Control Strategies in Guanine Biocrystallization

ABSTRACT

Highly reflective assemblies of guanine crystals generate a diverse array of optical phenomena in animals, from camouflage reflectors to image‐forming mirrors. The optical functions of guanine crystals and their related animal behaviors have been studied for over a century. However, only recently have we begun to unravel how organisms precisely control the structure and morphologies of these crystals which underly their photonic properties. In this minireview, we outline strategies employed by animals to control the formation and properties of guanine crystals, integrating both in vitro and in vivo studies. We frame these advances using analogies to ‘design’ principles in inorganic biomineralization—from the use of dopants and macromolecular templates in crystal morphology control to non‐classical crystallization pathways. We show that, despite their distinct biology, organic and inorganic biocrystallization share strikingly similar crystal design strategies. These strategies are a rich source of inspiration for bio‐inspired materials, especially in the development of sustainable optical materials. Finally, we outline key challenges for this exciting new field of organic biomineralization.

Biological guanine crystals produce spectacular photonic phenomena in animals and hold great promise as new, sustainable optical materials. We review how organisms precisely control the structure, morphologies, and resulting optical properties of these crystals using a set of ingenious ‘design’ strategies, including control of pH, template‐directed nucleation, crystal growth additives, and non‐classical crystallization.

1. Introduction

A historical perspective: Over 540 million years of evolution [1, 2], biology has mastered the art of constructing complex crystalline materials (biominerals) from simple inorganic ions, under ambient conditions [3, 4]. From the scaffolds of bone [5] to the spines of sea urchins [6, 7] to magnetic compasses in bacteria [8, 9], the impressive materials performances of biominerals are achieved through precise control over their crystal composition and structure. However, another world of biocrystallization, possibly just as widespread, and probably even more ancient [10], remains largely unexplored—molecular crystallization in biology. The material that dominates this world is guanine.

Since its discovery in the 1800s in silvery fish [11], guanine crystals caught the attention of biologists because of their role in animal coloration. From the 1960's to 80's, Land [12, 13], Denton [14], Nicol [15], Herring [16], Lythgoe [17], and others, rationalized how guanine crystal assemblies produce broad‐ and narrowband structural colors and image‐forming mirrors. The focus here was on optics and ecology and not on the crystals—which typically appeared as ‘holes’ in electron micrographs [18].

In the late 2000's, Addadi,Weiner and collaborators at the Weizmann Institute revisited the guanine story, focusing now on the crystal itself. Their work led to many important contributions, including; (i) the discovery that the platelet morphologies of biogenic guanine crystals may be ‘designed’ to enhance reflectance via the expression of high refractive index crystal faces [19], (ii) electron microscopy of photonic systems in their native state, with crystals intact [20, 21, 22] and, (iii) the determination of the crystal structure of the biogenic, β‐polymorph of guanine (Figure 1A,B) [23]. This inspired a search for other biogenic molecular crystals, leading to the characterization of ‘new’ biogenic crystals (isoxanthopterin [24, 25, 26], 7,8‐dihydroxanthopterin [27] and xanthine [28, 29, 30]) and partially crystalline materials (isoxanthopterin [31], uric acid [32, 33] and leucopterin [34, 35]), together with their exotic optical properties [36].

FIGURE 1.

Structure and morphologies of β‐guanine. Crystal structure viewed (A) perpendicular to the H‐bonded (100) plane along a and (B) parallel to the π‐stacked layers along b. Refractive indices; 1.83 in bc plane, 1.46 along a. (C) Simulated β‐guanine morphologies grown in vacuum and water, adapted from Ref. [48] CC BY 4.0. TEM images of biogenic guanine crystals from (D) scallop eyes (Pecten maximus), (E) white widow spider (Latrodectus pallidus) integument, (F) copepod (Sapphirina metallina) integument, (G) salmon (Salmo salar) skin, (H) wolf spider (Lycosa piochardi) eyes and (I) lizard (Acanthodactylus beershebensis) skin.

Recent discoveries of guanine crystals in unicellular eukaryotes [10] and bacteria [37] have greatly expanded the scope of this field—guanine crystallization has now been documented in 4 of the 6 kingdoms of life [38, 39]. Pilatova et al. [10] found that purine crystallization is widespread across all the major eukaryotic supergroups and was likely present in the last eukaryotic common ancestor. In dinoflagellates [40, 41] and green algae [42], nitrogen‐rich guanine crystals are primarily used as nitrogen‐storage reservoirs, enabling organisms to circumvent periods of nutrient deficiency. This gives rise to an intriguing evolutionary scenario—that guanine crystals originally evolved for fundamental cell metabolism functions in microorganisms and were later co‐opted, because of their high refractive index [43, 44] (Figure 1) for optical functions in higher organisms.

The fruit of all these endeavors was a new field—organic biomineralization. Interestingly, the emergence of this field was predicted 20 years prior in Weiner and Dove's seminal review on biomineralization: [45] “We suspect that many of these (organic) ‘minerals’ remain to be discovered, and exploring the functions they perform will be fascinating….” and, “…these are crystalline phases formed by organisms probably by the same underlying strategies used for ‘normal’ mineral formation.” While progress has been made on the first of these two points—the ‘optical functions’ (reviewed in Refs. [38, 46, 47]), and ‘discovery’ of new molecular crystals (reviewed in Refs. [18, 46, 47]), the final question of crystal formation, remains very much open.

Aims: Here, we outline chemical strategies employed by animals to control the formation and properties of guanine crystals. On one hand, the expectation that organic crystals would form in the same manner as inorganic biominerals seems improbable. Guanine is not a mineral at all, but a metabolite produced by every cell. Thus, the underlying questions are different: In the case of biomineralization—how do organisms sequester ions from their environment into hard mineralized tissues? In the case of organic biocrystallization—how do organisms upregulate the biosynthesis of metabolites, then transport, concentrate, and crystallize them? The anisotropic intermolecular bonding and solubilities of molecular crystals also contrast with those of ionically bonded solids. Indeed, the rationale behind the name ‘organic biomineralization’, itself a contradiction in terms, has more to do with the field's history, than any fundamental biological connection to biomineralization. Yet, despite these differences, on the surface at least, the design strategies underlying organic and inorganic biocrystallization are strikingly similar. Here, we frame advances in our understanding of organic biomineralization, with analogies to corresponding principles in inorganic biomineralization: the composite nature of bio‐crystals and the influence of intracrystalline additives on polymorph and morphology (Section 3.1), the role of confinement and macromolecular templates in crystallization and morphology control (Section 3.2) and non‐classical crystallization pathways (Section 3.3).

Why is this subject important? Guanine crystals are used in a wide array of coloration and visual phenomena in animals [49] and are responsible for regulating nitrogen‐metabolism [40] in superabundant and ecologically important microalgae [50, 51]. Understanding how organisms precisely control the nucleation, growth and morphology of highly insoluble N‐heterocyclic crystals will inspire new, bio‐inspired strategies for the design and formulation of synthetic molecular materials. The refractive indices of guanine [43, 44] (Figure 1A,B) and isoxanthopterin [52, 53] approach those of metal oxides [54] (e.g., ZnO) used in industrial optics and can exhibit similar or even superior optical performances [31, 55, 56] to man‐made optical devices. As such, biogenic molecular crystals represent extremely promising alternatives to toxic and environmentally harmful inorganic pigments in paints, pharmaceuticals, and food [31, 57].

2. Structure and Morphologies of Biogenic Guanine Crystals

All known biogenic guanine crystals, except those formed extracellularly by bacteria [37], adopt the metastable β‐polymorph, composed of planar, H‐bonded molecular layers within the bc plane(Figure 1A), π‐stacked along the a‐axis (Figure 1B) [23]. We note, that in our descriptions of the β‐guanine structure, we use the conventional P2₁/c space group setting (P2₁/c, a = 3.6 Å, b = 18.3 Å, c = 9.8 Å, β = 118°) [58]. Guanine crystals are highly birefringent, possessing a refractive index of 1.83 within the H‐bonded (100) plane and an out‐of‐plane refractive index of 1.46 (Figure 1A,B). The stable (by <2 KJ/mol) α‐polymorph [48] differs only from the β‐polymorph in the relative offset of adjacent layers. Given the similarity in their structures, morphological, and optical properties, neither polymorph has an obvious functional advantage. Many in vitro studies reported the initial formation of the β‐polymorph, suggesting that it is kinetically favored [59, 60]. Its formation in biogenic systems likely relates to the crystallization environment and kinetics.

In water, the thermodynamically preferred morphology of guanine is a prism or needle, elongated along the a‐axis, expressing the (100) face as a relative minor facet (Figure 1C) [48, 61]. This morphology is a product of relatively weak in‐plane H‐bonding, caused by the cost of desolvating H‐bonding interactions, and relatively strong π‐stacking, driven by the negative energy of desolvating this hydrophobic interaction (Figure 1C.) [48]. In contrast, animals form plate‐like crystals, where the high refractive index, but hydrophobic (100) plane is preferentially expressed to maximize the optical response of the crystals. Control over the shape of the (100) face allows organisms to generate a range of polygonal morphologies, enabling their assembly into different photonic superstructures (Figure 1D‐I).

3. Control Strategies in Guanine Biocrystallization

3.1. The Composite Nature of Biogenic Guanine Crystals

A characteristic of inorganic biominerals is that they are composites—occluding organic [62] and inorganic [63] additives within the crystals. For example, Mg²⁺ ions are frequently occluded within calcite biominerals [64], where they stabilize amorphous phases (e.g., amorphous calcium carbonate (ACC) and amorphous calcium phosphate (ACP)) [65, 66, 67], favor the formation of metastable aragonite [68], enhance biomineral hardness and toughness [69, 70] and influence crystal morphology [71, 72]. Additive occlusion is enabled by the relative flexibility of calcite's 6‐fold coordination geometry [64].

A priori, one would not anticipate such occlusion in guanine crystals where rigid molecules are connected by highly directional intermolecular bonds. Yet, many early studies reported the presence of hypoxanthine and xanthine in guanine‐forming, iridophore cells [14, 43, 73, 74, 75]. However, these studies could not determine whether these molecules were present as ‘solutes’ or included within the guanine structure. Pinsk's PXRD and solid‐state NMR analyses confirmed that many biogenic guanine crystals are solid solutions of guanine, hypoxanthine and sometimes xanthine, with hypoxanthine dopant concentrations reaching ∼20 mol. % in some species [76]. Are these dopants functional? Aihara et al. [77] originally proposed that variations in hypoxanthine doping in guanine crystals in fish may relate to their differing mechanical properties. Later, Gur et al. [38] hypothesized that hypoxanthine and xanthine might inhibit growth along guanine's π‐stacking direction, promoting the formation of highly reflective (100) plate crystals. However, Pinsk found no clear correlation between doping concentration and morphology—either in the thickness of the crystals or in the shape of the (100) face [76]. Instead, it was hypothesized that crystal doping enables animals to build physiologically “cheaper” crystals from mixtures of purine metabolites present in forming iridophore cells.

3.1.1. The Influence of Purine Additives on Crystal Morphology and Lattice Energy

In a series of crystallization experiments, Ma combined purine and polymeric additives (usually in organic solvents) to generate an impressive array of biogenic‐like guanine morphologies [60, 78, 79]. The presence of certain purines was also found to favor the formation of the metastable β‐polymorph [80]. This provided a hint that purine dopants may indeed affect guanine morphology and polymorph, the question was, how?

Lattice energy and polymorph preference: A rationalization for these effects was provided by Wagner and Hill [48] who used in vitro crystallization and DFT calculations, to show that: (i) across a wide compositional range, the free energy of metastable β‐guanine‐hypoxanthine solid solutions are within +4 kJ/mol of the physical mixture energy of the most stable pure solids (α‐guanine/α‐hypoxanthine, Figure 2A). This indicates that solid solution formation is feasible under kinetically driven, high supersaturation conditions, (ii) at low hypoxanthine doping (<10 mol.%), the lattice energy of β‐guanine solid solutions drops below that of pure β‐guanine and falls within 1.5 KJ/mol of pure α‐guanine (Figure 2A). Thus, solid solutions are likely to form under conditions in which β‐guanine crystallizes, and dopants may have the potential to stabilize the β‐polymorph under certain crystallization conditions.

FIGURE 2.

The influence of dopants on guanine crystal lattice energy, structure and morphology. (A) Calculations of the free energies of α‐ and β‐guanine‐hypoxanthine solid solutions, and the free energy of a physical mixture of α‐guanine and α‐hypoxanthine (dashed line). Inset; the solid solution energies relative to the physical mixture. For low doping levels, the free energy of β‐guanine‐hypoxanthine dips below that of pure β‐guanine. Adapted from Ref. [48] CC‐BY 4.0. (B) 3‐ and 2‐point H‐bonding interactions within the β‐guanine crystal. Hypoxanthine incorporation (yellow) leads to a loss of 2 H‐bonds. (C) Simulated guanine crystal morphologies upon varying the ΔG _cryst for the 3H‐bond versus 2H‐bond interactions. (D) Left: guanine crystals from the operculum (upper), eye (middle) and skin (lower) of zebrafish, containing increasing hypoxanthine levels. Middle; synthetic β‐guanine crystals with increasing hypoxanthine doping (from top to bottom). Adapted from Ref. [83] with permission 2024, Springer Nature. Right: changes in guanine crystal morphologies upon decreasing ─NH₂ interaction strength from Monte Carlo simulations. Adapted from Ref. [84] CC‐BY 4.0.

Crystal morphology: Crystal growth calculations show that the guanine morphology is dictated by 3 supramolecular synthons; [81, 82] the π‐π stacking (homosynthon) along a, the triple H‐bond (heterosynthon) along c, and the two‐point H‐bond (homosynthon) which links guanine dimers in the bc plane (Figure 2B). Since hypoxanthine has two less H‐bond donors than guanine (Figure 2B, yellow), its incorporation weakens in‐plane H‐bonding and increases the relative growth rate along the stacking direction, leading to a minimization of the (100) facet [48]. Thus, in water, hydrophobic capping layers, rather than purine additives, are required to inhibit growth along the π‐stack. Simulations also showed that by varying the relative strength of in‐plane H‐bond interactions along the [020] and [002] directions, dopants can modify the cross‐sectional shape of the (100) face (Figure 2C). This led to a hypothesis that dopant incorporation may be utilized to generate the range of cross‐sectional guanine shapes in biogenic systems.

An elegant study by Deis et al. [83] tested this hypothesis in a real biological system—finding that different guanine morphologies, in specific tissues in zebrafish (Danio rerio), correlate with their hypoxanthine content. High aspect ratio hexagons in the operculum displayed lower dopant content than low aspect ratio hexagons in the skin (Figure 2D). Genetic mutants, lacking pnp4a, a key enzyme in guanine biosynthesis, produced almost square crystals compared to elongated hexagons in the wild type. This change in morphology was associated with a two‐fold increase in hypoxanthine, caused by a compensatory upregulation of paralog enzymes which synthesize hypoxanthine. The authors also generated synthetic guanine crystals with increasing hypoxanthine doping, showing they could recreate zebrafish morphologies by controlling crystal composition (Figure 2D). Rothkegel et al., [84] then used Monte Carlo simulations to show that weakening of the ─NH₂ interaction upon hypoxanthine incorporation leads to preferential expression of the (012) crystal face over the (001) face, producing square morphologies (Figure 2D).

3.1.2. Unknowns and Perspectives

Like their inorganic counterparts, biogenic guanine crystals are often composites, occluding other purines in solid solutions. Purine dopants appear to play an analogous role to Mg²⁺ ions in calcium‐bearing biominerals, modifying crystal morphology, altering the energetic landscape of the crystal and potentially, the polymorph preference. While a compelling case has been made for the effect of hypoxanthine on crystal morphology within zebrafish [83, 84], other studies show that guanine crystals in different fish species, with ostensibly similar elongated hexagonal morphologies, have very different compositions [76, 77]. Comparing the compositions and morphologies of crystals in fish and other organisms would shed light on the generality of this control strategy. Applying new techniques for measuring the composition of individual crystals (rather than bulk chemical analyses) might be crucial in this regard. Other important but unexplored questions are the role of tautomerization in the occlusion of additives [85, 86], their spatial distribution within host crystals and their possible role in stabilizing amorphous phases, akin to Mg²⁺ in ACC and ACP. Additive inclusion clearly represents one of several biological strategies employed to control crystal shape, with templates (see Section 3.2) and crystal twinning [87, 88] also utilized in certain systems.

3.2. Crystallization Under Confinement and Macromolecular Templates

Many of the unique properties of biominerals originate from their formation in spatially constricted compartments [89]. These compartments are delimited by organic interfaces (the ‘matrix’), composed of lipid membranes, collagen and chitin [4]. Confinement has been attributed to the stabilization of amorphous phases [90] and control of crystal morphology [91], nucleation, and polymorph [92]. Confinement effects can either be intrinsic volume effects (e.g., the sluggish induction of nucleation in small volumes) [89] or extrinsic, relating to the proximity of the interfaces to the forming mineral in contrast to bulk crystallization.

Like inorganic biominerals, biogenic molecular crystals form within nanometric, membrane‐bound organelles [93]. One hypothesis was that these compartments may, in some way, shape or mold guanine into its characteristic morphologies, as observed in inorganic biominerals [94].

3.2.1. Iridosomes: The Crystal Forming Organelles

In animals, guanine crystals are formed in specialized reflective pigment cells (‘iridophores’), within subcellular organelles called ‘iridosomes’—which give rise to one or two, guanine crystals. Iridophore cells, together with melanin‐bearing, melanophore and pteridine‐bearing, xanthophore pigment cells [95], derive from a multipotent progenitor cell in the neural crest [96, 97, 98]—a transient tissue formed during vertebrate development [99, 100, 101, 102]. Since Bagnara first hypothesized the ‘common origin of pigment cells’ model [93], work on frogs (Xenopus laevis) [103, 104, 105] and zebrafish (Danio rerio) [106, 107, 108] identified key genetic and transcriptomic cues determining the differentiation, specification, and migration of iridophores at the cell level.

However, much less is known about the chemistry and morphogenesis of the crystal‐forming, iridosome. Bagnara proposed that all pigment types (including guanine crystals) derive from a common, ‘primordial organelle’ [93] which, depending on the biochemical cues, has the potential to produce all the various pigment types. This endosome‐derived organelle is thought to originate from the endoplasmic reticulum (ER), which fuses with Golgi‐derived vesicles carrying guanine precursors and biosynthetic enzymes.

3.2.2. Iridosome Morphogenesis: Template‐directed Nucleation and Morphology Control

Impressive but little‐known early electron microscopy studies of iridosome morphogenesis in fish (Poecilia reticulata) [109] and lizards (Sceloporus graciosus) [110] by Gundersen [109] and Morrison [110] observed that: (i) Early iridosomes are ‘double‐membrane bounded’ vesicles (Figure 3A‐i,B‐i), with the outer membrane likely derived from the ER and the inner membrane produced by fusion and invagination of Golgi‐derived vesicles. Ullate‐Agote's [111] work on corn snakes (Pantherophis guttatus) showed that iridosomes are lysosome related organelles (LROs), confirming their endosomal character. (ii) Prior to crystallization, Morrison [110] noted an ‘electron‐dense’ but non‐crystalline material within the inner vesicle (“crystal chamber”, Figure 3B‐i) hinting at the existence of disordered guanine precursor phase (Section 3.3). (iii) As the crystal develops within the crystal chamber (Figure 3A‐ii, B‐ii), the initially ‘corrugated’ inner vesicle membrane becomes straightened and ‘distended’ as it attaches to the growing facets of the crystal—a template‐directed nucleation (Figure 3A‐iii, B‐iii). (iv) The growing crystal then ‘pushes’ at the inner and outer iridosome membranes, molding them to the shape of the crystal (Figure 3B‐iv). The close fusion of inner and outer membranes makes it difficult to distinguish them in mature crystals (Figure 3A‐iii).

FIGURE 3.

Iridosome formation in fish, lizards and scallops. (A) TEM images of developing P. reticulata iridosomes (fish,). Adapted from Ref. [109] with permission 2005, Wiley (i) The elongated preiridosome is double membraned. The inner membrane (1) forms the crystal chamber (orange) with the outer membrane (2) delimiting the iridosome boundary. Intermembrane space; blue. (ii) Iridosome with partially formed guanine crystals. The crystal chamber membrane (arrow heads) closely binds to the growing crystal. Dm; granular material, possibly containing disordered guanine. (iii) Mature iridosome with displaced guanine crystals during sectioning. (B) TEM images of developing S. graciosus iridosomes (lizard). Adapted from Ref. [110] with permission 2005, Wiley (i) The preiridosome is double membraned (arrow; crystal chamber) and contains a non‐crystalline electron‐dense material (dm), possibly disordered guanine. (ii) A guanine crystal (pc) nucleates at the inner membrane edge as the electron dense disordered feature (dm) diminishes. (iii) Partially formed guanine crystal (pc). (iv) Mature guanine crystal with tightly bound crystal chamber membrane. (C) Cryo‐SEM images of developing iridosomes in P. maximus (scallop). Adapted from Ref. [112] CC‐BY 4.0. (i) A spherical preiridosome containing small intraluminal vesicles (yellow arrows) and disordered fibrils (white arrow). (ii) An ellipsoidal iridosome with two macromolecular sheets (white arrows) and larger intraluminal vesicles (yellow arrow). (iii) A guanine crystal nucleates between the macromolecular sheets, which template the growth of the crystal along the H‐bonding direction while capping growth in the π‐stacking direction. (iv) A mature guanine crystal stretching and reshaping the iridosome membrane, which ultimately condenses on the surface of the crystal. Scale bars: 200 nm. (A) Fish image Adapted from Ref. [116] (B) Lizard image Adapted from Ref. [117] (C) Scallop image Adapted from Ref. [118].

Later work on scallops (P. maximus) [112] and zebrafish (Danio rerio) [113] confirmed many of these features, revealing a fairly consistent cross‐taxa picture of guanine morphogenesis. In scallops, spherical pre‐iridosomes (Figure 3C‐i) elongate into ellipsoidal vesicles (Figure 3C‐ii), concomitantly with the formation of two intraluminal sheets—reminiscent of the crystal chamber observed by Morrison [110] and Gunderson [109]. An immature guanine crystal, then nucleates between the two sheets of the ‘crystal chamber’, which template nucleation and cap the (100) face of the guanine crystals (Figure 3C‐iii). These sheets appear to inhibit crystal growth along the preferential π‐stacking direction, directing the crystal to grow along the templates in the perpendicular, H‐bonding direction. Similarly, to Gunderson [109] and Morrison [110], the sheets straighten as they attach to the crystal. Eventually, the crystal contacts both sides of the iridosome membrane, which is pulled around the growth front of the crystal, adopting its faceted shape (Figure 3C‐iv). Eyal et al. [113], demonstrated a template‐directed crystal growth of guanine in zebrafish, where the straight, preassembled fibers are thought to be composed of amyloids. Revisiting Bagnara [93], both studies suggested that iridosome morphogenesis bears a striking resemblance to melanosome organellogenesis.

Despite differences in the details of iridosome architecture, studies on different species converge on a general principle—that preassembled macromolecules act as templates to generate thin, reflective guanine plates. This is also consistent with in vitro crystallization experiments which generated (100) plates in the presence of polymeric additives [60].

Another important observation from the scallop model was that the preassembled sheets align with the curvature of the eventual image‐forming mirror [49]. Thus, macromolecular templates may perform a secondary function in pre‐programing crystal orientation prior to nucleation [112]. Similar strategies are common in inorganic biomineralization, for example the mineralization of bone, where hydroxyapatite crystals nucleate and grow along prealigned collagen fibers [114, 115].

3.2.3. Unknowns and Perspectives

Much like inorganic biominerals, guanine crystallization is templated by a macromolecular matrix within the iridosome compartment. Crystal thickness is primarily controlled by the template's inhibition of crystal growth along the preferred π‐stacking direction rather than a physical ‘molding’ by the iridosome membrane. Indeed, re‐shaping of the iridosome by the growing crystal and not vice versa has been observed in many biological systems. Kimura [119] provided further evidence for the importance of interfaces in morphology control, demonstrating the formation of biogenic‐like β‐guanine plates when guanine was constricted at the air–water interface. The chemical identity of the biogenic templates is still a mystery with Morrison [110] and Gundersen's [109] work implying their lipidic nature and Wagner [112] and Eyal's [113] their proteinaceous character in their respective models. Determining the composition of these templates would elucidate whether specific lipids or protein sequences may direct nucleation, orientation and polymorph—analogous to the acidic proteins of biomineralization [120]. Another key unknown is how organisms control crystal size. The volumes of guanine crystals in different fish species vary by several orders of magnitude, presumably necessitating tight regulation of ER‐ and Golgi‐vesicle blebbing to modulate iridosome volume and guanine transport.

While confinement does not appear to have an ‘intrinsic’ effect on crystal morphology, further studies will elucidate the intriguing possibility that confinement may influence the stabilization of metastable β‐guanine or disordered phases confined in transport vesicles. Li. et al. [121] showed that guanine polymorphism/morphology could be controlled by modifying the local supersaturation in double emulsion compartments, providing a plausible mechanism for how this may be achieved in biology.

3.3. Crystallization From Transient Disordered Phases

A well‐known feature of inorganic biominerals is their crystallization from transient disordered phases [94, 122]. The formation of stable prenucleation clusters [123, 124] and metastable amorphous phases [66, 125] is afforded by the entropic gain resulting from the release of water molecules from the hydration shell [126]. The use of metastable amorphous precursors (e.g., ACC and ACP) enables; (i) compact storage of minerals facilitating rapid transport to the crystallization site (e.g., in vesicles), (ii) metastable amorphous solids can form from supersaturated biological fluids with low energy barriers and their crystallization is often fast [127] and easily triggered by environmental or biological cues, (iii) their deformable nature allows molding of biominerals into complex shapes [6], and (iv) influences mechanical properties [69].

The conformational flexibility of organic macromolecules often favors the formation of amorphous phases. However, for small, rigid organic biomolecules, like guanine, which possess strong, directional intermolecular interactions, and high diffusivity [128], a key question was whether they also might form through similar, non‐classical crystallization pathways.

3.3.1. Evidence for a Disordered Guanine Phase

Morrison [110] provided the first circumstantial evidence for a disordered guanine precursor, observing an electron‐dense, but noncrystalline material inside the crystal chamber in lizards, which diminished concomitantly with the nucleation of a guanine crystal (Figure 3B‐ii). The nucleation of the crystal some distance away, at the crystal chamber membrane, aligns with a dissolution‐reprecipitation pathway rather than a direct transformation from a disordered phase, observed in inorganic biomineralization [6]. Gundersen[109] similarly observed electron‐dense particles between the crystal chamber and outer vacuolar membrane (Dm, Figure 3A‐ii) in fish, with densities resembling the crystallizing material. Wagner [112] also frequently observed intraluminal vesicles (80–100 nm) closely bound to intraluminal sheets during iridosome development in scallops (Figure 3C i‐iii). We speculate these are transport vesicles trafficking disordered guanine precursors.

In their cryo‐SEM study, Levy‐Lior et al. [22] observed that crystal doublets in silver spiders (Tetragnatha extensa) were intercalated with a material which, unlike cytoplasm, was resistant to water sublimation. A similar material was present in spherical vesicles in white widow spiders (L. pallidus) [129]. PXRD analysis on a variety of spiders, exhibited a single broad peak at 2θ = 10° (λ = 1.54 Å), possibly originating from disordered guanine. Gur et al. [130] also observed a broad XRD peak at 3.2 Å in developing Koi fish (Cyprinus carpio) scales. Similarly, to the spiders, vesicles resistant to sublimation were observed close to the forming guanine crystal in cryo‐SEM. The authors also made an insightful hypothesis that the disordered phase may be “a nascent nematic‐like phase with short‐range order induced by the stacking forces” [130]. A recent study by Eyal, Deis, and Gorelick–Ashkenazi [131] found that immature iridosomes in developing zebrafish are comprised of a nitrogen‐rich material which is nondiffracting when interrogated by electron diffraction—further evidence of a disordered guanine or guanine precursor phase.

More insights on the structural nature of this phase were provided by Wagner [129] who used in situ synchrotron PXRD to show that immature, hatchling white widow spiders (L. pallidus) exhibit broad amorphous peaks centered at the (011), (012), and (100) positions of β‐guanine. This suggested that the disordered phase has short‐range order reminiscent of β‐guanine [78]. Finally, Goodenough [42] reported a nonetching, dense‐liquid‐like or amorphous phase in guanine‐forming vacuoles in green algae (Chlamydomonas reinhardtii), suggesting disordered guanine may also be used as precursor outside the animal kingdom.

This collection of observations on different organisms provided strong evidence that a disordered phase was somehow implicated in guanine crystallization, but the nature of this phase and the mechanism of its transformation into a crystal remained unknown.

3.3.2. Non‐classical Crystallization From a Disordered Precursor

In their PXRD work on spiders, Wagner et al. [129] observed that as organism development advances, sharp diffraction peaks of β‐guanine grow at the expense of the disordered component (Figure 4A). Cryo‐SEM showed that at the earliest development stage, iridosomes contain a granulated material together with separated guanine plates (Figure 4B). The ∼ 25 nm (100) plates, nucleate from the iridosome membrane (or more likely, crystal chamber [109, 110], Figure 4B) and grow in the H‐bonding direction toward the vesicle center (Figure 4B). The early granulated crystals reveal intense {h00} reflections in electron diffraction but few other Bragg maxima (Figure 4B), indicative of a nematic phase, possessing 1‐dimensional ordering along the π‐stacking direction, predicted by Gur et al. [130]. The arcing of the {h00} reflections indicates orientational disorder with preferred orientation along the stacking direction. This 1‐dimensional ordering or preordering is analogous to the protocrystallinity in synthetic ACC [134] and short‐range ordering in biogenic ACC [94, 135].

FIGURE 4.

The non‐classical guanine crystallization mechanism. (A) In situ μ‐spot WAXS patterns from L. pallidus hatchling to adults, showing emergence of β‐guanine Bragg maxima on a partially ordered phase. (B–D) Cryo‐SEM (upper) and electron diffraction (lower) of guanine crystals in iridosomes at different developmental stages of L. pallidus. (B) An immature granulated crystal, with platelets nucleating from iridosome membrane. ED pattern of an immature crystal with diffuse {h00} reflections. (C) A more mature crystal with more distinct platelet texture. The ED pattern exhibits 3D periodicity with better resolved {h00} reflections, though twinning and stacking faults exist. (D) Mature crystal with an intraluminal vesicle (yellow star). The crystal diffracts as a single crystal. (E) Time‐resolved Raman spectroscopy of guanine crystallizing from solution [132]. Plot of normalized peak areas of π‐stacking mode (105 cm^{− 1}, green) and H‐bonding mode (397 cm^{− 1}, blue). (F–I) Cryo‐SEM images of the platelet texture of developing guanine crystals in different organisms (F) L. pallidus (spider) [129], (G) P. maximus (scallop) [112], (H) A. beershebensis (lizard) [133] and (I) D. rerio (zebrafish) [113]. Insets in B‐D: TEM images of the respective crystals from which the ED was collected. Scale bars: 200 nm. (B‐D,F) Adapted from Ref. [129], CC‐BY‐NC 4.0. (G) Scallop image Adapted from Ref. [118] and Cryo‐SEM image Adapted from Ref. [112] CC‐BY 4.0. and (I) Adapted from Ref. [113] CC‐BY 4.0.

Later, the granulated material diminishes as the platelets grow (Figure 4C) and the {h00} reflections are better resolved (Figure 4C) due to increased orientational ordering along the π‐stacking. Other Bragg peaks, indicating the onset of 3D periodic ordering also appear but are streaked along a* due to disorder between adjacent H‐bonded layers, from stacking faults and crystal twinning.

Finally, in mature iridosomes the crystal platelets coalesce. Small intralumenal vesicles, presumably derived from colocalized Golgi, are still seen within the iridosome [129]. At this stage, the crystals diffract as coherent single crystals (Figure 4D). The relaxation of stacking faults and twinning defects suggests re‐orientation of crystal plates into crystallographic registry. Thus, in this system, guanine crystallizes in a non‐classical, multistep ordering process from a partially ordered precursor where periodicity first emerges in the stacking direction and, later in the H‐bonded plane. In vitro studies show this may be an intrinsic feature of guanine crystallization, with low‐frequency Raman modes from π‐stacking interactions emerging before H‐bonding modes during crystallization (Figure 4E) [132].

The platelet texture seen in white widow spiders recurs in guanine crystals across many taxa [112, 113, 129, 133] (Figure 4F‐I), where 10–20 nm H‐bonded plates coalesce and anneal into single crystals in an oriented attachment‐like process. This suggests a general crystallization mechanism which likely depends on the intrinsic structural properties of the guanine crystals rather than any biological regulation. Finally, thin‐tissue TEM images of mature crystals from the white widow spider reveal that the crystal plates are intercalated by an organic material (a meso‐crystal) suggesting a template directed process as observed in scallops and fish [112, 113].

3.3.3. The Nucleation Mechanism

With evidence of guanine's amorphous phase and crystallization mechanisms in hand, the next question was how this phase initially nucleates. Interrogating this question in situ in a biological system is currently impossible. Indri and Dietrich [132] thus used cryo‐TEM and MD simulations to elucidate a molecular level mechanism of nucleation from aqueous solutions. Guanine initially self‐assembles into flexible, 1D fibers, hundreds of nanometers long and ∼5 nm wide (Figure 5A). The fibers align, straighten and fuse in the lateral direction by release of intercalated hydration water (Figure 5B) in a ‘nearly oriented attachment’ [136, 137, 138, 139, 140] process. These fibers closely resemble those seen in calcium sulfate crystallization [141]. Further aggregation of fibers produces fibrous bundles (Figure 5C) where guanine molecules are π‐stacked along the long fiber‐axis but lack 3D periodic ordering. These bundles gradually develop a faceted character, displaying the characteristic (100) face‐twinning of β‐guanine (Figure 5D). Long‐rage 3D periodicity is only observed at relatively late stages of crystallization (Figure 5E).

FIGURE 5.

Multistep nucleation of β‐guanine. Adapted from Ref. [132] CC‐BY 4.0. (A–E) Cryo‐TEM of a crystallizing guanine solution. Flexible 1D fibers (A), align and straighten for form fiber assemblies (B). Fibrous bundles (C) form by nearly oriented attachment of coalesced fibers and begin to develop some faceted character (left inset). Right inset; electron diffraction patterns from the blue and yellow starred bundles exhibit strong (100) reflections. At later stages chevron‐faceted crystals (yellow arrows) emerge from the residual fibrous morphology (blue arrow). Finally, mature prismatic crystals exhibiting characteristic twinning (yellow line) are formed. (F–I) MD simulation snapshots of the largest guanine clusters at (F) 0.5 ns; single‐column stacks, (G) 2.5 ns; two‐column nuclei. (H) 20 ns; fiber elongation via particle attachment. (I) 650 ns; 25 nm long fiber. (H’) In‐plane molecular arrangement comparison of guanine fiber at 20 ns (containing a “defective” guanine in gold‐green), and β‐guanine (right). (J) 650 ns fiber cross‐section (orange layer) overlaid on β‐guanine (transparent). Orange dashed lines: domain boundary. Scale bars in cryo‐TEM micrographs (unless specified otherwise): 100 nm, and in MD simulations 1 nm.

MD simulations of earlier nucleation stages show that in the first few nanoseconds, short‐lived single‐column stacks of 5–6 guanine molecules assemble (Figure 5F), driven by hydrophobic interactions. These stacks then merge laterally, to form two molecule wide columns with alternating triple and double H‐bonded pairs (Figure 5G). The two‐column nuclei grow along the π‐stacking by single‐molecule addition and, in the H‐bonded plane, by particle attachment (Figure 5H). Further ripening and attachment events result in the formation of 1D fibers, similar in size to those seen experimentally (Figure 5I). During this attachment and ripening process, the distinct two‐point H‐bond motif (different from that in β‐guanine, Figure 5H′), in the early clusters progressively relaxes to that of crystalline guanine in a healing process. This indicates that different H‐bonding arrangements are favored in different size regimes − a key feature of multistep nucleation. The final fiber (Figure 5I) displays the same H‐bonded arrangement as β or α‐guanine but is composed of misoriented domains (Figure 5J) which originate the various particle attachment events—a possible source of twinning defects seen in mature guanine crystals.

3.3.4. Unknowns and Perspectives

Like inorganic biominerals, guanine nucleates and crystallizes in a multistep ordering process involving particle attachment, the evolution of intermolecular bonding interactions [132], healing of structural defects [132], oriented attachment and fusion of crystal plates [129]. In contrast to inorganic minerals though, disordered phases are not directly responsible for morphology control [112, 113]. The more likely function of the disordered phase is in facilitating the transport and accumulation of large quantities of insoluble guanine to the crystallization site, where crystallization from aqueous solutions would seem to unfeasibly slow and energetically costly. Given that amorphous guanine is extremely unstable and rapidly crystallizes in aqueous media [78], a key goal is to understand how it is stabilized in biological systems.

How do organisms overcome the extreme insolubility of guanine enabling its transport to the crystal or precursor phase? One intriguing solution was provided by Eyal, Deis, and Gorelick‐Ashkenazi [131] who showed the early iridosomes contain protonated guanine molecules in a disordered phase—inferred from the π*¹ to π*² peak ratios in XANES spectra. Confocal microscopy using pH‐sensitive dyes revealed that acidic iridosomes progressively neutralize during crystallization, suggesting that pH control may be used to accumulate guanine and trigger its crystallization. Perturbing proton pumps disrupted iridosome acidity and blocked crystal formation. Quantifying the pH would determine whether the lumen pH can attain a value close to the pK_a of guanine (3.3), where its solubility increases markedly. We note that even above this value, a fraction of guanine molecules will be protonated, and its solubility is elevated compared to neutral pH. Many outstanding questions remain regarding the influence of pH on guanine solubility, precursor composition, and crystallization kinetics, and these are poised to become the focus of future studies.

An alternative solution was proposed by Roy et al. [142] who showed in an in vitro study that the enzyme purine nucleoside phosphorylase (PNP) can be used to continuously convert (more soluble) guanosine to guanine in aqueous solutions, maintaining supersaturation and driving guanine crystallization. They suggest that a similar process operates in iridosomes, where soluble precursors may be enzymatically converted to guanine in situ—analogous to the oxidation of tyrosine to melanin (by tyrosinase) in melanosomes [143]. However, unlike tyrosinase, which is inside melanosomes, Deis et al. showed that PnP4a is only present in the cytoplasm [83]. The question arises, how is insoluble neutral guanine transported into iridosomes following its synthesis in the iridophore cytoplasm?

4. Concluding remarks

Tremendous progress has been made in understanding how organisms control the composition, polymorphism, morphology, and crystallization of guanine which dictate its fascinating optical properties. A synopsis of guanine biocrystallization and associated control strategies are schematically illustrated in Figure 6. As organic biomineralization enters its next phase, exciting challenges remain, including elucidating the intricate biochemical and genetic regulation of guanine trafficking and crystallization and the functions of molecular crystals in micro‐organisms. Another goal will be to develop suitable methods for the synthesis and scale‐up of bio‐inspired molecular crystals, as alternatives to toxic inorganic scattering materials in pigments, pharmaceuticals, and cosmetics. This has become especially important since new EU legislation preventing the used of TiO₂ in foods [144]. More widely, biology's crystallization tricks provide inspiration for controlling the properties of crystalline molecular materials—from the use of small‐molecule H‐bond additives and π‐stack blockers to the harnessing of amorphous phases. Such insights could also aid the search for drugs that prevent the formation of pathological molecular crystals.

FIGURE 6.

Schematic overview of guanine biocrystallization pathways and control strategies. Iridosomes form by the merging of ER and Golgi vesicles. These lysosome‐related organelles (LROs) have acidic lumens that facilitate guanine supersaturation for non‐classical crystallization from disordered precursors. Confinement, template‐directed growth, and purine dopants then generate diverse crystal morphologies with preferentially expressed (100) facets, aiding a variety of photonic phenomena.

Conflicts of Interest

The authors declare no conflict of interest.

Biographies

Shashanka S. Indri obtained his dual BS‐MS degree majoring in chemical sciences from the Indian Institute of Science Education and Research (IISER) Kolkata (master's thesis with Prof. Rahul Banerjee) in 2022. His research during this period laid the foundation for his interest in physical chemistry and self‐assembly. In the same year, he started his PhD in Chemistry under the supervision of Prof. Benjamin Palmer at Ben‐Gurion University of the Negev (BGU). His research focuses on fundamental aspects of guanine biocrystallization.

Avital Wagner completed her BSc (2018) and MSc (under supervision of Prof. Nachum Frage, 2020) in Materials Engineering at Ben‐Gurion University of the Negev (BGU), graduating summa cum laude in both. She earned her PhD in Chemistry at BGU (2024) under the supervision of Prof. Benjamin Palmer as an Azrieli Graduate Fellow. Her PhD investigated the formation mechanisms and morphology control of biogenic guanine crystals, for which she was awarded the Israel Chemical Society Prize for Excellence. Currently, Avital is a postdoctoral researcher in the Medical Biosciences Department at Radboud University Medical Center with Prof. Nico Sommerdijk.

Benjamin A. Palmer received his PhD from Cardiff University with Prof. Kenneth Harris before undertaking postdoctoral research with Profs. Lia Addadi and Steve Weiner at the Weizmann Institute, exploring biogenic molecular crystals. Benjamin started his independent group in 2019 as the Nahum Guzik Assistant Professor in Ben‐Gurion University. He is currently an Associate Professor in Chemistry at the University of Bristol and a Full Professor at BGU. His groups explore organic biomineralization and bio‐inspired materials. He is the recipient of the 2019 Azrieli Foundation Faculty Fellowship, the 2024 Israeli Chemical Society Tenne Prize and the 2025 Blavatnik Award.

Data Availability Statement

The data that support the findings of this study are available on request from the corresponding author. The data are not publicly available due to privacy or ethical restrictions.