides derived from CsgA and contribute to surface attachment of L. monocytogenes but is absent from many FapC modify the viscoelastic properties of biofilm model matrices. Biofouling 30, field isolates. Appl. Environ. Microbiol. 74, 5451–5456 (2008). 415–426 (2014). 35. Huber, B. et al. Genetic analysis of functions involved in the late stages of biofilm 7. Van Gerven, N., Van der Verren, S. E., Reiter, D. M. & Remaut, H. The role of development in Burkholderia cepacia H111. Mol. Microbiol. 46, 411–426 (2002). functional amyloids in bacterial virulence. J. Mol. Biol. 430, 3657–3684 (2018). 36. Theunissen, S. et al. The 285 kDa Bap/RTX hybrid cell surface protein (SO4317) of 8. Chapman, M. R. et al. Role of Escherichia coli curli operons in directing amyloid Shewanella oneidensis MR-1 is a key mediator of biofilm formation. Res. Microbiol. fiber formation. Science 295, 851–855 (2002). 161, 144–152 (2010). 9. Dueholm, M. S. et al. Fibrillation of the major curli subunit CsgA under a wide 37. Lasa, I. & Penadés, J. R. Bap: a family of surface proteins involved in biofilm range of conditions implies a robust design of aggregation. Biochemistry 50, formation. Res. Microbiol. 157, 99–107 (2006). 8281–8290 (2011). 38. Cherny, I. et al. The formation of Escherichia coli curli amyloid fibrils is mediated 10. Dueholm, M. S. et al. Functional amyloid in Pseudomonas. Mol. Microbiol. 77, by prion-like peptide repeats. J. Mol. Biol. 352, 245–252 (2005). 1009–1020 (2010). 39. Wang, X., Smith, D. R., Jones, J. W. & Chapman, M. R. In vitro polymerization of a 11. Romero, D., Aguilar, C., Losick, R. & Kolter, R. Amyloid fibers provide structural functional Escherichia coli amyloid protein. J. Biol. Chem. 282, 3713–3719 (2007). integrity to Bacillus subtilis biofilms. Proc. Natl Acad. Sci. USA 107, 2230–2234 40. Lembré, P., Vendrely, C. & Patrick, Di. Martino Amyloid fiber formation by syn- (2010). thetic peptides derived from the sequence of the protein CsgA of Escherichia coli. 12. de Jong, W., Wösten, H. A. B., Dijkhuizen, L. & Claessen, D. Attachment of Strep- Protein Pept. Lett. 20, 942–946 (2013). tomyces coelicolor is mediated by amyloidal fimbriae that are anchored to the cell 41. Lembré, P., Vendrely, C. & Patrick, Di. Martino Identification of an amyloidogenic surface via cellulose. Mol. Microbiol. 73, 1128–1140 (2009). peptide from the Bap protein of Staphylococcus epidermidis. Protein Pept. Lett. 21, 13. Schwartz, K., Syed, A. K., Stephenson, R. E., Rickard, A. H. & Boles, B. R. Functional 75–79 (2014). amyloids composed of phenol soluble modulins stabilize Staphylococcus aureus 42. Patrick., Di. Martino Bap: a new type of functional amyloid. Trends Microbiol. 24, biofilms. PLoS Pathog. 8, e1002744 (2012). 682–684 (2016). 14. Taglialegna, A. et al. Staphylococcal Bap proteins build amyloid scaffold biofilm matrices in response to environmental signals. PLoS Pathog. 12, e1005711 (2016). 15. Arrizubieta, M. J., Toledo-Arana, A., Amorena, B., Penadés, J. R. & Lasa, I. Calcium inhibits bap-dependent multicellular behavior in Staphylococcus aureus. J. Bac- ACKNOWLEDGEMENTS teriol. 186, 7490–7498 (2004). We thank Alejandro Toledo-Arana and Cristina Solano for fruitful discussions. This 16. Toledo-Arana, A. et al. The enterococcal surface protein, Esp, is involved in research was supported by grants RTI2018-096011-B-I00 and BIO2017-83035-R from Enterococcus faecalis biofilm formation. Appl. Environ. Microbiol. 67, 4538–4545 the Spanish Ministry of Science, Innovation and Universities, and Proyecto Intramural (2001). Incorporación-2018 CSIC. Published in partnership with Nanyang Technological University npj Biofilms and Microbiomes (2020) 15 A. Taglialegna et al. 12 AUTHOR CONTRIBUTIONS Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims A.T., L.M.-C., I.L., J.V. conceived and designed the experiments. A.T., L.M.-C., P.D.-M., S. in published maps and institutional affiliations. N., J.A.G., J.V. conducted experiments. S.N., S.V., J.A.G. assisted with the biophysical experiments. J.A.G., S.V., I.L., J.V. analyzed the data and wrote the manuscript. All authors read and approved the final manuscript. Open Access This article is licensed under a Creative Commons COMPETING INTERESTS Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give The authors declare no competing interests. appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons license, unless ADDITIONAL INFORMATION indicated otherwise in a credit line to the material. If material is not included in the Supplementary information is available for this paper at https://doi.org/10.1038/ article’s Creative Commons license and your intended use is not permitted by statutory s41522-020-0125-2. regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this license, visit http://creativecommons. Correspondence and requests for materials should be addressed to J.V. org/licenses/by/4.0/. Reprints and permission information is available at http://www.nature.com/ reprints © The Author(s) 2020 npj Biofilms and Microbiomes (2020) 15 Published in partnership with Nanyang Technological University