**1. Introduction**

Collagens constitute a superfamily of long-lived extracellular matrix structural proteins of fundamental evolutionary significance, found in both invertebrate and vertebrate taxa. They are among the most studied proteins due to their important functions in mammals, including humans. In addition to their structural function in cartilage and skin formation [1,2], as well as in the biomineralization of hard tissues [3] including bone [4] and dentine [5], collagens are involved in the regulation of diverse cellular functions and processes. During the last 60 years, research into collagens has evolved from the discovery of the structure of collagen [6,7], through studies on its chemistry and biochemistry [8–10], to present-day applications in cell therapy [11], biomedicine [12–14], cosmetics [15], and the food industry [16]. A rod-like triple-helical domain is the typical structural element in all collagens. However, they differ in their size, dislocations of the globular domains and imperfections within the triple helix, self-assembly behavior, and functional roles. The classification of collagens is based on structural and functional features of vertebrate collagens. For example, 28 collagen types have so far been identified and characterized at the molecular level in mammals (see for review [1,17]). Collagens are also divided into subfamilies based on their supramolecular assemblies: fibrils, beaded filaments, anchoring fibrils, and networks [11]. Usually, the amino acid sequences in collagens are responsible for the corresponding functional properties: energy storage capacity, stiffness, or elasticity [18]. Even the type of amino acid motif within the tropocollagen molecule of a collagen can significantly affect its mechanical properties. Consequently, it can be hypothesized that the diversity of collagen polyforms determines their future functions, even within the same organism.

Marine vertebrate collagens have attracted scientific attention, mostly as products of fisheries [19]. In particular, fish-sourced collagens from skins and scales [20–22] have been studied and used as alternative collagen sources to avoid the potential risks associated with mammalian collagen due to bovine spongiform encephalopathy and the swine influenza crisis [23].

In contrast to marine vertebrate collagens, similar structural proteins found in marine invertebrates represent one of the most ancient protein families within Metazoa. Marine invertebrate collagens arose earlier than their vertebrate analogs, and possess diverse unique structural features, including very special structure–function interrelations. Collagens from poriferans, coelenterates, annelids, mollusks, echinoderms, and crustaceans have been discussed in detail in several review papers (e.g., [24–32]) and books (e.g., [33,34]). The limiting factors that have hindered progress in this field of research are the difficulty of purifying marine invertebrate collagens and their relative species-dependent complexity. However, there are more than enough examples in practically every order of marine invertebrates to inspire experts in materials science and biomedicine, especially because the similarities in structure and biosynthesis between vertebrate and invertebrate collagens appear to be more impressive than the differences [24].

Sponges (Porifera) are the most simple and ancient multicellular organisms on our planet, and mostly live attached to a suitable substratum (rock, sandy sediments) on the seabed. Poriferans diverged from other Metazoans earlier in evolutionary history than any other known animal phylum, extant or extinct [35], with the first fossilized sponge remnants found in 1.8 billion-year-old sediments [36–41]. The phylum Porifera is divided into four classes: Hexactinellida, Demospongiae, and Homoscleromorpha, with silica-based skeletons; and Calcarea, with a skeletal network made of calcium carbonates [42]. According to Exposito et al. [27], before the divergence of the sponge and eumetazoan lineages took place, the genes which were responsible for the synthesis of some kind of ancestral fibrillar collagen arose at the dawn of the Metazoa. The duplication events leading to the formation of the A, B, and C clades of the fibrillar collagens occurred before the eumetazoan radiation. Interestingly, the similarity in the modular structure of sponges and humans is preserved only in the B clade of fibrillar collagens. This phenomenon correlates well with the hypothesis of the primordial function of type V/XI fibrillar collagens in initiating the formation of collagen fibrils [27].

Different systems of terminology relating to poriferan collagens are found in the literature, as sponges also display considerable polymorphism with respect to their collagenous structures. The insolubility of most poriferan collagens has been the main obstacle to carrying out any detailed biochemical analysis. Studies on the morphology and nanotopography of the collagenous fibrils have shown that they are dispersed throughout the intracellular matrix within the skeletons of sponges. Cuticular structures have been found in some sponges, but their molecular composition has not been determined [43].

It was accepted very early that collagen fibers in sponges can possess quite different morphological features [44]. Gross et al. isolated two distinct forms of collagen from *Spongia graminea*, which they called spongin A and spongin B [29]. The first corresponds to fine intercellular collagen fibrils, visible only by electron microscopy. The second, spongin B, forms macroscopically-visible rigid fibers which are characteristic of keratosan demosponges [43]. This was probably the moment when the terminological divergence arose with regard to the term *spongin*, which was initially proposed by Städeler [45] to denote the skeletal fibrous matter of bath sponges, and was also used for spongins A and B defined by Gross et al. [29]. Up to the present, the authors of numerous publications—especially those on applications of spongin-based scaffolds in tissue engineering [31,46–51]—have used the term collagen for spongin, or even defined spongin as "collagenic skeleton" [52]. Very recently, Tziveleka et al. [53] studied collagen from the marine demosponges *Axinella cannabina* and *Suberites carnosus*, and proposed three different terms: insoluble collagen (InSC), intercellular collagen (ICC), and spongin-like collagen (SlC). It is worth noting that the isolation of each form of collagen from demosponges is based on the selectivity of the method used. Data on collagen extraction methods from diverse mineralized sponges (Hexactinellida, Demospongiae) and sponges which lack mineralized skeletons (the subclass Keratosa)—including yields of the extracted collagens—may be found in the relevant papers.

In this review, we focus on the structural diversity of collagens and collagen-like proteins in selected sponges, with particular focus on their origin, structural features, and applications in biomedicine and technology, including materials science and biomimetics. The review has the ambitious goal of providing thorough and comprehensive coverage of poriferan collagens (Figure 1) as a multifaceted topic with controversial hypotheses and numerous open questions. We begin with a brief description of spongins and their practical applications. Next, we examine the collagen IV-related proteins in diverse representatives of Porifera. Special attention is paid to *Chondrosia* sp. collagens and their applications in marine biotechnology, biomedicine, and cosmetics. Finally, we discuss the current state of work related to the unique hydroxylated collagen discovered in anchoring siliceous spicules of psychrophilic deep-sea glass sponges. We are optimistic that both the attempts to establish implications for poriferan collagens and the numerous open questions raised in this review will inspire the scientific community to carry out research into collagens and collagen-related proteins from sponges, as ancient and intriguing structural biopolymers.

**Figure 1.** Schematic overview of the collagens and collagen-like structural proteins of poriferan origin described in this review.

#### **2. Spongins as Enigmatic Structural Proteins in Sponges**

It is recognized that so-called spongioblasts—derived from the epithelium of sponges—are responsible for the formation of spongin. Minchin claims that the fibers of skeletal spongin are formed extracellularly; however, the cuticular spongin fibrils are of intracellular origin [19]. In contrast to such structural proteins as collagen, fibroin (silk), elastin, resilin, and keratin, the chemistry and molecular biology—including the sequences—of spongins so far remain unknown. It seems that spongin is the last enigmatic proteinaceous biopolymer, although it is of very ancient origin and has undergone more than 300 years of investigations. Spongin in the form of cell-free skeletons of diverse bath sponges (Figures 2 and 3) has been used for more than 3000 years [54,55] for painting, bathing, and cleaning, as padding for battle armor, for medical purposes, and as a vessel for drinking water [56]. A brief overview of the practical applications of spongin from bath sponges in biomedicine and technology in recent times is given in the next section.

A suggestion of a similarity between silk and bath sponge skeletal fibers was reported for the first time by Geoffroy in 1705 [57], and was based on his chemical experiments. After that, attention was paid to practical applications of sponges in pharmacology due to the presence of iodine in their skeletons. For example, in 1819, Andrew Fyfe—a professor of chemistry in Aberdeen—identified large quantities of iodine in the marine sponge *Spongia usta*, the "Coventry Remedy", which was used even in ancient China [58]. In 1841, bath sponges were described as those in which the essential base of the skeleton consists of keratose fibrous matter. At that time, the structural and chemical similarity between horny fibers of sponges and silk was again suggested by Croockewit [59]. It would appear that the horny matter of sponge is closely analogous to silk and related proteins, differing from them only in that it contains additional halogens. According to Croockewit, the chemical formula of horny matter must be as follows: 20(C39H62N12O17)+I2S3P10 [59]. Schlossberger [60], however, reported the very slight solubility of the fibrous matter in ammoniacal solution of copper hydroxide. Additionally, treatment with diluted sulfuric acid leads to the identification of leucine and glycocoll, in contrast to the isolation of tyrosine and serine from sericin under similar conditions. Städeler in 1859

obtained similar results [45] and introduced for the first time the scientific term *spongin* for this horny matter. Then, in 1864, von Kölliker [61] carried out the first histological studies on sponges, including investigations of the structural features of fibrous spongin. Diverse iodine-containing sponges and the matter termed as "*Jodospongin*" were discussed by Hundeshagen in 1895 [62]. The organic origin of iodine in bath sponge skeleton was suggested by Harnack [63]. He estimated the concentration of iodine in spongin at 1.1–1.2%, and demonstrated that superheated steam destroys the organic portion of spongin fibers completely, liberating iodine.

**Figure 2.** The mineral- and cell-free skeleton of commercial *Hippospongia communis* bath sponge is an example of a 3D spongin scaffold.

In 1898, Harnack isolated the "*Jodspongin*" and characterized it as an albuminoid-like product, containing over 8.5% iodine and 9.4% nitrogen [63]. In 1926, Clancey carried out a critical analysis of the literature to-date relating to the identification of spongin by other authors. In contrast to other physiologists, he suggested that the origin of the skeletal spongin in Euceratosa was not the same as that of the spongin which surrounds the spicules in the Pseudoceratosa [64]. At that time, the common bath sponge *Hippospongia equina* and the "Turkey cup sponge" *Euspongia officinalis* were the sponges most studied with respect to spongin. The results published in various papers [65,66] showed remarkable differences, due to insufficiently effective analytical methods and the use of commercial sponges that had been variously prepared and bleached. Consequently, different results on the chemical nature of spongins from particular species were obtained.

**Figure 3.** Scaning electron microscopy (SEM) image of anastomosed spongin fibers from the demosponge *H. communis*, which are organized as sets of unconnected structures with dendritic architecture.

For example, Clancey [64] isolated up to 7% of iodogorgonic acid besides the other amino acids in acidic hydrolysates of spongin. Clancey [64] did not identify hydroxyproline in spongin fibers of *Hippospongia equina* which had been treated with acid and alkali. It should be noted here that in natural collagen, a 3(S)-hydroxy-L-proline (3-Hyp) residue occurs together with a 4-Hyp residue, which is known to markedly increase the conformational stability of the collagen triple helix [67]. Hydroxyproline is found almost exclusively in collagen [8]. Thus, Clancey found a remarkably high amount of glutamic acid (18.4%), as well as 14% glycine, 5.7% proline, 2.8% tyrosine, 11% tryptophan or histidine, and a trace of cystine. Block and Bolling [68] presented the following results on the chemistry of spongin (Table 1).


**Table 1.** Amino acid composition of spongin.

The content of glycine (about 14%) in this spongin is significantly lower than in collagen (between 25% and 33%) [8]. Thus, until the identification of two different spongins by Gross et al. in 1956, spongin was recognized, for the most part, as a halogenated scleroprotein (see Table 2) [69–71] or neurokeratin-like protein [68] due to the presence of cystine.

**Table 2.** The chemistry of spongin according to Ackerman and Burkhard [69].


Consequently, it is very curious that the two morphologically-distinct forms of spongin fibers—termed spongin A and spongin B—were classified by Gross et al. [29] as members of the collagen family. This was probably because such an analysis was supported by electron microscopy and X-ray diffraction, and by their general amino acid pattern, including corresponding glycine and hydroxyproline content. Ratios of glycine to hydroxyproline were 1.6 and 1.8 for spongins A and B, respectively. The results obtained with small-angle X-ray diffraction and electron microscopy showed the diameter of the spongin A unbranched fibril to be on the order of 20 nm, with an axial period of about 650 Å. The large and branched fibers of spongin B were 10–50 μm in width and composed primarily of bundles of thin unbranched filaments less than 10 nm wide [29]. Both fiber types and the amorphous matrix contain hexosamine, hexose, pentose, and uronic acid. Glucosamine, galactosamine, glucose, galactose, mannose, fucose, arabinose, and uronic acid were identified chromatographically in both spongin A and in the amorphous substance. It was shown that spongin B contains a small amount of amino sugar plus glucose and galactose. In contrast to mammalian collagen, neither spongin can be dissolved at all by collagenase (*Clostridium hystolyticum*) or pepsin, nor were they dissolved to any appreciable extent in alkali solutions or dilute acid [29]. In a paper by Katzmann et al. [72], it was reported that spongin B accounts for over 70% of the dry weight of the bath sponge *H. gossypina*, and contains approximately 7% by weight of glucosylgalactosylhydroxylysine but a negligible amount of other sugars.

Recently, Langasco et al. [52] isolated glycosaminoglycans (GAGs) from sponginous skeletons of selected bath sponges. Total GAG content—expressed as μg hexuronate/mg dry weight—shows some variability among the tested species, being 0.171 ± 0.021, 0.367 ± 0.028, and 0.460 ± 0.081 for *H. communis*, *Spongia officinalis*, and *S. lamella*, respectively. The data obtained suggest that these sponge GAGs are structurally divergent from vertebrate GAGs [52].

Thus, it seems that spongin chemistry is made very complex by the presence of diverse halogens (I, Br) which have never been reported in natural collagens or keratins. This may explain the very high resistance of this proteinaceous biopolymer to enzymatic treatment. Its unique resistance to various enzymes—including amylases, lysozymes, trypsin, pronase, collagenases, and other proteases—is well reported [44,72]. On the other hand, in the natural environment diverse bacteria are able to destroy spongin enzymatically and lead to extremely high levels of damage to the structure of the spongin-based skeletal fibers (see for details [73]). The isolation and purification of such special "*sponginases*" remain a challenge for future research, and will provide a key way to obtain peptides that will be useful for detailed proteomic analysis and the sequencing of spongin.

Understanding of the nature and origin of spongins—especially in keratosan demosponges (the orders Verongiida, Dictioceratida, and Dendroceratida)—changed dramatically after the discovery of chitin as a second structural component of the skeletal fibers of demosponges in the order Verongiida by the Ehrlich Group in 2007 [74–76]. It was shown that anastomosing and macroporous skeletons of diverse verongiids are made of some kind of spongin–chitin biocomposites. The content of chitin in such composites ranges between 10% and 60% depending on the sponge species [76]. The isolation and characterization of chitin in these composites was possible due to the well-known resistance of chitin to dissolution in alkaline solutions [77–79], in contrast to spongins, which are quickly dissoluble in alkali [72,80]. Consequently, all publications prior to 2007 on spongins found in Verongiida sponges must be re-examined. The only existing and up-to-date classification of spongins is that proposed by Garrone in 1978 [43]. He states that the following types of spongins can be defined and discussed (see Figure 4). The first spongin is to be found in the form of spiculated fibers. These structures are associated with the endogenous mineralized skeleton of the sponge. It is also responsible for the formation of wide fibers which include only a very thin mineral element in the core. This kind of spongin is also resistant to mild acid or alkaline hydrolysis, as well as to pepsin and diverse bacterial collagenases. However, this spongin can be partially destroyed by cuprammonium hydroxide treatment at room temperature.

**Figure 4.** Diversity of spongins according to [43].

Second are the spongin fibers which form the skeleton of the keratosan demosponges: the abundance and compactness of the spongin and the almost complete lack of its own inclusions—which are replaced with foreign particles—testify to the originality of the spongin in this group. A typical example of such spongin can be found in the genus *Ircinia*, characterized by spongin fibers cored with foreign debris (sand microparticles) [81]. Recently, Castritsi-Catharios et al. [82–84] described the chemical elements and the physical properties of such skeletal spongin from diverse commercial sponges before and after chemical treatment.

The importance of the so-called basal spongin is evident for all sponges as sessile animals. In sponges with no organized internal skeleton, the organism is attached to the substratum by a more or less continuous layer of external spongin. This spongin is secreted by the basopinacocytes. The basal spongin is continuous with the internal spongin only in poriferans with an organized skeleton, formed either of spongin fibers or spiculated fibers. Due to the function of the basal spongin in such demosponges as *Chondrosia reniformis* (a species lacking spicules and internal spongin), the animal is attached strongly to its substratum. The basal spongin is discontinuous in erect sponges, where it forms the starting points of the internal organized skeleton. However, in the endemic fresh water demosponge *Lubomirskia baicalensis*, the holdfast which is responsible for attaching the sponge body to the hard substratum contains both basal spongin and chitin [77].

The extremely flexible and elastic organic structures which are morphologically similar to mineralized spicules are known as spiculoids [85]. They have been described in representatives of the genera *Darwinella* and *Igernella* (order Dendroceratida), where they are either free or partly joined to the fibers of the skeleton. They are compressible and can be easily torn apart. Finally, spongin may be responsible for the protection of gemmule shells. Gemmules are formed within the tissues of most freshwater and some marine sponges, and represent morphologically diverse asexual reproductive spherical bodies a few tenths of a millimeter to more than 1 mm in diameter, composed of a dense mass of identical cells and surrounded by an organic coat called the shell. The shell of gemmules is fortified with siliceous spicules and gemmoscleres, which are embedded into a matrix composed of both chitin and a collagenous protein. This collagen has been referred to as spongin [43].
