Nanofibre production in spiders without electric charge

We all have experienced the effects of rubbing a plastic balloon against our hair: the single hairs are repelled from each other, but attracted towards the balloon. This is induced by Coulomb forces, also called electrostatic forces, which are caused by frictional charging in this example (triboelectrification). Even without friction, electrons drift from one material to the other whenever two surfaces are in contact, and after separation, the materials become oppositely charged. Most materials are easily discharged again by protons and electrons occurring in the air. The final charging of an object hence depends not only on the physical and chemical nature of the given material but also on the conductivity of the medium surrounding the object (Chang et al., 1995).

Many biological systems use the attraction between oppositely charged objects: this charging facilitates, for example, the landing of pollen on the stigma during pollination by wind (Bowker and Crenshaw, 2007) and also attracts the gluey capture threads of spiders to positively charged flying insects (Ortega-Jimenez and Dudley, 2013). Another type of capture thread, the cribellate capture thread, is even suggested to employ electrostatic forces not for the attraction of prey but for the formation of its own structure, using the repulsion between two identically charged objects (Opell, 1995a; Joel et al., 2015; Kronenberger and Vollrath, 2015). Prey is captured here by a combination of hygroscopic forces, van der Waals' forces and an entanglement of the prey in a wool-like mat of nanofibres (cribellate fibres) surrounding two larger stabilizing axial fibres (Opell, 1994; Hawthorn and Opell, 2003).

The cribellate fibres shape the outer structure of the capture thread with two characteristic regions: the puff with a larger diameter than the intermediate zones, clearly separating two puffs with a constriction (Fig. 1) (Peters, 1984; Opell, 1989; Joel et al., 2015). It is commonly assumed that the adhesive cribellate fibres are kept separate within the puff by being charged uniformly during their extraction, leading to a repulsion of the fibres (Peters, 1984; Sahni et al., 2011; Kronenberger and Vollrath, 2015). During the production of the thread, the nanofibres are extracted from the cribellum, a spinning plate with up to 40,000 spigots anterior to the spinnerets (Bertkau, 1882; Foelix and Jung, 1978). To assemble all these fibres to one functional thread, the spiders deploy a highly sophisticated movement of the spinnerets as well as the fourth pair of legs, finally wrapping the sheet of nanofibres around the axial fibres (Joel et al., 2015, 2016). On the metatarsus of the fourth leg, cribellate spiders bear a comb-like row of specialized setae, the calamistrum. It is assumed that the charge is transferred to the cribellate fibres by triboelectrification when they are brushed over a continuous area composed of the specialized setae of the calamistrum (Fig. 1B) (Peters, 1984; Kronenberger and Vollrath, 2015; Joel et al., 2016).

This kind of nanofibre production involving only low electric charge has attracted the interest of researchers, as technical nanofibres are typically produced by electrospinning, involving high tension (Teo and Ramakrishna, 2006; Kronenberger and Vollrath, 2015). From the biological perspective, evidence for a charging of cribellate fibres by the action of the calamistrum is missing so far. However, there are indications that electrostatic charges are involved in the formation of the puffy structure: high humidity can increase the conductivity of air and hence remove charge. Indeed, the thread loses its characteristic structure after being exposed to the fine mist of a nebulizer (Zheng et al., 2010; Elettro et al., 2015). Furthermore, spiders without calamistra produce capture threads without puffs (Joel et al., 2015). This could indicate missing electrostatic charges to keep the fibres separate, though the use of a nebulizer, i.e. fine droplets of water, could also destroy the structure of the thread as a result of capillary forces or a reaction with the spider silk. The aim of this study was hence to resolve the involvement of Coulomb forces in the nanofibre production of cribellate spiders.

There are several ways to determine whether cribellate nanofibres are indeed electrostatically charged and whether such charging helps formation of a puffy structure: (1) one can try to regenerate the puffy structure of threads by charging them; (2) one can remove any charge by (i) making the surrounding conductive or (ii) making the fibres conductive; and (3) one can determine whether they are attracted to or repelled from differently charged objects. On the basis of these three requirements, we aimed to determine whether electrostatic forces are deployed during the formation of the cribellate thread. All experiments were performed with Uloborus plumipes (Uloboridae) whose capture thread formation has been best described to date (Fig. 1A). To evaluate the general validity of the findings, threads of another Uloboridae, Zosis geniculata, as well as the distantly related Amaurobiidae Amaurobius ferox, Desidae Badumna longinqua and the Filistatidae Kukulcania hibernalis were examined (Bond et al., 2014; Fernández et al., 2014; Garrison et al., 2016). Capture threads of these species differ greatly in shape from those of Uloboridae (Fig. 2). Comparable results would indicate a conserved cribellate fibre production system in all cribellate spiders, whereas differences would hint at a diverging mechanism involved in the formation of cribellate capture threads.

The species used in the experiments (Amaurobius ferox, Kukulcania hibernalis, Uloborus plumipes and Zosis geniculata) are not endangered or protected species and special permits were not required. All applicable international, national and institutional guidelines for the care and use of animals were followed.

Adult female Uloborus plumipes Lucas 1846 and Zosis geniculata (Olivier 1789) were raised in larger terraria shared by several spiders of the same species as a lab colony under room temperature (∼21°C), room humidity (∼30%) and northern European diurnal rhythm. Each spider was able to build a web of its own. Once a week, spiders were fed with Drosophila melanogaster or juvenile Acheta domestica. Water was provided once or twice per month by sprinkling the web. Threads of such sprinkled webs were not used for further research.

Kukulcania hibernalis (Hentz 1842) were raised separately in 1 litre containers covered with gauze under elevated room temperature (∼26°C), normal room humidity and northern European diurnal rhythm. Once a week, spiders were fed with juvenile A. domestica or bean weevils. Water was provided once or twice per month by applying droplets near the burrow.

Thread samples of Amaurobius ferox were collected near the Institute of High Frequency Technology in Aachen (Germany). Thread samples of Badumna longinqua were provided by the Queensland Museum (Brisbane, Australia). Samples were stored in the dark and protected against dust in a box at room temperature and room humidity. Because of their origin, threads of B. longinqua were stored for a longer period of time (about a year) until use, whereas A. ferox threads were stored for no more than 2 weeks after collection. These samples were used only for observation in the scanning electron microscope, where we can exclude distorted thread behaviour due to the age of the threads.

Thread samples were taken from the web by picking them up gently with two strips of conductive foil on a sample holder. Special care was taken not to stretch the sample with this procedure. The samples could be observed without any further preparation (native), or after coating with carbon or a 10 nm gold layer (Hummer, Technics Inc., Alexandria, VA, USA) before examination in a scanning electron microscope (SEM 525 M, Philips AG, Amsterdam, The Netherlands).

Threads of U. plumipes without the typical puffy structure were taken as samples for these experiments. This lack of puffs can be achieved by either exposing normal threads to fine mist (see ‘Reaction to high humidity’, below) or taking the capture threads of U. plumipes from which calamistra were previously removed according to Joel et al. (2015). Samples were transferred into the SEM without any previous treatment. Each sample was exposed directly to the electron beam (15 kV and a spot size of 30 nm) for 5–15 min to observe thread structure and potential changes in the structure. Extensive exposure to an electron beam might damage nanofibres, but no indications of such damage were observed.

Thread samples of U. plumipes were picked up with two arms of a metal filament and thread shape, i.e. puff width, was initially controlled with light microscopy. Afterwards the thread was placed in a climate chamber with a constant temperature of 28°C. Humidity was raised to 80% either by letting tap water evaporate or by controlling humidity with the help of a nebulizer (Super Fog Nano, Lucky Reptile, Import Export Peter Hoch GmbH, Waldkirch, Germany). Because of the size of the climate chamber (volume about 10 l), the impact of the nebulizer is omnipresent. After an hour, the samples were taken from the climate chamber and again measured under a light microscope. To observe whether thread shape changes after drying the thread again, samples were stored in a desiccator for a week, regularly monitoring the shape of the threads.

To observe the influence of high humidity on the production of the threads, U. plumipes were taken from the terrarium and placed overnight in a climate chamber at room temperature but 95% humidity, kept high by the evaporation of tap water. The next morning, samples of freshly spun capture threads were taken as described above and examined with the SEM.

Control threads of U. plumipes and K. hibernalis were taken from three webs of each species and their configuration characterized with a video microscope (VW 9000C, Keyence Corporation, Osaka, Japan) whilst ionizing the surrounding air with the help of Milty Zerostat 3 (Armour Home Electronics Ltd, Bishop's Stortford, UK). This anti-static device prevents the formation of static charges and is used, for example, to protect sensitive electronic devices during handling.

To observe the influence of ionizing the air during the production of threads, the diurnal rhythm of one U. plumipes was shifted 12 h by turning white light on from 06:00 h to 07:00 h. For observation, the spider was illuminated from 07:00 h to 18:00 h using red LED (Paulmann Licht GmbH, Springe, Germany). During production of the web, the air surrounding the spider was ionized. After the spider had finished building the web, thread samples were taken and measured. These experiments were performed under room humidity.

If the puffs are established by a repulsion of the nanofibres, the removal of any charge should lead to the collapse of this structure, producing a uniformly structured thread with a constant diameter, resembling the intermediate zone. Based on this premise, we performed a Power Analysis (G*Power version 3.1.9.2) to calculate how large our sample has to be to get reliable data with a power above 0.95. Because of the large difference between the diameter of the puff and the diameter of the intermediate zone as well as the low standard deviation of the data, achieved by only taking threads of adult spiders, two samples per experiment would be enough to determine any significant differences (assumed if P<0.05) using a two-tailed t-test with a power of 0.999. Nevertheless, if not indicated otherwise, we calculated the mean between three spiders, with four puffs per spider measured. p_puff refers to the comparison between the measured diameter of the sample and the typical diameter of a native puff, whereas p_iz refers to the comparison between the measured diameter of the sample and the typical diameter of the intermediate zone of a native thread.

Whole webs (about 230 m²) of U. plumipes and K. hibernalis were taken as a sample. A positively charged glass rod or a negatively charged piece of foamed plastic was brought near to single capture threads without touching them (the distance was always >0.5 cm). Because of the elasticity of these threads, any deformation could be easily detected with the naked eye (>1 cm). Please note that Movie 1 was recorded with a single thread and not a whole web and hence deflection is not as pronounced.

When feeding U. plumipes and Z. geniculata with live D. melanogaster, we observed that cribellate capture threads attract flying fruit flies, copying the effect described for gluey capture threads (Movie 1). Because flying insects charge themselves positively, the potential charge leading to a repulsion between the cribellate nanofibres should be negative (Ortega-Jimenez and Dudley, 2013). Hence, if only electrostatic forces are keeping the cribellate fibres apart, the exposure of threads without a puffy structure to the electron beam of a SEM should lead to recovery of the structural features.

Threads of spiders with previously removed calamistra or threads after exposure to fine mist both lack the puffy structure. Taking these threads as samples and charging them negatively with an electron beam, however, did not lead to a recovery of the puffy structure (n≥3). The puffy structure of cribellate capture threads of U. plumipes cannot be retroactively generated by negatively charging the fibres after thread production.

If the protruding structure of cribellate threads was only maintained by a repulsion of the nanofibres, the removal of charge should lead to a collapse of the puffy structure, producing threads of uniform structure with a constant diameter of the intermediate zone (in case of U. plumipes: 65±10 μm, n=6). To test whether cribellate fibres are indeed repelled by one another, the fibres were made conductive by coating the complete thread with either gold or carbon. Both coatings are typically used to make biological samples electrically conductive for electron microscopy. Threads changed shape after being coated with gold: untreated thread of U. plumipes had a puffy structure resembling beads on a string with a puff diameter of 168 μm (Fig. 1D, Table 1) but gold coating led to a significant collapse and a more tooth-shaped structure of the puffs (Fig. 1C, Table 1). In contrast, threads coated with carbon had a structure resembling that of the native state (Fig. 1E, Table 1). This collapsing of the thread after gold coating, but not after carbon coating, was reproducible for the Uloboridae Z. geniculata as well as for the Amaurobiidae A. ferox, the Desidae B. longinqua and the Filistatidae K. hibernalis (Fig. 2). Hence, making the cribellate fibres conductive alone does not lead to a change of thread shape.

This experiment does not exclude the possibility that electrostatic forces are deployed to form the structure in the first place, i.e. during extraction of the fibres. Therefore, the surroundings of the spider have to be made conductive during thread production. Raising the humidity to 80% with the help of a nebulizer led to the collapse of the puffs, which was not reversible by drying the threads (U. plumipes; Table 1). However, we found that the cribellate threads lost their puffy structure only when raising the humidity with the help of a nebulizer, not when letting water simply evaporate, despite reaching the same humidity (U. plumipes; Fig. 3, Table 1). Furthermore, the threads did not change their structure when ionizing the surrounding air [tested for K. hibernalis (n=3) and U. plumipes; Table 1].

Using this knowledge, we placed U. plumipes in a climate chamber with a humidity of 95% or ionized the air during thread production and hence made the air conductive. Uloborus plumipes was indeed able to build the same structured capture threads at room humidity (∼30%) and in a climate chamber with elevated humidity (Table 1). The same was true for threads produced by a spider in ionized air (Table 1, Fig. 4). Hence, reducing the spider's ability to electrostatically charge fibres in the first place has no impact on cribellate thread production.

Calculating the forces needed to keep two fibres separate, we found these to be rather small (0.01–0.3 elementary charges per μm; for details concerning the calculation, see the  Appendix). Assuming the nanofibres are indeed charged with just 0.01–0.3 elementary charges per μm, removal of this low charge is not simple. Our further mathematical calculations revealed, however, that nanofibres with such a low charge do accumulate in a denser outer layer as observed in reality (Figs S1 and S2); however, single fibres are also uniformly distributed within the cross-section of the capture thread. But this does not fit the observed structural assembly of the capture thread, building a hollow structure around the two axial fibres (Joel et al., 2015, 2016). Any employment of Coulomb forces to generate the structure of the cribellate thread is therefore very unlikely.

Finally, to survey whether cribellate capture threads are charged at all, we approached webs of U. plumipes and K. hibernalis with a negatively charged piece of plastic or a positively charged glass rod. However, capture threads were attracted to both and the deflection could be seen with the naked eye (Movie 1). This demonstrates that the thread as a whole structure is not charged. In addition, the cribellate thread behaves like a dipole.

Although it is commonly assumed that cribellate fibres are electrostatically charged, our data do not support this hypothesis. Reviewing three previously proposed requirements to identify the impact of electrostatic forces on the capture thread structure, we could neither (1) regenerate the puffy structure of threads lacking puffs by charging them, nor (2) detect any structural change of the capture thread by making either the thread or its surrounding conductive (Fig. 4). Hence, the puff neither can be retroactively generated nor is preserved by electrostatic forces. Instead, when investigating our third condition, we found that the cribellate capture thread indeed acts as a dipole and is attracted to both positively and negatively charged objects. As we were not able to detect any differences between the threads of distantly related species, we suggest no cribellate spider deploys Coulomb forces to establish the structure of its capture thread.

Opell (1993) suggested that no electrostatic forces are involved in the formation of the puffy structure, because he likewise observed no collapsing of puffs when raising humidity only by evaporation. This exclusion of electrostatic forces in the formation of capture thread by cribellate spiders fits the described occurrence of cribellate spiders not only in deserts (like Uloborus diversus) but also in humid areas with 60–90% relative humidity (like U. plumipes, Octonoba sinensis, Stegodyphus pacificus or Waitkera waitakerensis) (Eberhard, 1971, 1972; Kullmann et al., 1971; Kumhof et al., 1992; Opell and Bond, 2000). In habitats with high humidity, fibres would more easily be discharged again. The same would be true for the spider, which should be oppositely charged after capture thread production. In contrast, in arid areas, not only the fibres but also the spiders are, and stay, more easily charged. Without any special behaviour to discharge themselves again, the spiders would accumulate electric charge during thread production and afterwards attract their own capture threads. Because neither a compensatory behaviour nor an attraction between spiders and their capture threads was observed, we conclude that cribellate spiders cannot employ permanent larger charge through triboelectrification during capture thread production. Although we calculated that extremely low charges are sufficient to repel two nanofibres, the simulation of their arrangement within the thread only as a result of repulsion does not fit the characterized structure of the cribellate capture thread (Joel et al., 2015, 2016). Hence, the employment of any electrostatic charge to establish the thread's structure can be excluded. Being charged could actually counteract the purpose of the capture threads as insects can sense electric fields (Clarke et al., 2013; Greggers et al., 2013; Sutton et al., 2016). Hence, a charged capture thread would be detected and avoided by potential prey. The observed dipole behaviour of cribellate capture threads instead benefits the capturing of prey, because, like ecribellate capture threads, cribellate capture threads attract approaching flying charged prey (Ortega-Jimenez and Dudley, 2013).

Two experimental setups nevertheless led to the collapse of the puffs: coating them with gold and contact with fine mist. Both results are more difficult to explain when excluding the involvement of electrostatic forces in structure formation. Collapse of puffs after coating threads with gold not only is visible for the four species studied here but can also be found in the literature on other Uloboridae and Deinopidae (Opell, 1989, 1990; Peters, 1992). Data comparing native threads and gold-coated ones of other spiders are missing so far, but we presume a similar effect would be observed there. Although this treatment should remove any charge, other effects must have led to the collapse of the puffs, as puffs and intermediate zones are still discriminable in the gold-coated samples. In contrast, following exposure of the threads to fine mist, the puffs were no longer discriminable from the intermediate zones. If both methods removed only electrostatic charge, threads should look alike following treatment. Because coating the thread with carbon is a thermal process, any distortion of the thread due to heat formation during the gold-coating process can be excluded. We suggest that coating with gold leads to artefacts by covering the cribellate fibres with too much material, leading to compression of the puff. The conserved structure after coating the thread with the typically thinner carbon layer supports this hypothesis.

The exposure to fine mist must have another impact on the fibres. High humidity alone indeed improves the adhesion force, whereas exposure to fine mist (droplets of liquid water) annihilates the adhesion force (Hawthorn and Opell, 2003; Opell, 2013; Elettro et al., 2015). Hence, liquid water has to influence the spider silk directly. Water has the effect of resetting the protein conformation of other silks into its unprocessed state (Blamires et al., 2012). Because the structure of capture threads after contact with water is very similar to that of threads produced by spiders without calamistrum, the calamistrum might modify the protein conformation instead of charging the fibres. This could lead to an autonomous curling, finally forming the puff. Though we tried to eliminate any electrostatic charge during capture thread production, it is possible that local Coulomb forces caused a change in protein folding as a result of the polar amino acids, which are more abundant in cribellate silk than in ecribellate silk (Perutz, 1978; Liao et al., 2011). A modification at the protein level would also explain the irreversibility of the loss of puffs. Further studies have to specify how the calamistrum processes the cribellate fibres to form the puffy structure and whether this crimping is indeed a modification at the protein level.

The finding that no electrostatic forces are involved in the nanofibre production of cribellate spiders annuls the possibility of transferring a corresponding biological model to a technical application involving only low electric charge. Nevertheless, the capture thread production system of cribellate spiders remains of interest for biomimetic approaches: the nanofibre production of cribellate spiders can be used as an inspiration to produce nanofibres without any electric charge at all, and additionally to reduce the typical diameter of technical nanofibres from 100–1000 nm to 10–30 nm (Friedrich and Langer, 1969; Opell, 1995b; Nayak et al., 2012).

Clearly, a lot of fibres form a rim at radius 1 but other fibres are distributed uniformly within the cross-section. The fibre density for 2000 fibres according to the radius is shown in Fig. S2B. Evidently, there is a high density close to the outermost radius of 1 in our case and uniform distribution inside this circle. This would be expected if in fact electrostatic repulsion led to the arrangement of charged fibres in space, given the length limitation. This contradicts the observation that in the centre of the puff hardly any fibres can be observed (Fig. S3; Joel et al., 2015).

We want to thank Robert Raven, Barbara Baehr and Robert Whyte from Queensland Museum for the supply of capture threads of B. longinqua. Many thanks to Moritz to Baben and Jan-Ole Achenbach from the Institute for Material Chemistry at RWTH Aachen University for coating the threads with carbon. Furthermore, we want to thank Peter Bräunig and Björn Kampa for supporting this study with lab equipment.