Question

Please read the article below, (Spinning of Spider Silk) and write a critical summary report of your reading.

The ability to make silk has made spiders successful; over

45,000 species have been characterized (http://www.wsc.

nmbe.ch/), and they can be found in most habitats on Earth1.

Spider silk fibers are formed in a fraction of a second from highly

concentrated protein solutions (known as ‘dope’) in sophisticated

spinning apparatuses. A female spider can spin up to seven types

of silk that are used for a variety of purposes, ranging from prey

capture to reproduction (Fig. 1a). All silks are composed of proteins

that typically consist of an extensive repetitive central part flanked

by smaller nonrepetitive domains (Fig. 1b).

Spider silk is an astonishing material that shows a combination

of high tensile strength and extensibility, which allows some

of the fibers to absorb more energy per weight than the strongest

of man-made materials, for example, Kevlar (Table 1)2. More

surprising, considering the biological functions of spider silks

(Fig. 1a), is the fact that the fibers are well tolerated when

implanted; for example, they have successfully been used for

regeneration of peripheral nerves in vivo3,4. These features make

spider silk one of the most fascinating materials known, and if

we can learn how to produce it on a large scale there will be

numerous possibilities for its application, from the construction

industry to medicine.

Making artificial spider silk fibers with mechanical properties

similar to the natural material, however, has been a major unmet

goal in material science for decades. What are the reasons for this

shortcoming? Large partial spidroins can be produced in heterologous

hosts5, but they, like many short recombinant spidroins6–8,

require the use of denaturing conditions during purification and/or

fiber formation, which probably explains why most fibers show disappointing

mechanical properties (Supplementary Table 1). How

to keep the spidroins correctly folded during production and soluble

at high concentrations without the use of organic solvents or other

chaotropic agents and how to make them form silk fibers without

the use of nonphysiological coagulation baths are major issues that

remain to be solved. Moreover, problems in making fibers with

mechanical properties similar to those of the natural material may

also arise due to the fact that the exact chemicophysical conditions

of the glands are unknown and that the rheological properties of

the dope and physical stresses are difficult to reproduce in vitro.

Recent advances in the understanding of the natural silk spinning

process9,10, in combination with detailed molecular studies of

partial spider silk proteins9,11–15, have revealed that spider silk

formation is governed by mechanisms that have not yet been

observed in any other physiological system. Scientific evidence

points to the importance of the terminal domains in the regulation

of spidroin assembly into fibers9,11–13,15,16 and suggests that current

spinning procedures are inadequate because they rely on nonphysiological

conditions and/or recombinant spidroins that lack one or

both of the terminal domains. Herein, we review recent advances

about the chemical biology of the natural spinning process, from

which we outline new ways to design optimal protein constructs

and biomimetic spinning devices.

Spidroins are conserved but also highly diverse

All spiders spin silk, and some can spin up to seven different types

of silks or glues (for example, female orb-weaving spiders)17 (Fig. 1a

and Table 1). The dragline silk is most studied and is well known for

its high tensile strength, extreme toughness, light weight and favorable

properties when implanted in living tissue3,4. This fiber is used

to make the framework of the web and is used as a safety line during

falls. The flagelliform silk fiber is coated with glue (aggregate silk)

and forms the extremely extensible capture spiral of the orb web.

The minor ampullate silk is used to reinforce the web. For making

the egg sac, female spiders use two types of silks: cylindriform (also

referred to as tubuliform) silk forms the durable outer shell of the

sack, and aciniform silk is used as a soft inner lining. Aciniform silk

is also used for wrapping prey. To attach silk fibers to a surface, the

spider uses pyriform silk17.

The different mechanical properties of the silks make them

attractive for a variety of applications, from superglues to extremely

strong and extendible ropes and fibers. The inherent differences of

the silks also open up the possibility of designing composites with

combined features for advanced technical and medical applications,

for example, exact matching of the mechanical properties of a specific

tissue for tissue engineering applications.

Each silk type is made in a specific gland (Fig. 1a) and mainlyconsists of spidroins whose repetitive motif has a characteristic primary

structure. However, this general view is complicated by the following

factors: the presence of variable repetitive motifs within the

same spidroin from different species18, non-gland-specific expression

of some spidroins19,20 and the facts that spidroins can even be

expressed in venom glands21 and that some spider silks containThough the repetitive parts and mechanical

properties vary a lot between silk types2,

the regulatory domains (NT and CT)11,12 and

the conditions of the silk glands9 are highly

conserved (details below), which means that

they have important roles in the regulation of

fiber formation. Conditions that allow artificial

spinning procedures are therefore likely

to be found in the natural counterpart and

should be independent of what specific type of

silk is to be made.

How do the silk glands work?

Because most data available are from experiments

on major ampullate glands and their

product, the dragline silk, we will focus on this

particular silk. The major ampullate gland has

a long, winding and narrow tail and a wider

and shorter ampulla or sac (Fig. 2a). The tail

and sac contain three different single layered

epithelial cell types that all contain numerousgranules. The epithelial cell types have distinct

distributions and thereby divide the gland into

three zones, A–C10 (Fig. 2a). The cells in the

A and B zones produce spidroins that form

two separate layers in the dope10. Surprisingly,

the C-zone epithelium does not produce spidroins

but was recently found to be important

for the physiological regulation of silk production

(described below)9. The S-shaped narrowing

duct is lined by a cuticular intima layer

that probably contributes structural support to

the duct and protects the underlying epithelial

cells from the fiber, but it has also been suggested

to act similarly to a hollow fiber dialysis

membrane, thereby allowing dehydration of

the dope25. The sac is connected to the duct via

the funnel, which has an unknown function

but appears to be associated to the cuticular

intima of the duct10.

The making of spider silk poses challenges

at molecular biological, biochemical

and physiological levels. The spidroin genes

are large, repetitive and very GC-rich, as glycine and alanine can

constitute more than 60% of spidroins26 and are coded for by thenonspidroin proteins22. Spidroin genes and proteins are most

commonly abbreviated by two letters indicating the gland where

they are mainly expressed, followed by Sp for spidroin and a number

referring to the different paralogs (for example, MaSp1 for Major

ampullate spidroin 1).

Spidroins vary in size21 but generally share a tripartite composition

of a nonrepetitive globular NT (~130 residues23), an extensive

region made up of silk-specific alanine- and/or glycine-rich repeat

units and a nonrepetitive globular CT of ~110 residues (Fig. 1b)12.

The NT and CT are evolutionarily conserved and have important

roles in the regulation of spider silk formation (described in detail

below). The spidroin repetitive parts, in contrast, show great variability

between silk types. The repetitive part is usually large and

can encompass up to hundreds of repeat units. The nature of the

repeat units correlates with the mechanical properties of the fiber;

for example, long polyalanine segments give strong fibers, and long

glycine-rich segments give more extendible fibers24. However, there

is no linear correlation between protein length and mechanical

properties5. In fact, the extensive arrays of almost-identical repeats

may be inadvertent consequences of unequal crossing over and

homologous recombination events during replication18, which

would implicate that shorter and less repetitive spidroins may be

useful for artificial production of spider silk.codons GGX and GCX. This puts high demands on spidroin gene

replication and transcription26–28. To meet the high need of glycine

and alanine during translation, the epithelial cells of the glands have

unusually large alanyl- and glycyl-tRNA pools29. Despite the spidroins

being able to assemble into a solid fiber within a fraction of a

second, the spider manages to keep them soluble at very high concentrations

(30–50%, w/v) for long-term storage30,31. This implies

that the tertiary and quaternary structures of the soluble spidroins

are optimized to prevent aggregation. There are currently two different

models for how spidroins are organized when stored in the

gland: in micelles, where the terminal domains form a hydrophilic

outer shell and the repetitive region is shielded in the center32, or as

a liquid crystalline feedstock33. These two models are not mutually

exclusive and may both occur. The structures of soluble native

spidroins are difficult to determine as the proteins are prone to

change their conformation upon being manipulated as required to

experimentally study the gland content. In spite of this, by using

13C NMR on whole major ampullate glands, the repetitive part of

the proteins was found to adopt a random34 and/or helical conformation31.

Polyproline type II segments have also been observed by

vibrational CD spectroscopy35. In the final silk fiber, the repetitive

part has converted to polyalanine β-sheet crystals flanked by more

amorphous glycine-rich repeats36. Although often referred to as

amorphous, the glycine-rich regions in the fiber contain 31-helical

Gly-Gly-X motifs37–40 and Gly-Pro-Gly-X-X motifs that form type II

β-turns41. Finally, the fiber-forming process enables fiber formation

in a defined segment of the duct, thus avoiding the fatal spread of

the assembly process to the dope in the gland.

Understanding how silk formation is regulated requires determination

of how the composition of silk dope varies throughoutthe gland, but the small size of the gland makes this technically

challenging. The duct is particularly difficult to work with experimentally

owing to the narrow lumen and tough cuticular intima,

which is hard to penetrate. Despite these difficulties, the pH from

the proximal part of the tail to halfway through the spinning duct

was recently determined to span from 7.6 to at least 5.7 (ref. 9). To

uphold this broad and steep pH gradient, the spiders must have

developed efficient methods to generate protons, and the almostnonexistent

variation in pH between individual silk glands9 suggeststhat the underlying mechanisms are tightly controlled. By using

enzyme activity staining of histological sections, we recently discovered

that carbonic anhydrase is responsible for generating and

maintaining the pH gradient9. Carbonic anhydrases are ubiquitous

enzymes found in all animals and photosynthesizing organisms42

and catalyze the following chemical reaction with an exceptionally

high turnover rate of up to 106 s−1 (ref. 42):

CO2 + H2O↔H+ +HCO3−

In the silk gland, active carbonic anhydrase is present from zone

C to the end of the duct (Fig. 2a,b). The pH decreases along the gland

(Fig. 2c), and, surprisingly, the concentrations of both HCO3

− and

CO2 rise9. The hydrophobic nature of CO2 implies it should diffusefreely across cell membranes, but the spider glands have apparently

developed a mechanism to uphold a high intraluminal pCO2

(ref. 9). A similar phenomenon has been described for parietal cells

in the ventricle43, but the underlying mechanisms are poorly investigated.

The high pCO2 in the silk gland is probably of physiological

importance, as in vitro studies show that acidic pH in combination

with a high CO2 concentration leads to marked structural changes

of the terminal domains that can trigger fiber formation (described

further below). In the sac, the concentrations of Na+, K+ and Cl−

are 199 mM, 6 mM and 164 mM, respectively9. The intraluminal

concentrations of Na+ and Cl− are slightly higher than they are in the

hemolymph (about 140 mM and 125 mM, respectively, though these

values were measured in two different orb-weaving spider species44),

whereas the silk gland K+ concentration is lower (about 20 mM in

hemolymph44). The concentrations of these ions have been suggested

to change along the silk production pathway45, but the concentrations

in the duct and hence at the site of fiber formation remain

unknown.

Coordinated molecular events govern silk formation

The terminal domains of spider silk proteins act as regulatory elements

that control spidroin solubility and assembly9,11,12. They are

structurally conserved, unique to spiders and present in almost

all spidroins, even though spidroins diverged in evolution several

hundred million years ago23,46. Both domains are bundles of five

α-helices, but they do not share any primary or tertiary structure

similarities, which indicates that they fulfill different functions.

CT is a constitutive, often disulfide-linked homodimer9,12,47 and

thereby probably links the spidroins in pairs already in the oxidizing

environment of the endoplasmic reticulum of the glands’ epithelial

cells. At pH levels greater than 6.5, as in the lumen of the synthesis

and storage parts of the gland (zones A and B; Fig. 2a–c), the CT is

stable and highly soluble9,47. When the dope travels down the duct, the

decreased pH and simultaneously increased pCO2 markedly affect

the CT’s structure and stability9. The low pH destabilizes the CTSimultaneously, CO2 interacts with partly buried

residues in CT, which can facilitate unfolding9.

The physiological use of CO2 to destabilize a

protein has not been described before and may

even be a unique feature of the spiders. At pH

levels less than 5.5, the CT loses its native helical

structure and forms β-sheet fibrils. This

behavior resembles that seen in the formation

of amyloid fibrils, where proteins that have lost

(or never gained) their native conformation

instead adopt a very regular structure of cross-

β-sheets (amyloid fibrils)48–50. The kinetics of

amyloid fibril formation are greatly acceleratedby the addition of preformed fibrils, i.e., seeds (nuclei)51. The spiders

apparently have adopted the seeding phenomenon to trigger ultrafast

fiber formation. The structural transition of CT into β-sheet

fibrils in the duct of the silk gland results in nuclei that may trigger

the conversion of the repetitive region into β-sheet polymers, a new

and fascinating functional use of amyloid fibrils9 (Fig. 2e,f).

The NT is the most conserved part of spidroins and confers solubility

to recombinant spidroins at pH 7 and rapid fiber formation

when the pH is lowered to 6 (refs. 11,23). It is mainly monomeric at

pH levels greater than 7 (refs. 16,52–54) but dimerizes when the pH

is lowered below ~6.4 (the exact pKa of dimerization depends on the

salt concentration)13,16,52. A relocation of helix 3 closer to helix 1 is a

prerequisite for the formation of the NT dimer interface53. The residues

responsible for sensing pH were recently identified for MaSp1

NT from the Euprosthenops australis spider13. During storage in the

proximal sac (pH ~7.5; Fig. 2), the NT is mainly monomeric, but

the dipolar subunits can interact by long-range electrostatic interactions

(Fig. 3). Dimerization is not possible at this stage because the

tilting of helices 3 and 5 sterically hinders close subunit interactions.

When the pH is lowered to around 6.5, as seen in the most distal

part of the sac near the funnel (Fig. 2), two conserved glutamic

acid residues will each pick up one proton. The loss of these two

negative charges allows the rearrangement of mainly helices 3 and 5

and a transition into a subunit conformation that is compatible

with dimerization53. To accomplish the structural transition, a tryptophan

residue (or phenylalanine in some NTs) will leave its wedged

interhelical position in the monomer and swing out to a position

where its side chain is more solvent exposed. This allows helices

1 and 3 to be more tightly packed and leads to the formation of a

rather flat dimer interface (Fig. 3). Although NT is dimeric at this

stage, it is not until the pH reaches 5.7 and below, correspondingto the pH halfway through the duct and beyond, that a structurally

defined and fully stable NT dimer is formed by protonation of a

third glutamic acid residue (Figs. 2d,e and 3)13. Some of the residues

that control NT dimerization and stabilization in E. australis MaSp1

are replaced by nontitratable residues in other spidroins, but all NTs

appear to harbor acidic residues at the dimer interface11,46,55. It seems

that the NT has conserved the ability to control dimerization and

stabilization in response to a pH gradient, but that the exact residues

that titrate during this process are variable is a supposition that

must be verified by analyses of wild-type and site-directed mutants

of NTs from additional spidroins.

The dimerization and stabilization process of NT may seem

overly complicated, but the multistep mechanism is probably vital

for the control of silk polymerization: the silk fiber is pulled from the

spider, and pulling forces can be propagated via the protein chains

when they are firmly interconnected via the stable NT dimers and

constitutive CT dimers (Fig. 2f). The pulling, as such, may promote

refolding of helical or random repetitive segments into extended

β-sheet conformations. As the spidroins flow through the duct, the

pulling also causes shearing, which has been shown to contribute to

the structural transition of the CT into β-sheet nuclei9,12, and this

transition can be further accelerated by lowered pH and elevatedleading to rapid polymerization of the spidroins.

Fast polymerization kinetics are a prerequisite for the spiders’ ability

to spin silk at speeds >1 m s−1, but for the spider it is also vital to

ensure that the polymerization process is confined to the duct and

prevented from spreading up to the sac, as this would prematurely

coagulate the contents of the gland. The dynamically associated NT

dimers present in the beginning of the duct apparently provide the

solution to both these problems: they ensure prealignment of the

NTs so that the interlocking of the silk proteins in the distal parts of

the duct is independent of diffusion (i.e., their association is ultrafast)

15, and, at the same time, they act as a safety mechanism that

keeps the pulling forces from propagating up to the gland Shortcomings of current methods to spin artificial silk

The insights into the ion and pH gradients of the silk gland and

importance of the terminal domains as regulators of spider silk

assembly have not yet been translated into a biomimetic spinning

process. As outlined in Supplementary Table 1, several recombinant

spidroin variants have been produced and spun or self-assembled

into fibers. The nature of the produced proteins varies greatly, and

none correspond to a full-length natural spidroin. Typically they are

much smaller than their natural counterparts and are composed of a

repetitive part with or without a CT. There is only one reported case

of fibers made from a recombinant spider silk protein that includes

all three parts of a canonical spidroin—the NT, a repetitive part and

the CT—but their mechanical properties were not analyzed11.

Most techniques used for fiber formation require a highly concentrated

spinning dope, typically 10–20% (w/v), and to achieve

this, the proteins are commonly dissolved in organic solvents such

as hexafluoroisopropanol. Notably, even when harsh solvents are

used, the solubility of the recombinant spidroins does not match

that seen in the native dope56.

The techniques used for fiber spinning include wet spinning

(ejection of the dope into a coagulation bath, most often containing

methanol or isopropanol)5, electrospinning57,58, self-assembly at airwater

interfaces59,60 and spinning in microfluidic devices8. Organic

solvents or coagulation baths are used to obtain both electrospun

(nonwoven) silk fibers and wet spun fibers. Fiber formation through

denaturation and aggregation will probably not lead to truly silk-like

replicas as the spidroins’ functions are lost. The natural spinning

process involves intricate molecular mechanisms and results in fibers

with specific secondary structures, i.e., the spidroins are assembled

into fibers, not aggregated. Self-assembly of spider silk proteins intometer-long fibers under nondenaturing conditions can be achieved

in tubes that are tilted from side to side, which induces shear forces

in the dope59 (Fig. 4). However, the method is difficult to scale

up, ion and pH gradients are not easily achieved, and the fibers

are variable in structure and mechanical properties. Interestingly,

though, by including the NT in the recombinant spidroins, the rate

of the fiber-forming process can be controlled by pH11. Microfluidic

spinning is perhaps the most attractive fiber spinning method as

it has successfully been used to obtain fibers without the use of

denaturing steps and can be used to create ion and pH gradients

as well as shear forces. Fiber formation of a MaSp2 analog has been

achieved by applying elongational flow, dropping the pH to 6 and

increasing the K+ concentration to 500 mM8. Apart from the high K+

concentration, this approach represents the most biomimetic spinning

procedure presented so far. The resulting fibers were not tested

for mechanical properties, and their structural properties were not

described, making it difficult to evaluate whether this particular

biomimetic spinning technique gives rise to functional fibers. Observed

effects of K+ on spidroin aggregation are not unequivocal; native spidroin

dope has been shown to aggregate into nanofibrils in response

to increased potassium concentration30, whereas a recombinantspidroin composed of repetitive segments and CT did not respond

to potassium concentrations up to 300 mM (ref. 61).

Taken together, much effort has been put into spinning artificial

spider silk fibers, but few or none of the produced fibers are usable for

practical purposes owing to poor mechanical performance, low reproducibility

or both. We believe that these shortcomings are most likely

due to insufficient mimicking of the natural spinning processes.

How to spin biomimetic spider silk?

To establish a bimimetic spinning technique, the first problem to

solve is how to obtain a stable aqueous solution of recombinant spider

silk proteins that matches the native solution (30–50% w/w), and

the second is how to mimic the conditions of the spider’s spinning

apparatus. Microfluidics or similar methods appear well suited for

mimicking the gradients of the silk gland and are therefore highly

interesting for future development of the spinning technique (Fig. 5).

In such devices, laminar flow may be created, allowing for diffusion

across the liquid interfaces and thus a gradual change of the conditions

in the spinning dope as it travels through the spinning apparatus.

Furthermore, increased flow rates generate shear forces, another

important factor needed for biomimetic spinning. A lowering of pH

from around 7.5 to 5.0–5.5 will be required, and increasing pCO2 will

most likely facilitate the nucleation of the polymerization process.

An increasing concentration of K+ could lead to faster polymerization

of the spidroins. For this biomimetic spinning device to have

the intended impact on the spidroins to be spun, the spidroins

must encompass both terminal domains. Including the terminal

domains in the recombinant spidroins could actually also solve the

first problem, as both the NT and the CT, at least in certain species,

are very stable and can be concentrated to around 300 mg ml−1 inaqueous buffers11,47. However, when these domains are coupled to

repeats from the minor ampullate spidroin, the solubility markedly

decreases47. This is surprising as the repetitive segments of spidroins

are generally not hydrophobic or very aggregation prone as such. For

example, alanine has the highest α-helix propensity of all of the residues62,

and this in combination with its biological hydrophilicity63

suggest that polyalanine segments are unsuited to spontaneously

aggregate into crystalline β-sheets. Along the same line, the repetitive

segments of the aciniform spidroins are α-helical and highly

soluble and form globular, folded domains that in the fiber are at

least partially transformed into β-sheet conformations64,65. We propose

that a main feature of the repetitive segments is to be soluble

to avoid premature aggregation of the spidroins and that the uniqueNT and CT have evolved to allow rapid conversion of the repetitive

segments into β-sheet aggregates at a confined place in the spinning

duct. If this hypothesis is correct, it implies that the exact sequence

of the repetitive segments may be less important than their

solubility, suggesting that recombinant spidroin analogs designed

from simple repeat units capped by NT and CT could be used for

production of silk-based biomaterials.

Future perspectives

During the last few years, important progress has been made toward

the accomplishment of generating artificial spider silk mimics: fulllength

spidroins have been characterized, two spider genomes have

been sequenced, and the milieu in different parts of silk glands

and how it affects spidroin domains at a molecular level have been

revealed. Hopefully, these advances in basic knowledge will allow

the design of proteins, methods and devices that can be used to produce

biomimetic spider silk for various purposes in the near future.

A challenge that remains is how to keep recombinant spidroins

soluble at high concentrations without the use of harsh solvents. We

also need to understand, in detail, the unprecedented effects of CO2

on spidroins to be able to harness them for biomimetic spinning.

Finally, we envision customized microfluidic spinning devices; for example, carbonic anhydrase, an efficient and robust enzyme, could

be introduced in a spinning apparatus to achieve a mimic of the

glands’ physiology.