Self-assembly as a route to nanotechnology has been described as the “creation of material from its constituent components in a spontaneous ‘natural’ manner, i.e. by an interaction between the components or by a specific rearrangement of them, that proceeds naturally without any special external impetus” [1]. Biology is good at this, as exemplified by our very existence, and is seen in lipid bilayers surrounding each of our cells or the cytoskeleton within them. So, it might be thought that using biological molecules would provide a straightforward route to building novel nanoscale devices. Although this ultimately may be achievable, straightforward it is not.

One route being investigated is to use amyloid fibrils as nanowires, rendering them conductive by suitable functionalization. Amyloid fibrils are familiar as the constituents of plaques found in diseases such as new-variant Creuzfeld-Jacobs Disease (v-CJD) and Alzheimer's Disease. They consist of misfolded protein molecules in a cross-ß-sheet structure (reflecting the packing of neighboring ß-sheets). It has been suggested that all proteins can, under the right conditions and when usual biological control has been lost, misfold and form such fibrils [2]. So, one should not assume that amyloid fibrils are necessarily toxic, as they can turn up in situations as innocent as yogurt (formed by the milk protein ß-lactoglobulin) [3]. The fibrils are very stable and can be made to conduct via various projected routes, including covalently linking colloidal Au to the fibril [4].

Amyloid fibrils might therefore seem excellent candidates for self-assembly of useful devices. But there are drawbacks. Owing to large surface area, any fibrillar system is prone to nonspecific aggregation, since van der Waals' attraction between surfaces close to each other causes sticking. Aggregation contributed to the failure 20 years ago to develop ‘molecular composites’, despite much funding from US defence agencies. The idea was to disperse individual molecules of inherently stiff liquid-crystalline polymers in a thermoplastic matrix. If molecular dispersion had been possible, calculations of the theoretical strength indicated far superior properties. In reality, the molecules clumped together into substantial aggregates, and the resultant mechanical properties were insufficiently improved over the unreinforced composite to warrant the substantial increase in production costs. More recently, similar problems have been plaguing attempts to use carbon nanotubes, which have the same infuriating propensity to aggregate in solution and resist subsequent dispersion. Some success has been achieved by treating the nanotube surfaces with surfactant, but it remains a nontrivial issue.

In both these cases, aggregates form through nonspecific interactions. In the case of amyloid fibrils, the aim is to use the biological propensity to self-assemble into fibrils initially, but then to use them as nanowires without any higher-order aggregation. Recent work suggests that, both in vivo and in vitro, aggregation into larger units may in fact take place, but whether this is necessarily the case is so far unclear. Also it seems that such aggregation is not, in fact, nonspecific, although the factors that control it are yet to be established. In a wide range of protein systems, large, quasi-spherical entities have been seen, including in sections from diseased animal brains and in benign systems such as insulin and milk proteins [5]. Under a polarized light microscope, these entities exhibit the Maltese Cross pattern of extinction familiar to polymer scientists as the signature of melt-crystallized polymers such as polyethylene [6]. Such spherulites occur in many crystallizable systems including metals, and are a manifestation of polycrystalline growth. It has recently been proposed that this ubiquitous spherulitic structure results from competition between the local crystallographic symmetries and the randomization of the local orientation because of growth-front nucleation [7]. However, it is not immediately obvious how this translates into spherulites in amyloid-fibril-containing systems, where it appears that spherulitic growth occurs simultaneously with fibril elongation, with a presumed equilibrium being established between the two types of aggregate.

Despite uncertainty about the detailed mechanisms by which suprafibrillar aggregation occurs, it is clear is that such aggregation must be considered when trying to self-assemble nanowires based on amyloid fibrils. It is too much to hope that assembling such fibrils, functionalizing them in solution, and then permitting them to self-assemble (e.g. on a patterned surface) will immediately lead to a workable device. There are many different branches of science, from physics to biology, that need to be brought to bear if such aspects of bionanotechnology are to deliver.

Further reading
[2] C.M. Dobson. Philos. Trans. R. Soc. London, Ser. B Biol Sci, 356 (2001), p. 133
[3] W.J. Gosal et al. Langmuir, 18 (19) (2002), p. 7174
[4] T. Scheibel et al. Proc. Natl. Acad. Sci. USA, 100 (8) (2003), p. 4527
[5] M.R.H. Krebs et al. Proc. Natl. Acad. Sci. USA, 101 (40) (2004), p. 14420
[6] D. Bassett. Principles of Polymer Morphology, Cambridge University Press (1981)
[7] L. Granasy et al. Nat. Mater., 3 (9) (2004), p. 645

[1] Athene M. Donald is professor of experimental physics at the Cavendish Laboratory, University of Cambridge, UK.


Read full text on ScienceDirect

DOI: 10.1016/S1369-7021(05)00913-2