Biological nanocoatings have evolved to serve multiple roles9,12-14. One is light manipulation, such as minimization of the light reflectance by the biological surfaces. Indeed, due to the difference in the refractive index of air and biological tissues, a significant amount of incoming light is reflected, which restricts biological functions such as vision. To cope with this problem, insects’ and many other arthropods’ eyes are covered by nanocoatings with the wavelength lower than that of the light of the visual spectrum, creating a gradient of the refractive index and minimizing light reflectance9,15,16. Initially discovered in moths as paraboloid pseudo-orderly packed nano-protrusions (nipple arrays), this type of coating was named moth-eye nanostructures and has found numerous technological applications13,15,17-19. In addition to the anti-reflectivity, other functionalities have been attributed to bionanocoatings, such as anti-wetting / self-cleaning, bactericidal, or anti-fouling9,13. Noteworthy, a trade-off appears to exist between these major functionalities, so that the better anti-reflective surfaces are poorer in providing the anti-wetting function, and vice versa8. The concept of the biological trade-off is omnipresent in evolution and ecology4,5,20, and our findings8 represent one of the few described examples of the biological trade-off at the nanoscale.
Arthropods possess diverse types of anti-reflective corneal nanocoatings, represented, in addition to the nipple arrays, by parallel strands, maze-like structures, irregularly spaced protrusions, dimples, and various transitions among them7,10,11,21,22. In our ongoing exploration of the diversity of insect corneal nanocoatings, we here assessed 97 species of the order Coleoptera (beetles, Supplementary Table S1). This massively expands the previous studies (7 species7,23,24, Supplementary Table S2), providing a compendium of beetle corneal nanocoatings represented by the maze-type, dimples-type, and nipples-type coatings (and transitions between these structures, further to the smooth unprotected surfaces seen in several species, Supplementary Tables S1 and S2). In addition to these previously seen types of corneal surfaces, our attention was attracted to the novel type of nanocoatings that is reminiscent of paving stones that we discovered in Luciola lusitanica fireflies (Fig. 1a, b).
Fireflies (Lampyridae) is a family of Coleoptera, represented by ca. 150 genera having the capacity to communicate through emission and reception of specific light signals, which is key for the insects’ sexual behavior25. Fireflies have evolved two mating behavior programs26. In one, employed by multiple genera including the genus Luciola, both males and females are capable of flight and both emit light signals that mediate their dialogue prior to mating27. In the other, represented among others by the genus Lampyris, non-flying females sit in the grass and continuously shine, attracting males that fly in the search of females without emitting light signals28 (Fig. 1a, Supplementary Fig. S1a). In previous works7,21,22, we showed that the corneal nanocoatings are adjusted to the insect's lifestyle. Thus, we could expect different nanostructures’ topography for the fireflies with different lifestyles. Indeed, we find that similarly prominent ‘paving stones’ structures coat the eyes of males and females of Luciola lusitanica fireflies (will be addressed as Luciola henceforth, Fig. 1b) – both of which need to have good vision for the mating program. In contrast, while elaborated nanocoatings (nipples-like) are present on the corneae of male Lampyris noctiluca fireflies (will be addressed as Lampyris henceforth), the non-flying females of this species that do not receive luminescent signals harbor much shorter, maze-type corneal nanocoatings (Fig. 1c). As the height and shape of the nanostructures determine their antireflective properties22,29, we performed a spectroscopic analysis of these four types of firefly nanocoatings. This study reveals that the antireflective properties of Luciola male, Luciola female, and Lampyris male corneae are comparable to each other across the visual spectrum (Fig. 1d). In contrast, the lenses of Lampyris females, that display less prominent nanocoatings, reflect 2-3 times more light across the spectrum tested (Fig. 1c, d). These findings confirm the inverse relationship between the height of nanostructures and the light reflection22,29, mirroring the behavioral characteristics of these insects at the nanostructural level.
Lantern – the light-emitting organ of fireflies – may also be expected to harbor anti-reflective nanocoating, in order to minimize the loss of the emitted light due to reflectance from the cuticle-air interface on the light’s way out. Previously, such antireflective nanocoatings have indeed been reported on the cuticle of the lantern of the firefly Aquatica lateralis, absent from the cuticle of the adjacent abdominal segments10,30. Here, we expand this analysis to the two other species of fireflies. In agreement with the ecological sexual dimorphism of the Luciola and Lampyris fireflies, and in reciprocity to the dimorphism we observed for their corneal nanocoatings (Fig. 1a-c), we find that the lantern abdominal segments of male and female Luciola, as well as of female Lampyris, are covered by parallel strand-type nanocoatings, ca. 50nm in height. In contrast, the same segments of the male Lampyris fireflies (that do not send light signals, Fig. 1a) are devoid of these structures (Supplementary Fig. S1b).
Given the previously established trade-off between the anti-reflectivity and anti-wettability8, we next wondered if corneae of female Lampyris fireflies, conducting the residentiary lifestyle28 and thus possibly needing more anti-wetting / self-cleaning protection, could be superior in these characteristics to the vision-biased Lampyris male corneae. We measured the contact angle of water droplets on the eyes' cuticles, finding the impressive hydrophobicity of Lampyris female eyes (contact angle of ca. 90°), while the corneal surfaces of Lampyris males were rather hydrophilic (contact angle about 60°, Fig. 1e). Thus, for Lampyris fireflies, we see the expected trade-off between the two main functionalities of the bio-nanocoatings, anti-reflectivity vs. anti-wetting, with the more vision-dependent males ‘choosing’ the first functionality on the expense of the second one, and the females demonstrating the opposite ecologically-justified selection.
As for the Luciola fireflies, whose corneal nanocoatings (for both sexes) are as prominent and as antireflective as those of Lampyris males (Fig. 1b-d), we expected the reciprocal loss of the anti-wetting properties. Surprisingly, we find that the hydrophobicity of Luciola male and female structures remain at the Lampyris female level (Fig. 1d). At the same time, the Luciola surface roughness, adhesiveness, and elasticity were close to those of Lampyris males (Fig. 1b, c and Supplementary Fig. S2a-c). These findings contrasted the Wenzel model and suggested that the Cassie-Baxter model, that takes into consideration air bubbles trapped between the water droplet and the surface9, should be considered instead to explain the anti-wetting effect of Luciola nanocoatings (Supplementary Note 1). Indeed, analysis of cross-sections of the Luciola and Lampyris corneae reveals that the paving stones-type of the nanostructures in Luciola are characterized by tighter contacts between them, as opposed by the more relaxed spacing intermingling the nipple-type nanostructures in Lampyris (Supplementary Fig. S2d). With additional analysis (Supplementary Note 1) we conclude that the air trapping in the tight spaces between the paving stones nanostructures in Luciola, occupying ca. 10% of the surface, is the main driver of the strong hydrophobicity provided by these corneal nanocoatings.
Thus, we have identified unusual natural nanocoatings that functionally excel in their functionalities the more common types, such as the nipples- and the maze-like structures. But why do certain species, such as Lampyris or Drosophila fruit flies8, have to choose between two types of nanostructures, each of which is adapted to only one functionality, while Luciola fireflies can benefit from both functionalities? We assume that some limiting factor prevents other insects from using this paving stones-like nanocoatings. We have excluded the idea that this factor is the strength of nanostructures and their resistance to mechanical damage: 2D Young's modulus measurements show that Luciola nanocoatings are at least as stable as nipple- and maze-like nanostructures (Fig. 1b, c, Supplementary Fig. S2c).
To seek insights into the possible reasons of the uniqueness of the bi-functional nanocoatings of Luciola, and for the lack of existence of such nanocoatings in other insects such as Lampyris, we performed simulations of the process of the corneal nanocoatings’ formation. Components of nanostructures (proteins and waxes8) are secreted by cone cells prior to formation of the other cuticle layers9,31, engaging in the Alan Turing’s reaction-diffusion interactions leading to nanopatterning in the cuticle7,8,32,33 (Fig. 2a, b). Simulations (Supplementary Notes 2 and 3, Supplementary Fig. S3, Supplementary Table S3) can reproduce topography for paving-stones-like, nipples-like, and maze-type nanostructures (Fig. 2c). Numerical investigation of the patterns within the parameter space predicts that the paving stones-like structures appear only in the strict area that is limited by narrowly defined parameters (Fig. 2c, Supplementary Fig. S4). In contrast, the nipples- and the maze-type patterns dwell in the region within the parameter space where formation of these patterns is permitted in the broad range of parameters. These simulations indicate that the paving stones-like nanocoatings, although possessing the dual functionality and thus initially appearing as an attractive evolutionary solution, may in fact be highly sensitive to the fluctuations in the reaction-diffusion conditions, and that even slight environmental disturbances could block the formation of these nanostructures.
For example, temperature changes, affecting the reaction and diffusion parameters, could decrease the Turing region size and stop the nanostructure formation34,35 (Supplementary Note 4). We showed that Luciola’s paving stones-like structures are much more sensitive to temperature variations than the Lampyris-like structures. Simulations at different average temperatures and daily temperature ranges DT reveal that the Lampyris nanostructures can be formed in a much wider range of temperature conditions than those of Luciola (Fig. 2d, Supplementary Table S4). Further, modeling the initiation of the nanostructures’ growth at different time points during a day with conditions reminiscent of the spring season, we find time periods when Luciola – but not Lampyris – nanocoatings fail to form (Fig. 2e, Supplementary Fig. S5). Although qualitative, these analyses suggest that the corneal nanostructure formation in Luciola is possible only in mild climatic conditions that are characterized by low temperature fluctuations. Analyzing regularly replenished collections of Lampyris and Luciola sampled at different locations from 1886 till nowadays (Supplementary Table S5), we find that, in agreement with the simulations, the areal of Luciola falls within the tempered climate zones, encompassing the Mediterranean and southern Europe and avoiding the altitudes >2000m. In contrast, Lampyris inhabits the entire territory of Europe including high altitudes27,36 (Fig. 2f, Supplementary Table S5). As nanocoatings form before the other layers of the cuticle, failure in their formation might negatively impact the whole process of cuticularization leading to its ultimate collapse9,31,32 with detrimental effects on the firefly viability and fitness.
Thus, two different ecoevolutionary strategies emerge in the formation and functionalities of insect corneal nanocoatings. One is represented by Luciola fireflies, which develop bi-functional paving stones-like nanocoatings that appear to be an ideal solution for anti-reflectivity and anti-wetting, but this solution is sensitive to environmental variations. The other strategy is represented by Lampyris fireflies that build anti-reflective nanocoatings in males and anti-wetting nanocoatings in females; these mono-functional nanocoatings are stable over a broad range of environmental conditions, permitting this species to inhabit diverse geographic and climatic zones. The Lampyris ‘nanotechnological’ strategy – a less perfect solution that is more stable to environmental variability – appears to be more generally applied in the insect kingdom, while the Luciola solution is likely an exception, given the unique nanocoatings of this species that we do not find elsewhere7,8 (Supplementary Table S1-S2). The evolutionary choice within this dichotomy – an ideal solution for a narrow set of environmental conditions vs. a less ideal (yet still functional) solution stable over a broader range of conditions – is described by the theory of evolutionary bet-hedging. Evolutionary bet-hedging studies the trade-off between the mean and the variance of fitness and explains how in varying environments, lower-quality adaptations to a broader diversity of conditions (reduced mean fitness with low variances) can be evolutionary favored over an ideal adaptation to a specific narrow condition (high mean fitness with higher variances)1-5. While numerous examples of evolutionary bet-hedging have been described in plants, birds, bacteria, etc.2,3,6, our findings reveal the first example of evolutionary bet-hedging at the nanoscale.
We next sought to provide a reverse engineering proof of the thermally limited Turing instability as the driver of the different fireflies’ nanocoatings. While genome sequencing and assembly of some fireflies are in process37, Luciola and Lampyris genomes have not yet been sequenced. This prevented us from a direct assessment of the corneal proteome and identification of the Turing Activator candidates – the approach we have previously applied to several insects with sequenced genomes8,11,21. However, earlier we showed that admixing of waxes (Turing Inhibitor) and Drosophila Retinin (Turing Activator) on glass surfaces recapitulates diverse insect-like nanocoatings such as nipples-like, maze-like and dimpled patterns, with different light- and liquid-handing functionalities8,11. Our simulations predict that the paving stones-like nanostructures can be obtained by decreasing the diffusion coefficient of the Turing Activator (Supplementary Note 2, Fig. 2c, Supplementary Fig. S3, Supplementary Table S3). We argued that this desired decrease in the diffusion coefficient could be achieved by increasing the size of Retinin. We cloned and purified a fusion protein of Retinin with nanoluciferase (NanoLuc)38 that we will refer to as RetLuc (38 kDa, 357 amino acids) and studied it alongside the parental Retinin (19 kDa, 179 amino acids), Supplementary Fig. S6a. To measure the diffusion coefficients of Retinin and RetLuc, we fluorescently labelled the two proteins with the ATTO565 dye and applied total internal reflection fluorescence (TIRF) microscopy. This yields the diffusion coefficients of 2.34±0.34⋅10-8 (mean ± SD) and 1.34±0.20⋅10-8 cm2/sec for Retinin and RetLuc, respectively (Fig. 3a), demonstrating that we have succeeded in obtaining the desired decrease in the diffusion coefficient upon the fusion of Retinin to NanoLuc. Using TIRF, we have further measured the diffusion coefficient of a BODPIY-labelled lipid C1, C12 (a decanoic acid, see Methods) as an example of the Turing Inhibitor we use further to create artificial nanocoatings. As required by the reaction-diffusion system from the Turing Inhibitor, it has the diffusion coefficient 3-5 folds higher than the Turing Activators – 7.29±0.32⋅10-8 cm2/sec (Fig. 3a). We further used wave-guide interferometry to show that the affinity of Retinin to lipids (decanoic acid in this experiment) did not change in the fusion protein: both Retinin and RetLuc reveal the equilibrium dissociation constant Kd = 34nM (Supplementary Fig. S6b, c).
Both Retinin and RetLuc were competent to build bionic artificial nanocoatings on glass upon admixtures with waxes (Fig. 3b, c). In a resemblance with the firefly corneal nanostructures where those of Luciola are narrower than those of Lampyris (Supplementary Fig. S3), the RetLuc structures were notably narrower and had the paving stones-like tight contacts in-between, unlike the structures built with Retinin (Fig. 3b-d, Supplementary Fig. S7a). The RetLuc nanocoatings were also more hydrophobic (Fig. 3e). From these data, we can consider that the artificial nanocoatings obtained with Retinin and RetLuc approach the natural corneal nanostructures of Lampyris and Luciola, respectively.
Such parameters as the joint diffusion coefficient (D), the reaction rate (f), and the broadness of nanostructures (ω) are related as follows: (Supplementary Note 2). Of note, D depends not only on the activator (Du) but also on the inhibitor diffusion coefficient (Dv) as . f is a function of reaction rates constants of all interactions, happening during nanocoatings process. Having measured the Du, Dv, and ω parameters, we can estimate the reaction rate. This analysis confirms that the fusion of Retinin with NanoLuc maintains unchanged the reaction rate of the protein (Fig. 3f).
We next tested the hypothesis of the differential thermal sensitivity of the Retinin- vs. RetLuc-based nanocoatings. Simulation of the artificial nanocoatings formation is achieved differently than that of the insect corneal nanostructures. In the case of artificial nanocoatings, there is no secretion component, and the main factors are the concentrations of the morphogens and their adsorption to the surface. The activation energy of wax and protein adsorption was taken as -30 and -55kJ/mol, respectively39. Using these parameters, we were able to simulate the formation of nanostructures on the glass (Fig. 4a), which reproduced to the good extent the experimentally observed artificial nanocoatings (Fig. 4b). Remarkably, simulations and experiment also agreed on the narrowed temperature conditions (only between 25°C and 30°C) where RetLuc was efficient to drive nanocoating, while the parental Retinin could induce robust nanostructuring in the whole spectrum of the tested temperatures (15-35°C, Fig. 4a, b, Supplementary Fig. S7b, Supplementary Table S6). At temperatures below 25°C, RetLuc-based nanostructures do not form; at temperatures above 30°C, in addition to the main layer of nanocoating, nanostructures start growing in solution, adhere to the main layer and create aggregates of varying sizes (Fig. 4a (brown structures) and Fig. 4b). Noteworthy, the topography differences can also be identified through the enzymatic activity of the NanoLuc, as RetLuc maintains the luciferase activity not only in solution, but within the nanocoatings after their washout (Fig. 4c). We found the highest level of luminescence from the nanocoatings self-assembled at 25°C, correlating with the total immobilized protein amount (Fig. 4c, Supplementary Fig. S7c). These data highlight the possibility of manufacturing of nano-patterned enzymatic activities.
Local changes in the temperature regime could be applied to precisely induce the formation of nanostructures, thus allowing to print designed nanostructures patterns from Retinin with different added functionalities – such as nanopatterned enzymatic activity, opening the possibility for technological applications. Such dimensionally precise temperature control is achievable by light interference patterns or holograms. We decided to use metal nanoparticles in our system to mediate the light energy conversion into heat. Metal nanoparticles can efficiently absorb visible light by the plasmon resonance phenomenon40. The plasmonic photothermal effect can be explained in the following manner: surface plasmons, which are collective oscillations of electrons near the metal-dielectric interface, become localized in the vicinity of curved surfaces of metal nanoparticles. These excited surface plasmons can act as highly damped oscillators and dissipate energy through electron-phonon interactions, converting light into thermal energy40 (Fig. 4d).
Nanoparticles could be formed from metals’ salts in Retinin or Retinin-like proteins' solutions11. We created gold nanoparticles through metal ion binding, reduction and coalescence, following the kinetics of their formation with light absorbance41 (Supplementary Fig.8a) and producing the nanoparticles ca. 100nm in diameter (Supplementary Fig. S8b). TEM shows that this size is made of the gold nanoparticle itself (ca. 20nm) and of a protein envelope or corona stabilizing the particle42 (Supplementary Fig. S9c). A feasibility test showed that, indeed, the solution with gold nanoparticles can be heated by the laser exposure (Fig.4d, Supplementary Fig. S8d). As an initial experiment in this direction, we used a laser to heat-induce the nanocoatings in the presence of gold nanoparticles (Fig. 4e). With further refinement, this approach will permit precise patterning of (enzymatically) active proteins under mild conditions, making it possible to work with sensitive surfaces such as living tissues. Further, we envision the possibility of applying two or more laser beams to obtain fine temperature-controlled nano-patterning (Supplementary Fig. S8e). Protein-based self-assembled nanocoatings have several advantages, including easy modification through chemical and biotechnological methods and scalable manufacturing. The approach we present permits immobilization of enzymes or peptides on a surface with the nanoscale precision. Applications may include creation of specialized nanocoatings for directed cell differentiation or laboratory test plasticware, treatments of living tissues for accelerated wound healing, or biocompatible continuous monitoring sensors.
In conclusion, we present here a multidisciplinary study that discovers the first nanoscale example of the evolutionary bet-hedging. Insects generally develop corneal nanocoatings that are stable to environmental perturbations, with the trade-off of these nanocoatings being not ideally adapted to serve their functionalities. In contrast, some insects rarely choose to develop highly functional nanocoatings that are, however, highly sensitive to environment variations, such as variations in temperature. Learning from these nature-offered solutions, we reverse-engineer temperature-sensitive, -controlled, and -operated soft nanocoatings with desired light and water-handing and enzymatic properties, that may find its way in developments of a variety of technological applications.