The amyloid protein conformation, a structure that propagates as insoluble fibrils of aggregated β-strand-containing peptides, is a common feature of many neurodegenerative diseases. In Alzheimer, Parkinson, and Huntington diseases, as well as in prion diseases, toxic amyloids or their precursors are blamed for all manner of neuropathology. In fact, most researchers have never met a nonpathogenic amyloid in higher animals and tend to view the protein fold as a bad seed, a carryover from evolution with no physiologically redeeming features.

That view may have to change, now that Jeffery Kelly and colleagues at the Scripps Research Institute in San Diego have discovered an abundant amyloid protein structure that functions in normal pigment production in mammalian cells. In the 29 November PloS online, the researchers report that amyloid fibers of the Pmel17 protein pack melanosomes, organelles that produce the protective pigment melanin. The amyloid provides a scaffold for the polymerization of tyrosine derivatives into melanin, accelerating the process and possibly protecting cells from the toxic effects of reactive melanin precursors. These results show for the first time a nonpathological role of amyloid, and suggest that this much-maligned conformation may, in fact, be an evolutionarily conserved motif with a number of physiological functions.

Melanin, the stuff of suntans and the body’s main defense against ionizing radiation, is produced inside membrane-bound melanosomes that are then transferred intact to neighboring cells. Melanosome maturation requires Pmel17, a transmembrane glycoprotein that is processed in melanosomes to release fragments (Mα), which then self-assemble into fibers. Present in many organisms, from fungi and insects to humans, Pmel17 is known to be essential for pigmentation in mice, chicken, and zebrafish.

Douglas Fowler and Atanas Koulov, co-first authors on the paper, used dye binding and biophysical techniques to show that Pmel17 fibers in melanosomes adopt an amyloid structure. In purified melanosomes from cow retina, Mα fibers bound both thioflavin S and Congo red, two amyloidophilic dyes. The fibers were detergent-resistant, but could be destroyed by boiling in 10 percent SDS, characteristic of a stable amyloid structure.

When the researchers expressed and purified recombinant Mα in bacteria, they found that the fragment spontaneously self-assembled into amyloid in vitro. The fibrillization was fast, at least four orders of magnitude faster than observed with the Alzheimer Aβ peptide or synuclein. Biophysical measurements, including electron microscopy, x-ray diffraction, and circular dichroism measurements were all consistent with the presence of the amyloid cross-β structure.

The function of the Pmel17 amyloid was revealed by experiments that simulated melanin production in vitro. Melanin is formed by polymerization of highly reactive tyrosine derivates, and the presence of Mα fibers sped up this polymerization. Interestingly, the structure of the melanin precursor indole-5,6-quinone (DHQ) resembles the core structure of thioflavin T, which might explain its avid binding to the Pmel17 amyloid. Also, Aβ and synuclein amyloids could also accelerate melanin production in vivo, suggesting that the presence of the cross-β sheet structure was the salient feature that enhanced polymerization.

Since amyloid formation is a highly toxic process for cells, why don’t melanocytes suffer consequences from harboring such a nasty player? The authors speculate that the secret lies in the way that functional amyloid formation occurs under optimized, controlled circumstances. The trafficking Pmel17 restricts its cleavage to melanosomes, thus sequestering the amyloidogenic Mα peptide inside a membrane-limited organelle. Also, the rapid kinetics of amyloid formation may serve to protect cells from damage by amyloid precursors. Compared to Pmel17’s tidy process, the extracellular formation of amyloid in neurodegenerative diseases seems downright disorderly, and more studies of this contrast may lead to new insights into pathogenic amyloid formation. Right off, the discovery of physiological amyloid in mammalian cells raises some new questions about possible pitfalls to using nonselective inhibitors of amyloid formation to treat such diseases.

The authors believe that Pmel17 is just the tip of a functional amyloid iceberg. They propose using the name “amyloidin” for nontoxic, useful amyloid, with the expectation that many more examples of this type of structure will be forthcoming.—Pat McCaffrey


  1. Functional amyloid and nonfunctional interpretations

    Fowler et al. report results of studies of the biophysical behavior of
    bovine melanosome protein Pmel17 and of the potential role of this protein in melanin biosynthesis. The data strongly support the conclusion that Pmel17, under appropriate conditions, rapidly assembles into amyloid-type polymers and that these polymers support melanin synthesis. The demonstration of a
    physiologically beneficial role of Pmel17 amyloid adds another example
    to a growing list of prokaryotic and eukaryotic proteins for which
    amyloid formation serves an important role in normal cellular
    function. In some respects, however, the authors' view of the
    relative importance of these findings to larger questions of
    prokaryotic and eukaryotic cell biology and the role of amyloid is
    unsupported experimentally and thus must be considered

    One of the most interesting observations of Fowler et al. was the rate
    of amyloid formation by Pmel17 (a half-time of roughly 1 second). This
    rate was four orders of magnitude larger than the rates observed for
    the Alzheimer and Parkinson disease-linked proteins amyloid
    β-protein and α-synuclein, respectively. This is a very
    interesting and important result, because investigation of its
    mechanistic basis has the potential to reveal structural factors
    controlling amyloid formation not only by Pmel17 but also by many
    others of the approximately 20 known amyloid proteins. The authors
    suggest that this rate may be a result of evolutionary pressures to
    minimize production of toxic intermediates in amyloid formation. This
    may be so, but this teleological interpretation remains unsupported

    This commentator is largely ignorant about melanosome biology, but one may postulate that factors other than "toxic intermediates" may have contributed with equal (or greater) importance to the protein evolution of Pmel17. For example, rapid assembly kinetics makes processes linked to Pmel17 assembly remarkably sensitive, and thus controllable, in temporal and
    concentration "regimes." Also, because Pmel17 assembly occurs within a
    specialized organelle—the early melanosome—any intermediates would
    be sequestered from other cellular compartments that might be
    sensitive to their toxic effects.

    The high rate of Pmel17 assembly raises questions about how this
    specific process can be integrated into the prevailing view that
    amyloid proteins share a common core structural organization. How does
    one reconcile a rate constant that is four orders of magnitude larger
    than those of other amyloid proteins with a "common structure"? Do the
    noncore regions of Pmel17 contribute to the population of the
    amyloid-competent conformational state by the nascent, proteolytically
    processed Pmel17 protein?

    Studies of melanin biosynthesis in vitro suggest that Pmel17 amyloid
    formation is linked to increased synthesis rates. The authors suggest
    that this "is the first example of an amyloid that functions in a
    chemical reaction." This may be so, but the argument would be
    strengthened if greater mechanistic insight existed. The suggestion
    that spatial organization of the substrate used in the experiments
    (DHQ) might be involved is quite reasonable and lends itself to future
    hypothesis testing.

    I question the proposition that a "general name amyloidin" be
    designated for "functional amyloid." Historically, terms have been
    created to define new entities and this process has helped scientists
    by standardizing meaning, facilitating communication and clarity of
    thought. The authors have done a superb job of characterizing the
    Pmel17 assembly product as amyloid, the classical fibrillar structure
    defined by its dye-binding characteristics, secondary structure, x-ray
    diffraction pattern, and morphology. Structurally, how does an
    amyloidin differ from an amyloid? How do you distinguish amyloids
    (which by definition have similar core structures) that have
    beneficial or detrimental effects that are context (intracellular,
    extracellular, or organ-specific milieu)-specific? This naming
    exercise becomes a semantic endeavor with little benefit, creating
    more obfuscation than clarification. How does one name a new amyloid
    protein that in the future may be found to have a beneficial effect?
    One would think initially that the protein is "nonfunctional" and
    name it amyloid, when in fact it is an amyloidin under some
    conditions. I suggest we maintain our current terminology and simply
    define under what conditions specific amyloid proteins have beneficial
    or detrimental physiologic effects.

    View all comments by David Teplow
  2. This paper and comment draw attention to other types of physiological amyloids or fibrillar structures that subserve normal funcitons. Examples include the elastins (e.g., Tamburro et al., 2005), as well as the complex network of filaments (from actin microfilaments to intermediate filaments) that also provide a scaffold and other functions needed by nearly all cell types. The existence of these normal filamentous poylmers that promote cell function and viablitiy will help dissect out how and why abnormal amyloid filaments exert deleterioius, pathological effects. It will also help us address the question of when, and by what mechanisms, they might serve a protective function in disease states.


    . Supramolecular amyloid-like assembly of the polypeptide sequence coded by exon 30 of human tropoelastin. J Biol Chem. 2005 Jan 28;280(4):2682-90. Epub 2004 Nov 18 PubMed.

Make a Comment

To make a comment you must login or register.


No Available References

Further Reading

No Available Further Reading

Primary Papers

  1. . Functional amyloid formation within mammalian tissue. PLoS Biol. 2006 Jan;4(1):e6. PubMed.