Decoding the secrets of

spider silk
Spider silks have been employed by man for several thousands of
years. Spider silks possess extraordinary mechanical properties due
to a combination of strength and extensibility that are superior to
most man-made fibers. Spider silk fibers are a protein-based material
produced in a highly sophisticated hierarchical process under mild
conditions. Here, we review the current understanding of spider silk and
its assembly process, as well as discuss the application of silk-based
materials to the fields of biomedicine and materials engineering.
Lukas Eisoldt, Andrew Smith, Thomas Scheibel*
Universität Bayreuth, Fakultät für Angewandte Naturwissenschaften, Lehrstuhl für Biomaterialien, Universitätsstraße 30, 95447 Bayreuth,
Germany
*E-mail: [email protected]

Silk has been produced by a large variety of arthropods for is built from up to five different silk types. A female orb weaving
hundreds of millions of years. Humans have used silk materials for spider is even able to produce up to seven different silks, including
several thousands of years for all kinds of applications, ranging one special silk for egg cases2 (Fig. 1). Spider silks are tailored
from textiles to wound dressings to uses in the military. The main for specific purposes and exhibit a great variation in mechanical
source of silk is the mulberry silkworm Bombyx mori, since it is properties. Between the different silk types their strength (the
easy to farm (farming started about 5000 years ago in China). stress needed to break the fiber) ranges from 0.02 to a remarkable
The silk of spiders cannot be collected as easily, but the nets 1.7 GPa, which exceeds steel (1.5 GPa), while the extensibility
of orb weaving spiders (Araneidae) have also been historically varies between 10 and 500 %3,4. Interestingly, most spider silks
employed for applications such as fishing or wound coverage, due have a sophisticated combination of strength and extensibility,
to their outstanding mechanical and biomedical properties. Orb which yields a very high toughness (the amount of energy
webs can be found all over the world, with web diameters ranging absorbed per volume before breakage) that exceeds most natural
from a few centimeters to several meters. Very recently spiders or man-made fibers. Also common for all silks is their viscoelastic
in Madagascar have been discovered to spin single threads which behavior, since upon stretching energy is dissipated in the form of
span across rivers, with a length of about 25 m1. Orb webs have heat, diminishing any elastic recoil5.
been adapted over millions of years to withstand and compensate The most investigated spider silk is dragline silk, the lifeline a spider
for the kinetic impacts of the large (in comparison to the fiber always drags behind, which shows the most remarkable combination of
diameter) prey of the spider. When examined by the naked eye strength and extensibility. It is produced in the major ampullate gland
all threads in an orb web appear to be the same, but in fact a web (individual glands exist for each silk type) of a spider, and therefore its

80 MARCH 2011 | VOLUME 14 | NUMBER 3 ISSN:1369 7021 © Elsevier Ltd 2011

 Decoding the secrets of spider silk REVIEW

Fig. 1 Schematic overview of different silk types produced by female orb weaving spiders (Araneae). Each silk type (highlighted in red) is tailored for a specific purpose.

correct biological name is major ampullate silk (MAS). MAS is used for Primary structure of spider silks –
the frame and radii of an orb web, as well as the aforementioned lifeline less (complexity) is more
by most spiders. The mechanical properties of dragline silk of several All spider silk threads are mainly composed of one or more
different species have been extensively analyzed4,6-8. In comparison to proteins, called spidroins, which tend to be large (up to 350 kDa per
man-made fibers, the maximum strength of up to 1.7 GPa (dragline monomer)14. Despite their different properties, all spidroins share a
of Caerostris darwini) is in the range of high-tech materials. Although, common primary structure pattern comprising a large central core of
Kevlar or carbon fibers have a higher strength and stiffness, spider silk repeated modular units, accounting for approximately 90 % of the
fibers have a much higher toughness due to their greater extensibility9. amino acids of the protein, flanked by non-repetitive domains (Fig. 2a).
However, dragline silk offers even more interesting features. For These non-repetitive terminal domains have a folded globular structure
example, a hanging spider hardly ever twists, indicating an interesting and are comprised of approximately 100 – 140 amino acids. Their
torsional dampening behavior of the dragline thread. When a dragline sequences are highly conserved throughout different spider species
thread is twisted it does not oscillate around the new position, like a and silk types15-17. The terminal domains are essential for storage
Kevlar fiber would. After a while, the fiber slowly returns to its initial of spidroins in the silk glands and fiber formation in the spinning
position, indicating that there is a sort of a shape memory within the duct, since it is here that they trigger crucial steps in the complex
fiber10. Another interesting feature of spider silk is its ability to undergo assembly of the spidroins upon environmental changes, like pH or ionic
supercontraction. When dragline silk is wetted, or when the relative composition and strength18,19.
humidity is above 60 %, a silk thread swells in diameter and shrinks in The sequence of the repetitive core is tailored for the individual
length by about 50 %11,12. Other spider silk types show this behavior as mechanical functions of the different silk types. In general, it consists
well, but in these cases this behavior is much less pronounced13. of modular units each with about 40 – 200 amino acids7,14,20,21,
Besides dragline silk the other silk types also show interesting and the modular units are repeated up to approximately 100 times
properties like an extraordinary extensibility of up to 500 % within the core domain14. In major ampullate, minor ampullate and
(flagelliform silk) or a glue-like behavior (aggregate silk), but in this flagelliform silk, the modular units are mainly comprised of a subset
review we will continue to focus on dragline silks. of the sequence motifs (Ala)4-14, (GlyAla)n, GlyGlyX and GlyProGlyXX,

MARCH 2011 | VOLUME 14 | NUMBER 3 81

REVIEW Decoding the secrets of spider silk

(a)

(b)

(c)

Fig. 2 The hierarchical setup of spider silk proteins. (a) Illustration of a spidroin comprised of up to 100 repetitive modules (blue) and two terminal domains
(red and green). The components are not drawn to scale. (b) The repetitive modules of the most prominent silk types comprise a distinct subset of amino acid
motifs. (c) Each motif is thought to fulfill a distinct structural role and contributes to the mechanical properties of the final fiber.

with X representing a variable amino acid (Fig. 2b). The subset, the has been shown that the polyalanine motif forms defined nano-
sequential order and the number of these motifs in each module are sized crystals (2 x 5 x 7 nm) based on tightly packed anti-parallel
important for the mechanical properties of the final fiber. In contrast β-sheets22,29-34, in which it is very likely that polyalanine stretches
to terminal domains, the core region is intrinsically unfolded, as long as from different silk molecules constituting very strong non-covalent
the spidroins are stored in the gland22-25. cross-links. The structure of the GGX motif in dragline silk is less
understood. Recent NMR studies provided evidence that these motifs
From sequence to structure might form β-sheets as well as less ordered helical structures33,35-37.
The dragline fiber is not only the best characterized spider silk in Meanwhile, the structure of the GPGXX motif from MaSp2 seems to
terms of mechanical properties, but also concerning its structure- be clear. Due to the proline residue, this pentapeptide likely forms
function relationship. Unlike other spider silks, dragline fibers can be β-turns which, when repeated, yield a spiral structure as suggested for
collected from orb webs or by “milking” (forced silking) of spiders. elastin38.
Electron, atomic force and light microscopy have revealed a core- It should be noted that in the case of Araneus diadematus both
shell structure for the dragline fiber26,27. The shell is quite thin and MAS components, named ADF 3 and 4 (Araneus diadematus fibroin)
contains different biomolecules like lipids, glycoproteins and other due to historical biological reasons, have similar proline contents
silk proteins28. The core consists mainly of two proteins produced (~ 13 %)21. Compared to other MAS fibers Araneus silk shows no
in the major ampullate gland named MaSp (Major ampullate Spidroin) significant differences concerning its mechanical behavior indicating
1 and 2. that the presence of low-proline-content spidroins is not necessary
A single modular unit of MaSp1 usually comprises a polyalanine for the mechanical performance. It is worth noting that a macroscopic
block and several GGX motifs. In modules of MaSp2 the GGX motif is effect of the proline content is only seen when dragline silk is wetted.
replaced by the motif GPGXX, increasing the proline content of MaSp2 The amount of proline influences the degree of supercontraction
dramatically. Several dozens of repeats of these modular units build and the mechanical properties of wetted silk39-41. However, it seems
the complete core region of these spidroins. The potential structures to be more important to have a pair of hydrophobic/hydrophilic
of the individual motifs and their contribution to the properties of the spidroins, since comparing the hydrophobicity of MA spidroins between
final fiber have been intensively investigated over the last two decades different species revealed that MaSp1/ADF4 are more hydrophobic and
(Fig. 2b). Using different NMR techniques and x-ray diffraction it MaSp2/ADF3 are more hydrophilic.

82 MARCH 2011 | VOLUME 14 | NUMBER 3

 Decoding the secrets of spider silk REVIEW

(a)

(b)

Fig. 3 The natural spinning process. (a) Illustration of a spider’s spinning gland divided into four parts. (b) Schematic model of the silk fiber assembly mechanism
occurring along the spinning apparatus.

The natural spinning process part of the spinning apparatus, an S-shaped duct, the liquid crystalline
The primary structure of silk proteins alone does not explain the fluid is exposed to a constant elongational flow. The liquid crystallinity
outstanding properties of a dragline silk fiber. Fibers technically spun from allows the molecules to flow in a pre-aligned manner and to further
reconstituted spidroins (i.e. spidroins obtained from chemically dissolved align along the flow axis during the passage through the duct. Due to
spider silk fibers) or from spidroins from the original spinning dope show its tapering, the shear forces increase along the duct leading to the
completely different mechanical properties compared to fibers spun by formation of β-sheet crystals19,40-50. In the last part of the spinning
spiders, indicating that the spinning process is also crucial42,43. apparatus the shear stress further increases, due to pulling of the fiber
The spiders spinning apparatus can be divided into four parts out of the spigot (Fig. 3b). The pulling forces lead to some structural
(Fig. 3a). In the first part, the tail, the spidroins are secreted from changes in the spidroins (especially their terminal regions) accompanied
specialized cells. In the second part, the sac or ampulla, the spidroins by exposure of hydrophobic areas followed by a phase separation
are stored, and further compounds are added later forming the between the solvent (i.e. water which is actively removed by epithelial
skin of the fiber. The protein concentration during storage in the cells51-53) and the spidroins54,55 (Fig. 3b). The structural changes in the
ampulla can reach up to 50 % (w/v)22,44. This is a remarkably high terminal domains re-arrange the position of the core regions within
concentration, since most proteins tend to aggregate irreversibly under the micellar-like structures, and this together with mechanical forces
such conditions. To prevent unspecific aggregation, spidroins form supports phase separation and spidroin assembly.
micellar-like structures controlled by their amphiphilicity19,45-48 and by As well as mechanical influences, chemical changes trigger a structural
the non-repetitive termini which induce or support the formation of arrangement, especially in the S-shaped duct. There the exchange of
such supramolecular assemblies18,19,48 (Fig. 3b). The stored spidroins sodium and chloride ions for the relatively kosmotropic potassium and
in the ampulla, constitute the so-called spinning dope which has phosphate ions is responsible for the exposure of hydrophobic areas
characteristics of a nematic liquid crystalline phase47,49. In the third within the C-terminal non-repetitive domain19,56. Upon lowering the

MARCH 2011 | VOLUME 14 | NUMBER 3 83

REVIEW Decoding the secrets of spider silk

(a) (b) very time-consuming and therefore not profitable, especially if a

specific silk type (other than dragline silk) is desired. In order to achieve
spider silk on a large scale, biotechnological production of spidroins
is the only feasible solution. For recombinant protein production, the
DNA sequence from a donor organism (here the spider) has to be
recombined with that from an acceptor (a so-called host) which is
used for the production. Most recombinant silk proteins are produced
in the bacterium Escherichia coli which is a well-established host for
the industrial scale production of proteins. It allows a fast and scalable
Fig. 4 Illustration of the hierarchical spider silk structure. (a) The fiber shows a
skin-core structure with fibrils forming the core. (b) On the nanostructural level production within 3 – 4 days from the initiation of production to the
the fibrils comprise small tightly packed β-sheet crystals (red arrow) and larger finally purified protein.
crystalline regions (black arrow) interconnected by an amorphous matrix .
Attempts to produce recombinant MA silk proteins based on
pH value from roughly 7 to 6, the amino-terminal non-repetitive domains fragments of native silk genes from spiders in bacteria resulted
dimerize, likely acting as additional physical cross-links between the in low yields, mainly due to the different codon usage of spiders
spidroins48,57,58. and E. coli67. However, in gene engineering approaches the codon
Before the fiber exits the spigot it passes the so-called “valve”, usage of the genes can be adapted to the host organism. In the
an organ that assists in restarting the spinning process after internal case of silk genes, the repetitive character allows a straight forward
rupture52. It twists the duct back and forth, thereby moving the approach. As depicted above, most spider silk proteins show a
ruptured fiber end into the opened valve. Once the fiber has left the simple primary structure composed of repeated modular units which
spider, the spider pulls and stretches the fiber leading to further water contain 100 – 200 amino acids based on smaller sequence motifs
evaporation and additional alignment of the molecules inside the fiber. (e.g. polyalanine). The arrangement and the exact sequence of the
All the mechanisms of this highly sophisticated spinning process motifs differ from silk type to silk type and from species to species,
enable the spider to efficiently produce a very tough fiber under mild but within an individual spidroin they are quite regular. Thus, by
conditions; a process yet unrivaled by any man-made fiber spinning analyzing a spidroin’s amino acid sequence, one can identify the
process. modular units and design artificial gene encoding for silk proteins
using host-specific codons. With this method it is possible to create
The structural organization of a recombinant silk proteins with a high sequence similarity (on a
dragline fiber protein level) to the natural blueprints. Additionally, completely
The structure of dragline silk fibers reveals three hierarchical levels59. On novel silk-proteins can be created by hybridizing motifs of different
the macroscopic level the fiber shows a core-shell structure (Fig. 4a). On silk types. This approach has been used by several groups to produce
the mesoscopic level, fibrils oriented along the fiber axis form the core of various engineered silk proteins. Most of these engineered proteins
the fiber, resembling the structure of a rope. On the molecular level small are based on the sequences of MaSp1/268,69 from various species
crystallites (2 x 5 x 7 nm) comprising tightly packed alanine β-sheets and or ADF3/470, but some also engineered recombinant flagelliform
larger crystalline regions (> 100 nm) based on glycine-alanine sequences spidroins71-74.
with high β-sheet content can be found. Both crystalline regions are The majority of the recombinantly produced silk proteins are
interconnected by a so-called amorphous matrix which contains loosely smaller (30 – 110 kDa) than natural spidroins (300 – 350 kDa).
arranged helical structures oriented along the fiber axis (Fig. 4b). In This is due to the biological limitation of E. coli, which lowers
this two-phase-model the crystalline regions mediate the strength of the protein yields dramatically with increasing protein size75,76.
the fiber, whereas the matrix is responsible for the elasticity of the However, this problem can be partly circumvented by modifying
fiber. However, results obtained from various x-ray and NMR methods the production host77. Strikingly, all artificial spidroins, except
indicate that parts of the amorphous regions are much more ordered and one protein based on the sequence of ADF346, show much
oriented, leading to a model with a third phase which interconnects the lower solubilities in aqueous solution than the natural spidroins
crystalline regions with the amorphous phase60-62. For further detailed (see68,69 and references therein). A lot of these proteins tend
information we recommend consulting additional literature63-66. to self-assemble or aggregate into small fibrillar structures with
sizes ranging from a few micrometers to centimeters78. Although
Biotechnological production of spider these spontaneously formed fibrils show β-sheet structures, they
silk proteins are only remotely reminiscent of natural silk fibers, since in this self-
In contrast to silkworms (Bombyx mori), most spiders cannot be farmed assembly process some crucial structural alignment steps do not take
due to their cannibalistic behavior. Collecting silks from webs is also place.

84 MARCH 2011 | VOLUME 14 | NUMBER 3

 Decoding the secrets of spider silk REVIEW

Fig. 5 Different non-natural assembly forms (bold text) of recombinant spider silk proteins and potential applications thereof.

Exploiting the application potential of Concerning applications, the aforementioned silk shapes can be
recombinant spider silks used as carriers for drugs or as scaffolds in tissue engineering80, since
So far, no biomimetic spinning process exists that would allow the natural silks show a good biocompatibility (cytocompatibility, low
formation of native-like spider silk fibers from recombinant silk immunogenicity, biodegradeability, low toxicity) in vivo and in vitro69,87-92.
proteins. However, despite the nonexistent spinning process most Recombinant spider silk proteins have a similar cytocompatibility to
recombinant spider silks can self-assemble into non-natural shapes natural silks making them suitable for biomedical applications93-95.
such as spheres, capsules, films, non-wovens or hydrogels (Fig. 5), Besides its biocompatibility recombinant spider silk structures share
which have a high application potential79-83. some other interesting features with natural spider silk fibers, like its
Depending on the processing conditions the assembly forms have smooth surface, making them suitable for technical applications as well.
different material properties linked to the molecular structure of These features can be used in technical applications, e.g. non-wovens
the employed silk protein. Morphologies with high β-sheet contents as air filtering devices or spheres, as additives for cosmetics to give the
(spheres, hydrogels) that are formed by fast salting-out or slow self- product a “silken” smooth feeling.
assembly processes are water-insoluble55,81,84,85. In contrast, forced
assembly (e.g. into films or non-wovens) out of fast evaporating Future perspective
organic solvents such as hexafluoroisopropanol yields α-helical rich Spider silks are impressive biopolymers that have evolved over
compositions rendering the structures water soluble. Interestingly, such millions of years. Over the last several decades a lot of progress has
structures can be post-treated with substances that induce β-sheet been made in unraveling some of its secrets. Nevertheless, certain
formation leading to the gain of water-insolubility80,82,86. Not only questions, especially concerning solubility, storage and assembly of
water solubility, but the mechanical properties of the silk materials the underlying spider silk proteins, which are crucial for the technical
are also linked to the β-sheet content. Since recombinant spider silk production of e.g. silk fibers, are still unanswered. The biotechnological
proteins are monodisperse (in comparison to synthetic polymers), production of recombinant spider silk in larger scales is a landmark in
high degrees of assembly control and, therefore, the final material’s spider silk research, since now investigations are enabled to answer
properties are possible. such questions. Furthermore, recombinant spider silk proteins can be

MARCH 2011 | VOLUME 14 | NUMBER 3 85

REVIEW Decoding the secrets of spider silk

processed into many different morphologies and shapes which have Acknowledgements:
great potential in various technical and biomedical applications. With This work was supported by the Bayrische Staatsministerium für Umwelt
recombinant production technologies it is possible to create tailor- und Gesundheit, project U8793-2008/11-2. We would to thank Dr.
made silk-based biopolymers for many different purposes on a large Martin Humenik for his creative input and technical support in creating
scale within a relatively short amount of time. the artwork.

REFERENCES 47. Knight, D. P., and Vollrath, F., Proc R Soc Lond Ser B-Biol Sci (1999) 266, 519.
1. Agnarsson, I., et al., PLoS ONE (2010) 5, e11234. 48. Hagn, F., et al., Angew Chem-Int Edit (2010) 50, 310.
2. Andersen, S. O., Comp Biochem Physiol (1970) 35 , 705. 49. Willcox, P. J., et al., Macromolecules (1996) 29, 5106.
3. Denny, M., J Exp Biol (1976) 65, 483. 50. Eisoldt, L., et al., J. Struct. Biol. (2010),
4. Blackledge, T. A., and Hayashi, C. Y., J Exp Biol (2006) 209, 2452. 51. Bonthrone, K. M., et al., Proc R Soc Lond Ser B-Biol Sci (1992) 248, 141.
5. Gosline, J. M., et al., Endeavour (1986) 10, 37. 52. Vollrath, F., and Knight, D. P., Int J Biol Macromol (1999) 24, 243.
6. Vollrath, F., J Biotechnol (2000) 74, 67. 53. Vollrath, F., et al., Proc R Soc Lond Ser B-Biol Sci (1998) 265, 817.
7. Hayashi, C. Y., et al., Mol Biol Evol (2004) 21, 1950. 54. Knight, D. P., et al., Int J Biol Macromol (2000) 27, 205.
8. Madsen, B., et al., Int J Biol Macromol (1999) 24, 301. 55. Rammensee, S., et al., PNAS (2008) 105, 6590.
9. Heim, M., et al., Angew Chem-Int Edit (2009) 48, 3584. 56. Knight, D. P., and Vollrath, F., Naturwissenschaften (2001) 88, 179.
10. Emile, O., et al., Nature (2006) 440, 621. 57. Landreh, M., et al., J Mol Biol (2010) 404, 328.
11. Work, R. W., and Morosoff, N., Text Res J (1982) 52, 349. 58. Gaines, W. A., et al., J Biol Chem (2010) 285, 40745.
12. Jelinski, L. W., et al., Int J Biol Macromol (1999) 24, 197. 59. Heim, M., et al., Chem Soc Rev (2010) 39, 156.
13. Gosline, J. M., et al., Elastomeric Network Models for the Frame and Viscid Silks 60. Thiel, B. L., et al., Biopolymers (1997) 41, 703.
from the Orb Web of the Spider Araneus diadematus. In Silk Polymers, American 61. Trancik, J. E., et al., Polymer (2006) 47, 5633.
Chemical Society, (1993), 544, 328. 62. Fossey, S. A., and Tripathy, S., Int J Biol Macromol (1999) 24, 119.
14. Ayoub, N. A., et al., Plos One (2007) 2, e514. 63. Porter, D., and Vollrath, F., Adv Mater (2009) 21, 487.
15. Motriuk-Smith, D., et al., Biomacromolecules (2005) 6, 3152. 64. Vollrath, F., and Porter, D., Soft Matter (2006) 2, 377.
16. Rising, A., et al., Biomacromolecules (2006) 7, 3120. 65. Ene, R., et al., Soft Matter (2009) 5, 4568.
17. Challis, R. J., et al., Insect Mol Biol (2006) 15, 45. 66. Papadopoulos, P., et al., Colloid Polym Sci (2009) 287, 231.
18. Askarieh, G., et al., Nature (2010) 465, 236. 67. Arcidiacono, S., et al., Appl Microbiol Biotechnol (1998) 49, 31.
19. Hagn, F., et al., Nature (2010) 465, 239. 68. Vendrely, C., and Scheibel, T., Macromol Biosci (2007) 7, 401.
20. Hinman, M. B., and Lewis, R. V., J Biol Chem (1992) 267, 19320. 69. Rising, A., et al., Cell Mol Life Sci (2010) 68, 169.
21. Guerette, P. A., et al., Science (1996) 272, 112. 70. Huemmerich, D., et al., Biochemistry (2004) 43, 13604.
22. Hijirida, D. H., et al., Biophys J (1996) 71, 3442. 71. Teule, F., et al., J Mater Sci (2007) 42, 8974.
23. Hronska, M., et al., Biomacromolecules (2004) 5, 834. 72. Miao, Y. G., et al., Appl Microbiol Biotechnol (2006) 71, 192.
24. Lefevre, T., et al., Biomacromolecules (2008) 9, 2399. 73. Zhou, Y. T., et al., Biomacromolecules (2001) 2, 111.
25. Lefevre, T., et al., Biomacromolecules (2007) 8, 2342. 74. Heim, M., et al., J Struct Biol (2010) 170, 420.
26. Li, S. F., et al., Biophys. J. (1994) 66, 1209. 75. Fahnestock, S. R., et al., J Biotechnol (2000) 74, 105.
27. Sponner, A., et al., PLoS ONE (2007) 2, e998. 76. Fahnestock, S. R., and Irwin, S. L., Appl Microbiol Biotechnol (1997) 47, 23.
28. Augsten, K., et al., Scanning (2000) 22, 12. 77. Xia, X.-X., et al., PNAS (2010) 107, 14059.
29. Warwicker, J. O., J Mol Biol (1960) 2, 350. 78. Huemmerich, D., et al., Curr Biol (2004) 14, 2070.
30. Grubb, D. T., and Jelinski, L. W., Macromolecules (1997) 30, 2860. 79. Hermanson, K. D., et al., Advanced Materials (2007) 19, 1810.
31. Simmons, A., et al., Macromolecules (1994) 27, 5235. 80. Spiess, K., et al., Macromol Biosci (2010) 10, 998.
32. Parkhe, A. D., et al., J Mol Recognit (1997) 10, 1. 81. Slotta, U. K., et al., Angew Chem-Int Edit (2008) 47, 4592.
33. van Beek, J. D., et al., PNAS (2002) 99, 10266. 82. Spiess, K., et al., Soft Matter (2010) 6, 4168.
34. Creager, M. S., et al., Biomacromolecules (2010) 11, 2039. 83. Hardy, J. G., and Scheibel, T. R., Prog Polym Sci (2010) 35, 1093.
35. Jenkins, J. E., et al., Biomacromolecules (2010) 11, 192. 84. Slotta, U., et al., Supramol Chem (2006) 18, 465.
36. Simmons, A. H., et al., Science (1996) 271, 84. 85. Lammel, A., et al., ChemSusChem (2008) 1, 413.
37. Lefevre, T., et al., Biophys J (2007) 92, 2885. 86. Huemmerich, D., et al., Appl Phys A-Mater Sci Process (2006) 82, 219.
38. Hayashi, C. Y., et al., Int J Biol Macromol (1999) 24, 271. 87. Vollrath, F., et al., In Vivo (2002) 16, 229.
39. Savage, K. N., and Gosline, J. M., J Exp Biol (2008) 211, 1937. 88. Allmeling, C., et al., J Cell Mol Med (2006) 10, 770.
40. Eles, P. T., and Michal, C. A., Biomacromolecules (2004) 5, 661. 89. Allmeling, C., et al., Cell Proliferat (2008) 41, 408.
41. Boutry, C., and Blackledge, T. A., J Exp Biol (2010) 213, 3505. 90. Gellynck, K., et al., J Mater Sci-Mater M (2008) 19, 2963.
42. Shao, Z., et al., Macromolecules (2003) 36, 1157. 91. Gellynck, K., et al., J Mater Sci-Mater M (2008) 19, 3399.
43. Seidel, A., et al., Macromolecules (2000) 33, 775. 92. Hakimi, O., et al., J Biomed Mater Res A (2010) 92A, 1366.
44. Vollrath, F., and Knight, D. P., Nature (2001) 410, 541. 93. Agapov, I., et al., Dokl Biochem Biophys (2009) 426, 127.
45. Jin, H. J., and Kaplan, D. L., Nature (2003) 424, 1057. 94. Widhe, M., et al., Biomaterials (2010) 31, 9575.
46. Exler, J. H., et al., Angew Chem-Int Edit (2007) 46, 3559. 95. Fredriksson, C., et al., Materials (2009) 2, 1908.

86 MARCH 2011 | VOLUME 14 | NUMBER 3