We find them here, we find them there: Functional bacterial amyloid

Review

We find them here, we find them there:Functional bacterial amyloid

D. Otzena,* and P. H. Nielsenb

a Interdisciplinary Nanoscience Centre, Department of Molecular Biology, Aarhus University,Gustav Wieds Vej 10C, 8000 Aarhus C (Denmark), Fax: +45 96 12 31 78, e-mail: [email protected] Department of Biotechnology, Chemistry and Environmental Engineering, Aalborg University,Sohngaardsholmsvej 57, 9000 Aalborg (Denmark)

Received 4 September 2007; received after revision 22 October 2007; accepted 29 October 2007Online First 24 November 2007

We seek them here, we seek them there,We scientists seek them everywhere:Are they our heaven, are they our hell?Those damned elusive strands in that fibril!

Adapted from The Scarlet Pimpernel by Baroness dOrczy

Abstract. Protein amyloid is often deposited in con-nection with neurodegenerative diseases. Such depos-its generally possess three principal drawbacks: cyto-toxicity, lack of spatial control in their deposition andstructural polymorphism. These are typical features ofbiologically non-optimized systems which have notbeen exposed to evolutionary pressure. Nevertheless,Nature uses the cross-b self-organizing principle inmany structural contexts where a strong but pliablematerial is needed. Functional amyloid is found inhumans, invertebrates, fungi and, not least, bacteria, in

which amyloid may be the rule rather than theexception. Detailed case studies reveal how directednucleation can use tailor-made proteins optimized toassume a specific amyloid conformation, leading toremarkably robust assemblies. This makes it highlychallenging to purify and analyze the products formedin vivo. We contrast pathogenic and in-vitro-formedamyloid with functional amyloid, paying particularreference to bacterial amyloid, and discuss challengesand perspectives in identifying and characterizing thisclass of protein.

Keywords. Protein, bacteria, aggregate, CsgA, functional amyloid, cytotoxicity, polymorphism.

Introduction: a historical preamble

It is an abiding passion of human beings to attempt toread and re-view history as an orderly progression ofevents running according to a well-coordinated and

rational script. Such views generally do not hold up tocloser inspection. Nevertheless, in the experimentalsciences, there is a steady development in technicalabilities which engage in intense dialogue with scien-tific theories and views to generate some broad frontsalong which many of the most exciting events takeplace. Viewed retrospectively, a perception of thesefronts may make it easier to appreciate the current* Corresponding author.

Cell.Mol. Life Sci. 65 (2008) 910 9271420-682X/08/060910-18DOI 10.1007/s00018-007-7404-4Birkhuser Verlag, Basel, 2007

Cellular and Molecular Life Sciences

research climate and perhaps even gain an inkling ofthe scientific activities just over the horizon.

Functional amyloid: from folding to misfolding andbackThe case of functional bacterial amyloid lends itselfwell to this historical perspective. Functional amyloid(FA), as we will detail below, is in essence a bio-logically useful example of what was originallythought to be a regrettable lapse in Mother Naturespainstaking attempts to fold proteins properly. Assuch, the intellectual origins of FA are intimately tiedto the understanding of a protein as an autonomouslyfolding biological entity, that is, a molecule whoseamino acid sequence intrinsically encodes a singlebiologically active structure that is also the thermo-dynamically most stable state. This stunning conceptwas introduced by Anfinsen [1] and colleagues in thelate 1950s, sparking off a flurry of activities over thenext two to three decades that led to the presentationof the Levinthal paradox [2] and the search for proteinintermediates that could explain how proteins couldreach the folded state on a biologically meaningfultime scale [3]. Thanks to the advent of proteinengineering in the early 1980s, it then became possibleto map out protein-folding pathways, including thestructures of both intermediates and more elusivetransition states, at residue- and even atomic-scaleresolution using Fershts f-value approach [4, 5]. Thismade the decade from the late 1980s to the late 1990s ahotbed for the study of proper in vitro proteinfolding, focusing on the process leading from thedenatured state to the native state, generating sophis-ticated models for protein folding [6] which could bebacked up by in silico simulations [7]. Most of thesestudies used simple single-domainmodel proteins thatfolded rapidly and reversibly but were not represen-tative of the large and lumbering multi-domainproteins making up most of the cell. Such proteinswere often found to be difficult to persuade to jumpblithely from a chemically denatured state to thenative state, but tended to collapse in an undignifiedheap as misfolded aggregates, although the nature ofthese aggregates was generally subjected to closerscrutiny (Fig. 1). In parallel, it became apparent thatliving organisms had developed a whole arsenal ofproteins to ensure that folding of such proteins in vivohappened in an orderly and controlled fashion in thecell, namely the molecular chaperones [8]. Concom-itantly, proteins were identified which formed cellularaggregates in neurodegenerative diseases such asParkinsons disease [9]. In the late 1990s, attentionthus shifted from folding to misfolding, the lattersparkling with the glitter of medical relevance. A boldgeneralization by C. M. Dobson, F. Chiti and their

colleagues on the generic ability of proteins tofibrillate to amyloid has helped to make the study ofprotein misfolding as amyloid a very prominent topicin protein science in the last decade and has estab-lished interesting predictive models of aggregation.While much remains to be discovered within dysfunc-tional amyloid formation, it is also becoming increas-ingly clear that such well-ordered aggregates serve auseful purpose in Nature in an immense number ofcircumstances, making FA not a curiosity but almost abedrock of biological structures. Thiswill undoubtedlymake studies of FA formation a very active front in thecoming decade. In a sense, therefore, we havereturned to the proper folding studies of the 1980sand 1990s with the twist that these studies involveintermolecular rather than just intramolecular associ-ations.In this review, we will begin by reviewing some basictenets of amyloid structure, formation and pathoge-nicity, as well as some attempts by researchers toharness these properties, before we lift the curtain onhow Nature itself has exploited amyloid for its ownends and avoided the pitfalls associated with theuncontrolled formation of this state. We will alsohighlight aspects of the purification and identificationof these components which, by their very nature, aredifficult to mobilize into an aqueous phase. We alsorefer the reader to a number of excellent more generalor more specialized reviews of pathological and FAwhich have recently appeared [1017].We will not discuss the rich and complex field of prionproteins, which deserves a review in its own right.Prion proteins are defined as being able to exist in twodifferent stable monomeric states, one of which is ableto actively convert the other to its own form. If thecomplex is stable, it can form the basis for a growingaggregate. This conformational imprinting is analo-gous to that of amyloid, except that conventionalamyloid usually relies on a oligomeric structure toconvert and incorporate monomers; furthermore, theoligomeric structure does not usually accumulate tothe same extent as the infectious prion state. With thepossible exception of the CPEB protein fromAplysia,there is no evidence for a functional use of the prionaggregate, although the loss-of-function associatedwith aggregation can serve purposes summarized inTable 1.

General amyloid features

Amyloid structureAmyloid can be defined as orderly repeats of proteinmolecules arranged as a fiber of indefinite length in across-b structure, in which the b strands are perpen-

Cell.Mol. Life Sci. Vol. 65, 2008 Review Article 911

dicular to the fiber axis. Amyloidwas first described asinsoluble deposits in 1854 byVirchow, who consideredit amylose- or cellulose-like due to its ability to stainwith iodine [18]. Friedrich and Kekul [19] identifiedprotein components 5 years later, but the name hasstuck. Although amyloid is somewhat misleading,the name is a salutary reminder of some structuralproperties it shares with cellulose and amylose,namely a repetitive sequence of units which form anorganized structure with regular binding pockets forvarious dyes. Thus, amyloid in human tissue is tradi-tionally visualized by pathologists using the sulfonat-ed azo dye Congo Red, which lends a distinct apple-green birefringence to amyloid deposits under cross-polarized filters. Similarly, the fluorescent dye Thio-flavin T (ThT) undergoes a spectral shift upon bind-ing. Despite their availability and ease of handling,none of these dyes has absolute specificity for amyloid[20] and they are best used against isolated proteins; invivo it is more appropriate to use conformationallyspecific monoclonal antibodies [21, 22]. These anti-bodies, originally developed against Ab fibrils, bind toall hitherto tested amyloid fibrils, and illustrate that allfibrillated proteins share an underlying structuralmotif [23].Indeed, the ability to make a large group of com-pletely unrelated proteins fibrillate to the same over-all end structure has led to the suggestion thatessentially all proteins can form such structuresunder appropriate conditions [24]. This is confirmed

by atomic-level structural details about amyloid whichare emerging at a breathtaking pace. Conventionalamyloid is non-crystalline and insoluble, and thereforedoes not lend itself to X-ray crystallography orsolution-state nuclear magnetic resonance (NMR).However, solid-state NMR is currently making veryimpressive advances and has already been used toreveal structures of both peptides [2527] and, in lessdetail, full-length proteins [28]. In parallel, the recentavailability of highly focused synchrotron X-raybeamlines (Fig. 2) has made it possible to elucidatethe structure of microcrystals of a large number ofdifferent fibrillating peptides [29, 30].Although all thepeptides in this highly impressive work are organizedas standard b sheets with b strands perpendicular tothe fibril axis, the individual b strands contact eachother in different ways for different peptides. This hasled to a unified structural scheme with eight differenttypes of interface or zipper to which all amyloidstructures, including the functional ones, are likely toadhere [29, 30].The current challenge is to link such atomic-leveldetails with the overall fibrillar architecture. Here theemerging method of choice is synchrotron-basedsmall-angle X-ray scattering, a solution techniquewhich makes it possible to monitor the overallstructure of different aggregative states which decayand accumulate during the fibrillation process [31].The technique was put to impressive use in the case ofinsulin, allowing the authors to identify the coarse

Figure 1. Different pathways inthe folding of a proteinmolecule.In addition to folding to thenative state (in casu the F. solanipisi cutinase, PDB code 1CEX),the protein can aggregate to astate devoid of persistent struc-ture (amorphous aggregate) orform well-ordered but unwantedassemblies rich in b-sheet struc-ture. Chaperones can guide theprotein to form such b-sheetstructures fulfilling certain bio-logical functions, as described inthis review.

912 D. Otzen and P. H. Nielsen Functional bacterial amyloid

Table 1. Overview of functional amyloid.

Organism Protein Purpose Ancillary components References

Eukaryotic amyloid

Humans Ma (component ofPmel17)

by fibrillating, Ma forms a scaffold forbinding and cross-linking to the thioflavinT-like substance melanin, thus protectingagainst DNA-damaging UV light

furin (cleavage and subsequentliberation of Ma from the cellmembrane)

6567

Silk moth and fish chorion proteinswith conservedcentral amyloid-forming domain

forms a protective coating for the oocyteand embryo

unknown; difficult to identifyindividual proteins

137139

Spiders spidroins spider silk for webs; stored up to 50% asliquid crystals in specific glands;assembled by passage through spinningduct and extraction of water and salt

spinning duct organ 140

Aspergillus andother mycelialfungi

hydrophobins form an amphipathic amyloid membraneat the air-water surface to reduce surfacetension and allow hyphal growth out ofwater; provide spores and appressoriawith a hydrophobic coat for hostattachment at the cuticle surface as well asevasion of host macrophages

unknown 10, 89

Prion proteins1

Mammals ECTO-NOX involved in both [NAD(P)H] oxidationand protein disulfide-thiol interchange;oscillates between these two states in asynchronized (entrained) fashion, viastructural changes (a helix to b sheet andvice versa); ECTO-NOX can, like theother prions, fibrillate and bind copperions; mice lacking the PrP protein haveproblems with their circadian rhythm

probably spontaneous 141

Aplysia (sea slug) CPEB (cytoplasmicpolyadenylationelement-bindingprotein)

homologous to yeast prions and behaveslike prions when expressed in yeast; theprion form may function as a kind ofmemory storage, leading to long-termsynaptic changes


Saccharomycescerevisiae (bakersyeast)

Ure2p the native state is involved in nitrogencatabolism; aggregation of the prion stateleads to upregulation of the Dal5ureidosuccinate and allantoatetransporter


S. cerevisiae Sup35 Sup35 is a termination factor in the nativestate; aggregation of Sup35 occursspontaneously in 107 to 105 of all yeastcells, leading to stop codon read-throughand (random) emergence of newproperties

spontaneous aggregation 144

Podosporaanserina(filamentousfungus)

HET-s the native state serves to form andmaintain barriers between heterokaryonsof different prion strains (HET-S versusHET-s)


Bacterial amyloid

Escherichia coliand otherEnterobacteriaceae

CsgA CsgA monomers form curli amyloidfibrils; these are involved in attachment tosurfaces, biofilm formation and invasioninto host cells

CsgB (nucleator), CsgG (membranepore for export of CsgA and CsgB)

68, 69, 79

Streptomycetescoelicolor andotherstreptomycetes

ChpAH chaplins modulate the surface tension ofwater and the surface of streptomycetes,allowing production of aerial hyphae;they are organized in rodlets at the surfaceof aerial hyphae and spores

self-assemble to amyloid fibrils;ChpAC are probably anchored to thecell wall; unknown nucleator; rodlinshelp form rodlets

73


structure not only of the starting monomer and end-point fibrils, but also of the otherwise elusive oligo-meric state [32]. Let us therefore now briefly turn tothe actual mechanism of fibrillation in which thisimportant state also features.

The kinetics of amyloid formationMost proteins only fibrillate if they are exposed toconditions such as extremes of pH, elevated temper-ature or intermediate concentrations of denaturants[33]. This typically induces partially structured stateswhich remain flexible enough to form contacts withsimilar proteinmolecules but not so unstructured as tomake ordered contacts unfeasible. Proteins whichfibrillate under physiological conditions typically doso because they have either suffered a destabilizingmutation [34] and/or have been proteolyticallycleaved [35]. Both of these events can make themmore disposed to form such a flexible state. A typicaltime profile for fibrillation (Fig. 3), monitored by thechange in ThT fluorescence which reports on theaccumulation of fibrillar structures, will contain a lagphase ofminutes to days (characterized by the lag timetlag), in which there is no formation of amyloid. The lagphase is followed by the elongation phase, in whichaggregation rapidly commences, and within a shorttime (relative to tlag) most of the monomer populationis incorporated into fibrils, precluding further growthand leading to a plateau in ThT fluorescence. The lagphase is conventionally ascribed to the build-up of afibril nucleus [36] although alternativemodels are alsoavailable [37]. The nucleus is typically thermodynami-cally unstable, because it requires many interactionsto be formed cooperatively between a number ofdifferent protein molecules (anywhere between 2 and40). This leads to an apparent activation energy for thefibrillation reaction. However, once formed, it can actas a template for the subsequent deposition of addi-tional protein molecules to form the mature fibril.There is still no consensus about how the nuclei areelongated to form fibrils and how multi-strandedfibrils form. Usually it is supposed that monomers

Table 1 (Continued)

Organism Protein Purpose Ancillary components References

Mycobacteriumtuberculosis

MTP pili have adhesive properties and areantigens for infection and recognized byIgG from patients

74

Xanthomonas spp.and other plantpathogens

HpaG cytotoxin; harpin causes a hypersensitiveresponse in plant cells, leading to celldeath

forms amyloid fibrils under plantapoplast-like conditions

76

Klebsiellapneumoniae

Microcin E492 cytotoxin; microcin killsEnterobacteriaceae bacteria by formingpores in the cytoplasmic membrane

103

1 Note that in the case of prion proteins, the functionality appears mainly to be derived from loss-of-function associated with aggregation.

Figure 2. Atomic structure of the GNNQQNY peptide whichfibrillates below 1 mM but forms diffracting microcrystals at 10100 mM. (a) The figure emphasizes the regular hydrogen bondingbetween individual b-strands which are in perfect register andperpendicular to the long axis. (b) The crystal is stabilized by closeside chain contacts between each sheet. Reproduced with permis-sion from Nelson et al. [29].


simply add on to the growing ends of the fibrils [38], ascan be seen directly by seeding fibrils with monomers[39]. However, there is also evidence that oligomerscan join up directly [32, 40] to form mature fibrils.Alternative models of lateral association of fibrilshave also been proposed [41].

Multiple prefibrillar stages and fibrillarpolymorphismIn practice, a protein will often aggregate in severalstages, and in this process, oligomers of various sizescan accumulate. This is not least due to the fact thatthese intermolecular interactions are at least in thepathological state not evolutionarily optimized, andtherefore a number of different states with similar butnot identical structure and stability will coexist in adynamic equilibrium, some of them closer to thefibrillar state than others. However, it is mechanisti-cally simplest to lump them all together into aprefibrillar aggregate class with an overall activationenergy. In some cases, these states have very distinc-tive kinetics and structural properties, and thus it ispossible to separate individual steps such as initialformation of amorphous aggregates, followed byrearrangements to crystalline or other ordered statesand then finally the fibrillar state [42, 43].The fact that most amyloid (apart from the functionalkind) is not biologically optimized means that wecannot expect that each aggregate is necessarily awell-defined species which is well separated fromother aggregates in terms of structure and stability.

This goes for both prefibrillar aggregates and maturefibres. We call this phenomenon fibrillar polymor-phism. For example, the peptide hormone glucagoncan form different fibrils with remarkably differentstructural and thermodynamic properties dependingon solute composition, temperature and concentra-tion [4446], and fibrils formed by the Ab peptideinvolved in Alzheimers disease vary significantlydepending on whether they are shaken or not duringthe fibrillation process [47]. Thus, in the absence ofbiological optimization, it is left to the fibrillatingprotein itself to undergo an in situ evolutionaryprocess in which the most stable aggregate under theprevailing conditions is selected [48].

Predicting amyloid formationSignificant efforts have been made to model thekinetics of aggregation computationally based onsimple physico-chemical properties such as secon-dary-structure propensities, hydrophobicity, chargeand patterns of amphiphilicity [49, 50]. Currentmodels [49, 51, 52] are able to predict with reasonablesuccess not only the effect of various mutations onthe change in the growth rate of fibrils but also theabsolute rates of elongation. Like most pioneeringefforts, these models are not without limitations;they presuppose that aggregation occurs from thedenatured state rather than partially structuredstates with various idiosyncratic features, they can-not take into account the pervasive but elusivephenomenon of fibrillar polymorphism and they donot predict the lag time of fibrillation. This is notsurprising, since lag times are very sensitive to asimple experimental variable such as agitation whichis very difficult to model quantitatively or vary in asimple and systematic fashion. However, it is some-what ironic that the lag phase is predictively the mostelusive parameter, since it is also biologically themost important.

Why is amyloid toxic?Accumulation ofmature fibrils can be detrimental if itoccurs on a large scale in the systemic amyloidoseswhere kilograms of protein can accumulate andultimately lead to organ failure and death by simplephysical blockage [53]. However, biologically, thefibrillated state is relatively inert and does not appearto play a major role in neurodegenerative diseases;instead it is the prefibrillar aggregates which havebeen shown to be cytotoxic [5456]. The reasons forthis are hotly contested and an intense focus of currentinvestigation but are likely to be related one way orthe other to the accumulation of unwanted proteinsurface, which would otherwise be sequestered in thecore of the correctly folded protein. Much evidence

Figure 3. Stages in protein fibrillation in vitro. The monomerassembles into a prefibrillar aggregate such as an oligomer (in thiscasea-synuclein visualizedby electronmicroscopy [104])which arethen assumed to be elongated to protofilaments that mature tothicker fibrils consisting of interwoven bundles of protofilaments(in this case fibrils of glucagon visualized by electron microscopy[C. B. Andersen, G. Christiansen and D. E. Otzen, unpublishedresults].


has beenmarshalled to support the hypothesis that theprefibrillar aggregates interact with cell membranes,leading to e.g. pore formation and uncontrolled ionefflux [5759] or generally increased membranepermeabilization [60]. Such pore formation has evenbeen suggested to have a physiological role in assistingCa2+ homeostatis in bone tissue [61]. It has also beensuggested that oligomers could react with metabolitesand proteins in unpredictable ways, leading to un-desirable compounds [55].

Amyloids in Nature

Repetitive elements in NatureNature provides us with many examples of how toexploit repetitive protein elements in structural con-texts. These include collagen, fibrin, keratin andmyosin, to name but a few. As if to highlight ourown present inadequacy in the molecular-assemblybusiness, each case provides an example of howimportant it is to be able to control proper alignmentof these units. Thus, for actin, monomer- and fiber-binding proteins regulate the monomer pool, orches-trate the formation of fibers, organize fibers intoarrays and depolymerize fibers formonomer recycling[62]. The three polypeptide chains forming thecollagen helix are aligned through helical extensionswhich are cleaved off before assembly [63]. Tropoe-lastin (the soluble precursor of mature elastin)assembles on a preformed template of fibrillin-richmicrofibrils [64] and is subsequently cross-linkedenzymatically.

Using amyloid with a purpose: FAs and the lessonsthey teachIn view of the widespread use of such repetitiveelements in living organisms and the universal abilityof proteins to aggregate under appropriate conditions,it should come as no surprise that Nature has alsofoundwidespread use of the cross-b structure. In orderto form a biologically optimized system, however,Naturemust overcome three weaknesses observed foramyloid formed in vitro and in deposition diseases:pathogenicity, uncontrolled formation (i.e. randomdeposition) and polymorphism. In all cases where FAhas been studied in molecular detail, it has emergedthat the pitfalls can be avoided by the use of ancillaryproteins and/or the spatial separation provided bydifferent cellular compartments. Let us look at a fewinstructive examples.The protein Ma, which is the main component ofhuman melanosomes, is so aggregation-prone that,left to its own devices, it aggregates within a fewseconds even in the presence of 5 M guanidinium

chloride [65]. Yet in vivo, Ma is synthesized as thelumen-exposed component of the membrane proteinPmel17, and this association suppresses its aggregativebehavior until it is released in a post-Golgi compart-ment by furin cleavage [66]. A disulfide bond betweenMa and the membrane-bound moiety Mb, mayfurther regulate against fibrillation until Mb is de-graded, after which Ma can presumably aggregaterapidly [67].The protein CsgA is the main component in Escher-ichia coli curli structures, but although it aggregatesreadily by itself upon incubation as a purified one-component solution in vitro [68], in vivo it is secretedout of E. coli and only starts to form curli fibrils ifexposed to the protein CsgB, which remains attachedto the surface of the outer membrane where it mayform a nucleating structure [69]. Furthermore, in vitrostudies suggest that when CsgA aggregates by itself,the fibrillation nucleus is a dimer [70] which by virtueof its small size is presumably not cytotoxic. Thissuggests that E. coli may even have two complemen-tary systems to prevent formation of cytotoxic aggre-gates.Undoubtedly many more exquisite mechanisms willbe revealed as we uncover details of the formation ofother FAs. In the following sections, we mainlyaddress bacterial FA, but for overview purposes wehave also included in Table 1, a list of amyloidoccurring in eukaryotic tissue.

Prokaryotic amyloids: an overviewAmong the prokaryotes, amyloids were not detectedbefore the late 1980 s, where the Gram-negative E.coli was shown to produce amyloid fimbriae calledcurli [71] . Since then, several other enteric bacteriasuch as Salmonella spp. have also been shown toproduce amyloid fimbriae, known as tafi [72] . Thesefilamentous structures were thought to be involved inadhesion to surfaces, cell aggregation and biofilmformation and have been studied in detail in terms ofamyloid structure, involved genes and their regula-tion, and their possible function (see below). In 2003,another family of bacterial amyloids, known aschaplins, were found in the Gram-positive filamen-tous Streptomyces coelicolor by Claessen et al. [73].These amyloids are very different from curli and tafiin amino acid composition, structure and functionand are involved in production of aerial hyphae anddispersal of spores. Within the past 3 years, otherbacterial FAs have been described: pili from Myco-bacterium tuberculosis [74] , Microcin 492 fromKlebsiella pneumoniae [75] , and harpins from Xan-thomonas axonopodis [76], the last two acting ascytotoxins.


However, amyloids are much more widely distributedamong bacteria than those isolated discoveries sug-gest. Our recent study of biofilm samples from avariety of different habitats, including freshwaterlakes, brackish water, drinking reservoirs and wastewater treatment plants, shows that amyloids arepresent in all habitats and are often associated with alarge fraction of the bacteria [22]. Using amyloid-specific antibodies and dyes, we have found thatamyloids are expressed by 540% of all bacteria inthese habitats. The list embraces bacteria from severalphyla, including Proteobacteria, Actinobacteria, Fir-micutes, Chloroflexi and Bacteriodetes [22]. Further-more, several earlier reports of pili and other cellappendages and cell envelope components are likelyto be amyloid, although it was often not realized in theoriginal studies. Some of these results are also brieflyreviewed below. The methods used for their isolationare summarized in Table 2.Bacterial amyloids seem to be generally present inmost bacteria with multiple functions, and we havejust scratched the surface to reveal the huge diversityin structure and function of amyloid in Nature.

E. coli and Salmonella typhimurium curli :attachment cablesE. coli, S. typhimurium and other Enterobacteriaceaeproduce thin fimbriae called curli or tafi (thinaggregative fimbriae) [12, 68, 72, 77], which can bevisualized as extracellular material by atomic forcemicroscopy (AFM) and light microscopy [78]. Theyare 312 nmwide andmay be up to a fewmicrometerslong. Although the major protein component in thefinal curli fibril is CsgA, at least six proteins areinvolved in the formation of curli. Of these, CsgAforms a central outer membrane pore in complex withCsgE andCsgF [79], which probably plays a role in theexport of CsgA and CsgB to the surface. Theimportant contribution of CsgB has been describedabove. Interestingly, the CsgA core amyloid-formingdomain (i.e. the part integrated into the amyloidstructure) can be divided into five repeating units;peptides representing three of these repeating units

also form amyloid [70], indicating that CsgA isoptimized for fibrillation.The operons in E. coli (csgBA and csgDEFG) havehomologous operons in Salmonella (agfBA and agf-DEFG) and the fimbriae are biochemical and func-tional analogs, so we will mainly focus on curli fromE.coli. Together with cellulose fibers, which are alsoproduced bymany of theses bacteria, curli (or tafi) areof key importance for adhesion to surfaces and biofilmformation and thus in pathogenesis by attachment,invasion of host cells, interaction with host proteinsand activation of the immune system [10, 12, 80]. Curlican also bind to the cytokine-releasing factor flagellinin intestinal epithelial cells and thus elicit an immuneresponse [81]. Firm adhesion to the surface only takesplace after 816 h when curli are expressed, so curliare believed to be important for aggregation and formaking strong attachment to the surfaces [78]. Thus,curli promote binding to a variety of other surfaces,including plant cells [82, 83], stainless steel [84, 85],glass and plastics [85, 86], and can substantiallyenhance the resistance to chlorine [84].

Chaplins fromGram-positive S. coelicolor: formationof hydrophobic surfacesThe filamentous bacterium S. coelicolor and manyother streptomycetes undergo complex morphologi-cal differentiation, forming a submerged myceliumthat may grow into the air and septate into spores. S.coelicolor produces chaplins which are a class of eighthydrophobic proteins (ChpAH) that spontaneouslyself-assemble into amyloid fibrils forming an amphi-pathic membrane [73]. This hydrophobic membranemediates attachment to other hydrophobic surfaces,facilitates penetration of the liquid-air interface bylowering surface tension and thus allows formation ofaerial hyphae. The chaplins are also responsible formaking the spores hydrophobic, which probablyfacilitates wind dispersal. Three groups of proteinare responsible for formation of aerial filaments andspores: the spore-associated protein SapB, chaplinsand rodlins [73, 87, 88]. The chaplin genes have onlybeen discovered in sporulating Actinomyces, a group

Table 2. Methods used to isolate functional amyloid.

Organism Protein Method Reference

S.coelicolor

chaplins removal of other components by boiling 2% SDS and extraction with TFA 73

M.tuberculosis

MTP pili mechanical shearing, differential centrifugation and extraction to remove lipids 74

P.fluorescens

25-kDaprotein

homogenization and freeze-thaw cycles, followed by removal of contaminatingmacromolecules by 2% boiling SDS

M.S. Dueholm et al.,unpublished results

E. coli CsgA E. coli cells blended to free curli, followed by pelleting of curli by centrifugationand purification by SDS-PAGE

68


of Gram-positive Actinobacteria. In function, thechaplins are very similar to the hydrophobins pro-duced by many filamentous fungi [89] . The proteinsChpAC are around 225 amino acids and CHpDHare up to 63 amino acids in length. They are onlyknown from Streptomyces, except for the aerialmycelium-forming Termobifida fusca, which has asingle chp gene [90]. Chaplins assemble in solutionwhen seeded with the assembled form of the protein.Chaplin fibrils are organized in rodlets at the surfaceof aerial hyphae and on spores from S. coelicolor. Thepresence of rodlin proteins may help organize thechaplin fibrils in a rodlet layer which consists of pairedrods 812 nmwide and up to 450 nm long [91, 92]. It isstill not clear whether the chaplins assemble primarilyon the surface of the filaments or at the water-airinterphase after release from the cells. Claessen andcoworkers [73, 91] suggested an initial release to thewater during submerged growth where they lower thesurface tension by assembling at the water-air inter-face. This allows the bacteria to form aerial hyphae.The amphipathic membrane is then formed at the cellwall-air interface and may induce the rodlet layer.However, new AFM studies [93] on the surface ofliving S. coelicolor and S. avermitilis indicate that thechaplins can also be assembled to form amyloid fiberdirectly on the cell surface during submerged vegeta-tive growth. These studies also showed that aerialgrowth is characterized by a thick fibrous layerconsisting of chaplin fibrils and rodlins. The longchaplins (ChpAC) are probably anchored to the cellwall, because they contain a sequence which isspecifically recognized by sortase enzymes [73, 94].In contrast, the short chaplins, encoded by genesexpressed 10 to 25-fold more than those for the longchaplins [92] are not recognized, making them lessstable and more easily sheared off from growingaereal hyphae. It is not yet known whether Strepto-mycetes form nucleation proteins as is known for E.coli, but it is possible that a smaller degree ofbiological control is required to form two-dimensionalfilms at the water-air interface, as opposed to longfibrils growing from a bacterial surface.

M. tuberculosis pili: adherence via piliThe Gram-positive M. tuberculosis causes tuberculo-sis in humans. The bacterium produces thin pili (MTP,23 nm in width) that are antigens for infection andrecognized by IgG from patients [74]. MTP areassumed to be critical for adherence, colonizationand infection of the host. They bind to the humanextracellular matrix protein laminin and share anumber of properties with curli, although theiramyloid nature is not yet conclusively proven. Thepurified pili were clearly fibrillar by transmission

electron microscopy but, like curli, could not besolubilized sufficiently to run on SDS-PAGE. Instead,LC-MS was used to identify a peptide sequence andhence the gene mtp (H37Rv ORF Rv3312A), whichcontains a putative transmembrane (TM) sequence(aa 1030). Mutants without this gene demonstratedreduced laminin-binding capabilities, consistent withthe fact that pili mediate specific recognition of hostcell receptors facilitating close contact and tissuecolonization [95]. M. tuberculosis preferentially at-taches to and invades damaged areas of the humanrespiratory mucosa in an organ culture system [96]. Inthese instances of tissue damage, extracellular matrixproteins such as laminin may be more highly exposedthan in healthy tissue.The mtp gene is present in only a few mycobacteriaand seems to be limited to the pathogenic species,underlining its role in cell invasion. An investigationby Dahl [97] reported surface projections in the formof extracellular fibrils from carbon-starved M. tuber-culosis cells. The fibrils could branch and adherebetween cells and to surfaces, forming meshes andclumping. The fibrils did not detach from the cell andwere clustered, suggesting a cellular source. They arevery similar to fibrils seen in M. avium subsp. para-tuberculosis [98] and might well be amyloid. Appen-dages termed pili are present on many bacteria thatcause disease in the human respiratory tract includingthe Gram-positive pathogens group B Streptococcusand Corynebacterium diphtheriae [99, 100]. Futurestudies will show whether some of these also areamyloid.

Amyloids in Gram-positive bacteriaAmong theGram-positive bacteria, only the amyloidsof S. coelicolor, a few other streptomycetes and M.tuberculosis have been studied in detail. However,beside these and the Mycolata, other Gram-positivebacteria belonging to the phyla Actinobacteria andFircimutes produce amyloids. Phylogenetically distantnon-Mycolata actinobacteria (from genera Actino-spica, Geodermatophilus and Streptosporangium)produce amyloid surface structures [P. L. Jensen,P. H. Nielsen and D. Otzen, unpublished results].Among Firmicutes, bacteria from the genera Bacillus,Paenibacillus and Enterococcus produce amyloid [22;P. L. Jensen, P. H. Nielsen and D. Otzen, unpublishedresults]. Several Bacillus species (e.g. Bacillus cereus)often form biofilms at the air-liquid surface whenreleasing spores to the air [101], so this could indicatea streptomycetes-like function of the amyloids on thecell-water surfaces by these species. Furthermore,spores from B. atrophaeus are covered with rodletswhich are structurally very similar to amyloid fibrils[102], so it seems that the spores can also be covered


by amyloids and this will explain their extremeresistance. These results strongly indicate that pro-duction of amyloids is widely distributed among mostGram-positive bacteria, and they form an integratedpart of the cellular and spore surface.

Amyloids in natural biofilmsWe have recently demonstrated the widespread dis-tribution of amyloids among microorganisms growingin biofilms from a lake, seawater, drinking water andwaste water treatment systems [22]. We initiallydetected surface-associated amyloids by ThT staining,following this up by two conformationally specificantibodies targeting amyloid directly (Fig. 4a). Thedetection methods were each combined with fluores-cence in situ hybridization using fluorescently labeledoligonucleotide probes in order to link phenotypewith identity of uncultured microorganisms. Repre-sentatives from Proteobacteria (Alpha-, Beta-,Gamma- and Deltaproteobacteria), Bacteriodetes,Chloroflexi and Actinobacteria, and most likely alsofrom other phyla, produced amyloids in these naturalbiofilms. Quantification of the microorganisms pro-ducing amyloid adhesins showed that they constituted

at least 540% of all prokaryotes present in thebiofilms (Fig. 4b). Most of the bacteria formedmonospecies microcolonies within the biofilms andthe amyloids could be visualized on the cell surfacesand in the extracellular matrix between the cells.Interestingly, some filamentous bacteria, e.g. from thephylum Chloroflexi, also expressed amyloids as a partof the sheath or close to the septum between theindividual cells in the filaments. Production of amy-loids was confirmed by environmental isolates be-longing to the Gammaproteobacteria, Bacteriodetes,Firmicutes and Actinobacteria.Some of these have been investigated in more detail.A 25-kDa protein constituting amyloid fimbriae couldbe purified from an isolate closely related to Pseudo-monans fluorescens (Fig. 5a, Table 2) [M. S.Dueholm,P. Larsen, J. L. Nielsen, P. H. Nielsen and D. E. Otzen,unpublished results]. Fibrillation could be reproducedin vitro. The amyloid-like structure was verified byFourier-transform infrared spectroscopy, circular di-chroism and ThT fluorescence (Fig. 5b). Partial se-quencing by MS/MS revealed that the fimbrin con-tained at least one repeated motif. This motif wasdifferent from those previously found for other known

Figure 4. (a) Simultaneous stain-ing of microcolonies with Thiofla-vinTand conformationally specificamyloid antibody [from ref. 22].Bar represents 20 mm. (b) Fractionof microorganisms from differentnatural biofilms that contain amy-loid as determined using the amy-loid-specific antibodies WO1 andWO2 [reproducedwith permissionfrom ref. 22].


amyloids, but it was not possible to find the corre-sponding gene in genomes from various pseudomo-nads. The diversity of species and morphologiesexpressing amyloids in natural biofilms suggest thattheir function is very diverse and not only directlyrelated to adhesion and biofilm formation.

Microcin E492 fromK. pneumoniae and harpins fromplant pathogensSome bacteria purposely produce cytotoxic proteins,some of which have recently been shown to haveamyloid properties. Microcin E492 is a small peptidebacteriocin naturally produced byK. pneumoniae thatkills Enterobacteriaceae bacteria by forming pores inthe cytoplasmic membrane [103]. Microcin activity ishighest during the exponential phase and disappearsin the stationary phase although production continues[103]: this is probably because it assembles intoamyloid-like fibrils, as shown to occur in vitro [75].The fibrils have the same structural, morphological,tinctorial and biochemical properties as the aggre-gates observed in the disease conditions. Aggregationis seeded by preformed aggregates. Amyloid forma-tion also occurs in vivo, where it is associated with aloss of toxicity of the protein. The finding thatmicrocin E492 naturally exists both as functionaltoxic pores and as harmless fibrils suggests thatprotein aggregation into amyloid fibrils can be usedto park potentially explosive proteins when theiraggressive properties are no longer needed. Interest-ingly, small annular Ab and a-synuclein protofibrils[104, 105] resemble the cytolyticb barrel pore-formingtoxins from bacteria such as Clostridium perfringens

[106], suggesting that a common structure could beresponsible for toxic membrane permeabilization.In many Gram-negative plant pathogens, the hrp(hypersensitive response and pathogenicity) genesencode proteins involved in the secretion of harpins.Harpins are heat-stable glycine-rich type III-secretedproteins which elicit a hypersensitive response inplants, i.e. an early defense response restricting growthof plant pathogens by causing cell death (similar toapoptosis) [76]. Mutants of these bacteria unable toproduce harpins lose their plant pathogenicity [107].Many Xanthomonas species produce harpins and theprotein HpaG fromX. axonopodis pv. glycines 8ra hasrecently been shown to form amyloid under plantapoplast conditions [76]. When expressed in E. coli,both oligomers, protofibrils and fibrils lead to ahypersensitive response in the plant cells (Fig. 6),which is unexpected in view of the conventionalwisdom that only the oligomers are the cytotoxicspecies [76]. It is possible that this may be related tothe fact that the harpin sequence has a region which ishomologous to the yeast prion protein [108], and theprion toxicity mechanism is not clearly understood.Harpins from other plant pathogens, such as Erwiniaamylovora and P. syringae, also form amyloid-likeprotofibrils [76]. It is presently not clear whether theformation of amyloid takes place on the surface of thebacteria or whether the monomers only or primarilyself-assemble in the plant cells. To our knowledge, onthe surface of these bacteria has not yet beeninvestigated whether amyloidic appendages are pres-ent, and we hope that future studies will address theseissues.

Figure 5. (a) Transmission electronmicroscopy picture of amyloid produced byPseudomonas species [Dueholm et al., unpublished data].(b)Circular dichroism spectra of amyloid producedbyPsedomonas sp. (squares) andCsgA fromE. coli (circles) resuspended inPBSbuffer(full lines) and dissolved in 80% formic acid (dashed lines) [Dueholm et al., unpublished data].


Model studies of the aggregation of bacterial proteinsThere are cases where aggregation of bacterialproteins under physiological conditions may be anunwanted spin-off of an otherwise useful propensity toassociate. Thus, the five Salmonella flagellar rodproteins normally associate in a-helical coiled coilsto form a straight rod structure [109]. However,overexpression of these proteins inE. coli leads four ofthem to self- or cross-aggregate under physiologicalconditions, rather than cross-reacting with flagellalcomplex [109]. A large part of these proteins is madeup of natively unfolded regions. The chaperone FliJappears to prevent unwanted aggregation by bindingto the unfolded regions [110].There exist reports of the aggregation of numerousother bacterial proteins such as HypF-N [111] and S6[112], but these studies have taken place in thepresence of organic solvent or extreme pH, and haveserved more to decipher the biophysical process ofaggregation per se. In the same vein, several studieshave focused on the fibrillation of peptides derivedfrom soluble bacterial proteins, e.g. CsgB [113] andOspA [114]. In general, the behavior of isolatedsequences which form part of an otherwise globularprotein may bear little resemblance to the propertiesof the globular protein, making such studies ofquestionable biological relevance (despite fascinatingproperties in their own right) unless the sequence canbe considered representative for the protein as such.Thus, already in 1992, Jarrett and Lansbury [115]studied the E. coli protein OsmB which mainlyconsists of repeat motifs showing a resemblance to

the Ab peptide; a 17-residue peptide representingsuch a motif fibrillated readily in vitro in a nucleation-dependent manner, similar to the fibrillar nature ofthe repeat motifs in CsgA [116]. In view of ourdiscovery of the widespread role of amyloid inbacterial biofilm [22], it is interesting that OsmB islinked to capsular polysaccharide synthesis whichplays a role in the later stages of biofilm development[117]. Despite their lack of homology, bothOsmB andthe bona fide bacterial amyloid protein CsgA consistof multiple repeats of a simple repeat [70].Heterologous overexpression of protein in bacteria inmany cases leads to its accumulation in insolubleinclusion bodies. Interestingly, proteins in these in-clusion bodies are not simply amorphous collectionsof unfolded proteins but possess genuine amyloidstructure, as seen by their high b sheet content as wellas their ability to seed deposition of soluble versions ofthe same protein and bind Congo Red and ThioflavinTCR/ThT [118]. Furthermore, aggregates from inclu-sion bodies can be cytotoxic to mammalian cells, incontrast to thermal aggregates of the same protein[118]. The presence of amyloid structures in inclusionbodies could stem from the generic ability of proteinsto form amyloid when forced to undergo foldingthrough partially structured intermediates at highconcentrations. They also illustrate yet again howNature can use amyloid as a storage state for surplusunwanted protein.

Figure 6. (a) Formation of oligomer structures by the harpin protein HpaG overexpressed with a His6 tag in E. coli. (b) Ability ofrecombinant HpaG to induce a hypersensitive response in tobacco leaves. An inactive variant L50P is included for comparison.[Reproduced with permission from ref. 76.]


Detecting, purifying and characterizing bacterialamyloidIt is clear from the many previous sections that muchhas yet to be discovered about the formation andaction of bacterial amyloid. Central to these efforts isan ability to identify, purify and analyse these proteins,and we will therefore in the following section focus onhow this may best be accomplished.Detection of bacterial amyloids in pure culture studiesis usually based on the binding of the dye Congo Redto the b sheets of the proteins [119]. In this way,bacteria that produce amyloids are generally easilyidentified by red colonies when growing on agar platescontaining Congo Red [120]. Amyloid fibrils can alsobe visualized by Congo Red and Sirius Red withbright-field or polarized light microscopy [121], whilestaining with fluorescent dyes can be carried out withCongoRed [122], ThT [123], 2-(4-methylaminophen-yl) or benzothiazole (BTA-1) [124]. ThT and CongoRed are among the most frequently used dyes foramyloid staining and ThT is considered to be the mostsensitive [121]. The specificity of ThTand Congo Redis not as good as the antibodies, because other organicmolecules, such as DNA and cellulose, can also bestained by ThT [125, 126] and Congo Red [127]. Thiskind of dye promiscuity reflects a profound truthabout the nature of their binding: although there islittle direct structural information on the nature of thedye binding [128], these types of dyes probably targetgrooves provided by regularly repeating structures,which may also allow them to dimerize and thusenhance their spectroscopic properties [129]. Inaddition, Congo Red can also react specifically withcertain non-amyloid proteins [20].We have found antibodies to be more reliable fordetection of amyloid [22, 121, 130]. The use ofconformationally specific antibodies (WO1 andWO2) is particularly efficient, because they specifi-cally target the fibril structure described above, with-out binding to monomers of the same proteins [21].Furthermore, antibodies can be combined with fluo-rescence in situ hybridization with oligonucleotideprobes for detection of microbes directly in mixedbiofilms [22]. Nevertheless, antibodies are not perfect,either: due to their large size, they may find it difficultto access amyloid fibrils if they are covered andmasked by other extracellular polymers, leading tofalse-negative results. We have found that saponifica-tion of Gram-positive bacteria by prolonged exposureto an alcoholic alkali solution at elevated temper-atures can improve antibody binding without destroy-ing the overall cellular architecture [P. L. Jensen, P. H.Nielsen and D. Otzen, unpublished results].The real proof for the presence of bacterial amyloids isusually regarded as transmission electron microscopy

images and purification in hot/boiling SDS, with averification of the structure using Fourier-transforminfrared spectroscopy, circular dichroism and ThTfluorescence. Furthermore, treatment with strong acidsuch as 90100% formic acid should release mono-mers that can be further purified and characterized bySDS-PAGEand various spectroscopic andmicroscopictechniques and sequenced by MS/MS. This requirespure cultures, vigorous expression of amyloids whichtypically only takes place under certain growth con-ditions and an efficient purification protocol, which inour experience needs optimization for each species dueto the very robust not to say recalcitrant amyloidstructures involved (cf. Table 2).In view of these difficulties, we believe that amyloidshave in many former studies been overlooked simplybecause the purification and detection methods werenot available or suitable. In future studies there will bea great challenge to refine and develop the aboveprocedures to reveal the true structure of these cellenvelopes.

Bacterial amyloids and diseasesBacterial amyloid may be associated with diseases innumerous ways. As previously mentioned, curli, piliand other amyloid states are implicated in the abilityof bacteria to attach to and invade human cells, withdetrimental effects for the host cell, while harpinsinduce plant cell apoptosis though the latter may beviewed as a defensemechanismby the plant to preventbacteria spreading to other parts of the organism.Other bacteria, e.g., the spirochete Borrelia burgdor-feri, induce Alzheimer disease-like symptoms (for-mation of hyperphosphorylated tau and amyloiddeposits of the Ab peptide as well as upregulation ofthe Ab precursor protein) in mammalian cells [131].The mechanism is unclear, but may involve exposureto the outer membranes lipopolysaccharide [131],possibly even in a templating process where theplaque forms around the spirochete [132]. Borrelia isnot known to form amyloid by itself. However, inother cases, it is possible that bacterial amyloid mayplay a direct role in disease induction. Nocardiaotitidiscaviarum can induce Parkinsons disease symp-toms in the brain [133]. Broxmeyer [134] hypothezisedthat several bacteria in the Mycolata (Nocardia andMycobacterium) may induce amyloidosis, as they areoften correlated with Parkinsons. As we have men-tioned, all these bacteria produce amyloid, whichcould be envisaged to seed amyloid formation in thebrain. This type of mechanism is supported by thework of Lundmark et al. [135], who injected curlifibrils into mice, leading to accelerated formation ofthe disease-associated amyloid proteinAA.However,it was not directly demonstrated that amyloid fibrils


can act as direct templates for further amyloidformation. If this is the case, amyloid-producingmicrobes may constitute a rich potential source either at infections or as environmental amyloids thatmay affect human health. The same research groupalso demonstrated that amyloid fibrils added to thedrinking water of a mouse model for amyloidosisdecreased the lag time to onset of disease [136],suggesting that the added amyloid can penetrate thegut and act as a seed for amyloid growth and diseaseinduction. Thus, amyloid-producing microbes consti-tute a source of environmental amyloids which mightpotentially act as seed for formation of amyloid inhumans.

Perspectives

This review has provided a glimpse of the great varietyof bacterial amyloids which play central roles inmaintaining structural integrity of individual cells aswell as whole bacterial communities. Apart from somevery general guidelines, nothing indicates that there isany simple unifying principle in the biological build-up and consolidation of amyloid; rather, its formationreflects the immediate functional demand of theamyloid-producing organism. There is likely to bejust as much variation among bacterial amyloidassembly as between bacterial and eukaryotic amy-loid. In this sense, bacterial amyloid nicely illustrateshow the principle of (more or less) self-organizedassembly can be tailored to meet specific require-ments. In the future, we expect that insights into themolecular assemblies of FAwill be very important touncover more of these biological principles. Someaspects that are of particular relevance include thefollowing:

the existence of sequence motifs for amyloidformation (e.g. repetitive elements and the phys-ico-chemical properties of the sequence)

the details of the nucleation mechanism (forexample, what determines whether it involves oneor several proteins)

fibril growth (directionality and polarity of growth,branching properties, rate of growth under differentconditions)

the molecular composition of the amyloid (singlecomponent or hybrid and how this affects materialproperties such as pliability and mechanicalstrength)

the molecular structure of the cross-b elements thatconstitute the core of the amyloid

the ultrastructure of the in vivo amyloid (fibrilquaternary structure, diameter, extent of coiling).

All these questions will ensure that FA research willbe an active and fertile research area formany years tocome.

Acknowledgements. We are grateful to P. Larsen, M. SimonsenDueholm,A. Yde, P. Lttge Jensen and J. L. Nielsen for theirmanyexperimental and intellectual contributions to our research onbacterial amyloid. D.O. acknowledges support from the VillumKannRasmussenFoundation and the research networkBioNETaswell as the Danish Research Foundation. P.H.N. acknowledgessupport from the Danish Research Foundation and AalborgUniversity.

1 Anfinsen, C. B. (1973) Principles that govern the folding ofprotein chains. Science 181, 223230.

2 Levinthal, C. (1968) Are there pathways for protein folding?J. Chim. Phys. 85, 4445.

3 Tanford, C. (1968) Protein denaturation. B. The transitionfromnative to denatured state.Adv. Prot.Chem. 23, 218282.

4 Fersht, A. R. (1999) Structure and Mechanism in ProteinScience: A Guide to Enzyme Catalysis and Protein Folding.Freeman, New York.

5 Matouschek, A., Kellis, J. T., Serrano, L. and Fersht, A. R.(1989) Mapping the transition state and pathway of proteinfolding by protein engineering. Nature 340, 122126.

6 Itzhaki, L. S., Otzen, D. E. and Fersht, A. R. (1995) Thestructure of the transition state for folding of chymotrypsininhibitor 2 analysed by protein engineering: evidence for anucleation-condensation mechanism for protein folding.J. Mol. Biol. 254, 260288.

7 Ladurner, A. G., Itzhaki, L. S., Daggett, V. and Fersht, A. R.(1998) Synergy between simulation and experiment indescribing the energy landscape of protein folding. Proc.Natl. Acad. Sci. USA 95, 84738478.

8 Ellis, J. E., Laskey, R. A. and Lorimer, G. H., (Eds.) (1993)Molecular chaperones. Chapman & Hall, London.

9 Spillantini, M. G., Schmidt, M. L., Lee, V. M.-Y., Trojanow-ski, J. Q., Jakes,R. andGoedert,M. (1997)Alpha-synuclein inLewy bodies., Nature 388, 839840.

10 Gebbink,M. F. B. G., Claessen,D., Bouma,B.,Dijkhuizen,L.andWsten, H. A. B. (2005) Amyloids a functional coat formicroorganisms. Nat. Rev. Microbiol. 3, 333341.

11 Chiti, F. and Dobson, C. M. (2006) Protein misfolding,functional amyloid, and human disease. Annu. Rev. Bio-chem. 75, 333366.

12 Barnhart,M. M. and Chapman,M. R. (2006) Curli biogenesisand function. Annu. Rev. Microbiol. 60, 131147.

13 Kelly, J. W. and Balch, W. E. (2003) Amyloid as a naturalproduct. J. Cell Biol. 161, 461462.

14 Fowler, D. M., Koulov, A. V., Balch, W. E. and Kelly, J. W.(2007) Functional amyloid from bacteria to humans. TrendsBiochem. Sci. 32, 217224.

15 Talbot, N. J. (2003) Aerial morphogenesis: enter the chaplins.Curr. Biol. 13, R696698.

16 Morozova-Roche, L. and Malusauskas, M. (2007) A falseparadise mixed blessings in the protein universe: theamyloid as a new challenge in drug development. Curr.Med. Chem. 14, 12211230.

17 Fndrich, M. (2007) On the structural definition of amyloidfibrils and other polypeptide aggregates. Cell. Mol. LifeSci. 64, 20662078.

18 Virchow, R. (1854) Weitere Mittheilungen ber das Vorkom-men der pflanzlichen Cellulose beim Menschen. VirchowsArch. 6, 268271.

19 Friedrich, N. and Kekul, A. (1859) Zur Amyloidfrage.Pathol. Anat. Physiol. Klin. Med. 16, 5065.

20 Khurana, R., Uversky, V. N., Nielsen, L. and Fink, A. L.(2001) Is Congo Red an amyloid-specific dye? J. Biol.Chem. 276, 2271522721.


21 ONuallain, B. and Wetzel, R. (2002) Conformational Absrecognizing a generic amyloid fibril epitope. Proc. Natl. Acad.Sci. USA 99, 14851490.

22 Larsen, P., Dueholm, M., Christiansen, G., Nielsen, J. L.,Otzen, D. E. and Nielsen, P. H. (2007) Amyloid adhesin areabundant in natural biofilms. Env. Microbiol. 9, 30773090.

23 Sunde, M., Serpell, L. C., Bartlam, M., Fraser, P. E., Pepys,M. B. and Blake, C. C. F. (1997) Common core structure ofamyloid fibrils by synchrotron X-ray diffraction. J. Mol. Biol.273, 729739.

24 Chiti, F., Webster, P., Taddei, N., Clark, A., Stefani, M.,Ramponi, G. and Dobson, C. M. (1999) Designing conditionsfor in vitro formation of amyloid protofilaments and fibrils.Proc. Natl. Acad. Sci. USA 96, 35903594.

25 Ferguson, N., Becker, J. , Tidow, H., Tremmel, S. , Sharpe,T. D., Krause, G., Flinders, J. , Berriman, J. , Oschkinat, H.and Fersht, A. R. (2007) General structural motifs ofamyloid protofilaments. Proc. Natl. Acad. Sci. USA 103,1624816253.

26 Balbuch, J. J. , Ishii, Y. , Antzutkin, O. N. , Leapman, R. D. ,Rizzo, N. W. , Dyda, F. , Reed, J. and Tycko, R. (2000)Amyloid fibril formation by Ab16 22, a seven-residuefragment of the Alzheimers b-amyloid peptide, andstructural characterization by solid state NMR. Biochem-istry 39, 13748 13759.

27 Iwata, K., Fujiwara, T., Matsuki, Y., Akutsu, H., Takahashi,S., Naiki, H. and Goto, Y. (2006) 3D structure of amyloidprotofilaments of beta2-microglobulin fragment probed bysolid-state NMR. Proc. Natl. Acad. Sci. USA 103, 1811918124.

28 Heise, H., Hoyer,W., Becker, S., Andronesi,O. C., Riedel,D.and Baldus, M. (2005) Molecular-level secondary structure,polymorphism, and dynamics of full-length alpha-synucleinfibrils studied by solid-stateNMR.Proc. Natl. Acad. Sci.USA102, 1587115876.

29 Nelson, R., Sawaya, M. R., Balbirnie, M., Madsen, A. .,Riekel, C., Grothe, R. and Eisenberg, D. (2005) Structure ofthe cross-beta spine of amyloid-like fibrils. Nature 435, 773778.

30 Sawaya, M. R., Sambashivan, S., Nelson, R., Ivanova, M. I.,Sievers, S. A., Apostol,M. I., Thompson,M. J., Balbirnie,M.,Wiltzius, J. J. W., McFarlane, H. T., Madsen, A. ., Riek, R.and Eisenberg, D. (2007) Atomic structures of amyloid cross-b spines reveal varied steric zippers. Nature 447, 453457.

31 Koch, M. H., Vachette, P. and Svergun, D. I. (2003) Small-angle scattering: a view on the properties, structures andstructural changes of biological macromolecules in solution.Q. Rev. Biophys. 36, 147227.

32 Vestergaard, B., Groenning, M., Roessle, M., Kastrup, J. S.,Van de Weert, M., Flink, J. M., Frokjaer, S., Gajhede, M. andSvergun, D. I. (2007) A helical structural nucleus is theprimary elongating unit of insulin amyloid fibril. PLoSBiol. 5,e134.

33 Frokjaer, S. and Otzen, D. E. (2005) Protein drug stability aformulation challenge. Nat. Rev. Drug. Deliv. 4, 298306.

34 Chen, C.-D., Huff, M. E., Matteson, J., Page, L., Phillips, R.,Kelly, J. W. and Balch, W. E. (2001) Furin initiates gelsolinfamilial amyloidosis in the Golgi through a defect in Ca2+stabilization. EMBO J. 20, 62776287.

35 Kim, S. H., Wang, R., Gordon, D. J., Bass, J., Steiner, D. F.,Lynn, D. G., Thinakaran, G., Meredith, S. C. and Sisodia, S. S.(1999) Furin mediates enhanced production of fibrillogenicABri peptides in familial British dementia. Nat. Neurosci. 11,984988.

36 Harper, J. D. and Lansbury, P. T. J. (1997) Models of amyloidseeding in Alzheimers disease and scrapie: mechanistictruths and physiological consequences of the time-dependentsolubility of amyloid proteins. Annu. Rev. Biochem. 66, 385407.

37 Ferrone, F. (1999) Analysis of protein aggregation kinetics.Methods Enzymol. 309, 256274.

38 Pallitto, M. M. and Murphy, R. M. (2001) A mathematicalmodel of the kinetics of b-amyloid fibril growth from thedenatured state. Biophys. J. 81, 18051822.

39 Ban, T., Hamada, D., Hasegawa, K., Naiki, H. and Goto, Y.(2003) Direct observation of amyloid fibril growth monitoredby thioflavin T fluorescence. J. Biol. Chem. 278, 1646216465.

40 Shahi, P., Sharma, R., Sanger, S., Kumar, I. and Jolly, R. S.(2007) Formation of amyloid fibrils via longitudinal growth ofoligomer. Biochemistry 46, 73657373.

41 Khurana, R., Ionescu-Zanetti, C., Pope, M., Li, J., Nielson,L., Ramirez-Alvarado,M., Regan, L., Fink, A. L. and Carter,S. A. (2003) A general model for amyloid fibril assemblybased on morphological studies using atomic force micro-scopy. Biophys. J. 85, 11351144.

42 Otzen, D. E. andOliveberg, M. (2004) Transient formation ofnanocrystalline structures during fibrillation of anAlzheimer-like peptide. Prot. Sci. 13, 14171421.

43 Calamai, M., Canale, C., Relini, A., Stefani, M., Chiti, F. andDobson, C. M. (2005) Reversel of protein aggregationprovides evidence for multiple aggregated states. J. Mol.Biol. 346, 603616.

44 Pedersen, J. S., Dikov, D., Flink, J. L., Hjuler, H. A., Chris-tiansen, G. and Otzen, D. E. (2006) The changing face ofglucagon fibrillation: structural polymorphism and conforma-tional imprinting. J. Mol. Biol. 355, 501523.

45 Pedersen, J. S., Dikov, D., Flink, J. L. and Otzen, D. E. (2006)Sulfates dramatically stabilize a salt dependent type ofglucagon fibrils. Biophys. J. 90, 41814194.

46 Pedersen, J. S., Dikov, D. and Otzen, D. E. (2006) N- and C-terminal hydrophobic patches are involved in fibrillation ofglucagon. Biochemistry 45, 1450314512.

47 Petkova, A. T., Leapman, R. D., Guo, Z., Yau, W.-M.,Mattson, M. P. and Tycko, R. (2005) Self-propagating, molec-ular-level polymorphism in Alzheimers b-amyloid fibrils.Science 307, 262265.

48 Pedersen, J. S. and Otzen, D. E. (2008) Amyloid a state inmany guises: survival of the fittest fibril fold. Protein Sci. 17,19.

49 Chiti, F., Stefani, M., Taddei, N., Ramponi, G. and Dobson,C. M. (2003) Rationalization of the effects of mutations onpeptide and protein aggregation rates. Nature 424, 805808.

50 DuBay, K. F., Pawar, A. P., Chiti, F., Zurdo, J., Dobson, C. M.and Vendruscolo, M. (2004) Predicting absolute aggregationrates of amyloidogenic polypeptide chains. J. Mol. Biol. 341,13171326.

51 Pawar, A. P., DuBay, K. F., Zurdo, J., Chiti, F., Vendruscolo,M. and Dobson, C. M. (2005) Prediction of aggregation-prone and aggregation-susceptible regions in proteinsassociated with neurodegenerative diseases. J. Mol. Biol.350, 379392.

52 Fernandez-Escamilla, A. M., Rousseau, F., Schymkowitz, J.and Serrano, L. (2004) Prediction of sequence-dependent andmutational effects on the aggregation of peptides and protein.Nat. Biotechnol. 22, 13021306.

53 Pepys, M. B., Hawkins, P. H., Booth, D. R., Vigushin, D. M.,Tennent, G. A., Soutar, A. K., Totty, N., Bguyen, O., Blake,C. C., Terry, C. J., Feest, T. G., Zalin, A. M. and Hsuan, J. J.(1993) Human lysozyme gene mutations cause hereditarysystemic amyloidosis. Nature 362, 553557.

54 Kayed, R., Head, E., Thompson, J. L., McIntire, T. M.,Milton, S. C., Cotman, C. W. and Glabe, C. G. (2003) Com-mon structure of soluble amyloid oligomers implies commonmechanism of pathogenesis. Science 300, 486489.

55 Bucciantini, M., Giannoni, E., Chiti, F., Baroni, F., Formigli,L., Zurdo, J., Taddei, N., Ramponi, G., Dobson, C. M. andStefani, M. (2002) Inherent toxicity of aggregates implies acommon mechanism for protein misfolding diseases. Nature416, 507511.

56 Malisauskas, M., Ostman, J., Darinskas, A., Zamotin, V.,Liutkevicius, E., Lundgren, E. and Morozova-Roche, L. A.(2005) Does the cytotoxic effect of transient amyloid oligom-


ers from common equine lysozyme in vitro imply innateamyloid toxicity? J. Biol. Chem. 280, 62696275.

57 Volles, M. J. and Lansbury, P. T. (2003) Zeroing in on thepathogenic form of alpha-synuclein and its mechanism ofneurotoxicity in Parkinsons disease. Biochemistry 42, 78717878.

58 Kourie, J. I., Culverson, A. L., Farrelly, P. V., Henry, C. L. andLaohachai, K. N. (2002) Heterogeneous amyloid-formed ionchannels as a common cytotoxic mechanism. Cell Biochem.Biophys. 36, 191207.

59 Lashuel, H. A. and Lansbury, P. T. (2006) Are amyloiddiseases caused by protein aggregates that mimic bacterialpore-forming toxins? Q. Rev. Biophys., 135.

60 Sokolov, Y., Kozak, J. A., Kayed, R., Chanturiya, A., Glabe,C. G. and Hall, J. E. (2006) Soluble amyloid oligomersincrease bilayer conductance by altering dielectric structure.J. Gen. Physiol. 128, 637647.

61 Diociaiuti, M., Polzi, L. Z., Valvo, L., Malchoidi-Albedi, F.,Bombelli, C. and Gaudiano, M. C. (2006) Calcitonin formsoligomeric pore-like structures in lipid membrane. Biophys. J.91, 22752281.

62 Welch, M. D. and Mullins, R. D. (2002) Cellular control ofactin nucleation. Annu. Rev. Cell. Dev. Biol. 18, 247288.

63 Silver, F. H., Freeman, J. W. and Seehra,G. P. (2003)Collagenself-assembly and the development of tendon mechanicalproperties. J. Biomech. 36, 15291553.

64 Sato, F., Wachi, H., M., I., Nonaka, R., Onoue, S., Urban, Z.,Starcher, B. C. and Seyama, Y. (2007) Distinct steps of cross-linking, self-association, and maturation of tropoelastin arenecessary for elastic fiber formation. J. Mol. Biol. 369, 841851.

65 Fowler, D. M., Koulov, A. V., Alory-Jost, C., Marks, M. S.,Balch, W. E. and Kelly, J. W. (2005) Functional amyloidformation within mammalian tissue. PLoS Biol. 4, 18.

66 Berson, J. F., Theos, A. C., Harper, D. C., Tenza, D., Raposo,G. and Marks, M. S. (2003) Proprotein convertase cleavageliberates a fibrillogenic fragment of a resident glycoprotein toinitiate melanosome biogenesis. J. Cell. Biol. 161, 521533.

67 Berson, J. F., Harper, D. C., Tenza, D., Raposo, G. andMarks,M. S. (2001) Pmel17 initiates premelanosome morphogenesiswithin multivesicular bodies. Mol. Biol. Cell 12, 34513464.

68 Chapman, M. R., Robinson, L. S., Pinkner, J. S., Roth, R.,Heuser, J., Hammar, M., Normark, S. and Hultgren, S. J.(2002) Role of Eschericia coli curli operons in directingamyloid fiber formation. Science 295, 851855.

69 Hammer, N. D., Schmidt, J. C. and Chapman, M. R. (2007)The curli nucleator protein, CsgB, contains an amyloidogenicdomain that directs CsgA polymerization. Proc. Natl. Acad.Sci. USA 104, 1249412499.

70 Wang, X., Smith, D. R., Jones, J. W. and Chapman, M. R.(2007) In vitro polymerization of a functionalEscherichia coliamyloid protein. J. Biol. Chem. 282, 37133719.

71 Olsn, A., Jonsson, A. and Normark, S. (1989) Fibronectinbinding mediated by a novel class of surface organelles onEscherichia coli. Nature 338, 652655.

72 Rmling, U., Bian, Z., Hammar, M., Sierralta, W. D. andNormark, S. (1998) Curli fibers are highly conserved betweenSalmonella typhimurium and Escherichia coli with respect toopen structure and regulation. J. Bacteriol. 180, 722731.

73 Claessen, D., Rink, R., de Jong, W., Siebring, J., de Vreughd,P., Boersma, F. G. H., Dijkhuizen, L. and Wsten, H. A. B.(2003) A novel class of secreted hydrophobic proteins isinvolved in aerial hyphae formation in Streptomyces coeli-color by forming amyloid-like fibrils. Genes Dev. 17, 17141726.

74 Alteri, C. J., Xicohtencatl-Cortes, J., Hess, S., Caballero-Olin,G., Giron, J. A. and Friedman, R. L. (2007) Mycobacteriumtuberculosis produces pili during human infection. Proc. Natl.Acad. Sci. USA 104, 51455150.

75 Bieler, S., Estrada, L., Lagos, R., Baeza, M., Castilla, J. andSoto, C. (2005) Amyloid formation modulates the biological

activity of a bacterial protein. J. Biol. Chem. 280, 2688026885.

76 Oh, J., Kim, J.-G., Jeon, E., Yo, C.-H., Moon, J. S., Rhee, S.and Hwang, I. (2007) Amyloidogenesis of type III-dependentharpins from plant pathogenic bacteria. J. Biol. Chem. 282,1360113609.

77 Zogaj, X., Bokranz, W., Nimtz, M. and Rmling, U. (2003)Production of cellulose and curli fimbriae by members of thefamily Enterobacteriaceae isolated from the human gastro-intestinal tract. Infect. Immun. 71, 41514158.

78 Jonas, K., Tomenius, H., Kader, A., Normark, S., Rmling,U., Belova, L. M. and Melefors, O. (2007) Roles of curli,cellulose and BapA in Salmonella biofilm morphologystudied by atomic force microscopy. BMC Microbiol. 7, 70.

79 Robinson, L. S., Ashman, E. M., Hultgren, S. J. and Chap-man,M. R. (2006) Secretion of curli fibre subunits ismediatedby the outer membrane-localized CsgG protein. Mol. Micro-biol. 59, 870881.

80 Kanamaru, S., Kurazono, H., Terai, A., Monden, K., Kumon,H., Mizunoe, Y., Ogawa, O. and Yamamoto, S. (2006)Increased biofilm formation in Escherichia coli isolatedfrom acute prostatitis. Int. J. Antimicrob. Agents 28 (Suppl.1) S21S25.

81 Rochon, M. and Rmling, U. (2006) Flagellin in combinationwith curli fimbriae elicits an immune response in the gastro-intestinal epithelial cell line HT-29. Microbes Infect. 8, 20272033.

82 Barak, J. D., Gorski, L., Naraghi-Arani, P. and Charkowski,A. O. (2005) Salmonella enterica virulence genes are requiredfor bacterial attachment to plant tissue. Appl. Env. Micro-biol. 71, 56855691.

83 Jeter, C. and Matthysse, A. G. (2005) Characterization of thebinding of diarrheagenic strains ofE. coli to plant surfaces andthe role of curli in the interaction of the bacteria with alfalfasprouts. Mol. Plant Microbe Interact. 18, 12351242.

84 Ryu, J. H. and Beuchat, L. R. (2005) Biofilm formation byEscherichia coli O157:H7 on stainless steel: effect of exopo-lysaccharide and Curli production on its resistance tochlorine. Appl. Environ. Microbiol. 71, 247254.

85 Ryu, J. H., Kim, H., Frank, J. F. and Beuchat, L. R. (2004)Attachment and biofilm formation on stainless steel byEscherichia coli O157:H7 as affected by curli production.Lett. Appl. Microbiol. 39, 359362.

86 Prigent-Combaret, C., Prensier, G., Le Thi, T. T., Vidal, O.,Lejeune, P. and Dorel, C. (2000) Developmental pathway forbiofilm formation in curli-producing Escherichia coli strains:role of flagella, curli and colanic acid. Environ. Microbiol. 2,450464.

87 Elliot, M. A., Karoonuthaisir, N., Huang, J., Bibb, M. J.,Cohen, S. N., Kao, C. M. and Buttner, M. J. (2003) Thechaplins: a family of hydrophobic cell-surface proteinsinvolved in aerial mycelium formation in Streptomyces coeli-color. Genes Dev. 17, 17271740.

88 Capstick, D. A., Willey, J. M., Buttner, M. J. and Elliot, M. A.(2007) SapB and the chaplins: connections between morpho-genetic proteins in Streptomyces coelicolor. Mol. Micro-biol. 64, 602613.

89 Wsten, H. A. B. and de Vocht, M. L. (2000) Hydrophobins,the fungal coat unravelled. Biochim. Biophys. Acta 1469, 7986.

90 Chater, K. F. and Chandra, G. (2006) The evolution ofdevelopment in Streptomyces analysed by genome compar-isons. FEMS Microbiol. Rev. 30, 651672.

91 Claessen, D., de Jong, W., Dijkhuizen, L. and Wsten,H. A. B. (2006) Regulation of Streptomyces development:reach for the sky! Trends Microbiol. 14, 313319.

92 Claessen, D., Stokroos, I., Deelstra, H. J., Penninga, N. A.,Bormann, C., Salas, J. A., Dijkhuizen, L. and Wsten,H. A. B. (2004) The formation of the rodlet layer of strepto-mycetes is the result of the interplay between rodlins andchaplins. Mol. Microbiol. 53, 433443.


93 Del Sol, R., Armstrong, I., Wright, C. and Dyson, P. (2007)Characterization of changes to the cell surface during the lifecycle of Streptomyces coelicolor : atomic force microscopy ofliving cells. J. Bacteriol. 189, 22192225.

94 Elliot, M. A. and Talbot, N. J. (2004) Building filaments in theair: aerial morphogenesis in bacteria and fungi. Curr. Opin.Microbiol. 7, 594601.

95 Klemm, P. and Schembri, M. A. (2000) Bacterial adhesins:function and structure. Int. J. Med. Microbiol. 290, 2735.

96 Middleton, A. M., Chadwick, M. V., Nicholson, A. G.,Dewar, A., Groger, R. K., Brown, E. J., Ratliff, T. L. andWilson, R. (2002) Interaction ofMycobacterium tuberculosiswith human respiratory mucosa. Tuberculosis (Edinb.) 82,6978.

97 Dahl, J. L. (2005) Scanning electron microscopy analysis ofagedMycobacterium tuberculosis cells. Can. J. Microbiol. 51,277281.

98 Merkal, R. S., Rhoades, K. R., Gallagher, J. E. and Ritchie,A. E. (1973) Scanning electron microscopy of mycobacteria.Am. Rev. Respir. Dis. 108, 381387.

99 Ton-That, H. and Schneewind, O. (2003) Assembly of pili onthe surface of Corynebacterium diphtheriae. Mol. Micro-biol. 50, 14291438.

100 Lauer, P., Rinaudo, C. D., Soriani, M., Margarit, I. , Maione,D., R., R., Taddei, A. R., Mora,M., Rappuoli, R., Grandi, G.and Telford, J. L. (2005) Genome analysis reveals pili inGroup B Streptococcus. Science 309, 105.

101 Wijman, J. G., de Leeuw, P. P., Moezelaar, R., Zwietering,M. H. and Abee, T. (2007) Air-liquid interface biofilms ofBacillus cereus : formation, sporulation and dispersion. Appl.Environ. Microbiol. 73, 14811488.

102 Plomp, M., Leighton, T. J., Wheeler, K. E., Hill, H. D. andMalkin, A. J. (2007) In vitro high-resolution structural dy-namics of single germinating bacterial spores. Proc. Natl.Acad. Sci. USA 104, 96449649.

103 Destoumieux-Garzon, D., Thomas, X., Santamaria, M.,Goulard, C., Barthelemy, M., Boscher, B., Bessin, Y., Molle,G., Pons, A. M., Letellier, L., Peduzzi, J. and Rebuffat, S.(2003) Microcin E492 antibacterial activity: evidence for aTonB-dependent inner membrane permeabilization on Es-cherichia coli. Mol. Microbiol. 49, 10311041.

104 Lashuel, H. A., Hartley, D., Petre, B. M., Weals, T. andLansbury, P. T. (2002) Amyloid pores from pathogenicmutations. Nature 418, 291.

105 Dang, T. X., Hotze, E. M., Rouiller, I., Tweten, R. K. andWilson-Kubalek, E. M. (2005) Prepore to pore transition of acholesterol-dependent cytolysin visualized by electronmicro-scopy. J. Struct. Biol. 150, 100108.

106 Tweten, R. K., Parker, M. W. and Johnson, A. E. (2001) Thecholesterol-dependent cytolysins. Curr. Top. Microbiol. Im-munol. 257, 1533.

107 Wei, K., Tang, D. J., He, Y. Q., Feng, J. X., Jiang, B. L., Lu,G. T., Chen, B. L. and Tang, J. L. (2007) hpaR, a putativemarR family transcriptional regulator, is positively controlledby HrpG and HrpX and involved in the pathogenesis,hypersensitive response, and extracellular protease produc-tion of Xanthomonas campestris pathovar campestris. J. Bac-teriol. 189, 20552062.

108 Kim, J. G., Jeon, E., Oh, J., Moon, J. S. and Hwang, I. (2004)Mutational analysis of Xanthomonas harpin HpaG identifiesa key functional region that elicits the hypersensitive responsein nonhost plants. J. Bacteriol. 186, 62396247.

109 Saijo-Hamano, Y., Uchida, N., Namba, K. and Oosawa, K.(2004) In vitro characterization of FlgB, FlgC, FlgF, FlgG, andFliE, flagellar basal body proteins of Salmonella. J. Mol. Biol.339, 423435.

110 Minamino, T. and Macnab, R. M. (1999) Components of theSalmonella flagellar export apparatus and classification ofexport substrates. J. Bacteriol. 181, 13881394.

111 Marcon, G., Plakoutsi, G., Canale, C., Relini, A., Taddei, N.,Dobson, C. M., Ramponi, G. and Chiti, F. (2005) Amyloid

formation from HypF-N under conditions in which theprotein is initially in its native state. J. Mol. Biol. 347, 323335.

112 Pedersen, J. S., Christiansen, G. and Otzen, D. E. (2004)Modulation of S6 fibrillation by unfolding rates and gate-keeper residues. J. Mol. Biol. 341, 575588.

113 Gro, M., Wilkins, D. K., Pitkeathly, M. C., Chung, E. W.,Higham,C., Clark,A. andDobson,C. M. (1999) Formation ofamyloid fibrils by peptides derived from the bacterial coldshock protein CspB. Prot. Sci. 8, 13501357.

114 Ohnishi, S., Koide, A. and Koide, S. J. (2000) Solutionconformation and amyloid-like fibril formation of a polarpeptide derived from a b-hairpin in the OspA single-layer b-sheet. J. Mol. Biol. 301, 477489.

115 Jarrett, J. T. and Lansbury, P. T. (1992) Amyloid fibrilformation requires a chemically discriminating nucleationevent: studies of an amyloidogenic sequence from thebacterial protein OsmB. Biochemistry 31, 1234512352.

116 Cherny, I., Rockah, L., Levy-Nissenbaum, O., Gophna, U.,Ron, E. Z. and Gazit, E. (2005) The formation of Escherichiacoli curli amyloid fibrils is mediated by prion-like peptiderepeats. J. Mol. Biol. 352, 245252.

117 Majdalani, N. and Gottesman, S. (2005) The Rcs phosphor-elay: a complex signal transduction system. Annu. Rev.Microbiol. 59, 379405.

118 Gonzlez-Montalbn, N., Villaverde, A. and Aris, A. (2007)Amyloid-linked cellular toxicity triggered by bacterial inclu-sion bodies. Biochem. Biophys. Res. Commun. 355, 637642.

119 Klunk, W. E., Jacob, R. F. and Mason, R. P. (1999) Quantify-ing amyloid by Congo Red spectral shift assay. MethodsEnzymol. 309, 285305.

120 Collinson, S. K., Doig, P. C., Doran, J. L., Clouthier, S., Trust,T. J. and Kay, W. W. (1993) Thin, aggregative fimbriaemediate binding of Salmonella enteritidis to fibronectin.J. Bacteriol. 175, 1218.

121 Westermark, G. T., Johnson, K. H. andWestermark, P. (1999)Staining methods for identification of amyloid in tissue.Methods Enzymol. 309, 325.

122 Linke, R. P. (2000) Highly sensitive diagnosis of amyloid andvarious amyloid syndromes using Congo red fluorescence.Virchows Arch. 436, 439448.

123 Krebs,M. R. H., Bromley, E. H. C., Rogers, S. S. andDonald,A. M. (2005) Themechanismof amyloid spherulite formationby bovine insulin. Biophys. J. 88, 20132021.

124 Klunk, W. E., Wang, Y., Huang, G. F., Debnath, M. L., Holt,D. P. and Mathis, C. A. (2001) Uncharged thioflavin-Tderivatives bind to amyloid-beta protein with high affinityand readily enter the brain. Life Sci. 69, 14711484.

125 Raj, C. R. and Ramaraj, R. (2001) Emission of thioflavin Tand its off-on control in polymer membranes. Photochem.Photobiol. 74, 752759.

126 Ilanchelian,M. andRamaraj, R. (2004) Emission of thioflavinT and its control in the presence of DNA. J. Photochem.Photobiol. A 162, 129137.

127 Wood, P. J. and Fulcher, R. G. (1978) Interaction of some dyeswith cereal 3-glucans. Cereal Chem. 55, 952966.

128 Turnell, W. G. and Finch, J. T. (1992) Binding of the dyeCongo Red to the amyloid protein pig insulin reveals a novelhomology amongst amyloid-forming peptide sequences.J. Mol. Biol. 227, 12051223.

129 Groenning, M., Olsen, L., Van de Weert, M., Flink, J. M.,Frokjaer, S. and Jorgensen, F. S. (2007) Studyon the bindingofthioflavin T to beta-sheet-rich and non-beta-sheet cavities.J. Struct. Biol. 158, 358369.

130 Glabe, C. G. (2004) Conformation-dependent antibodiestarget diseases of protein misfolding. Trends Biochem. Sci.29, 542547.

131 Miklossy, J., Kis, A., Radenovic, A., Miller, L., Forro, L.,Martins, R., Reiss, K., Darbinian, N., Darekar, P., Mihaly, L.and Khalili, K. (2006) Beta-amyloid deposition and Alzheim-ers type changes induced by Borrelia spirochetes. Neurobiol.Aging 27, 223236.


132 MacDonald, A. B. (2006) Plaques of Alzheimers diseaseoriginate from cysts ofBorrelia burgdorferi, the Lyme diseasespirochete. Med. Hypoth. 67, 592600.

133 Daz-Corrales, F. J., Colasante, C., Contreras, Q., Puig, M.,Serrano, J. A., Hernndez, I. and Beaman, B. I. (2004)Nocardia otitidiscaviarum (GAM-5) induces parkinsonian-like alterations in mouse. Braz. J. Med. Biol. Res. 37, 539548.

134 Broxmeyer, L. (2002) Parkinsons: another look. Med.Hypoth. 59, 373377.

135 Lundmark,K.,Westermark,G.,Olsn,A. andWestermark, P.(2005)Protein fibrils in nature can enhance amyloid proteinAamyloidosis in mice: cross-seeding as a disease mechanism.Proc. Natl. Acad. Sci. USA 102, 60986102.

136 Lundmark, K., Westermark, G. T., Nystrm, S., Murphy,C. L., Solomon,A. andWestermark, P. (2002)Transmissibilityof systemic amyloidosis by a prion-like mechanism. Proc.Natl. Acad. Sci. USA 99, 6976984.

137 Iconomidou, V. A., Vriend, G. and Hamodrakas, S. J. (2000)Amyloids protect the silkmoth oocyte and embryo. FEBSLett. 479, 141145.

138 Iconomidou, V. A., Chryssikos, G. D., Giorius, V., Galarius,A. S., Cordopatis, P., Hoenger, A. and Hamodrakas, S. J.(2006) Amyloid fibril formation propensity is inherent intothe hexapeptide tandemly repeating sequence of the centraldomain of silkmoth chorion proteins of theA-family. J. Struct.Biol. 156, 480488.

139 Podrabsky, J. E., Carpenter, J. F. and Hand, S. C. (2001)Survival of water stress in annual fish embryos: dehydrationavoidance and egg envelope amyloid fibers. Am. J. Physiol.Regul. Integr. Comp. Physiol. 280, R123131.

140 Kenney, J. M., Knight, D., Wise, M. J. and Vollrath, F. (2002)Amyloidogenic nature of spider silk. Eur. J. Biochem. 269,41594163.

141 Morr, D. J. and Morr, D. M. (2003) Cell surface NADHoxidases (ECTO-NOX proteins) with roles in cancer, cellulartime-keeping, growth, aging and neurodegenerative diseases.Free Radic. Res. 37, 795808.

142 Si, K., Lindquist, S. L. and Kandel, E. R. (2003) A neuronalisoform of the aplysia CPEB has prion-like properties, Cell115, 879891.

143 Baxa, U., Cheng, N., Winkler, D. C., Chiu, T. K., Davies,D. R., Sharma, D., Inouye, H., Kirschner, D. A., Wickner,R. B. and Steven, A. C. (2005) Filaments of the Ure2p prionprotein have a cross-b core structure. J. Struct. Biol. 150, 170179.

144 True, H. L. and Lindquist, S. L. (2000) Ayeast prion providesa mechanism for genetic variation and phenotypic diversity.Nature 407, 477483.

145 Coustou, V., Deleu, C., Saupe, S. and Begueret, J. (1997) Theprotein product of the het-s heterokaryon incompatibilitygene of the fungus Podospora anserina behaves as a prionanalog. Proc. Natl. Acad. Sci. USA 94, 97739778.

To access this journal online:http://www.birkhauser.ch/CMLS


Documents

We find them here, we find them there: Functional bacterial amyloid