Sponges are very simple and ancient animals, but they have a remarkably complex genetic and molecular organisation (Sebé-Pedrós et al., 2018; Musser et al., 2021). The presence of sponge fossils from more than 600 million years makes palaeontologists consider this phylum as the oldest animal group still living on the earth (Erwin et al., 2011). Sponges are pluricellular organisms with a simple body structure not organised in specialised tissues or organs (Simpson, 1984). Nevertheless, the evolution from unicellular to multi-cellular animals needed a relevant implementation of gene regulation mechanisms (Brunet and King, 2017). Amazingly, there are many molecular analogies between the highly evolved multi-cellular animals and the simple multi-cellular sponges. The presence of RNA based regulatory factors like long noncoding RNA (Gaiti et al., 2018), or signal cascade translation systems based on inflammatory pathway response (Pozzolini et al., 2016; Pozzolini et al., 2017) and growth factors (Pozzolini et al., 2019) are examples showing the great molecular similitudes between sponges and other higher animals. The simplicity of sponges’ body, the direct contact of their cells with the external environment (sponges filter water that permeates in all the body, through a very well organised channels system) and their sessile status induced on the other hand the need to protect from micro-organisms infections and from other animals’ predation. This requirement is managed by the production of a great amount and variety of natural compounds, mainly supplied by symbiotic microorganisms (Brinkmann et al., 2017). All these features make sponges “unique” in the animal kingdom: a fascinating balancing of body simplicity and molecular complexity.

One of the most peculiar aspects of sponges is the structural organisation of their skeleton. The sponge body is supported by organic extracellular matrix (mainly collagen or/and “spongin” in some cases mixed with chitin) (Pozzolini et al., 2012; Ehrlich et al., 2018; Jesionowski et al., 2018) and in many species by inorganic structures called spicules, formed by amorphous silicon dioxide or, in Calcispongiae, calcium carbonate (Simpson, 1984; Ehrlich et al., 2011). Siliceous spicules have different shapes and dimensions with an extremely ordered organisation at a nanometre level (Şen et al., 2016). Biosilica is produced by Hexactinellida, by many Demospongiae and by some Homoscleromorpha (Simpson, 1984). There are many studies on the structural organisation and chemical composition of the biomaterial forming sponge spicules (see for overview Schoeppler et al., 2017; Wysokowski et al., 2018; Görlich et al., 2020; Pisera et al., 2021) and at present time, it seems clear that the features and the properties of biosilica depend on the composition of this material. Sponge biosilica is, in fact, a composite material where organic biomolecules contribute to its synthesis in environmental conditions, to the shape determination of spicules and to their technological properties (Khrunyk et al., 2020), but it is not easy at all to study the dynamics of spicule building. To satisfy this need, a multidisciplinary approach is necessary. Advanced material science technology to analyse in detail the structure and the properties of biosilica and good animal and cellular models to explain the molecular and genetic rules of spicule building are both necessary. At present, Hexactinellida have been mainly studied by material scientists: the production of large and long spicules (up to 2 meters long in Monoraphis chuni) (Müller et al., 2007; Ehrlich et al., 2008; Pisera et al., 2021) and their intriguing technological properties as optic fibre (Cattaneo-Vietti et al., 1996), were the main reasons for the choice of this class of animals. Demospongiae are instead approached by sponge cell biologists and molecular biologists, for the possibility to manage in controlled environment living animals and their sponge cell culture (Valisano et al., 2006; Pozzolini et al., 2014). It is also now clear that Hexactinellida and Demospongiae composite silica-based spicules apparently have common features but also very relevant differences in their composition (Müller et al., 2011; Shimizu et al., 2015). Definitively, many questions are still open to understand exhaustively the biological rules of biosilica production. Specifically, in Demospongiae, siliceous spicules are most commonly formed by monoaxonic or tetraxonic needle-like structures, from some μm to 1-2 mm long (Hooper and Van Soest, 2002). In the Demospongiae Tethya aurantium was for the first time sequenced and cloned silicatein, one key protein involved in the biosilicification process (Shimizu et al., 1998; Cha et al., 1999). In this last two decades, many studies approached from different points of view the role of silicatein in Demospongiae silica biosynthesis as well as its biotechnological applications (for an exhaustive and complete review see Shimizu and Morse, 2018) and animals and 3D cells models have been also developed to study the dynamics of spicule formation (Müller et al., 2006; Valisano et al., 2012; Müller et al., 2013a). In all these works, silicatein was indicated as the main molecular actor of biosilica formation. Silicatein is indeed a very peculiar protein: it has a remarkable sequence homology with the lysosomal protease family of cathepsins, but it is water-insoluble, and it has an enzymatic active site with serine in substitution to cysteine (Shimizu and Morse, 2018). Its insolubility, due to a higher number of hydrophobic regions and different post-translational modifications compared to cathepsins (Armirotti et al., 2009), determines its self-assemblage properties (Murr and Morse, 2005). This evidence clearly supports the assumption of the structural role of silicatein in the formation of the axial filament of spicules as well as its being a key component of biosilica composite structure. The enzymatic active site suggests, in the meantime, the catalytic role of this protein, even if there is not a definitive agreement among scientists on the mechanism. At present time, in fact, two main hypotheses are supported in literature: silicatein as an enzyme able to catalyse the biosilica precipitation from some organosilicon compound (Cha et al., 1999; Povarova et al., 2018a; Povarova et al., 2018b) and silicatein as enzyme able to catalyse the biosilica polymerisation from orthosilicic acid (Müller et al., 2013b). Both models are apparently sustainable and silicatein can be considered one of the major contributors for the biosilica condensation in environmental conditions. It is much more difficult to understand the molecular keys of spicule growth and spicule shape control. Silicatein probably is involved in growth control and shaping, mainly in Demospongiae (Müller et al., 2009), but many other components like collagen, chitin and actin seem to play some fundamental roles (Ehrlich, 2010; Ehrlich et al., 2010; Ehrlich et al., 2016; Ehrlich et al., 2022).

A recent work (Riesgo et al., 2015) further considered the scenario concerning silicatein: a comparative analysis of transcriptomes and genomes datasets puts clearly in evidence that not all biosilica forming sponges have/express silicatein genes. Mainly Hexactinellida and Homoscleromorpha sponges do have not this protein and some silica forming Demospongiae apparently do not have silicatein (Chondrilla nucula). In addition, sponges that do not produce biosilica seem to express silicatein genes. Another interesting and relevant aspect is the evidence of a remarkable expression of cathepsins (specifically cathepsins B and L) in biosilica forming sponges, whose functional role is to be better clarified (Riesgo et al., 2015). In this key work the authors, on the basis of transcriptome analysis results, hypothesized a lysosomal origin of the “silica deposition vesicle” of Demospongiae (Riesgo et al., 2015). It is evident that the biosilicification mechanisms in sponges are far to be understood, and one of the major gaps of knowledge to be filled is to improve the description of the organic content of Demospongiae spicules.

To accomplish this objective, in the present work, an integrated transcriptome and proteome analysis is proposed as a strategy. Here, for the first time, using as experimental model the species Petrosia ficiformis, is described the protein content of Demospongiae spicules by a proteomics analysis followed with protein sequence identification based on the transcriptome dataset matching.

The biomineralization process in Porifera still has several unsolved questions. Most of the scientists studying biosilicification in demosponges attribute to silicatein a key role in the skeleton forming process, (see the recent review, Shimizu and Morse, 2018). Indeed, silicatein is very abundant inside spicules and its distribution is intimately mixed with the inorganic part to form a composite (Müller et al., 2013; Shimizu and Morse, 2018). Nevertheless, at present, this peculiar protein, with its proposed double role (enzymatic and structural), in our opinion is not enough to fully explain all the aspects of the biosilicification mechanism in these animals. In biological controlled biomineralization, the molecular actors are very numerous, involving proteins and other biomacromolecules (polyamines for example), whose role in the process is often very difficult to identify (Wysokowski et al., 2018). Proteomic analysis in these cases give a relevant contribution to identify the various component of an organic compound and in the past, it has already been used to describe the post-translational modification of silicatein (Armirotti et al., 2009). In the present study we have improved the strategy looking again inside spicules using an integrated approach based on proteome and transcriptome analysis. Recently, similar methodologies were used to describe the organic matrix content of diatoms and corals (Kotzsch et al., 2016; Peled et al., 2020).

Using this technique, we have found together with silicatein a list of proteins until now never described before inside biosilica. The P. ficiformis transcriptome availability (Riesgo et al., 2014) allowed in the present work a fine identification of proteins inside spicules ( Figure 2 and Table 1 ) but, for many of them, it is not easy to attempt a possible explanation of their role in the process. Previous scientific publications demonstrate that the spicule formation in demosponges begins inside cells, but it continues outside, and the process pass through the production of biosilica building units inside cells successively secreted outside (Wang et al., 2011). A last century study (Lévi, 1963) and a more recent description of spicule formation by Schröder and co-workers (Schröder et al., 2007) show the formation of specific intracellular vesicles in which biosilica is precipitated. In accordance with what is described in many other biomineralization processes, the building blocks are organized intracellularly in appropriate compartments where the chemical conditions for mineral precipitation are created and maintained. The ultrastructural analysis reported in the work of Schröder and co-workers gives some explanation of the process: silica is released in extracellular granules outside the sclerocite; these granules are rich in silica, and they are named silicasomes. Morphological analysis showed that silicasomes positively react with antibodies against silicatein, and this reaction allowed to follow the intra- and extracellular evolution of these organelles (Schröder et al., 2007). This evidence suggests a subcellular process involving some systems related to endosome/exosome or microvesicles pathways. Furthermore, it is known that the acidic environment favours biosilica precipitation (Shimizu et al., 2001) and, for this reason, also lysosome or lysosome-related organelles may be involved in the formation of silicasomes. In this conceptual direction it is also placed the co-expression of silicatein and cathepsins recently described in sponge tissues by Riesgo and co-authors, suggesting some specific roles of these proteins in the biomineralisation process as well as clearly remarking the hypothesis of a silica deposition environment of lysosomal origin (Riesgo et al., 2015). In this study Riesgo and co-authors made an accurate analysis of the evolution of cathepsin genes among metazoan and many of them have been described in sponges. Specifically, in P. ficiformis transcriptome they documented the expression of Cathepsin B, L, and D as well as 4 different silicateins (Riesgo et al., 2015). This information derives from comparative transcriptomic analysis of sponges, without any further insights into protein localisation. Our findings demonstrate, for the first time, the presence of cathepsins and silicatein inside the biosilica, fully confirming the lysosome hypothesis. Additionally, in our research results many other proteins are for the first time identified inside P. ficiformis spicules after inorganic silica removal, almost half of them is provided with a signal peptide and shows a strong similarity with proteins normally found inside lysosomes (see Table 1 ). The scenario is however very complex: from one hand the common lysosomal origin of the majority of proteins listed in our work is something coherent and easy to discuss on the basis of the morphological evidence published (Schröder et al., 2007) and of the evolutionary study cited (Riesgo et al., 2015); on the other hand, it becomes very difficult to attribute a specific role in the process for many of the proteins here identified.

The first aspect emerging from the analysis of biosilica protein content shown in Table 1 of our results is the demonstration of presence of some isoforms of cathepsin family together with silicateins, data strongly congruent with the previously published hypothesis of Riesgo (Riesgo et al., 2015). Riesgo and co-authors already suggested a possible role of cathepsins in the biosilification process (Riesgo et al., 2015). In particular, the fact that one silicatein cited by Riesgo has cysteine inside the active site is suggestive of a more complex scenario where some cathepsins can play some role (Riesgo et al., 2015). In addition, it also known that mutants of L cathepsin produced in recombinant systems confirm the existence of structural issues other than active sites in the biosilica biosynthesis in silicateins and cathepsins (Fairhead et al., 2008). On these bases, the changing of the active site from Cys to Ser is not the only molecular key of the mechanism, but also other regions of the protein can have a role in the catalysis and/or in the biosilica organisation and/or in other proteins assembling (Fairhead et al., 2008). Despite all these considerations, definitively it is not easy to define the exact role of various cathepsins in the biosilica precipitation environment: one additional possibility we suggest is that they can be involved in the specific activation of other proteins relevant in the process (i.e., cleaving signal peptides and\or regulatory peptides) but more studies must be performed to better understand if these roles are related only to catalysis or also to structural organisation. Only future experimental demonstrations based on selective enzymatic inhibition or gene silencing approach in cellular models can give a final answer.

In Figure 3 and in Supplementary File 2 are listed some details of domain and sequence homology also of the other proteins found and, in the results, we have underlined the enzymatic roles of the homologous polypeptides as described in other organisms. But, if it is difficult the possible explanation of cathepsins and silicateins roles, the discussion of other proteins here found in the biosilica matrix only on the basis of their structural similarity with other well-known proteins of lysosomal origin is much more speculative. A further interesting aspect is that many of these proteins have well defined enzymatic domains (see Figure 3 ), differently from other situations (corals and diatoms for example), where a considerable numbers of proteins with structural functions and apparently lacking catalytic domains have been found (Ramos-Silva et al., 2013; Kotzsch et al., 2016). In our research results we conversely identified aryl sulphatase, glycosyl hydrolase, beta-hexosaminidase, all typical proteins found inside lysosomes. Also, proteins involved in the lysosome cholesterol transport have been found (see Table 1A and Figure 2 ). Only three proteins are of unknown origin, (with some low similarities to acetylglucosamine transferase, polysaccharide lyase and one lipoprotein).

The presence of enzymes related to glycosaminoglycans (GAGs) and polysaccharides catabolism inside biosilica is indeed intriguing. The sugars, the polyol and the GAGs interaction with silicon recently suggested by Nielsen (Nielsen, 2014) and the regulation of connective tissue formation in mammals influenced by silicon is suggestive of some biochemical mechanism involving silicon and GAGs and other polysaccharides in mammalians (Götz et al., 2019). The presence of these enzymes inside of silica sponge spicules suggest a possible sugars-silicon interaction originally involved in biomineralization. A future explanation of their role can give important answers not only in biosilicification process in sponges but also in cartilage and bone formation in mammalians and humans. The results here shown undoubtedly confirm the lysosome hypothesis since the list of proteins here found are in their majority clearly related to lysosomes, but at the same time opens the door to many unanswered questions, specifically on the metabolic management of silicon inside cells. The silicon uptake in sponges is under the control of specific transporters (Schröder et al., 2004; Maldonado et al., 2020), but the intracellular destiny of silicon is not clear. The hypothesis of one organosilicon compound already emerged from the beginning of studies of silicatein (see Shimizu et al., 1998) but, at the same time no specific ideas have already been confirmed. Bio-organic chemistry works suggested the general possibility to have silica derivatives with organic molecules of biological relevance (Kinrade et al., 1999; Vis et al., 2020). The presence of many enzymes related to sugar metabolism could be suggestive of silicon-organic compounds specifically involved in this process, while cathepsins and silicateins can play a role both in the enzymatic control of biosilica precipitation and in the structural organisation of spicules.

Much more difficult is to give some functional hypothesis of other proteins. Another finding is a protein with pyrrolo-quinoline quinone (PQQ) dehydrogenase domain. This family of proteins are normally present in prokaryotes but recently they have been described also in Eukaryotic, involved in sugars oxidoreductases process (Matsumura et al., 2014).

In this scenario, the apparent absence of silintaphins (Wiens et al., 2009) and galectins (Schröder et al., 2006) is surprising. In the first case, also inside the template transcriptome (Riesgo et al., 2014) we have not found silintaphins. We cannot currently confirm if this is due to incompleteness in transcriptome database used in this study or to some specific genetic differences that occurred in this sponge species. Different considerations should be made regarding galectins. In this second case, in the transcriptome database used, the presence of galectin transcript is confirmed, but any traces of these proteins in the proteomic analysis was not found, despite the presence of galectins in the biosilica matrix was proposed using specific antibodies in a previous work on S. domuncula (Schröder et al., 2006). In our experimental conditions, we performed a strong acid spicules cleaning before silica dissolution. Differently, in the study of Schröder and co-workers the extraction procedure was completely different and milder (Schröder et al., 2006) and maybe some proteins involved in the biosilica building not intimately mixed inside biosilica in our approach can be missed. The methodological approach in the extraction procedure without doubt can affect the results. Here we have used this very strong cleaning process (cold HNO₃/H₂SO₄) of spicules in order to avoid contaminations. MS analysis, in fact, is highly sensitive and it is not easy to evaluate proteins at the quantitative level. This sensitivity on the other hand allows to identify many proteins not found before with different analytical approaches. Here we decide to focus our attention to those proteins very intimately mixed with the inorganic silica lattice.

Biomineralization in sponges is a process not completely explained. In this work, using a strong spicules cleaning procedure, we have extracted the organic proteinaceous material embedded inside biosilica spicules from the marine sponge P. ficiformis and analysed by MS analysis. The results obtained confirmed the hypothesis of lysosome origin of silicasomes, due to the identification of various cathepsins together with silicateins and surprisingly we found many proteins similar to lysosome enzymes, several of them related to polysaccharide and sugar metabolism. Taken together, these results open a new perspective in the study of biosilica forming in demosponges, suggesting a more complex scenario where sugars and maybe lipids can have key roles.

The original contributions presented in the study are publicly available. This data can be found here: https://doi.org/10.5281/zenodo.6033766.

Conceptualization: MP and MG. Methodology: MP, IM, SG, CO, ET, and MG. Investigation: MP, IM, SG, CO, and ET. Writing: MP and MG. Review and editing: MG and ET. Supervision: MP and MG. Funding acquisition: MP, IM, and MG. The manuscript was written through contributions of all authors. All authors have given approval to the final version of the manuscript.

This work was partially supported by University of Genova funds to MP and MG. IM acknowledges funding from the Czech Science Foundation (grant number 20-03899S).

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.