Biofoams and natural protein surfactants

Naturally occurring foam constituent and surfactant proteins with intriguing structures and functions are now being identified from a variety of biological sources. The ranaspumins from tropical frog foam nests comprise a range of proteins with a mixture of surfactant, carbohydrate binding and antimicrobial activities that together provide a stable, biocompatible, protective foam environment for developing eggs and embryos. Ranasmurfin, a blue protein from a different species of frog, displays a novel structure with a unique chromophoric crosslink. Latherin, primarily from horse sweat, but with similarities to salivary, oral and upper respiratory tract proteins, illustrates several potential roles for surfactant proteins in mammalian systems. These proteins, together with the previously discovered hydrophobins of fungi, throw new light on biomolecular processes at air–water and other interfaces. This review provides a perspective on these recent findings, focussing on structure and biophysical properties.


Introduction
Fluid-based froths and foams are inherently unstable entities and relatively rare in biology. Formation of bubbles in liquids, and particularly in water, requires considerable energy input to overcome high surface tension and increased surface energy at the exposed gas-liquid interface. Consequently, foams are energetically expensive to make and difficult to maintain, with a tendency to collapse over time unless stabilised mechanically or kinetically by additional processes. Stability at the molecular level is an additional major issue for foams made out of biological macromolecules, since the surface tension forces at the air-water interface are often sufficient to disrupt macromolecular conformations . Most proteins are potentially susceptible to surface effects and inadvertent foaming will often lead to denaturation.
As a result, denatured proteins often display surfactant properties, presumably due to aberrant exposure of hydrophobic groups. In the biological context, resistance to microbial degradation, predation, and other environmental challenges is a significant issue. Moreover, for biofoams that come into contact with delicate biological tissues, damage to cell membranes that might result from conventional surfactant activity must somehow be avoided or averted. It is perhaps therefore not surprising that biological foam and surfactant activity is relatively uncommon, except in special instances. The purpose of this review is to examine some such special instances, and to summarize recent developments in the study of the structure(s) and function(s) of a range of proteins with natural foam and surfactant activities in instructive biological contexts. Relevant earlier work on surfactant proteins and peptides has been reviewed by others elsewhere and will be only briefly summarized here, allowing us to focus on more recent studies, mainly from our own group together with the beginnings of similar investigations in other systems.

The general physics of foams and surfactants
Soap bubbles, foams and related surfactant activities are not only a source of childhood (and childish adult) fascination, but also a rich source of intriguing physics and physical chemistry . William Thomson (Lord Kelvin) speculated that the structure of the ether could be likened to that of a foam . Using geometric arguments, based partly on earlier empirical rules of Plateau, he showed that an ideal 3-dimensional foam would comprise body-centred cubic packing of 14-sided polyhedra A C C E P T E D M A N U S C R I P T

ACCEPTED MANUSCRIPT
(tetrakaidecahedron, or "Kelvin cell" lattices) which minimises the total surface area in a space packed with identical units. More recently, using computational methods, Weaire and Phelan showed that a more complicated polyhedral arrangement satisfied this criterion slightly better. But reality is a little more complicated, and Kelvin cells or Weaire-Phelan structures are rarely seen in practice. When first formed, "wet" foams are usually made up of spherical (air) bubbles separated by relatively thick films of liquid (water), giving the traditional "kugelschaum" (spherical bubble) structure. As the liquid drains (under gravity or by capillary action) the lamellae between bubbles get thinner and the "dry" foam takes on a more irregular polyhedral ("polyederschaum") structure. But such foams are physically unstable. The excess pressure inside a bubble is inversely proportional to its radius. For a spherical bubble the excess pressure (∆P) is given by: ∆P = 2γ/r , where γ is the surface tension of the liquid. Also, and not unrelated, because surface area-to-volume ratios are higher, the excess surface free energy of liquid (water) molecules exposed to the air interface is higher for small bubbles compared to large. Consequently, given the opportunity, bubbles will tend to burst (at the surface of the foam), and smaller bubbles (with higher excess pressure and surface energies) will tend to coalesce to form larger bubbles. Air may also diffuse across thin liquid films from smaller to larger bubbles, likewise leading to eventual collapse of the foam.
Foam stability therefore depends on numerous kinetic and non-equilibrium processes related to viscosity, surface tension, drainage, diffusion, capillarity, and so forth.
Initial formation of foams or bubbles is facilitated by reduction in surface tension, and this is the basis for the everyday experience with soaps and detergents. Water is acknowledged to be an unusual liquid in almost all respects, and has an unusually high surface tension (excess surface energy) compared to most other fluids, related to the characteristic tetrahedral hydrogen-bonded structure in the bulk liquid. Water molecules at the air-water interface must adopt a less satisfactory packing arrangement that, crudely speaking, leaves fewer intermolecular H-bonds intact.
Soaps, detergents, lipids and other amphiphilic molecules can reduce this effect by forming (mono)layers at the interface, exposing less polar functional groups to the air whilst presenting a more water-compatible surface to the bulk liquid. Slightly more molecular detail can be provided by surface spectroscopy techniques such as infrared reflection absorption spectroscopy (IRRAS) . In favourable circumstances this can give information about the structure and orientation of molecules in the surface layers, based on characteristic IR absorption bands. With peptides and proteins, for example, this can give estimates of secondary structure (helix, sheet) content and relative orientation at the air-water interface.

Foams and surfactants in biology
Probably the largest foam masses of natural biological origin are those seen on the seashore or in turbulent freshwater streams, usually resulting from the adventitious agitation of natural organic materials or detritus, but without any obvious function relevant to the organisms involved in producing the foam constituents . One exception is the recently described foam accumulations associated with the synchronous reproductive stages of a species of marine tunicate (sea squirt) that appears to enhance fertilization of eggs and assist in the settling and retention of the larvae during A C C E P T E D M A N U S C R I P T

ACCEPTED MANUSCRIPT
spawning . These foams are found on rocky beaches and tidal channels in inter-tidal regions in Chile, where tunicate eggs and larvae would otherwise be dispersed by wave activity. The foam appears to be formed by the action of turbulent aerated seawater on materials released in large quantities by the tunicate colonies during spawning, though the precise composition of this material has not yet been described.
Consequently, the energy invested by the tunicates is confined to the synthesis and release of the foam precursor materials, whilst the foaming stage itself relies on wave energy. This is in contrast to most other cases of biofoam production, where both stages of foam production require direct action by the animals concerned.
The largest foam masses created by land or semi-aquatic animals are the foam nests of various species of tropical and sub-tropical frog (see Figure 1 for example). These are remarkable biological materials that, depending on the species, are adapted to persist intact in underground burrows, floating on microbe-infested temporary pools, or higher up in vegetation overhanging water. These foams are stable, resilient to physical and biological environmental challenges yet must be compatible with the membranes of delicate reproductive stages. Because frogs are external fertilisers, this biocompatibility must apply to both naked eggs and spermatozoa, as well as to developing embryos. As we shall see, these frog nest foams are not based on conventional small molecule surfactants, but rather depend on specialised surfactant proteins in synergy with a range of other proteins that can act together to protect the foams against microbial and parasitic attack, as well as providing structural stability for the foam. Surfactant (as opposed to foam) activity is seen in cases where wetting of non-polar surfaces is required, with protein-based systems including hydrophobins, latherin and lung surfactants to be described below. Other natural surfactants, mostly of plant or microbial origin, are predominantly lipid-based or conform to established concepts of low molecular weight detergents and have been extensively reviewed elsewhere .
Non-specific or adventitious foam or froth formation is often seen with mucins, for example in saliva, slimes and egg jellies , but here we will concentrate on more definitive protein surfactants produced by vertebrate animals.

Protein foams and surfactants
The relatively non-specific foaming of denatured proteins is commonplace and widely exploited in food technology and other processes . However, this usually requires much higher protein concentrations (typically > 10 mg ml -1 ) and much more vigorous physical treatment (whipping and sparging) than is the case with the specialised surfactant and related proteins to be described here. The process is generally acknowledged to be associated with the higher hydrophobicity and/or increased viscosity of denatured protein in which physical entrapment of air bubbles is facilitated in concentrated viscous mixtures . This is usually the dominant mechanism in common culinary processes such as the whipping of cream or preparation of meringue from egg white, for example. Other familiar examples include the use of protein products to stabilise "instant whips", beer foam, and other products. somewhat controversial and is complicated by significant calcium-binding activity .
Hydrophobins have been described as the most powerful surface-active proteins known and, like the proteins from frogs and horses that we discuss below, their activity is intrinsic to the proteins themselves, independent of any obligatory association with lipids or carbohydrates. They are small proteins (7-9kDa), unique to filamentous fungi where they are secreted during growth and spread of these fungi.
By lowering the water surface tension, the hydrophobins make it easier for the growing hyphae to penetrate through the air-water interface , subsequently forming a protective coating on the aerial structures and spores. Hydrophobins are also involved in attachment of fungi to surfaces such as plant leaves or insect cuticles . The proteins exhibit a characteristic four-disulphide bridge motif and a distinct amphipathic tertiary structure related to their self-assembly and surfactant properties . Each monomer has a discrete hydrophobic patch that is thought to be involved in interaction with an identical partner protein that obscures the hydrophobic region. This permits miscibility with the bulk water phase until reaching an air-water interface or other non-polar surface, where, they probably dissociate and re-orient with the hydrophobic surface exposed to the interface, in a process similar to the micellar rearrangement mode of action assumed for conventional small molecule amphiphilic detergents. This

ACCEPTED MANUSCRIPT
process would not necessarily require any significant conformational change in the protein, which in any case would likely be precluded by the stability imposed by intra-molecular disulphide bridging in these proteins .
Although not strictly a surfactant in terms of this current review, it is worth making reference to the flocculant activity of Moringa oleifera seed protein that is attracting interest as a means of water treatment in deprived areas . The seeds of this tropical tree have been used in traditional water cleaning processes in parts of Africa, and recent biophysical work has characterized one of the small (6.5-13 kDa) proteins present in the seed extracts that can adsorb to hydrophilic surfaces .

Foam nest proteins
As might be gathered from the brief summary so far, with the notable exception of the Microbiological analysis (unpublished) shows that the foams produced in the wild are contaminated with a rich variety of organisms, growth of which seems to be inhibited in the foam. The compatibility of the foam with eggs and sperm suggests that this microbial resistance is not due to the membrane disruption that might be expected from simple detergent activity, but is a more complex property of the foam components. Non-covalent binding of carbohydrate chains to lectins in the interface layer could provide a stabilizing matrix that would aid both foam stability and water retention, as pictured in Figure 2. Other macromolecules (in addition to the lectins) with which the carbohydrates may be associated have not yet been identified, but could be mucins, which would serve also to increase the viscosity of the matrix of foam nests, particularly in those of frogs that produce the more rigid aerial nests. Mucins are common to frog egg jellies, and may additionally act to restrain microbial colonisation.
Carbohydrate binding proteins may also form part of the antimicrobial defence system in the nest. Lectins are normally unable to kill bacteria in the absence of accessory proteins, but can agglutinate particles bearing their target sugars. Consequently, their role in frog nest foams may be to restrain microbial dissemination and colonisation of the foam and eggs and to inhibit microbial activity by blocking cell surface receptors.
Exactly this function is thought to apply to the fish fucolectins, which are present in large amounts in gills, eggs and blood . However, as a possible exception to this rule, killing of bacteria by specific lectins in the absence of accessory proteins has recently been demonstrated for E. coli expressing human blood group antigens . This raises the possibility that foam lectins may have a more direct antimicrobial role than has currently been observed. Furthermore, it is also worth noting that plant lectins, most

ACCEPTED MANUSCRIPT
notably in seeds and beans, can also act as anti-feedants and deter parasitism and predation by insects, birds and mammals through the disruptive effects that lectins can have on the epithelium of the gut. The role of surfactants as a defence against predators, recently demonstrated for insects , may also be relevant here.
The structure of one of the proteins in this mixture (RSN-2) has been examined in detail, both in solution by high-resolution NMR and (to much lower resolution) by neutron reflectivity and IRRAS at the air-water interface . As mentioned above, More generally, the picture that is revealed here from the túngara frog foam is of a fascinating synergy involving a range of specialized proteins with a mix of useful properties that work together to meet the requirements of a biocompatible foam, sufficiently robust and biochemically stable to act as a temporary nest. As illustrated in Figure 2, the surfactant activity of ranaspumins is just one part of a possible mechanism in which initial foam formation is further stabilized by self-assembly of a protein-carbohydrate matrix at the interface, providing longer term physical stability

ACCEPTED MANUSCRIPT
and water retention. And several of these proteins appear to have dual/multiple functions including protease inhibition and other roles in inhibiting predation and microbial degradation.
Ranasmurfin -a blue protein with a new type of protein chromophore Foam nest components from other frog species have not yet been studied systematically. However, one Asian species, Polypedates leucomystax (striped tree frog, or Java whipping frog), has been examined in some detail and shows significant differences when compared to the ranaspumins described above. P. leucomystax seem to rely more on viscosity than surfactant activity for initial stability of the nest foam.
These frogs, common and widespread in south and east Asia, produce a sticky, syrupy fluid (from the female, together with eggs) that is whipped up by the mating pair in much the same way as described above for E. pustulosus to form a protective environment for developing eggs and tadpoles. However, unlike the túngara frog, Although full analysis of the protein components of the P. leucomystax nest foam is not yet complete, one particularly intriguing protein has been examined in some detail at the molecular level . An unusual feature is that, although unpigmented or pale creamy pink/orange when first produced, some of these nests subsequently develop a streaky blue/green pigmentation that is more pronounced when nests are physically disrupted. The purpose (if any) of this pigmentation is not yet known, nor is it clear why not all nests undergo this colour change in the wild, but the colour is associated with a specific and quite unusual protein, designated ranasmurfin. This was first observed by SDS PAGE analysis of natural nest material, during which a brilliant turquoise blue band, corresponding to a protein of around 28 kDa, was observed

ACCEPTED MANUSCRIPT
migrating on the (unstained) electrophoresis gels. Subsequent purification and characterization of this protein from natural material has confirmed its uniqueness.
Although cDNA encoding this protein has not yet been isolated, a combination of fortuitous circumstances allowed high resolution structure determination. The purified natural protein crystallised readily as bright blue crystals that diffracted well; the protein was found to contain a heavy metal atom -identified as zinc by X-ray fluorescence and metal analysis -that facilitated phasing of the X-ray diffraction data and structure determination. The resulting electron density map was of sufficient quality (1.1 Å resolution) that, together with mass spectrometry of peptide fragments, the amino acid sequence could be determined directly.
The structure of ranasmurfin is shown in Figure 4, revealing an unusual dimeric structure with a novel fold . Also apparent in the structure are several posttranslational modifications, including an unusual extended chromophoric co-factor confirmed by chemical and spectroscopic evidence to be an N-linked indophenol-type moiety of a type not previously observed and comprising a Lys-Tyr-Tyr-Lys crosslink that unites the dimeric protein structure. This, together with two histidine sidechains (one from each monomer), coordinates the zinc (presumably Zn 2+ ) to form the blue chromophore.
By analogy with similar co-factors found in other systems, it is possible that ranasmurfin is involved in an extensive protein crosslinking function, particularly at its exposed surface, to promote long-term stabilization of the foam nest. Alternatively, or simultaneously, because of its unusual spectral properties, it may also be part of a sunscreen mechanism that protects the unpigmented eggs and embryos in nests exposed to tropical sunlight, or simply provides camouflage for the otherwise highly noticeable nests.
The other components of this foam have yet to be analysed but, interestingly, the natural, unfractionated nest material shows protease inhibition (cystatin-like) activity as potent as that found in foam nests of the túngara frog . This suggests that a similar cocktail of anti-microbial and anti-feedant components may also be present in the foam, providing short-to medium-term protection against biochemical degradation as well as mechanical stability.

ACCEPTED MANUSCRIPT
Latherin -a mammalian surfactant protein One of the first proteins shown to have strong surfactant properties in its native state is latherin, which is found in the sweat of horses, and recently shown also to occur in horse saliva . It is latherin's significant surfactant activity that gives rise to the familiar foam in the horse pelt, formed by friction during vigorous exercise . The protein is characterised by an unusually high leucine content (ca. 24% leucine, compared to an average of around 10% for most other proteins), which may directly relate to its surface properties. Bacterial recombinant latherin is also strongly surface active, so this surfactance is intrinsic to the protein and not dependent on association

Evolutionary aspects
There is as yet insufficient sequence information available to allow sensible between an archaean and a species of frog, or even horizontal gene transfer from the former to the latter. cDNA for ranasmurfin has not been isolated, so we cannot yet determine whether it is encoded within the P. leucomystax genome. Interestingly however, there is at present no indication of a similar gene in the one species of amphibian whose genome has been sequenced (Xenopus tropicalis; ref ). This intriguing phenomenon clearly must be pursued further, as more information on the M. smithii and amphibian genomes and proteomes becomes available .

Implications and applications
There are already numerous applications of proteins as foams and surfactants in food processing and related activities , where the empirical science of soft matter has been exploited (mostly unawares) for centuries. The food science literature on this topic has been covered in numerous reviews and reports available elsewhere (e.g.