Current Protein and Peptide Science, 291-309 291 Natural...

Current Protein and Peptide Science, 2008, 9, 291-309 291

1389-2037/08 $55.00+.00 © 2008 Bentham Science Publishers Ltd.

Natural Protective Amyloids

Vassiliki A. Iconomidou and Stavros J. Hamodrakas*

Department of Cell Biology and Biophysics, Faculty of Biology, University of Athens, Panepistimiopolis, Athens 157 01, Greece

Abstract: Amyloidoses are a group of diseases including neurodegenerative diseases like Alzheimer’s disease and also type II diabetes, spongiform encephalopathies and many others, believed to be caused by protein aggregation and subse-quent amyloid fibril formation. However, occasionally, living organisms exploit amyloid fibril formation, a property in-herent into amino acid sequences, and perform specific physiological functions from amyloids, in differing biological con-texts. Some of these functional amyloids are natural protective amyloids. Here, we review recent evidence on silkmoth chorion protein synthetic peptide-analogues that documents the function of silkmoth chorion, the major component of the eggshell, a structure with extraordinary physiological and mechanical properties, as a natural protective amyloid. Also, we briefly discuss the reported function of other natural, protective amyloids like fish chorion, the protein Pmel17 which forms amyloid fibrils that act as templates and accelerate the covalent polymerization of reactive small molecules into melanin, the hydrophobins and the antifreeze protein from winter flounder. Molecular self-assembly is becoming an increasingly popular route to new supramolecular structures and molecular materials and the inspiration for such structures is commonly derived from self-assembling systems in biology. Therefore, a careful examination of these studies may set the basis for the exploration of new routes for the formation of novel biocompatible polymeric structures with exceptional physico-chemical properties, for potentially new biomedical and industrial applications.

Keywords: Amyloids, natural protective amyloids, silkmoth chorion protein peptide analogues, amyloid fibrillogenesis, liquid crystals, -pleated sheet, fish chorion, Pmel17, hydrophobins, antifreeze protein.

INTRODUCTION

Proteins or peptides convert under certain conditions from their soluble forms into ordered fibrillar aggregates, called amyloid fibrils. Protein aggregation and occasionally ensued amyloid fibril formation are believed to be the cause of an intriguing group of neurodegenerative diseases includ-ing Alzheimer’s disease, prion diseases, Parkinson’s and Huntington’s and, also, type II diabetes and many others, that are referred to as amyloidoses [1,2]. However, in addi-tion, occasionally, living organisms take advantage of the inherent ability of proteins and peptides, to form such struc-tures under certain conditions and generate novel and diverse biological functions [2-4 and references therein], which were noted following our proposal for the existence of natural protective amyloids [5]. There is no apparent similarity be-tween amyloidogenic proteins or peptides in aminoacid se-quence, molecular weight, morphology or surrounding con-ditions. Furthermore, many proteins not implicated in con-formational disease have also been shown to form amyloids in vitro, leading to the hypothesis that the potential for amy-loidogenesis may be a near universal feature of proteins [6].

However, recent evidence indicates that there is a se-quence propensity for amyloid formation [7,8 and references therein], sometimes inherent in sequence after millions of years of molecular evolution, as we have shown [5,9].

*Address correspondence to this author at the Department of Cell Biology and Biophysics, Faculty of Biology, University of Athens, Panepistimiopo-lis, Athens 157 01, Greece; Tel: +30-210-727 4931; Fax: +30-210-727 4254; E-mail: [email protected], [email protected]

The major component of the eggshell (90-95%) of many insect and fish eggs is chorion. Proteins account for more than 95% of its dry mass. This proteinaceous shell forms the outer layer of the eggshell and has extraordinary mechanical and physiological properties, protecting the oocyte and the developing embryo from a series of environmental hazards such as temperature variations, mechanical pressure, prote-ases, bacteria, viruses etc [10]. It also allows for sperm entry and fertilization and for the exchange of the respiratory gases [10]. Fig. (1a) shows an electron micrograph of a thin trans-verse section of a silkmoth chorion. A lamellar ultrastructure of packed fibrils is seen: silkmoth chorion is a biological analogue of a cholesteric liquid crystal [11,12]. The X-ray diffraction pattern of a silkmoth chorion shown in Fig. (1b)indicates that -sheet is the dominant secondary structure of its constituent proteins. ATR FT-IR (Fig. 1c) and laser-Raman spectroscopy [12 and references therein] suggest that the -sheets are antiparallel.

About 200 proteins have been detected in the silkmoth chorion [17]. These proteins have been classified into two major classes, A and B [10]. The gene families encoding these proteins are related and constitute a superfamily with two branches, the -branch and the -branch [18]. Sequence analyses and secondary structure prediction revealed that chorion proteins consist of three domains [19]. The central domain is conserved in each of the two classes. The flanking N- and C-terminal domains are more variable and contain characteristic tandem repeats (ref. 19; see also Fig. 2). A and B central domains show distant similarities suggesting that the chorion genes constitute a superfamily derived from a single ancestral gene [18].

Not For Distribution

292 Current Protein and Peptide Science, 2008, Vol. 9, No. 3 Iconomidou and Hamodrakas

Fig. (1). (a) Transmission electron micrograph of an oblique section through the helicoidal proteinaceous chorion of the silkmoth An-theraea polyphemus. Arrays of protein fibrils (ca. 100 Å, F, arrow-heads) are seen in each lamella (L). Arrows point towards the outer (O) and inner (I), closest to the oocyte, surfaces of chorion. Bar 0.4

m. Reproduced from [5] with permission. (b) X-ray diffraction pattern from an almost flat fragment of a silkmoth Antheraea poly-phemus chorion. The flat fragment of chorion was irradiated with the X-ray beam parallel to the outer and inner surfaces of chorion (Fig. 1a corresponds to a cut face of chorion that would be encoun-tered by the beam). The presence of reflections corresponding to periodicities of 4.6 and 9.1 Å suggests abundance of -sheet in chorion proteins. The elliptical scattering at ca. 30 Å-1 indicates a helicoidal architecture for silkmoth chorion and most probably arises from ca. 30 Å protofilaments, constituents of the ca. 100 Å fibrils [12]. Reproduced from [5] with permission. (c) ATR FT-IR spectra taken from the outer (red) and inner (blue) surfaces of an Antheraea polyphemus silkmoth chorion. 2nd derivatives are in-cluded. The amide I, II and III bands at 1626, 1514 and 1230 cm-1,respectively, clearly indicate a -sheet type of structure for silk-moth chorion proteins. The shoulder in the amide I region, at 1694 cm-1, suggests that the -sheets are antiparallel [25, 27-30].

The study of the properties of chorion proteins has long been hampered by the fact that it has proven very difficult to purify individual chorion proteins in large enough amounts of sufficient purity for structural studies. We therefore syn-thesized several chorion protein peptide-analogues and stud-ied their structural and assembly properties under various conditions. These studies, reviewed in detail in this work, suggest that silkmoth chorion is a natural protective amyloid. The function of other natural, presumably protective amy-loids is also briefly discussed.

AMYLOID CHARACTERISTICS OF SILKMOTH CHORION PEPTIDES

A. Family Peptide-Analogues

cA Peptide

Initially, a 51-residue peptide was synthesized [20], that can be considered as a generic central domain of the A class of silkmoth chorion proteins (Figs. 2a, 2b). This peptide, referred to below as cA peptide, is representative for about 20-30% of all the proteinaceous material in the eggshell. We chose this peptide because the central domains of the A class chorion proteins are highly conserved in both sequence and length and because this conservation indicates that this do-main plays an important functional role in the formation of chorion structure [5].

The cA peptide forms, structurally uniform, amyloid-like fibrils by self-assembly in an astonishing variety of various solvents, pH values, ionic strengths and temperatures (V.A.I and S.J.H., In preparation). The fibrils were judged to be amyloid-like from their tinctorial and structural characteris-tics: they bind Congo-red showing the characteristic for amyloids red-green birefrigence when seen under crossed polars [21 and refs. therein] (Figs. 3a, 3b) as well as Thioflavin-T (data not shown). Electron micrographs (Figs. 4a, 4b) show that they are straight, unbranched double heli-ces of indeterminate length and uniform in diameter (~90 Å). Each double-helical fibril consists of two protofilaments wound around each other. The protofilaments both have a uniform diameter of approximately 30-40 Å. The pitch of the double helix (Fig. 4a, arrows) is approximately 920 Å.


Natural Protective Amyloids Current Protein and Peptide Science, 2008, Vol. 9, No. 3 293

Fig. (2). (a) Alignment of several chorion proteins of the A class [12 and references therein]. The names of the proteins are to the left of each sequence. The designations Bm and Ap refer to B. mori and A. polyphemus respectively. By convention 'pc' numbers refer to sequences de-rived from cDNA clones, whereas the initials ERA and AL refer to 'early' A and 'late' A proteins correspondingly. Numbers to the right of each sequence are the total number of residues of each protein. The borders of the central conservative domain are marked at the top (by ar-rows). The numbering at the top is arbitrary. Black-boxed residues are identical and gray-boxed residues represent conservative substitutions. The cA peptide sequence is shown at the bottom. Asterisks at the bottom mark invariant Gly (G) residues. Reproduced from [5] with permis-sion. (b) A schematic representation of the tripartite structure of silkmoth chorion proteins of the A family. A highly conservative central domain of invariant length, and two more variable flanking “arms” constitute each protein. Characteristic, tandemly repeating peptides are present both in the central domain and in the “arms” [12 and references therein]. The aminoacid sequence and relative position of the syn-thetic cA peptide (one letter code), designed to be an analogue of the entire central domain of the A family, is shown. Invariant glycines (G) repeating every six-residues are boxed and marked with an asterisk below the sequence. Reproduced from [26] with permission.

Fig. (3). (a) Photomicrographs of cA peptide fibrils stained with Congo red: (a) Bright field illumination, (b) crossed polars. The red-green birefringence characteristic for amyloid fibrils is clearly seen. Bar 200 m. Reproduced from [26] with permission.

Suspensions of these fibrils form oriented fibres, which give characteristic “cross- ” X-ray diffraction patterns (Fig. 4c) [22]. In these oriented fibres, the long axes of the amy-loid-like fibrils seen in the electron micrographs of Figs. (3a)and (3b) are oriented more or less parallel to the fibre axis.

The oriented X-ray pattern (Fig. 4c) taken from these fibres indicates the presence of oriented -sheets in the amy-loid-like fibrils of peptide cA. The presence of reflections corresponding to periodicities of 4.66 and 10.12 Å indicates the existence of -sheets [23]. The strong meridional reflec-



tion at 4.66 Å suggests that the -sheets are oriented so that their -strands are perpendicular to the fibre axis and thus also to the long axis of the amyloid-like fibrils. The strong equatorial reflection at 10.12 Å, which corresponds to the inter-sheet distance, suggests that the packing of the -sheets is parallel to the fibre axis and preferentially oriented. This X-ray pattern closely resembles typical cross- patterns [22] taken from amyloid fibres [21 and refs. therein].

A list of the meridional and equatorial reflections ob-served in the X-ray pattern (Fig. 4c) is given in Table 1. Re-flections along the meridian are orders of the 115 Å spacing (116±1 Å) of the -helix of the amyloid protofilament model proposed by Blake & Serpell [24], which supports this

model. Reflections along the equator are orders of the 10.12 Å inter-sheet distance indicating a long range order of packed -sheets in the fibre.

The ATR FT-IR (900-1800 cm-1) spectrum of cA peptide amyloid fibrils, cast on a Au mirror (Fig. 4d), clearly sug-gests that the -sheets present in the structure of the cA pep-tide amyloid fibrils (evident from the existence of the amide I and III bands at 1628 and 1234 cm-1, respectively) are anti-parallel (shoulder at 1692 cm-1 in the amide I band) [25,26].

cA_m1 and cA_m2 Peptides

Further, we have designed and synthesized, two peptides, mutants of the cA peptide. The first, called hereinafter

Fig. (4). (a) Electron micrograph of amyloid-like fibrils derived by self-assembly, from a 9 mg.ml-1 solution of the cA peptide in a sodium acetate 50mM buffer, pH 5. Fibrils are negatively stained with 1% uranyl acetate. They are of indeterminate length (several microns), un-branched, approximately 90 Å in diameter and have a double helical structure. The pitch of the double helix is ~ 920 Å (marked with ar-rows). A pair of protofilaments each 30-40 Å in diameter are wound around each other, forming the double-helical fibrils. Bar 800 Å. (b)Electron micrograph of amyloid-like fibrils derived from a solution of the cA peptide (conditions as in Fig. 4a). Fibrils are rotary shadowed with Pt/Pd at an angle of 7 degrees under high vacuum. Bar 1000 Å. (c) X-ray diffraction pattern from an oriented fibre of cA peptide amy-loid-like fibrils. The meridian, M (direction parallel to the fibre axis) is horizontal and the equator, E, is vertical in this display. The X-ray diffraction pattern is a typical “cross- ” pattern showing a 4.66 Å reflection on the meridian and a 10.12 Å reflection on the equator. This indicates a regular structural repeat of 4.66 Å along the fibre axis (meridian) and a structural spacing of 10.12 Å perpendicular to the fibre axis. The structural repeat of 4.66 Å along the fibre axis corresponds to the spacing of adjacent -strands (which should be perpendicular to the fibre axis, a "cross- " structure) and the 10.12 Å spacing perpendicular to the fibre axis corresponds to the face-to-face separation (pack-ing distance) of the -sheets. (d) ATR FT-IR (900-1800 cm-1) spectrum of cA peptide amyloid fibrils, cast on a Au mirror. 2nd derivative spectra are included. Errorbar equals in the IR spectrum. Reproduced from [5] with permission.



cA_m1, is approximately half the size of the cA peptide (Fig. 5), whereas the second, called cA_m2, is a variant of the first, having three hydrophobic residues (two valines, V, and an alanine, A) replaced by glutamates (E) at specific posi-tions (Fig. 5). The logic behind this synthesis was the follow-ing: We were interested to find out (a) whether amyloid-like fibril formation is sustained by fractions of the cA peptide, which contains tandemly repeating hexapeptides, in other words to determine whether amyloid fibril formation ability is inherent into the hexapeptide repeating structure of the central domain of chorion proteins, and, (b) whether specific mutations in the model -strands of cA, at locations where -former residues exist, would influence the structure and as-sembly of the formed super-structures [9].

cA_m1 peptide (Fig. 5) folds and self-assembles, forming mature amyloid-like fibrils (Fig. 6) after 1-2 weeks incuba-tion, also in a variety of solvents and conditions as the cA peptide. In contrast, solutions of cA_m2 (Fig. 5), under the

same conditions and after incubation for several months, do not form amyloid-like fibrils and only very rarely show some tubular-like structures (data not shown). The fibrils formed by the cA_m1 peptide are very similar in structure and prop-erties to the fibrils formed by the cA peptide [5, 9]. They are of indeterminate length, they have a thickness of ca. 100 Å and they appear to be double helical in structure (Fig. 6). They also bind Congo red showing the characteristic red-green/yellow birefringence when seen under crossed polars (Fig. 7), and they give characteristic “cross- ”-like X-ray diffraction patterns (Fig. 8) from oriented fibers. Therefore, they display all the features that characterize amyloid fibrils as well.

The X-ray diffraction patterns of oriented fibers from cA_m1 exhibit several reflections (Fig. 8). Most of these reflections appear as rings due to rather poor alignment of the constituent fibrils. The strong reflections corresponding to periodicities of 4.67 Å and 10.45 Å may be attributed to

Table 1. Spacings of the Meridional and Equatorial Reflections, Observed in the X-Ray Diffraction Pattern taken from an Ori-ented Fibre of cA Peptide Amyloid-like Fibrils (Fig. 4c). Indexing was done Assuming a Repeat Period of 116±1 Å for the Meridional Reflections. The Equatorial Reflections are Orders of the 10.12 Å Inter-Sheet Packing Distance

Meridian Equator

d(obs) Å Index d(obs) Å Index

4.66 25 10.12

3.63 32 5.05 2

2.92 40 3.33 3

2.81 41 2.52 4

2.27 51

2.12 55

Fig. (5). Another schematic representation of the tripartite structure of silkmoth chorion proteins of the A family. A highly conservative cen-tral domain of invariant length, and two more variable flanking “arms” constitute each protein. Characteristic, tandemly repeating peptides are present both in the central domain and in the “arms” [12 and references therein]. The aminoacid sequence and relative position of the synthetic cA peptide (one letter code), designed to be an analogue of the entire central domain of the A family, is shown and also the se-quences of the designed synthetic mutant peptides cA_m1 and cA_m2. cA_m1, 24-residues in length, has approximately half the size of the cA peptide (51 residues). Peptide cA_m2 differs from cA_m1 at three positions, where glutamates (E) have replaced two valine (V) and one alanine (A) residues. Conserved residues are shaded. Invariant glycines (G) repeating every six-residues are marked with an asterisk below the sequence. Reproduced from [9] with permission.



Fig. (6). Electron micrograph of amyloid-like fibrils derived by self-assembly, from a 10 mg.ml-1 solution of the cA_m1 peptide in a sodium acetate 50mM buffer, pH 5. Fibrils were negatively stained with 1% uranyl acetate. They are of indeterminate length (several microns), unbranched, approximately 90 Å in diameter and have a double helical structure. A pair of protofilaments each 40-50 Å in diameter are frequently wound around each other, forming the double-helical fibrils. Bar 0.2 m. Reproduced from [9] with per-mission.

Fig. (7). Photomicrographs of cA_m1 peptide fibrils stained with Congo red: (a) Bright field illumination, (b) crossed polars. The red-green/yellow birefringence characteristic for amyloid fibrils is clearly seen. Bar 100 m. Reproduced from [9] with permission.

Fig. (8). X-ray diffraction pattern from an oriented fibre of cA_m1 peptide amyloid-like fibrils. The meridian, M (direction parallel to the fibre axis) is horizontal and the equator, E, is vertical in this display. The X-ray diffraction pattern resembles a “cross- ” pattern showing a 4.67 Å reflection as a ring and 10.45 and 11.5 Å reflec-tions on the equator. This indicates a regular structural repeat of 4.67 Å and structural spacings of 10.45 and 11.5 Å perpendicular to the fibre axis. These two equatorial reflections usually merge into one, in other diffraction patterns from oriented fibres of cA_m1 peptide amyloid-like fibrils. The structural repeat of 4.67 Å corre-sponds to the spacing of adjacent -strands and the 10.45 Å spacing parallel to the fibre axis corresponds to the face-to-face separation (packing distance) of the -sheets. Possible origin and measured spacings of the other reflections have been discussed previously [5]. Reproduced from [9] with permission.

the inter-strand and inter-sheet distances of -sheet arrange-ments, respectively. These reflections are characteristic of the "cross- " conformation [22], and are observed for several amyloid-like fibrils [5, 21 and references therein], in which the -strands (if oriented) are perpendicular to the fibre axis and the sheets are packed parallel to the fibre axis. This con-formation would produce oriented patterns from oriented samples with a meridional reflection at ca. 4.7 Å and the corresponding equatorial reflection at ca. 10 Å, i.e. very similar to the patterns obtained from fibres of the chorion cA peptide [5]. As seen in Fig. (8), only the 10.45 Å reflection of the oriented fibers produced from the amyloid fibrils of the cA_m1 peptide shows preferred orientation. Similarly, preferred orientation is observed for the very sharp 5.22 Å and 3.46 Å reflections, which are the second (10.45/2) and third (10.45/3) orders of the 10.45 Å reflection. This proba-bly indicates that the packed -sheets exhibit long range or-der in a direction perpendicular to the fibre axis, that is the sheets are packed parallel to the fibre axis. The fact that the 4.67 Å reflection is not preferentially meridional (perpen-dicular to the fibre axis) but instead is a ring, indicates a more random orientation of the packed -sheets (and con-comitantly of the -strands) with respect to the fibre axis. It is perhaps interesting to note that diffraction patterns ob-tained from not very well oriented fibres of the cA peptide



give almost identical X-ray diffraction patterns to the pat-terns obtained from fibres of the cA_m1 peptide, which im-plies almost identical secondary structures for both cA and cA_m1 peptides (data not shown).

Attempts to obtain oriented fibres from solutions of the cA_m2 peptide were not successful since this peptide did not form amyloid fibrils or suitable viscous solutions for fibre formation. Furthermore, solutions of this peptide do not bind Congo red and do not exhibit the characteristic for amyloids red-green/yellow birefringence.

The ATR FT-IR spectra of the cA_m1 and cA_m2 pep-tides are compared in Figs. (9a-c) and Table 2. The spectrum from cA_m1 peptide amyloid fibrils (Fig. 9a and Table 2)shows one prominent band at 1628 cm-1 in the amide I region and an amide III component at 1228 cm-1, which are defi-nitely assigned to -sheet [25, 27-30]. The low frequency of these amide I and III components results from the strong hydrogen bonds in the -sheets, whereas the very narrow width of the amide I band at 1628 cm-1 (ca. 20 cm-1) suggests that the distribution of the phi and psi angles in the sheets is narrow and implies a very uniform structure. Thus, ATR FT-IR supports the presence of uniform -sheets in the structure of cA_m1 peptide fibrils, in agreement with the existence of a -sheet structure suggested by X-ray diffraction.

It is interesting to note that, the ATR FT-IR spectrum of the cA_m1 peptide obtained immediately after dissolving the peptide in D2O (Fig. 9b and Table 2), before any formation of amyloid fibrils, exhibits components characteristic of -sheet structure as well: an amide I band at 1633 cm-1 and an amide II band at 1437 cm-1 [31]. Formation of mature amy-loid fibrils occurs after a period of approximately 4-5 days, as we have shown previously rather conclusively for solu-tions of the cA peptide [26] and for solutions of the cA_m1 peptide (our unpublished data) as well. Therefore, in this case, it appears that -sheet structure dictates formation of amyloid fibrils in a rather decisive way, since -sheet struc-ture is the structure that the peptide cA_m1 adopts, directly after solution.

On the contrary, thin films cast from cA_m2 peptide so-lutions on front-coated gold mirrors produce ATR FT-IR spectra characteristic of an unordered conformation of the peptide in these solutions (Fig. 9c). The infrared band at 1653 cm-1 in the amide I region and the absence of any fea-tures in the amide III region are indicative of unordered structure [25, 27-30]. A shoulder at ca. 1630 cm-1 may indi-cate a small fraction of -sheet structure.

In conclusion, the synthesis and study of the structural properties of peptides cA_m1 and cA_m2, provided convinc-ing evidence that a peptide with a length half of that of the central domain of the A family of silkmoth chorion proteins (namely cA_m1) folds and self-assembles into amyloid fi-brils, which are very similar in properties to those of the cA peptide (which corresponds to the entire length of the A fam-ily central domain). This probably implies that the underly-ing molecular substructure that dictates proper folding and self assembly of chorion fibrils into the superstructure of silkmoth chorion is encoded into the tandem hexapeptide repeats present in the aminoacid sequences of the central

Fig. (9). ATR FT-IR (1100-1800 cm-1) spectrum of : (A) cA_m1 peptide amyloid fibrils, cast as a thin film on a Au mirror, (B) a freshly made solution of cA_m1 peptide in D2O (10 mg.ml-1) cast as a thin film on a diamond ATR element, and (C) a cA_m2 peptide solution, cast as a thin film on a Au mirror. Errorbar equals in the IR spectra. Reproduced from [9] with permission.

domain of silkmoth chorion proteins. Apparently, it is in strong support of our proposal that silkmoth chorion is a natural protective amyloid [5, 26]. Furthermore, we have demonstrated that carefully designed mutations on the se-quence of these amyloidogenic peptides (cA_m2) can inhibit self-assembly and amyloid formation. Of course, it remains to be seen which is the shortest possible peptide from the sequences of chorion proteins that folds and self assembles forming fibrils similar to those appearing in vivo in the struc-ture of silkmoth chorion.



B. Family Peptide-Analogues

B Peptide

We have also synthesized an 18-residue peptide (B pep-tide) representative of a part of the central conservative do-main of the B family silkmoth chorion proteins (Fig. 10). The B peptide when dissolved in distilled water (pH=5.5) at a concentration of 10 mg ml-1 was found to produce gels [32]. Twisted and non-twisted ribbons are constituents of the gels (Fig. 11). Each ribbon consists of thin protofilaments (fibrils) that have the tendency to coalesce laterally among each other (Fig. 11). The protofilaments have a uniform di-ameter of approximately 30-40 Å. They associate sideways to form twisted and non-twisted ribbons of variable width. Similar twisted ribbons have been observed in amyloid-fibrils formed by fusion peptides [33], an SH3 domain [34], various Alzheimer's disease peptides [35] and a peptide from the adenovirus fibre shaft [36], to mention a few examples. "Beading" is evident along the protofilaments (fibrils), indi-cating that they probably have a helical structure [37]. The

"beads" have dimensions of the order of ca. 30 Å. Congo red stained B peptide gels showed the red-green birefringence characteristic of amyloid fibrils when viewed under crossed polars (Fig. 12).

Suspensions of the fibril-containing gels seen in Fig. (11), form fibres. X-ray diffraction patterns (Fig. 13) taken from these fibres showed four major reflections at 4.64, 5.20, 9.14 and 26.00 Å. These reflections appear as rings due to the poor alignment of the constituent fibrils. Several attempts were made to obtain oriented specimens, but they were not successful. This is probably due to random packing of the ribbons in the gels, which adopt all possible orientations (see Fig. 11) as confirmed by electron microscopy (data not shown).

The strong and relatively sharp reflection corresponding at a periodicity of 4.64 Å and the weaker one at 9.14 Å may be attributed to -sheet structure present in the fibrils (inter-strand and inter-sheet distances respectively). These reflec-tions are characteristic of the "cross- " conformation [22]

Fig. (10). A schematic representation of the tripartite structure of silkmoth chorion proteins of the B family. A highly conservative centraldomain of invariant length, and two more variable flanking “arms” constitute each protein. Characteristic, tandemly repeating peptides are present both in the central domain and in the “arms” [12 and references therein]. The aminoacid sequence and relative position of the syn-thetic B peptide (one letter code), designed to be an analogue of a part of the central domain of the B family, is shown. Invariant glycines repeating every six-residues are marked with an asterisk below the sequence. Reproduced from [32] with permission.

Table 2. Main ATR FT-IR (1100-1800 cm-1) Peak Maxima of: (a) cA_m1 Peptide Amyloid Fibrils Cast as a thin Film on a Au Mirror, (b) Soluble cA_m1 Peptide in D2O Cast as a thin Film on a Diamond ATR Element and (c) cA_m2 Peptide Solu-tion in Water Cast as a thin Film on a Au Mirror (Fig. 9). Tentative Assignments are Included. For details See Text

cA_ml fibrils cA_ml in D2O cA_m2 solution in H2O

Peak (cm-1) Assignment Peak (cm-1) Assignment Peak (cm-1) Assignment

1228 Amide III ( -sheet)

1279 Amide III ( -turns?)

1437 CH2-deformation

1543 Amide II 1437 Amide II’ ( -sheet) 1541 Amide II

1628 Amide I ( -sheet) 1633 Amide I’ ( -sheet) 1630 (sh) Amide I ( -sheet?)

1653 Amide I (unordered

structure)

1672 Amide I ( -turns?)

(TFA?)



Fig. (11). Electron micrograph of amyloid-like fibrils derived by self-assembly from a 10 mg/ml solution of the B peptide in distilled water, pH 5.5. Fibrils are negatively stained with 1% uranyl acetate. The protofilaments (fibrils) are 30-40 Å in diameter and they self-assemble laterally into twisted and non-twisted ribbons of indeter-minate thickness and length (several microns) forming gels (not shown). Individual protofilaments have a "beaded" appearance, suggesting a helical structure. The "beads" have a diameter of ca. 25-30 Å. Reproduced from [32] with permission.

observed for several amyloid-like fibrils [5,21 and references therein], in which the -strands are oriented perpendicular to the fibril axis and the sheets run mainly parallel to the fibril axis. This conformation would produce oriented patterns from oriented samples with the ca. 4.7 Å reflection meridi-onal and the ca. 10 Å reflection equatorial, as in the oriented patterns obtained from fibres of the chorion cA peptide [5].

In the case of the B peptide fibrils, all efforts to obtain ori-ented samples were not successful as described above, and thus the diffraction patterns contained diffraction rings in-stead of oriented reflections. A similar case to ours, where non-oriented X-ray diffraction patterns were observed, is clearly that of a peptide from the adenovirus fibre shaft [36], which also produces amyloid-like fibrils. The reflection at 26 Å may arise from the lateral packing of the fibrils when forming the ribbons or, perhaps, it is due to the repeating "beaded" structure along the protofilaments (Fig. 11). The origin of the 5.2 Å reflection is not clear. It might simply be the fifth order of the 26 Å reflection. However, no additional data are available to explain its presence.

Spectral acquisition by ATR FT-IR and FT-Raman spec-troscopy have been shown to yield rich information about the secondary structure of the B peptide, without the draw-backs associated with the more conventional vibrational techniques. Table 3 shows the most prominent bands and their tentative assignments in the FT-Raman spectrum of the B-peptide, utilizing as sample the fibre used for the X-ray diffraction experiment (Fig. 14).

In the amide I region of the Raman spectrum, the band at 1669 cm-1 (Fig. 14) suggests the presence of -sheet confor-mation in the structure of the B peptide, when it forms amy-loid-like fibrils in the fibre used for X-ray diffraction ex-periments [38,39 and references therein]. Analysis of the amide I band of the FT-Raman spectrum (Fig. 14) using the method of Hamodrakas et al. [40] suggests that there is 64% antiparallel -sheet and 30% -turns in the B peptide struc-ture. The presence of a strong band at 1229 cm-1 in the con-formationally sensitive amide III region (Fig. 14), a charac-teristic signature of -pleated sheet [39 and references therein], strongly supports the amide I evidence as well as the X-ray diffraction data. Bands at ca. 620, 1003, 1033 and 1603 cm-1 are most likely due to the two Phe residues present in the B-peptide (Fig. 10).

Fig. (12). Photomicrographs of B peptide fibrils stained with Congo red: (a) Bright field illumination, (b) crossed polars. The red-green bire-fringence characteristic for amyloid fibrils is clearly seen. Bar 150 m. Reproduced from [9] with permission.



Fig. (13). X-ray diffraction pattern from a fibre of B peptide amy-loid-like fibrils. Four reflections, which correspond to structural repeats of 4.64, 5.20, 9.14 and 26 Å, are marked by arrows. The two most prominent at 4.64 and 9.14 Å, which probably correspond to the spacing of adjacent -strands and to the packing distance of

-sheets respectively, indicate the presence of -sheets ("cross- "sheets) in the structure of the fibrils. Due to the poor alignment of fibrils these reflections appear as rings. The reflection at 26 Å is either due to the packing of the fibrils or originates from the "beaded" sub-structure along the fibrils (see Fig. 11). The origin of the 5.2 (=1/5 of 26) Å reflection is not known. Reproduced from [32] with permission.

Table 3. Main FT-Raman Bands (600-1800 cm-1) and Tenta-tive Assignments in the FT-Raman Spectrum of B Peptide

Band Assignment

620 Phe

832 v(C-C)

859 v

886 Skeletal stretch

925 v(C-C) of Pro ring

1003 Phe

1033 Phe

1087 v(C-N)

1229 Amide III ( -sheet)

1318 v(C-N)

1341 (CH)

1403 v5(COO-)

1453 (CH2, CH3)

1605 Phe

1669 Amide I ( -sheet)

Fig. (14). FT-Raman spectrum of B peptide amyloid-like fibrils in the range 600-1800 cm-1. Error bar equals the standard deviation ( )of the measurement. The spectrum has been acquired from the fibre used to obtain the X-ray diffraction pattern of Fig. 13. Reproduced from [32] with permission.

Concomitant evidence for the preponderance of -sheet in amyloid-like fibrils derived from the B peptide is given by ATR FT-IR spectroscopy (Fig. 15). The ATR FT-IR spec-trum shows one prominent band at 1623 cm-1 in the amide I region. This band is clearly due to -sheet conformation [28]. Its relatively low wavenumber may be attributed to strong hydrogen bonds in the -sheets [28], whereas the fact that it is sharp (its half-width is ca. 20 cm-1) suggests that the distribution of the phi and psi angles in the sheets is rather narrow, which means a very uniform structure. This is also true for the amide I band of the FT-Raman spectrum (Fig. 14). Its apparent half-width is ca. 17 cm-1, lower than that of the very uniform silk fibroin (ca. 27 cm-1) [39 and references therein]. The infrared shoulder at 1697 cm-1 is a strong indi-cation that the -sheets are antiparallel [27, 28, 41] in good agreement with the analysis of the amide I band of the FT-Raman spectrum. Thus, both FT-Raman and ATR FT-IR support the presence of uniform antiparallel -sheets in the structure of B peptide fibrils, apparently in agreement with the existence of a "cross- " structure implied by X-ray dif-fraction and Congo red binding data.

Fig. (15). ATR FT-IR spectrum of B peptide amyloid-like fibrils in the range 1100-1800 cm-1. Reproduced from [32] with permission.



MODEL STRUCTURE OF SILKMOTH CHORION PEPTIDES

Taking into account all experimental and theoretical evi-dence accumulated previously for silkmoth chorion proteins [12 and refs. therein] and their synthetic peptide analogues [5, 9, 32], and the hexapeptide periodicities in the sequences of the A and B families of silkmoth chorion proteins [42, 43], the models shown in Figs. (16a,b and 17a,b), for the peptide cA, are the most probable models for the structure of silkmoth chorion peptides.

Fig. (16). (a) Antiparallel twisted -sheet model proposed for the cA peptide. Sequence should be read continuously, beginning at the bottom. Invariant glycines (G) occupying the 2nd position in the -turns are black boxed. Tentative II' -turns alternate with four-residue -strands. (b) A ribbon representation of the antiparallel twisted -sheet model proposed for the cA peptide. View approxi-mately along the central -strands denoted as arrows. Reproduced from [5] with permission.

The model shown in Figs. (16a,b) is an antiparallel twisted -pleated sheet of four-residue -strands alternating with type II' -turns. Invariant Gly (G) residues occupy the second position of the -turns, a location especially favour-able for Gly in II' turns of twisted -sheets of globular pro-teins [44].

Another interesting possibility for the structure of the cA peptide might be that of the left-handed parallel -helix (Figs. 17a,b), similar to that found in the structure of UDP-N-acetylglucosamine acyltransferase [45] and other left-handed parallel -helical proteins [4]. This protein shows hexapeptide sequence motifs. It is interesting to note that right-handed parallel -helices similar to those found in the pectate lyases have been postulated as the main molecular components of amyloid protofibrils, although no detailed molecular models were presented [46]. Characteristic hexapeptide periodicities of both Gly and hydrophobic resi-dues also appear in the sequence of the cA peptide [12, 43]. Its sequence shows structural similarities with the sequence of UDP-N-acetylglucosamine acyltransferase (Fig. 17a), and the peptide would have a nice hydrophobic core when folded this way (Fig. 17b). Another attractive feature of the left-handed parallel -helix model of Fig. (17b) are the hydro-phobic faces of the triangular prism-like -helix. Neverthe-less, the "edges" of this prism are occupied by charged, polar residues and glycines and this makes 3-D packing difficult, unless there are very specific interactions. On the contrary, the hydrophobic faces of the antiparallel -sheet structure shown in Fig. (16b) facilitate uniform 3-D packing of the -sheets, leaving the polar and charged residues on both lateral "edges" of the sheet for favourable lateral interactions. Al-though we were the first, to our knowledge, to propose a detailed left handed parallel -helix structural model at atomic resolution, as a possible structure underlying amyloid fibrils [5], this publication remained unnoticed. However, several groups also proposed recently that -helices might dictate amyloid fibrillar structure [4, 47-52].

Fig. (17). (a) Structurally based alignment of the sequence of the cA peptide with the sequence of the N-terminal domain of UDP-N-Acetylglucosamine Acyltransferase (LpxA-E.coli) [45]. The one letter code is used. Sequences should be read continuously beginning at the top. The structure of the N-terminal domain of LpxA-E.coli is a left-handed parallel -helix [45]. C1 to C10 rows identify individual turns of the helix. PB1, PB2 and PB3 denote the parallel -strands of each turn. T1, T2 and T3 denote turn residues. Conserved hydrophobic residues which have their side-chains pointing towards the interior of the left-handed parallel -helix to form the hydrophobic core, are black boxed. (b) A ribbon representation of the cA peptide in a left-handed parallel -helix conformation. Arrows denote -strands. View almost perpen-dicular to a face of the left-handed parallel -helix. Reproduced from [5] with permission.



Preliminary calculations of X-ray diffraction patterns from the models presented in Figs. (16b and 17b) and com-parison with the experimental diffraction pattern of Fig. (4c)(data not shown), as well as analysis of the amide I band of FT-Raman spectra [32 and data not shown] and the presence of the high frequency component in the 1690-1700 cm-1 in the ATR FT-IR spectra taken from samples containing amy-loid-like fibrils formed from all, similar to natural, chorion peptides synthesized so far, clearly support the antiparallel twisted -pleated sheet model shown in Fig. (16b) (however, see also below). Unfortunately, as the study of Khurana and Fink [53] has conclusively shown, proteins that adopt a par-allel -helix structure do not exhibit a unique infrared signa-ture.

Fig. (18). (a) A schematic antiparallel twisted -sheet model ("cross- " structure) for the cA_m2 peptide, with the side-chains of the three glutamate (E) residues that have replaced two valines (V) and one alanine (A) residues, in the cA_m2 peptide (cf. Fig. 5) as ball and sticks. Arrows represent -strands. Four-residue -strands alternate with tentative type II' -turns. View along the -strands. It is clear that this structure is not favoured because of strong repul-sive electrostatic interactions of the glutamate side chains in close proximity. (b) A ribbon representation of the cA_m2 peptide in a left-handed parallel -helix conformation with the side-chains of the glutamates (as in (a) above) added as balls and sticks. Four-residue -strands alternate with tentative type II -turns (the two middle residues of the -turns comprise the Gly (G) residues, tan-demly repeating every six residues in Fig. 5, and the subsequent, usually polar or charged residue). Arrows represent -strands. View parallel to the axis of the helix. It is evident that this structure is also not favourable because the side chains of two glutamate resi-dues are packed in the hydrophobic interior of the left-handed par-allel -helix. Reproduced from [9] with permission.

In contrast, none of the models shown in Figs. (16b, 17b)seem to be favourable structures for peptide cA_m2, which is not similar to natural chorion peptides (it contains three non-conservative “mutations” (Fig. 5)). As shown in the lad-der model of Fig. (18a), glutamates would be very close to each other leading to unfavorable electrostatic interactions. Similarly, in the left-handed helix model of Fig. (18b), two glutamates would occupy the hydrophobic interior of the -helix producing a very unstable structure. These results ex-plain why cA_m2 adopts a random coil structure (Fig. 9c)and are clearly compatible with the fact that the cA_m2 pep-tide does not form amyloid fibrils even after very long incu-bation periods.

Fig. (19). Photomicrographs of a part of a silkmoth chorion from Bombyx mori stained with Congo red: (a) Bright field illumination, (b) crossed polars. The red-green/yellow birefringence characteris-tic for amyloids is clearly seen. Bar 400 m.

Thirty years ago, it was suggested that, because of the inherent twist of the -sheets in the monomer, the polymeric amyloid protofilaments might form long spacing helical structures in which the protofilaments are intertwined to pro-duce 100 Å doubly helical amyloid fibrils [54]. The verbal description of this model, reminiscent of the -helix structure of transthyretin amyloid protofilament produced twenty years later by high-resolution X-ray studies [24], fits well with our data. We postulate that successive cA units form continuous twisted antiparallel -sheets (the so-called -sheet helices following the nomenclature proposed by Blake and Serpell [25]), along the protofilaments, with their -strands perpendicular to the long axis of the protofilaments ("cross- " structures). The thickness of each individual cA unit is of the order of 30-40 Å, similar to the thickness of the individual protofilaments. Suspiciously, the pitch of the dou-ble helical amyloid fibril formed by the two intertwined pro-tofilaments is 920 Å (Fig. 4a), a multiple of the 115 Å spac-ing of the -helix in the transthyretin amyloid protofilament [24]. Furthermore, the antiparallel twisted -pleated sheet model of the cA peptide is an eight-stranded antiparallel -sheet, in contrast to the six-stranded -sheet of transthyretin. According to Sunde and Blake [55], an eight-stranded -sheet could well form a ' -sheet helix' instead of a six-stranded -sheet.

SIKMOTH CHORION: A NATURAL PROTECTIVE AMYLOID

We have shown that cA-peptide fibrils have an amyloid nature and also all the other synthesized peptide analogues of silkmoth chorion proteins, but we have so far silently as-sumed that the peptide fibrils are truly representative of the structure of chorion proteins in the eggshell. The cA peptide alone corresponds to about 25-30% of the total chorion mass. Its self-assembly mechanisms produce amyloid-like fibrils under a great and diverse variety of conditions (see above), which strongly suggests that it should fold in an amyloid fashion also in the physiological state. Concomitant evidence for this assumption can be found in Fig. (1a). Lamellae (lay-ers) of fibrils with the same dimensions (70-100 Å; see also [10] and [11]) as the cA peptide double-helical fibrils shown in Fig. (4a), constitute the helicoidal architecture of silkmoth



chorion. Chorion fibrils consist of 30-40 Å protofilaments with a helical structure [12 and references therein]. Further-more, antiparallel -pleated sheet is the dominant molecular conformation of silkmoth chorion proteins in vivo (12 and references therein and also Figs 1b,c). In addition, silkmoth chorion binds Congo red showing the characteristic red-green/yellow birefringence when seen under crossed polars (Figs. 19a,b). This, strengthens further our proposal that silkmoth chorion is a natural protective amyloid.

Amyloids are generally associated with diseases such as Alzheimer's, spongiform encephalopathies, type II diabetes etc: More than 20 types of human disease are associated with the deposition of protein fibrils forming amyloids and result-ing in tissue damage and degeneration [1, 2, 56, 57]. Amy-loidogenic proteins appear to be related by their ability to undergo a conformational change and adopt a new amyloi-dogenic conformation under partially denaturing conditions in vivo, which permits self-assembly into amyloid [57, 58]. Our study [5] was the first to show that not all amyloids are by definition harmful. In chorion protein amyloids, the amy-loidogenic conformation is, apparently, the native conforma-tion. Chorion proteins and peptide-analogues provide a model system for the study of amyloid formation, and per-haps we can even extract medically relevant information from the chorion destruction mechanisms [59] used by the embryo upon hatching.

Our proposal for the existence of natural, functional (in our case protective) amyloids [5] was followed by a number of examples, which undoubtedly confirmed the existence of several natural, functional amyloids, performing various im-portant functions [see for example 2, 3, 4 and references therein].

However, to our knowledge, this is the first well docu-mented case where amyloid-like fibrils are formed from pep-tides that have a sequence so clearly folded in an antiparallel

-pleated sheet type of structure of the "cross- " type. These amyloid-like chorion peptides play an important functional role, after millions of years of molecular evolution: protect the oocyte and the developing embryo from a wide range of environmental hazards [10, 12]. Chorion proteins self-assemble extracellularly to form the chorion of silkmoths, far-away from the follicle cells that synthesize and secrete them [10, 12].

AMYLOID FIBRILLOGENESIS OF SILKMOTH CHORION PEPTIDE-ANALOGUES

The phenomenon of the transformation of proteins into amyloid fibrils is of interest, firstly, because it is related to the protein folding problem, and secondly because it is con-nected to the so-called conformational diseases, the amyloi-doses. Consequently, various attempts have been directed towards an understanding of the fibrillogenesis pathway(s), with the aim of developing inhibitors-drugs of therapeutic benefit. These are summarized in several excellent recent reviews [6, 60-64]. However, the molecular and energetic factors affecting protein misfolding and amyloid fibrillo-genesis are still largely unknown [63, 65].

Recently, we presented data, which clearly show that the first main step of amyloid-like fibrillogenesis from silkmoth

chorion peptides is the formation of nuclei of liquid crystal-line nature [26]. Subsequently, these liquid-crystalline nuclei ‘collapse’ and they are transformed into amyloid-like fibrils in a time-period, which depends on several factors. The transformation is performed, most probably, as a result of a conformational transition to the structure of chorion pep-tides, from a left-handed parallel -helix to an antiparallel -pleated sheet. Apparently, chorion peptides play suitably this role, after millions of years of molecular evolution.

These data can be summarized as follows [26]:

cA peptide was found to spontaneously assemble into supramolecular spherical structures, after 1-2 hours incuba-tion, under a great variety of conditions. These structures, when viewed in a polarizing microscope under crossed po-lars, are seen to have a liquid crystalline texture. They are spherulites with ‘Maltese crosses’ (Fig. 20). Under a trans-mission electron microscope, after negative staining, non-transparent spherical structures are seen (Fig. 21), apparently spherulites with smaller diameters (compare with Fig. 20). Spherulites with larger diameters than those shown in Fig. (21) are also seen (not shown). The spherulites appear to have diameters ranging from 0.1 m to 200 m, combining evidence from light and electron micrographs (Figs. 20 and 21). They frequently coexist with readily formed (after an hour or less incubation) fibrils, 50-100 Å in thickness (Fig. 22). These fibrils frequently seem to be derived from spheru-lites that ‘collapse’ or ‘deteriorate’ (Figs. 23 and 24). Since it might be argued that the association of spherulites with fi-brils is not real, it should be mentioned that upon dilution and re-examination of the samples, fibrils are still seen to co-exist with micelle-like structures. Furthermore, as seen from Fig. (25), multiple fibrils radiate out from “collaps-ing”/”deteriorating” micelle in all different directions, which means that they are indeed derived from the “collaps-ing”/”deteriorating” micelles. The spherulitic structures are characteristic, (and for the first 2-3 days dominant) features in the electron micrographs, which were taken from the sam-ples, collected from the initial cA peptide incubations, on an every day basis, for approximately two weeks. Figs. (26a-e)show that in a period of one week micelle numbers decrease, with a concomitant increase of straight amyloid-like fibrils. However, quantification is not possible since it appears that there is no obvious statistical relationship between fibrils bound to micelles and free micelles. The exact time of trans-formation of these spherulitic structures to mature amyloid-like fibrils (Fig. 27) was found to depend mainly on tempera-ture, concentration and type of solution. At room tempera-ture, for most concentrations and types of solutions, the time is approximately one week. The fibrils formed from the cA peptide solutions exhibit all the hallmarks of amyloids, men-tioned above.

It is interesting to note that, solutions of peptide cA_m1 behave in an almost identical way: amyloid fibrillogenesis occurs also via a liquid-crystalline intermediate phase (our unpublished data).

The common structural properties of amyloid fibrils most probably imply similar mechanisms of amyloid fibril forma-tion and, perhaps, common features of amyloid disease pathogenesis. Therefore, much effort has been devoted to-wards understanding the pathway(s) of fibrillogenesis. Our



Fig. (20). Photomicrograph of cA peptide spherulites viewed in a polarizing microscope under crossed polars. These supramolecular spherical structures are formed spontaneously after 1-2 hours incu-bation, under a great variety of conditions (see Methods). ‘Maltese crosses’ are clearly seen. Bar 4 m. Reproduced from [26] with permission.

Fig. (21). Electron micrograph of a cA peptide spherulite derived by self-assembly, from a 6.5 mg.ml-1 solution of the cA peptide in distilled water, pH 5.5. The sample was negatively stained with 1% uranyl acetate. Bar 100 nm. Reproduced from [26] with permission.

work [26], has shown rather conclusively that the first main step of chorion peptide amyloid fibrillogenesis involves the formation of a liquid crystalline phase, immediately after dissolution of chorion peptides in a variety of solvents and environments (Figs. 20, 21). The spherulitic, liquid crystal-line phase formed is preserved for a period of 3-4 days. Dur-ing this period the spherulites convert gradually into amyloid fibrils in a rather consistent and spectacular manner (Figs. 22-26). They are seen to ‘explode’ producing several short fibrillar components, which, apparently, self assemble to form long fibrils. Apparently, in previous fibrillogenetic

Fig. (22). Electron micrograph of cA peptide spherulites along with readily formed fibrils, 50-100 Å in thickness, after 1-2 hours incu-bation. The samples were negatively stained with 1% uranyl ace-tate. Bar 100 nm. Reproduced from [26] with permission.

Fig. (23). Electron micrograph of cA peptide spherulites along with readily formed amyloid fibrils, which frequently seem to emanate from spherulites that ‘collapse’ or ‘deteriorate’. The samples were negatively stained with 1% uranyl acetate. Bar 100 nm. Reproduced from [26] with permission.

studies of the synthetic amyloid A peptide [66], micelles of much smaller dimensions (diameters of the order of 140 Å) have been observed to act as initial nuclei from which fibrils emerge. Nevertheless, more recently, the A peptide, at high concentrations of the order of 300-600 M, was found to assemble into clearly defined spheres called “ amy balls”, with diameters of ca. 20-200 m; however, it is not certain whether these spherical structures are spherulites [67]. It is also interesting to note that, AFM images of freshly dis-solved A (1-42) peptide after its adsorption onto a freshly cleaned mica surface show predominantly monomeric and dimeric globular structures, with diameters of 1.5 to 2.5 nm [68]. No fibrillar structures are observed in AFM images of the freshly dissolved A ’s. Repeated AFM imaging reveals that A ’s retain the globular and non-fibrillar shape for an extended period of time [68].



Fig. (24). A ‘snapshot’ of spherulite “collapse” / “deterioration” after 2 days incubation. The sample was negatively stained with 1% uranyl acetate. Bar 0.2 m. Reproduced from [26] with permission.

Fig. (25). A ‘snapshot’ whereby multiple fibrils radiate out in all different directions from a “collapsing”/”deteriorating” micelle (arrow), which means that they are derived from the “collapsing “/”deteriorating” micelle. Bar is 400 nm. Reproduced from [26] with permission.

In spiders, soluble proteins are converted to form insolu-ble silk fibres, stronger than steel and, recently, the amyloi-dogenic nature of spider silk has been documented [69]. An important lesson to learn from the spider is how it stores protein molecules in a highly concentrated liquid crystalline state and then extends these in the spinning duct to form a supremely tough thread [70]. Apparently, this is another clear case where formation of insoluble, very tough material

with amyloidogenic properties, initially passes through a liquid crystalline state. Observing closely both cases, the transformation of liquid crystalline soluble spider silk into fibres stronger than steel and also the transformation of solu-ble silkmoth chorion proteins into silkmoth chorion, a struc-ture with extraordinary mechanical and thermal properties, via a liquid crystalline phase, may help us to gain insight, through lateral thinking, into the sudden but unwanted as-sembly of other proteins into amyloids, in various amyloi-doses.

Following our proposal that amyloid fibrillogenesis of silkmoth chorion protein peptide-analogues proceeds via a liquid-crystalline intermediate phase [26], it has nicely been shown that the formation of amyloid-containing spherulite-like structures has been observed in several instances of amyloid diseases, as well as in amyloid fibril-containing solutions in vitro [71-73 and references therein].

At this stage, it is obviously natural to wonder what is the molecular denominator of the spherulitic and of the mature fibrillar state. The view that is currently valid is that amyloid fibrillogenesis requires partial unfolding of globular proteins or partial folding of disordered proteins [60]. However, in the case of chorion peptides this view may not necessarily be true. Apparently, when chorion peptides adopt their final structure forming mature amyloid fibrils, they have a charac-teristic antiparallel -pleated sheet structure strongly sup-ported by X-ray diffraction, ATR FT-IR, FT-Raman spec-troscopy and modelling data (Fig. 16b) and this is obviously their dominant molecular structure in the amyloid fibrillar state. It is tempting to speculate that in the spherulitic mi-celles formed directly after the solution of chorion peptides into various solvents, the peptides adopt a left-handed paral-lel -helix type of structure (Fig. 17b), which, as a model of chorion peptide structure, has several attractive features as well. We propose here that the transition from the spherulitic liquid crystalline phase to the mature fibrillar amyloid state involves a transformation of the left-handed parallel -helix type of structure to the antiparallel -sheet type of structure present in the mature fibrils of the amyloid state. This mo-lecular conformational “switch” can easily be achieved by a transition of a type II -turn (in the structure of the left-handed parallel -helix) to a type II’ -turn (in the structure of the antiparallel -pleated sheet) every six residues along the sequence of the cA peptide. Whether this is actually true remains to be seen by more refined experimental future work (In preparation). If it is true, it will be the first case whereby a peptide capable to form amyloid fibrils may adopt two, well defined, -sheet structures of a different kind. Appar-ently, after millions of years of molecular evolution, in a rather unique way, these dual-structure, amyloid-like chorion peptides play an important functional role: protect the oocyte and the developing embryo from a wide range of environ-mental hazards.

It is also clear that, the results of these studies may have important morphogenetic implications for silkmoth chorion structure as well: silkmoth chorion, a biological analogue of a cholesteric liquid crystal [12 and references therein] self-assembles from its constituent proteins a long distance away from the follicle cells which secrete these proteins on the surface of the oocyte. From our work [26], we have shown



Fig. (26). cA peptide dissolved in a methanol-distilled water 1:1 mixture, at a concentration of 6.5 mg.ml-1 observed for a period of approxi-mately one week. It is clearly seen that micelle numbers decrease gradually, with a concomitant increase of straight amyloid-like fibrils. (a) Day 0. Only micelles exist. Bar 10 m (b) Day 2. Co-existence of micelles with some fibrils. Bar 100 nm (c) Day 5. The numbers of fibrils have considerably increased whereas those of micelles decreased. Bar 100 nm (d) Day 6. Total disappearance of micelles. The field is full of immature amyloid like fibrils. Bar 100 nm (e) Day 7. Only immature amyloid-like fibrils are seen. Bar 100 nm. Reproduced from [26] with permission.

Fig. (27). Electron micrograph of mature cA peptide amyloid-like fibrils, after one week incubation. The sample was negatively stained with 1% uranyl acetate. Bar 0.2 m. Reproduced from [26] with permission.

rather conclusively that the cA peptide, a representative of the entire central domain of the A class of chorion proteins,

that is a representative of 25% of the entire chorion mass approximately, self-assembles under a great variety of condi-tions into spherulitic liquid crystalline structures, obviously precursor substructures of silkmoth chorion itself, which are transformed into amyloid-like fibrils constituting chorion. Thus, it appears that in vitro formation of such an important biological structure can be studied in detail. Since it might be argued that the peptide concentrations which lead to the liq-uid crystalline intermediates cannot occur in vivo, it should be mentioned that similar spherulitic structures are formed at very low (nM) concentrations of the cA peptide in the same solvents [26]. These structures seem to retain their shape for a prolonged, not very well defined, period of time before their conversion into amyloid fibrils.

It remains also to be seen whether the results of this work can be generalized into other pathological cases involving amyloidoses and their, relevant in each case, proteins.

OTHER NATURAL PROTECTIVE AMYLOIDS

The generic ability of polypeptide chains to form amy-loid fibrils, which can vary dramatically with sequence [2, 6], has been exploited by living organisms for specific pur-poses, and it has been found that certain organisms, during



their normal physiological life cycle, convert one or more of their proteins into amyloid fibrils with functional rather than pathological (disease-associated) properties [2-4 and refer-ences therein], following our proposal for the existence of natural protective amyloids [5]. However, although several examples have been documented of functional amyloids [2-4 and references therein], only certain other examples, except silkmoth chorion, can be considered as clear cases of natural protective amyloids [74-76].

The first is the fish chorion of Austrofundulus limnaeous,which helps diapausing embryos of this species to survive desiccating conditions by reducing evaporative water loss [74]. A high proportion of intermolecular -sheet in the pro-teins of this chorion, positive staining and green birefrin-gence with Congo red, and detection of long, unbranched fibrils with a diameter of 4-6 nm, provides the example of another natural protective amyloid. Apparently, the high resistance of diapausing embryos to water stress is not corre-lated with ontogenetic changes in the egg envelope (chorion) [74]. Although we have documented with certainty the exis-tence of antiparallel -sheet as the dominant molecular con-formation in the proteins of the chorions of at least two other fishes, Salmo gairdneri [12, 75] and Dentex dentex [76] by laser-Raman and ATR FT-IR spectroscopy and an underly-ing fibrillar ultrastructure, we have not yet concomitant evi-dence from Congo red staining or Thioflavin-T binding that these represent natural protective amyloids as well. Unfortu-nately, in all cases of fish chorions [74, 75, 76], there are no available amino acid sequences of fish chorion proteins.

The second example is related to the membrane protein Pmel17 [77, 78]. The melanosomes, lysosome-related organ-elles that differentiate in melanocytes to allow the epidermal production of the melanin pigment, are characterized by in-tralumenal fibrous striations upon which melanin granules form. This fibrous material, which resembles amyloid fibrils, is assembled from the intralumenal domains of the mem-brane protein Pmel17 that is proteolytically cleaved by a proprotein convertase [75]. Pmel17 amyloid, templates and accelerates the covalent polymerization of reactive small molecules into melanin, a critically important biopolymer that protects against a broad range of cytotoxic insults, in-cluding UV and oxidative damage. Intracellular Pmel17 amyloidogenesis is carefully orchestrated by the secretory pathway, utilizing membrane sequestration and proteolytic steps to protect the cell from amyloid and amyloidogenic intermediates that can be toxic [76]. Thus, in this case, it appears that amyloids indirectly act as protective amyloids.

The third example of a natural protective amyloid is that of hydrophobins. Hydrophobins are a family of low molecu-lar weight (7–9 kDa) cysteine-rich proteins that have unusual biophysical properties [79]. They are highly surfactive mole-cules that can polymerize at an air/water or water/oil inter-face, forming a tough, polymeric monolayer that is able to reverse the hydropathicity of surfaces, both in vivo and invitro. These proteins are produced and secreted by filamen-tous fungi and are distinguished by a conserved pattern of cysteines in a well defined consensus sequence [79].

More than forty hydrophobins have been shown to play several key roles in fungal development. Most significantly, they act to allow the emergence of aerial hyphae from the

mycelial growth medium and to provide a water-repellent coating to spores that are aerially dispersed. These functions are brought about through polymerization of the secreted hydrophobins at the air-liquid interface into fibrils with di-mensions of around 10 x 200 nm, which aggregate into an amphipathic monolayer on the hyphal or spore surface. The inner face of this monolayer is hydrophilic in nature, while the outer face is highly hydrophobic; it is this hydrophobicity that confers water-repellency on the hypha/spore. The fibrils are exceptionally robust, and hydrophobins have been di-vided into class I and class II types on the basis of their hy-dropathicity plots and the solubility characteristics of the fibrils that they form. They clearly have an amyloid-fibril like nature [79] and act as an extremely effective water re-pellent coating. It should be noted that the presence of an interface appears to be required for fibril formation and in this way these fibrils may be distinguished from other amy-loid fibres.

The fourth example of a natural protective amyloid can be considered that of type I antifreeze protein(AFP) of win-ter flounder. This protein also known as thermal hysteresis protein and as ice-structuring protein, was found at relatively high concentrations in the circulatory system (10–15 mg/ml) and in the skin of fish living in subzero seawater, and pro-tects the organism from macromolecular ice growth by ad-sorption inhibition [80]. It was found that the winter flounder type I AFP in solution forms a translucent gel upon freezing and thawing at physiological pH and below [80]. The gel was found to be amyloidotic in nature, by electron micros-copy, fluorescence staining by the amyloid-specific dye Thioflavin T, and solid-state 13C-NMR spectroscopy. It ap-pears that type I AFP may be used in applications such as cryopreservation, cryosurgery and as food additive [80].

The findings that native amyloid plays a natural protec-tive role in fungi, insects, fishes and mammals demonstrate that the conformation underlying amyloid fibrils is an an-cient, conserved structure vital to normal cell and tissue physiology. The study of the Chrysopa flava silk [22], which set the basis for the proposal of the “cross- ” structure, is another clear such paradigm.

ACKNOWLEDGEMENTS

We thank (in alphabetical order) Dr. B. Agianian, Dr. G. Chryssikos, Prof. P. Cordopatis, Mr. P. Everitt, Dr. A. Galanis, Dr. V. Gionis, Mr. K. Goldie, Prof. A. Hoenger, Dr. K. Leonard, Mr. C. Roome, Dr. L. Serrano, Dr. P. Tucker and Prof. G. Vriend for their help with the experiments. Spe-cial thanks are due to Prof. F.C. Kafatos for his unfailing interest and help. We also thank the University of Athens, the Greek Ministry of Research and Technology, the EMBL summer visitor's program and EMBO for financial support. We thank the anonymous reviewers of this manuscript for their useful comments.

REFERENCES

[1] Uversky, V.N. and Fink, A.L. (2004) Biochim. Biophys. Acta, 1698, 131.

[2] Chiti, F. and Dobson, C.M. (2006) Annu. Rev. Biochem., 75, 333. [3] Kelly, J.W. and Balch, W.E. (2002) J. Cell Biol., 161(3), 461. [4] Stevens, F.J. (2004) J. Protein Folding Disord., 11, 233.



[5] Iconomidou, V.A., Vriend, G. and Hamodrakas, S.J. (2000) FEBS Letts., 479, 141.

[6] Dobson, C.M. (1999) Trends Biochem. Sci., 24, 329. [7] Esteras-Chopo, A., Serrano, L. and de la Paz, M. L. (2005) Proc.

Natl. Acad. Sci. USA, 102 (46), 16672-16677. [8] de la Paz, M. L. and Serrano, L. (2004) Proc. Natl. Acad. Sci. USA,

101(1), 87.[9] Iconomidou, V.A., Chryssikos, G.D., Gionis, V., Galanis, A.S.,

Cordopatis, P., Hoenger, A. and Hamodrakas, S.J. (2006) J. Struct. Biol., 156, 480.

[10] Kafatos, F.C., Regier, J.C., Mazur, G.D., Nadel, M.R., Blau, H.M., Petri, W.H., Wyman, A.R., Gelinas, R.E., Moore, P.B., Paul, M., Efstratiadis, A., Vournakis, J.N., Goldsmith, M.R., Hunsley, J.R., Baker, B., Nardi, J. and Koehler, M. (1977) in: Results and Prob-lems in Cell Differentiation (W. Beerman, Ed.) vol., 8, pp., 45-145, Springer-Verlag, Berlin, Heidelberg, New York.

[11] Mazur, G.D., Regier, J.C. and Kafatos F.C. (1982) in: Insect Ultra-structure (Akai, H. and King R.C., Eds.) vol., 1, pp., 150-183, Ple-num Press, New York.

[12] Hamodrakas, S.J. (1992) in: Results and Problems in Cell Differen-tiation (Case, S.T., Ed.) vol., 19 (Ch., 6), pp.115-186, Springer-Verlag, Berlin, Heidelberg.

[13] Hamodrakas, S.J., Asher, S.A., Mazur, G.D., Regier, J.C. and Kafa-tos, F.C. (1982) Biochim. Biophys. Acta, 703, 216.

[14] Hamodrakas, S.J., Kamitsos, E.I. and Papanicolaou, A. (1984) Int. J. Biol. Macromol., 6, 333.

[15] Hamodrakas, S.J., Paulson, J.R., Rodakis, G.C. and Kafatos, F.C. (1983) Int. J. Biol. Macromol., 5, 149.

[16] Hamodrakas, S.J., Margaritis, L.H., Papasideri, I. and Fowler, A. (1986) Int. J. Biol. Macromol., 8, 237.

[17] Regier, J.C. and Kafatos, F.C. (1985) in: Comprehensive Insect Biochemistry, Physiology and Pharmacology (Gilbert, L.I. and Kerkut, G.A., Eds.) vol., 1, pp. 113-151, Pergamon Press, Oxford, New York.

[18] Lekanidou, R., Rodakis, G.C., Eickbush, T.H. and Kafatos, F.C. (1986) Proc. Natl. Acad. Sci. USA, 83, 6514.

[19] Hamodrakas, S.J., Jones, C.W. and Kafatos, F.C. (1982) Biochim. Biophys. Acta, 700, 42.

[20] Benaki, D.C., Aggeli, A., Chryssikos, G.D., Yiannopoulos, Y.D., Kamitsos, E.I., Brumley, E., Case, S.T., Boden, N. and Hamodra-kas, S.J. (1998) Int. J. Biol. Macromol., 23, 49.

[21] Sunde, M. and Blake, C. (1997) Adv. Prot. Chem., 50, 123. [22] Geddes, A.J., Parker, K.D., Atkins, E.D.T. and Beighton, E. (1968)

J. Mol. Biol., 32, 343-358 [23] Fraser, R.D.B. and MacRae T.P. (1973) “Conformation in fibrous

proteins and related synthetic polypeptides”. Academic Press, New York, London.

[24] Blake, C.C.F. and Serpell, L.C. (1996) Structure, 4, 989. [25] Jackson, M. and Mantsch, H.H. (1995) Crit. Rev. Biochem. Mol.

Biol., 30(2), 95. [26] Hamodrakas, S.J., Hoenger, A. and Iconomidou, V.A. (2004) J.

Struct. Biol., 145, 226. [27] Krimm, S. and Bandekar, J. (1986) Adv. Prot. Chem., 38, 181. [28] Surewicz, W.K., Mantsch, H.H. and Chapman, D. (1993) Biochem-

istry, 32(2), 389. [29] Haris, P.I. and Chapman, D. (1995) Biopolymers, 37, 251. [30] Cai, S. and Singh, B.R. (1999) Biophys. Chem., 80, 7. [31] Haris, P.I. and Severcan, F. (1999) J. Mol. Catal. B: Enzym., 7,

207. [32] Iconomidou, V.A., Chryssikos, G.D., Gionis, V., Vriend, G.,

Hoenger, A. and Hamodrakas, S.J. (2001) FEBS Lett., 499, 268. [33] Ulrich, A.S., Tichelaar, W., Forster, G., Zschornig, O., Weinkauf,

S. and Meyer, H.W. (1999) Biophys. J., 77, 829. [34] Jimenez, J.L., Guijarro, J.I., Orlova, E., Zurdo, J., Dobson, C.M.,

Sunde, M. and Saibil, H.R. (1999) EMBO J., 18, 815. [35] Fraser, P.E., Nguyen, J.T., Surewicz, W.K. and Kirschner, D.A.

(1991) Biophys. J., 60, 1190. [36] Luckey, M., Hernandez, J-F, Arlaud, G., Forsyth, V.T., Ruigrok,

R.W.H. and Mitraki, A. (2000) FEBS Lett., 468, 23. [37] Hamodrakas, S.J., Margaritis, L.H., Papassideri, I. and Fowler, A.

(1986) Int. J. Biol. Macromol., 8, 237.

[38] Iconomidou, V.I., Chryssikos, G.D., Gionis, V., Pavlidis, M.A., Paipetis, A. and Hamodrakas, S.J. (2000) J. Struct. Biol., 132, 112.

[39] Hamodrakas, S.J., Asher, S.A., Mazur, G.D., Regier, J.C. and Kafa-tos, F.C. (1982) Biochim. Biophys. Acta, 703, 216.

[40] Hamodrakas, S.J., Kamitsos, E.I. and Papanicolaou, A. (1984) Int. J. Biol. Macromol., 6, 333.

[41] Krejchi, M., Cooper, S.J., Deguchi, Y., Atkins, E.D.T., Fournier, M.J., Mason, T.L. and Tirrell, D.A. (1997) Macromolecules, 30,5012.

[42] Hamodrakas, S.J., Etmektzoglou, T. and Kafatos, F.C. (1985) J. Mol. Biol., 186, 583.

[43] Hamodrakas, S.J., Bosshard, H.E. and Carlson, C.N. (1988) Prot. Eng., 2(3), 201.

[44] Sibanda, B.L. and Thornton, J.M. (1985) Nature, 316, 170. [45] Raetz, C.R.H. and Roderick, S.R. (1995) Science, 270, 997. [46] Lazo, N.D. and Downing, D.T. (1998) Biochemistry, 37(7), 1731. [47] Wille, H., Mitchelitsch, M.D., Guénebaut, V., Supattapone, S.,

Serban, A., Cohen, F.E., Agard, D.A. and Prusiner, S.B. (2002) Proc. Natl. Acad. Sci. USA, 99(6), 3563.

[48] Wetzel, R. (2002) Structure 10, 1031. [49] Pickersgill, R.W. (2003) Structure 11, 137. [50] Williams, A.D. and Portelius, E., Kheterpal, I., Guo Jun-Tao, Cook,

K.D., Hu, Y., Wetzel, R. (2004) J. Mol. Biol., 335, 833. [51] Kishimoto, A., Hasegawa, K., Suzuki, H., Taguchi, H., Namba, K.

and Yoshida, M. (2004) Biochem. Biophys. Res. Comm., 315, 739. [52] Govaerts, C., Wille, H., Prusiner, S.B. and Cohen, F.E. (2004)

Proc. Natl. Acad. Sci. USA, 101(22), 8342. [53] Khurana, R. and Fink, A.L. (2000) Biophys. J., 78, 994. [54] Cooper, J.H. (1976) in: Amyloidosis (Wegelius, O. and Pasternak,

A., Eds.) pp., 61-68, Academic Press, London, New York. [55] Sunde, M. and Blake, C. (1998) Quart. Rev. Biophys., 31 (1), 1. [56] Pepys, M.B. (1996) in: The Oxford Textbook of Medicine

(Weatherall, D.J., Ledingham, J.G.G. and Warell, D.A. Eds.), 3rd

edn, vol., 2, pp.1512-1524, Oxford University Press, Oxford, UK. [57] Kelly, J.F. (1996) Curr. Opin. Struct. Biol., 6, 11. [58] Kelly, J.F. (1998) Curr. Opin. Struct. Biol., 8, 101. [59] Hagenmaier, H.E. (1974) Comp. Biochem. Physiol., 49b, 313. [60] Rochet, J.C. and Lansbury, P.T. Jr (2000) Curr. Opin. Struct. Biol.,

10, 60. [61] Walsh, D.M., Hartley, D.M., Kusumoto, Y., Fezoui, Y., Condron,

M.M., Lomakin, A., Benedek, G.B., Selkoe, D.J. and Teplow, D.B. (1999) J. Biol. Chem., 274, 25945.

[62] Kelly, J.W. (2000) Nat. Struct. Biol., 7, 824. [63] Soto, C. (2001) FEBS Lett., 498, 204. [64] erovnik, E. (2002) Eur. J. Biochem., 269, 3362. [65] Carel, R.W. and Gooptu, B. (1998) Curr. Opin. Struct. Biol., 8,

799. [66] Lomakin, A., Chung, D.S., Benedek,G.B., Kirschner, D.A. and

Teplow, D.B. (1996) Proc. Natl. Acad. Sci. USA, 93, 1125. [67] Westlind-Danielsson, A. and Arnerup, G. (2001) Biochemistry, 40,

14736-14743. [68] Parbhu, A., Lin, H., Thimm, J. and Lal, R. (2002) Peptides, 23,

1265. [69] Kenney, J.M., Knight, D., Wise, M.J. and Vollrath, F. (2002) Eur.

J. Biochem., 269, 4159. [70] Vollrath, F. and Knight, D.P. (2001) Nature, 410, 541. [71] Krebs, M.R.H., MacPhee, C.E., Miller, A.F., Dunlop, I.E., Dobson,

C.M. and Donald, A.M. (2004) Proc. Natl. Acad. Sci. USA, 101(40), 14420-14424.

[72] Krebs, M.R.H., Bromley, E.H.C., Rogers, S.S. and Donald, A.M. (2005) Biophys. J., 88, 2013.

[73] Krebs, M.R.H., Bromley, E.H.C. and Donald, A.M. (2005) J. Struct. Biol., 149, 30.

[74] Podrabsky, J.E., Carpenter, J.F. and Hand, S.C. (2001) Am. J. Physiol. Regul. Integr. Comp. Physiol., 280, R123-R131.

[75] Papadopoulou, P., Galanopoulos, V.K. and Hamodrakas, S.J. (1996) J. Struct. Biol., 116, 399.

[76] Iconomidou, V.A., Chryssikos, G.D., Gionis, V., Pavlidis, M.A., Paipetis, A. and Hamodrakas, S.J. (2000) J. Struct. Biol., 132, 112.

[77] Berson J.F., Theos, A.C., Harper, D.C., Tenza, D., Raposo, G. and Marks, M.S. (2003) J. Cell Biol., 161, 521.



[78] Fowler, D.M., Koulov, A.V., Alory-Jost, C., Marks, M.S., Balch, W.E. and Kelly, J.W. (2006) PloS Biol., 4(1), e6 (1-8).

[79] Mackay, J.P., Matthews, J.M., Winefield, R.D. Mackay, L.G., Haverkamp, R.G. and Templeton, M.D. (2001) Structure, 9, 83.

[80] Graether, S.P., Slupsky, C.M. and Sykes, B.D. (2003) Biophys.J., 84, 552.

Received: March 11, 2007 Revised: December 03, 2007 Accepted: December 13, 2007


Current Protein and Peptide Science, 291-309 291 Natural...

Documents

Transcript of Current Protein and Peptide Science, 291-309 291 Natural...