PEDS Advance Access originally published online on February 9, 2007
Protein Engineering Design and Selection 2007 20(3):109-116; doi:10.1093/protein/gzm001
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Conversion of a monodispersed globular protein into an amyloid-like filament by appending an artificial peptide at the N-terminal
1 Department of Protein Engineering, Cancer Institute, Japanese Foundation for Cancer Research and CREST, JST, Koto-Ku, Tokyo 135-8550 Japan 2 Department of Molecular Cell Biology, Institute of DNA Medicine, The Jikei University School of Medicine, Japan
3 To whom correspondence should be addressed. E-mail: kshiba{at}jfcr.or.jp
 |
Abstract |
|---|
 Top Abstract IntroductionMaterials and methodsResultsDiscussionAcknowledgementReferences |
|---|
The soluble, globular,
-helix-rich peptide SipA446–684 is a domain of a bacterial protein that binds to mammalian filamentous-actin and re-arranges the host cell's cytoskeleton. We show that adding two copies of NHBP-1, a carbon nanomaterial binding peptide, to its N-terminal can induce SipA446–684 to polymerize and assume a fibrillar structure under physiological conditions. The fibrils formed showed thioflavine T and Congo red staining profiles that are characteristic of and specific for amyloid-like structures. The -helical structure of the globular protein was retained in the fibrils, suggesting the appended NHBP-1 sequence plays a key role in the formation of cross-ß spines within the fibrils. Consistent with that idea, we observed that a synthetic NHBP-1 peptide can form an amyloid-like structure under appropriate conditions. Thus, our findings add a new subtype of amyloid-like structure formation and suggest this method of assembly could be exploited in nano-biotechnology.
Keywords: cross-ß structure/peptide aptamer/self-assembly, single wall carbon nanohorns/SipA
|
Introduction |
|---|
|
TopAbstractIntroductionMaterials and methodsResultsDiscussionAcknowledgementReferences |
|---|
Over the course of the evolution of biomacromolecules, filamentous structures have recurrently emerged as an architectural theme. Microtubules, flagella and actin are all examples of filamentous macromolecules in which a dynamic equilibrium between dispersed monomers and fibrous polymers is elaborately controlled to address various biological activities (Alberts et al., 2002
). By contrast, some fibrous biomacromolecules such as keratin and collagen, among others, are comparatively static in nature; their equilibrium conditions largely shift to a polymer state and the dissolution of their structures often requires a specialized biological strategy, e.g. proteolytic degradation. In extreme cases, the fibrils, once assembled, resist any cellular scavenging activities and accumulate as stable amyloid plaques within organs. These are referred to as amyloidoses and are associated with several fatal diseases (Westermark, 2005). Initially, amyloid formation was thought to be unusual and limited to very few protein species. However, recent studies have shown that formation of amyloid-like structures is a rather generic feature of polypeptide chains (Dobson, 1999, 2003), so that many proteins are able to form amyloid-like structures under appropriate (mostly harsh) conditions. It even has been proposed that the amyloid-like structure might represent a primordial shape of polypeptides (Pickersgill, 2003). It is now known, for example, that in some organisms amyloid-like proteins serve as part of the cellular machinery necessary for normal cellular activity (Chapman et al., 2002; Claessen et al., 2003; Gebbink et al., 2005; Fowler et al., 2006).
As the list of amyloid-like proteins grows, the consensus as to what constitutes an amyloid protein is becoming less clear. The term ‘amyloid’ is usually restricted to in vivo accumulated fibrils that can be stained by Congo red (Westermark, 2005); other fibrillar materials are called ‘amyloid-like’ proteins. These are generally long, straight and unbranched structures a few nanometers in diameter and are stained by Congo red and/or thioflavine T (ThT), which is believed to indicate the presence of cross-ß structures (ß-sheets whose strands run perpendicular to the long axis of the fibril) within the polypeptides. Despite these common structural characteristics, the formation of amyloid-like structures appears to proceed via several pathways. Nelson and Eisenberg (2006) have classified the modes of amyloid-like structure formation into three general types: (i) a re-folding class, (ii) a natively disordered class and (iii) a gain-of-function class. In the re-folding class (type-i), native protein (mostly soluble and globular) is first unfolded and then refolded into a fibrillar structure. With this mode of amyloid formation, the structure of the native protein is lost as the new cross-ß structure emerges. Under the appropriate conditions, insulin (Jimenez et al., 2002) and the SH3 domain (Jimenez et al., 1999), among others (Nelson and Eisenberg, 2006), form amyloid-like fibrils via this process.
Segments of natural proteins or artificial peptides often are able to form amyloid-like structures (type-ii). These peptides are usually unstructured in their monomeric form but become structured, forming left-handed ß-helix or cross-ß structures, upon fibril formation (Perutz et al., 2002; Pickersgill, 2003; Govaerts et al., 2004; Nelson et al., 2005). In the third class (type-iii), a conformational change within a limited region of a protein leads to the appearance of a binding site for another molecule, resulting in polymerization of the protein. This gain-of-function model includes the ‘3D domain swapping model’ proposed based on studies of cystatin C (Janowski et al., 2001) and ß2-microglublin (Eakin et al., 2004), and the ‘cross-ß spine model’ proposed based on studies of Sup35p fibrils (Nelson et al., 2005) and ß2-microglublin (Ivanova et al., 2004). With this mode of amyloid formation, most of the natural structure of the monomeric units is retained by the fibrils.
Here we report the formation of an amyloid-like structure via fusion of a type-ii amyloid forming peptide with a soluble, globular, -helix-rich protein. The nanofibers obtained were unbranched structures, 6–7 nm in diameter, and were stained by both ThT and Congo red. These nanofibers retained structural features of the globular protein in that they were comprised mainly of -helix. This work adds a new subtype to the gain-of-function class of amyloid-like proteins.
|
Materials and methods |
|---|
|
TopAbstractIntroductionMaterials and methodsResultsDiscussionAcknowledgementReferences |
|---|
Construction of expression plasmids
The expression vector, pET-SipA446–684, carries a PCR-amplified genomic fragment of Salmonella typhimurium, which codes for the C-terminal 239 amino acids of SipA (SipA446–684) inserted at the NdeI and BamHI sites of pET3c (Studier et al., 1990) (Sano, unpublished). In addition, into the NdeI site (which includes the first codon of SipA446–684) were inserted two copies of the synthetic oligonucleotide duplexes (KY-1341: 5'-phosphorylated TATGGATTATTTCTCGAGCCCGTATTATGAACAGCTGTT-3' and KY-1342: 5'-phosphorylated TAAACAGCTGTTCATAATACGGGCTCGAGAAATAATCCA-3'). The resultant expression plasmid, pKIS-7, encoded (N1)2-SipA446–684, which has tandem repeats of NHBP-1 situated between the first and second codons of SipA446–684.
To purify the unmodified and modified SipA446–684 proteins, Escherichia coli BL21(DE3) cells (Studier et al., 1990) harboring pET-SipA446–684 or pKIS-7 were inoculated into 1 l of LB medium (Sambrook and Russell, 2001) containing 50 mg/ml carbenicillin at 37°C. When the OD660 of the culture reached 
0.4, isopropyl-ß-thio-D-galactopyranoside was added to a final concentration of 1 mM to induce expression. After incubating an additional 4 h, the cells were collected by centrifugation and stored at –80°C until subjected to the purification processes.
For protein purification, thawed cells were first incubated for 10 min at room temperature in 20 ml of buffer [20 mM TrisHCl (pH 8.0), 0.5 mM ethylenediaminetetraacetic acid (EDTA), 5 mM dithiothreitol (DTT), 1 tablet of protease inhibitor cocktail (Complete Mini, EDTA free, Roche) and 0.2 mg/ml lysozyme], after which they were disrupted by sonication (using a SONIFER 250 equipped with microtip, duty cycle 50%, 2 min x 3 times, Branson). In the following steps, we used separate procedures to purify SipA446–684 and (N1)2-SipA446–684.
To purify SipA446–684, cell debris was removed by centrifugation (7720g x 30 min), and the resultant lysates were precipitated by adding ammonium sulfate to 25% saturation. After removing the precipitant by centrifugation (7720g x 30 min), the supernatant was precipitated again, this time by adding ammonium sulfate to 50% saturation. The precipitants were then collected by centrifugation (7720g x 30 min), resuspended in 10 mM TrisHCl (pH 8.0) and 0.5 mM DTT and dialyzed against the same solution. The dialyzed sample was loaded onto a Q-Sepharose HP ion exchange column (1.6 x 10 cm; Amersham), and SipA446–684 was fractionated using a linear 0–400 mM NaCl gradient in 10 mM TrisHCl (pH 8.0) and 0.5 mM DTT at a flow rate of 3 ml/min.
To purify (N1)2-SipA446–684, inclusion bodies were collected by centrifugation (7720g x 30 min) and re-suspended in 15 ml of buffer [50 mM TrisHCl (pH 8.0), 1 mM EDTA, 1 mM DTT and 1% (v/v) Triton X100] by sonication using a SONIFER 250 with a microtip (output 3–4, duty cycle 50%, 2 min x 2 times). After repeating this protocol three times, the remaining inclusion bodies were solubilized in 5 ml of 20 mM sodium phosphate buffer (pH 6.0) containing 7.5 M guanidine hydrochloride. This solution was then quickly diluted in 300 ml of 0.5 mM sodium bicarbonate at room temperature and stirred for 3 h, after which the resultant protein solution was concentrated using a Centriprep 10 (Millipore). The concentrations of proteins were determined by UV absorption at 280 nm with extinction coefficients calculated from the protein sequences.
Preparation of (N1)2-SipA446–684-SWNHs
In this study, we used the oxidized form of single walled carbon nanohorns (SWNHs) (Murata et al., 2001), which were first suspended in 0.5 mM sodium bicarbonate and 0.05% Tween 20 at a concentration of 0.25 mg/ml and then sonicated for 30 s. The resultant SWNH solution was stirred while 1/20 volume of 14 mg/ml (N1)2-SipA446–684 in 20 mM sodium phosphate buffer (pH 6.0) containing 6 M guanidine hydrochloride was added dropwise, after which the solution was stirred for an additional 1 h at room temperature and then dialyzed against phosphate-buffered saline (PBS). Samples were then transferred to vials and allowed to stand for 2 days, after that we collected the supernatant for transmission electron microscope (TEM) analysis.
Transmission electron micrographs were collected with a HITACHI H-7500 microscope operating with an accelerating voltage of 100 kV. Samples were stained with 4% aqueous uranyl acetate.
Dynamic light scattering (DLS) measurements were carried out using a DynaPro (Protein Solution). The samples contained 10 and 0.5 mg/ml of SipA446–684 and (N1)2-SipA446–684 in PBS, respectively. The radius of hydration (Rh) size distribution was calculated using Dynamics version 5 software.
Circular dichroisn (CD) spectra were obtained using a JASCO J-820 spectropolarimeter. Measurements were made at 20°C using a quartz cell with a 0.2 and 0.01 cm path lengths for proteins and peptides, respectively. The protein samples contained 0.15 mg/ml protein in 20 mM potassium phosphate buffer (pH 8.0). Lyophilized synthetic NHBP-1 was dissolved in 20 mM potassium phosphate buffer pH 8.0 at the concentration of 10 mg/ml, and then diluted for CD measurements. Because of the residual of trifluoroacetic acid in synthetic peptide, pH of the sample solution was around 6.5. For Fourier transform infrared spectroscopy (FTIR) measurements, proteins (1.0 mg) were dried on a calcium fluoride window and analyzed in a Shimadzu FTIR-8200A spectrometer.
Thioflavin T fluorescence assay
Assay solution was comprised of 2 mg/ml protein and 10 mM ThT in 20 mM potassium phosphate buffer (pH 8.0) or 8 mg/ml peptide and 10 mM ThT in 16 mM potassium phosphate buffer (pH 8.0) with indicated concentrations of NaCl. The fluorescence spectra were obtained using JASCO FP-6500. The excitation wavelength was 435 nm.
Assay solution was comprised of 1.2 mg/ml protein and 6 mM Congo red in 20 mM potassium phosphate buffer (pH 8.0). After incubation for 15 min, the absorption spectra were acquired using a Shimadzu UV-2550 spectrophotometer. NHBP-1 gel was stained with Congo red solution at a final concentration of 0.17%, washed three times with Milli Q and then spread on a slide glass. To remove unbound Congo red, the specimen was immersed in 99% EtOH. Birefringence of the specimen was observed by DMLP polarized light microscope (Leica).
|
Results |
|---|
|
TopAbstractIntroductionMaterials and methodsResultsDiscussionAcknowledgementReferences |
|---|
Filament formation from (N1)2-SipA446–684
(N1)2-SipA446–684 is the fusion protein in which two copies of NHBP-1, a peptide aptamer that binds SWNHs (Kase et al., 2004), are connected to the N-terminal of the actin binding domain of SipA, a Salmonella type III secretion protein (Fig. 1A).
|
We originally constructed this fusion protein with the aim of using it as a dispersant for SWNHs. (N1)2-SipA446–684 was designed to have a di-block structure: the N-terminal region contains binding sequences for hydrophobic carbonaceous materials, whereas the rest of the molecule (SipA446–684) is remarkably soluble and monodispersed in aqueous solution. In the past, similar amphiphilic molecules have been successfully used as dispersants for both carbon nanotubes (Arnold et al., 2005
) and carbon nanohorns (Murakami et al., 2006). However, when we mixed the purified (N1)2-SipA446–684 and SWNHs under aqueous conditions, the complex did not pass through dextran-based chromatographic media (Sephadex G25, Amersham Bioscience) equilibrated with PBS, which indicated that large aggregates had formed (data not shown). Observation of a sample in a TEM revealed that a filamentous material interconnected the spherical SWNH molecules to form large molecular networks (Fig. 2A and B). Because we had never seen filamentous structures form from SWNHs, we surmised that the fibrils were self-assemblages of (N1)2-SipA446–684. That idea was confirmed when we examined a sample of (N1)2-SipA446–684 solution. Although the solution appeared transparent, the TEM images showed that similar fibrillar structures were present in the protein sample (Fig. 2C). Because no such macromolecular assemblages formed from purified SipA446–684 (data not shown), we concluded that the N-terminal appended sequence had endowed soluble SipA446–684 protein with a capacity for self-assembly.
|
In contrast to the typical amyloid, which has a straight structure, the assembled filamentous structures were apparently flexible, and so exhibited numerous twists and turns. The radii of the fibrils measured were nearly identical at 6–7 nm, and at 50–300 nm, their lengths were shorter than typical amyloid fibrils, which are usually >1 µm in length (Fig. 2C). Incubation of a sample of protein solution for 4 weeks at 4°C had no effect on the diameters or lengths of the fibrils, or on the apparent transparency (data not shown). Although branched fibrils were never observed, we did encounter substantial numbers of ring-shaped fibrils (Fig. 3), which have been previously described for other amyloid-like proteins. Flexible nature of (N1)2-SipA446–684 fibrils may responsible for the formation of ring-shaped structure.
|
When we measured the Rh of both SipA446–684 and (N1)2-SipA446–684 using DLS (Fig. 2D), we calculated the mean Rh of SipA446–684 to be 3.3 nm, which is in good agreement with the Rh previously obtained using X-ray small angle scattering (2.7 nm) (Mitra et al., 2000
). On the other hand, the mean Rh of (N1)2-SipA446–684 was calculated to be 44.4 nm with a relatively large deviation, which is indicative of a polydispersed population of short filaments. Analysis of the secondary structure of (N1)2-SipA446–684
The bacterial protein SipA binds to a host cell's F-actin and re-arranges its cytoskeleton, thereby playing a key role in bacterial infection of mammalian cells (Zhou et al., 1999). The C-terminal fragment of SipA, SipA446–684 retains the ability to bind to F-actin and to inhibit depolymerization of actin filaments (Mitra et al., 2000). X-ray diffraction analyses of a shorter fragment, comprised of residues 513 to 657, revealed that the polypeptide makes up a single domain 30 by 40 Å in size and dominated by helical secondary structure (Fig. 1B) (Lilic et al., 2003). This globular peptide has a well-packed hydrophobic core and does not appear to undergo conformation change (Lilic et al., 2003). The fragment we used in the present study contains an additional 67 residues at its N-terminal and 27 residues at its C-terminal. TEM analysis suggests that these two extensions form non-globular ‘arms’ (Fig. 1B) that tether actin subunits in opposing strands (Mitra et al., 2000; Lilic et al., 2003).
SWNHs are spherical aggregates of elongated graphitic tubes (Iijima et al., 1999) and have the potential to serve as novel carriers in drug delivery systems (Murakami et al., 2004, 2006; Ajima et al., 2005). The 12-mer sequence of NHBP-1 is enriched in polar and aromatic residues (Fig. 1), which suggests that the peptide interacts with the graphene structure through 
– interactions and with the surface structure composed of carbon and other atom(s) that may have been introduced during acid treatment (Zhu et al., 2003). By adding functional groups such as NHBP-1, the ability to bind SWNHs can be transferred to various nanoparticles (Sano et al., 2005). Our earlier work with nuclear magnetic resonance (Kulp et al., 2005) and CD spectroscopy (Kase et al., 2004) showed that NHBP-1 is conformationally labile at neutral pH, exhibiting two-state conformational exchanges involving motion of the first five residues and the ring of the proline at position six. At acidic pH, however, or in the presence of TFE, the structure of the peptide is stabilized through formation of -helix. Thus, the inherent instability of the NHBP-1 sequence can be overcome by extrinsic factors (Kulp et al., 2005).
We next used far-UV CD spectroscopy to examine the secondary structure of fibrils assembled form SipA446–684 and (N1)2-SipA446–684. The spectrum obtained with parental SipA446–684 was in good agreement with the one described previously (Mitra et al., 2000) and indicated that the protein is rich in -helix (Fig. 4A, blue). Similarly, (N1)2-SipA446–684 also was rich in -helical structure, though the content has slightly less than in the parental peptide (Fig. 4A, red).
|
We also obtained FTIR spectra for the two proteins. In addition to maxima at 1655 and 1547 cm–1, which were derived from the
-helical structure, the spectrum obtained from (N1)2-SipA446–684 contained a maximum at 1630 cm–1, which corresponds to a ß-sheet structure (Fig. 4B, red). No such shoulder at 1630 cm–1 was seen in the spectrum from SipA446–684 (Fig. 4B, blue). Thus, CD and FTIR analyses both indicated that the (N1)2-SipA446–684 fibrils retained a large amount of the -helical structure of the parental SipA446–684, but they also gained ß-sheet structure that was not observed in SipA446–684. (N1)2-SipA446–684 fibrils are an amyloid-like protein
Whether or not a protein should be considered amyloid-like can be determined based on its ability to bind azo dyes such as ThT and Congo red (Kelenyi, 1967). Although the molecular mechanism is not yet fully understood, ThT is believed to be intercalated into the cross-ß region of amyloid and amyloid-like proteins, which leads to a characteristic fluorescence profile (
ex = 446 nm, em = 490 nm) (Naiki et al., 1989). When we carried out a ThT binding assay with the filamentous structures assembled from (N1)2-SipA446–684 at neutral pH, we observed a dramatic increase in fluorescence at 480 nm (Fig. 5A, red), whereas no fluorescence signal was observed with the parental SipA446–684 (Fig. 5A, blue). When similar experiments were performed at pH 2.5, no fluorescence was observed with (N1)2-SipA446–684 or SipA446–684 (data not shown), indicating that fibril formation from (N1)2-SipA446–684 proceeds at neutral but not acidic pH. We also observed an increase in absorbance and a red-shift in the absorption peak of Congo red when it was incubated with (N1)2-SipA446–684 (Fig. 5B). Taken together, these results demonstrate that the filamentous structure of (N1)2-SipA446–684 contains cross-ß structure, and so can be categorized as an amyloid-like protein.
|
Formation of amyloid-like fibrils from NHBP-1 peptide
We have shown that a soluble globular protein, SipA446–684, is able to form a filamentous structure when two copies of the NHBP-1 sequence are appended at its N-terminal. That the assembled fibrils retain -helical structure (Fig. 4A) suggests two possible processes by which the fibrils could be formed. One involves a re-folding mechanism (i.e. type-i according to Nelson and Eisenberg) in which part of (N1)2-SipA446–684 is first unfolded and then re-folded into a cross-ß structure. In this mode, the re-folding could be limited to a specific region, which would enable retention of the parental -helical structure in much of the fibril. However, we believe this type of fiber formation is unlikely for SipA446–684 because the core domain (residues 551–657) is well packed and does not appear to undergo a conformation change at neutral pH. We also confirmed that, after being unfolded in 6 M guanidine hydrochloride, SipA446–684 could be re-folded into a protein having Rh and CD profiles identical to the originally folded protein (data not shown).
The other possible scenario involves a ‘gain-of-function’ mechanism (type-iii according to Nelson and Eisenberg). Recently, Sambashivan et al. (2005) showed that insertion of a Gly-(Gln)10-Gly peptide after Gly112 of RNase A causes the soluble protein to be transformed into a fibrillar one. The inserted peptide forms a type-ii (natively disordered class) amyloid-like structure within a loop that separates the two subdomains of RNase A (residues 1–111 and 112–115). Normally, RNase A forms a domain-swapped dimer between these two subdomains. A proposed 3D domain-swapped zipper-spine model of fibrillar RNase A predicts that a spine is formed from hydrogen-bonded (Gln)10 ß-strands and is surrounded by the globular domain-swapped RNase A (Sambashivan et al., 2005). In short, RNase A gained the function of self-assembly through insertion of the Gly-(Gln)10-Gly peptide sequence.
We considered the possibility that a similar ability to mediate a gain-of-function could be inherent in the NHBP-1 peptide. For the most part, our earlier work with synthetic NHBP-1 produced no evidence of fibril formation under the conditions used (Kase et al., 2004; Kulp et al., 2005). However, in one experiment with a NHBP-1-displaying nanoparticle, the modified particle tended to aggregate under high salt conditions (Sano et al., 2005). Therefore, to re-evaluate amyloid-like structure formation by NHBP-1 in the presence of salt, we dissolved synthetic NHBP-1 peptide to a concentration of 10 mg/ml in phosphate buffer and then added NaCl to a final concentration of 1 M. After incubation for 1 h, the peptide solution had gelled, and strong fluorescence was observed when it was stained with ThT (Fig. 6A). The gelation was dependent on the presence of NaCl, and the ThT fluorescence intensity increased with higher concentrations of NaCl (Fig. 6A inset). The possibility that the observed fluorescence reflected inhibition of the molecular motion of ThT caused by the gelation was excluded by measuring ThT fluorescence in agar (data not shown).
|
TEM analysis confirmed the formation of a fibrous structure (Fig. 6B), and the fibrils showed birefringence after staining with Congo red (Fig. 6C). Thus, NHBP-1 peptide appears to be a type-ii amyloid-like structure forming peptide.
The findings summarized above suggest that NHBP-1 should form ß-strands under conditions where amyloid-like structure is formed. Our earlier work indicates that the peptide adopts a random state in the presence of neutral pH and low ionic strength. When we re-assessed the CD profiles of the peptide at neutral pH in the presence of 1 M NaCl, we found that at a concentration of 8 mg/ml the peptide showed a spectrum characteristic of random structure in the absence of NaCl (green), but showed a ß-structure type spectrum in the presence of 1 M NaCl (red) (Fig. 7). We also noticed that a low concentration of peptide (1 mg/ml) did not show a structural transition under the conditions used (Fig. 7, blue). Apparently, in addition to the previously reported -helix, NHBP-1 can adopt a ß-structure under the appropriate conditions.
|
|
Discussion |
|---|
|
TopAbstractIntroductionMaterials and methodsResultsDiscussionAcknowledgementReferences |
|---|
We have shown that addition of two copies of NHBP-1, a peptide aptamer that binds to carbon nanohorns, endows monodispersed SipA protein with the capacity to form filaments. Although the fibrils formed were relatively short and flexible (Fig. 2C), as compared with typical amyloid-like proteins, they should be classified as amyloid-like structures based on their ability to bind azo dye (Fig. 5), which indicates that the fibrils contain cross-ß structure somewhere in their overall structure. Consistent with that idea is the appearance of a ß-structure specific signal in FTIR measurements (Fig. 4B). Reflecting their small size, fibril formation did not cause the solution to gel, and the solution remained transparent; nor did longer incubation result in the growth of the fibrils. CD analyses showed that the fibrils contain a considerable amount of
-helix (25–30%, depending on the model used; data not shown), which indicates that most of the helical structure of SipA is retained in the fibrils. Models that can account for the retention of the structure of the monomeric units include (i) one in which only part of the protein is transformed to a cross-ß structure and (ii) one in which the appended NHBP-1 peptide forms a cross-ß spine that is decorated by the folded SipA protein. We believe the second gain-of-function type mechanism most likely underlies the formation of (N1)2-SipA446–684 fibrils. In fact, we demonstrated that the NHBP-1 peptide can form a typical amyloid-like structure in the presence of 1 M NaCl and a high concentration of the peptide (Fig. 5). Although we originally selected NHBP-1 as a binding peptide for carbon nanomaterials (Kase et al., 2004), it is not entirely surprising that it has an unexpected capacity for self-assembly. In fact, many peptides have now been shown to have the ability to self-assemble (Zhang et al., 1993; Tenidis et al., 2000; Reches et al., 2002), and it is interesting that these peptides, as well as NHBP-1, often have aromatic residues in their sequences (Gazit, 2005). At the moment, it is not clear whether or not its capacity for self-assembly is required for NHBP-1 to be recognized by carbon nanohorns. Interestingly, some of the solid-state materials that bind peptides derived from natural proteins also possess a capacity for self-assembly (He et al., 2003). We must carefully consider the functional relationship between natural peptide binders and artificial peptide aptamers, however.
TEM images showed the diameters of the fibrils formed in the present study to be roughly 6–7 nm (Fig. 2C). The size of the globular domain of SipA513–657 is roughly 3 by 4 by 4 nm (Lilic et al., 2003). Bearing that in mind, we propose the model for the structure of (N1)2-SipA446–684 fibrils shown in Fig. 8. In the center of the fibrils, a cross-ß spine is formed from the N-terminal NHBP-1 (x2) sequences, though the exact structure of the spine is difficult to predict from the available data. It could be a helical nanotube (Perutz et al., 2002) or, perhaps, it is a steric zipper structure such as that revealed by X-ray analysis of GNNQQNY fibrils (Nelson et al., 2005). We also do not exclude the possibility that a portion of SipA446–684, especially the N-terminal arm, is involved in the spine formation. In fact, we expect that part of SipA446–684 does underlie (N1)2-SipA446–684 self-assembly, as NHBP-1, itself, does not form fibrils under the conditions (N1)2-SipA446–684 does. We also noticed that two copies of NHBP-1 are needed for fibril formation: N1-SipA446–684, which has only one copy of NHBP-1, did not form fibrils under the conditions used (data not shown). Apparently, steric hindrance between SipA446–684 domains prevent fibril the formation of straight and long fibril of (N1)2-SipA446–684. Surrounding the spine should be the folded SipA446–684, which gives the fibrils their -helical structure (Fig. 8). The apparent flexibility of the fibrils suggests the SipA446–684 is loosely packed, which is consistent with the observation that SipA446–684 does not have the capacity for self-assembly (Nelson et al., 2005).
|
Our findings add a new subtype to the gain-of-function class of amyloid-like structure formation and strengthen the notion that a variety of molecular strategies can realize the emergence of filamentous structures in a biomolecular system. They also suggest the possibility of exploiting this method of assembly in nano-biotechnology.
|
Footnotes |
|---|
Edited by Haruki Nakamura
|
Acknowledgement |
|---|
|
TopAbstractIntroductionMaterials and methodsResultsDiscussionAcknowledgementReferences |
|---|
We thank Prof. Yuichiro Maéda for providing pET-SipA446–684. We acknowledge Prof. Sumio Iijima and Dr Masako Yudasaka for providing carbon nanohorns.
|
References |
|---|
|
TopAbstractIntroductionMaterials and methodsResultsDiscussionAcknowledgementReferences |
|---|
Ajima K., Yudasaka M., Murakami T., Maigne A., Shiba K., Iijima S. (2005) Mol. Pharmcol. 2:475–480.[CrossRef][Medline]
Alberts B., Johnson A., Lewis J., Raff M., Roberts K., Walter P. (2002) Molecular Biology of the Cell(Garland Publishing, New York).
Arnold M.S., Guler M.O., Hersam M.C., Stupp S.I. (2005) Langmuir 21:4705–4709.[CrossRef][ISI][Medline]
Chapman M.R., Robinson L.S., Pinkner J.S., Roth R., Heuser J., Hammar M., Normark S., Hultgren S.J. (2002) Science 295:851–855.
Claessen D., Rink R., de Jong W., Siebring J., de Vreugd P., Boersma F.G., Dijkhuizen L., Wosten H.A. (2003) Genes Dev. 17:1714–1726.
Dobson C.M. (1999) Trends Biochem. Sci. 24:329–332.[CrossRef][ISI][Medline]
Dobson C.M. (2003) Nature 426:884–890.[CrossRef][Medline]
Eakin C.M., Attenello F.J., Morgan C.J., Miranker A.D. (2004) Biochemistry 43:7808–7815.[CrossRef][Medline]
Fowler D.M., Koulov A.V., Alory-Jost C., Marks M.S., Balch W.E., Kelly J.W. (2006) PLoS Biol. 4:e6.[CrossRef][Medline]
Gazit E. (2005) FEBS J. 272:5971–5978.[CrossRef][Medline]
Gebbink M.F., Claessen D., Bouma B., Dijkhuizen L., Wosten H.A. (2005) Nat. Rev. Microbiol. 3:333–341.[CrossRef][ISI][Medline]
Govaerts C., Wille H., Prusiner S.B., Cohen F.E. (2004) Proc. Natl Acad. Sci. USA 101:8342–8347.
He G., Dahl T., Veis A., George A. (2003) Nat. Mater. 2:552–558.[CrossRef][ISI][Medline]
Iijima S., Yudasaka M., Yamada R., Bandow S., Suenaga K., Kokai F., Takahashi K. (1999) Chem. Phys. Lett. 309:165–170.
Ivanova M.I., Sawaya M.R., Gingery M., Attinger A., Eisenberg D. (2004) Proc. Natl Acad. Sci. USA 101:10584–10589.
Janowski R., Kozak M., Jankowska E., Grzonka Z., Grubb A., Abrahamson M., Jaskolski M. (2001) Nat. Struct. Biol. 8:316–320.[CrossRef][ISI][Medline]
Jimenez J.L., Guijarro J.I., Orlova E., Zurdo J., Dobson C.M., Sunde M., Saibil H.R. (1999) EMBO J. 18:815–821.[CrossRef][ISI][Medline]
Jimenez J.L., Nettleton E.J., Bouchard M., Robinson C.V., Dobson C.M., Saibil H.R. (2002) Proc. Natl Acad. Sci. USA 99:9196–9201.
Kase D., Kulp J.L. III, Yudasaka M., Evans J.S., Iijima S., Shiba K. (2004) Langmuir 20:8931–8941.[CrossRef][ISI][Medline]
Kelenyi G. (1967) J. Histochem. Cytochem. 15:172–180.[Abstract]
Kulp J.L. III, Shiba K., Evans J.S. (2005) Langmuir 21:11907–11914.[CrossRef][ISI][Medline]
Lilic M., Galkin V.E., Orlova A., VanLoock M.S., Egelman E.H., Stebbins C.E. (2003) Science 301:1918–1921.
Mitra K., Zhou D., Galan J.E. (2000) FEBS Lett. 482:81–84.[CrossRef][ISI][Medline]
Murakami T., Ajima K., Miyawaki J., Yudasaka M., Iijima S., Shiba K. (2004) Mol. Pharmcol. 1:399–405.[CrossRef][Medline]
Murakami T., Fan J., Yudasaka M., Iijima S., Shiba K. (2006) Mol. Pharmcol. 3:407–414.[CrossRef][Medline]
Murata K., Kaneko K., Steele W., Kokai F., Takahashi K., Kasuya D., Hirahara K., Yudasaka M., Iijima S. (2001) J. Phys. Chem. B 105:10210–10216.
Naiki H., Higuchi K., Hosokawa M., Takeda T. (1989) Anal. Biochem. 177:244–249.[CrossRef][ISI][Medline]
Nelson R. and Eisenberg D. (2006) Curr. Opin. Struct. Biol. 16:260–265.[CrossRef][ISI][Medline]
Nelson R., Sawaya M.R., Balbirnie M., Madsen A.O., Riekel C., Grothe R., Eisenberg D. (2005) Nature 435:773–778.[CrossRef][Medline]
Perutz M.F., Finch J.T., Berriman J., Lesk A. (2002) Proc. Natl. Acad. Sci. USA 99:5591–5595.
Pickersgill R.W. (2003) Structure 11:137–138.[Medline]
Reches M., Porat Y., Gazit E. (2002) J. Biol. Chem. 277:35475–35480.
Sambashivan S., Liu Y., Sawaya M.R., Gingery M., Eisenberg D. (2005) Nature 437:266–269.[CrossRef][Medline]
Sambrook J. and Russell D.W. (2001) Molecular Cloning A Laboratory Manual(Cold Spring Harbor, New York).
Sano K., Ajima K., Iwahori K., Yudasaka M., Iijima S., Yamashita I., Shiba K. (2005) Small 1:826–832.
Studier F.W., Rosenberg A.H., Dunn J.J., Dubendorff J.W. (1990) Methods Enzymol. 185:60–89.[Medline]
Tenidis K., Waldner M., Bernhagen J., Fischle W., Bergmann M., Weber M., Merkle M.L., Voelter W., Brunner H., Kapurniotu A. (2000) J. Mol. Biol. 295:1055–1071.[CrossRef][ISI][Medline]
Westermark P. (2005) FEBS J. 272:5942–5949.[CrossRef][Medline]
Zhang S., Holmes T., Lockshin C., Rich A. (1993) Proc. Natl. Acad. Sci. USA 90:3334–3338.
Zhou D., Mooseker M.S., Galan J.E. (1999) Science 283:2092–2095.
Zhu J., Kase D., Shiba K., Kasuya D., Yudasaka M., Iijima S. (2003) Nano Lett. 3:1033–1036.
Received October 19, 2006; revised December 8, 2006; accepted December 24, 2006.
CiteULike 
Connotea 
Del.icio.us What's this?
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||