Mechanically, silks are outstanding material systems. The toughness of silk fibers is superior to any of the best synthetic high-performance fibers available today, including Kevlar (Dupont Advanced Fiber Systems). Although important insights into silk protein self-assembly have been achieved over the past 10 years (
1,
2), the mechanisms by which these proteins achieve metastable states in the glands of the spinning organisms remain unclear. This is a remarkable processing achievement that allows the concentration of protein in the glands to reach >30 weight % in water (
1), whereas at this concentration most proteins, globular or fibrous, aggregate and precipitate. In the case of silk, this would result in the premature formation of β sheets (crystallization), resulting in insolubility of the spinning dope and blockage of the spinning apparatus, events that would be catastrophic to the silk-spinning organism (
2). The lack of full comprehension of these processing steps has limited the ability to spin reconstituted silk solutions into fibers with properties comparable to those of native fibers.
Relationships between silk protein processing and fiber properties suggest that material features are controlled by a combination of the chemistry and the spinning process. Mechanical properties of spider silks are modulated on the basis of spinning conditions, including temperature, reeling rate, and drawing rate, as well as the specific type of silk () (
3,
4). Silkworm silk fibers are also influenced by process conditions such that properties can be achieved to match those of spider silks (
5). Silks from
Bombyx mori obtained at different reeling rates result in stronger but more brittle fibers. Wet spinning with postspinning treatments can also be used to generate spider silk–like properties and microstructures from silkworm silks (
6).
The concentration of silk protein increases during the progression from synthesis to spinning in the glands, with changes in chemistry (cations and decrease in pH) and physical features (sol-gel phase transition), along with mechanical alignment in shear flow (
1). Chain alignment is involved in later processing stages, and the internal drawdown of fibers—perhaps including active transport of water, addition of hygroscopic polymers such as sericin (in the silkworm), or other unknownmechanisms—plays an important role in the critical final steps that lead to fiber formation. Silks are not “squeezed out” of the glands but are mostly pulled out, either by the legs of the spider or by the “figure eight” movement of the head of the silkworm during spinning. Rheological studies into the different structures in the spinning gel formed in vivo versus in vitro are instructive in terms of how much native silk material (harvested directly from the gland) differs from reprocessed (fibers resolubilized into solution) silks (
7,
8). This disparity in solution behavior, as discerned from rheological assessments, suggests that there are still substantial knowledge gaps in understanding how to reverse engineer silk protein fibers. Morphological differences in the fiber cross sections among species originate at the spinerettes of spiders and can provide important clues on how processing within the gland, married to the spinning system, can coexist to control morphological features of the final fibers. Thin ribbons, such as those produced by tarantulas, to circular cross sections and other morphologies have been reported depending on the specific spider species. The influence of gland-specific features and process variables have not been studied in detail related to fiber morphology, yet could provide interesting insight into modes to regulate such fiber features.
Silk chemistries have been selected through evolution, and therefore structure-function relationships have been optimized for survival related to the needs of the silk-producing organisms (
9–
11). Among silk variants, particular interest is devoted to the silkworm silk of
B. mori as the commodity silk protein used in textiles and medical sutures and to the dragline silks of spiders such as
Nephila clavipes and
Araneus diadematus. Silks are modular in design, with large internal repeats flanked by shorter (~100 amino acid) terminal domains (N and C termini). Silks have high molecular weight (200 to 350 kD or higher) with transcripts of 10,000 base pairs and higher and >3000 amino acids (
10,
12). The larger modular domains are interrupted with relatively short spacers with hydrophilic charge groups in the case of silkworm silk. N and C termini are involved in the assembly and processing of silks, including pH control of assembly (
12). The N and C termini are highly conserved, in spite of their relatively small size compared with the internal modules (N terminus around 130 amino acids and C terminus around 100 amino acids). Accordingly, the inclusion of such modules in cloned silks may be essential when the emulation of native properties is the goal ().
Because native silkworm and spider populations already display variations in the sizes of silk genes and, as a result, in the proteins generated, it is perhaps important to relate size and modular unit distribution to silk protein assembly and material functions. The roles of the N and C termini in processing related to pH, cations, or protein concentration may be responsible for only a subset of the functional features of these domains (
13). It is conceivable that their role is to help stabilize silk during synthesis and processing in the gland, further underscoring the critical need to include such domains in genetically engineered silks (
14,
15). Some dragline spider silks consist of two key proteins, such as Masp1 and Masp2 in the case of
N. clavipes or
A. diadematus, with one of the proteins usually containing proline as a putative entropic spring (
16). Yet silkworm silks generally have one main protein.
The competition for water between the modular domains in silk proteins has not been quantified. This has implications for processing silks in aqueous and organic solvents. The silk-spinning dope in the glands of spiders and silkworms is stable to perturbations, such as when spiders are flicked off the ceiling by an arachnophobe or during the handling of silkworms. In contrast, high concentrations of silk protein in water readily crystallize from their metastable state, with almost any energy input resulting in rapid crystallization. Under natural conditions, silks can be stored for days and weeks in the producing organisms at ambient temperatures in preparation for spinning without premature crystallization. Conversely, careful control of these processes is essential in laboratory conditions in order to avoid premature crystallization. Many of the above factors lead to questions about how silks can be used as design templates in order to build synthetic polymer chemistry analogs that would recapitulate some or all of the useful functional features that exist in naturally occurring silks (
17). For example, the ability to design synthetic block copolymers with modular designs that can be processed in water under ambient conditions, and yet achieve water-stable robust mechanical properties without the need for chemical crosslinking, would seem to be an enviable outcome from the lessons offered from the designs of silk chemistries and processing.
The main source of silkworm silk is through sericulture, the domesticated production of silk from
B. mori. This production approach provides the textile world and the medical suture world with an abundance of material for current demands. Additionally, this production process is a source of support for microeconomies in many parts of the world. In contrast to the established supply chain available for silkworm silk, there are no commercial options for spider silks, largely because of the more aggressive nature of spiders and the more complex and smaller quantities of silk mixtures generated in orb webs. Silks have been cloned and expressed in a variety of heterologous hosts, from bacteria and fungi to plants and mammals, as well as via transgenic silkworms (
18–
22), yet the material properties of these cloned silks are inferior to the native materials. This difference can be traced to molecular weight and the modular domains. With few exceptions, most silks when cloned and expressed include only some of the core modular domains; in some cases, these modular domains plus the C terminus are truncated. These truncated silks may be sufficient for some applications, but only where mechanical properties are not critical.
Some silks are encoded by single exons, whereas others have conserved introns, features that are unique because in most eukaryotes it is unusual to have proteins encoded by single exons, particularly when the encoded proteins are so large (>3000 amino acids) (
23). Silk genetic domains are remarkable for their repetitive and modular nature, their size, and the highly evolved host cell requirements to meet the extreme demands of just a few amino acids (e.g., glycine and/or alanine in the case of silkworm silk) (
24). Further, the level of silk protein generated in the silkworm during transition from the worm to the moth is remarkable, prompting research into the use of transgenic silkworms as production factories for cloned genes and the encoded proteins (
25).
Whether transgenic animals or plants, transgenic silkworms themselves,
Escherichia coli, or other hosts will become the most efficient system remains to be determined. Regardless, issues of secretion, stabilization of large modular gene structures, and determining metabolic demands during production still need to be resolved. These insights will not only drive improvements in silk expression but also provide important clues for the broader biotechnology community for protein expression needs. The role of processing, in combination with sequence chemistry, will play a critical role in silk material properties. Improved processing approaches for solutions and fibers, such as postspinning draw, water annealing, and the infusion of metals, can help to achieve outcomes that mimic native silks (
26–
28).
The publisher's final edited version of this article is available at 
