*2.4. Isoelectric Point Determination*

The SF-InSCs were subjected to titration for the determination of the acid-base properties and the isoelectric point. The titration curves are shown in Figure 3. The pH of freshly-prepared InSC dispersions were 3.48 and 3.54 for *A. cannabina* and *S. carnosus*, respectively. After the HF treatment, the pH of the SF-InSC dispersions were slightly altered (3.66 and 3.34 for *A. cannabina* and *S. carnosus*, respectively). These values are lower than those reported for *C. reniformis* [26], probably due to the higher content in acidic amino acids (aspartic acid or glutamic acid), as also supported by the high contents of Asx and Glx found in both sponges from the amino acid content analysis (Table 4). The isoelectric point was calculated to be approximately 6.7 and 6.3 for *A. cannabina* and *S. carnosus*, respectively. These results are in agreement with previously-reported data determining the isoelectric point of insoluble collagen at pH values around 7 [26].

**Figure 3.** Titration curves of insoluble collagen after spicules removal (SF-InSC) isolated from *A. cannabina* (**A**) and *S. carnosus* (**B**).

#### *2.5. Amino Acid Profile*

The amino acid composition of collagen is one of the key factors affecting its properties. Therefore, the amino acid profile of the ICCs from both sponges was determined (Table 4). Their composition was analogous to that described for the sponges *G. cydonium*, *C. reniformis*, and *I. variabilis* [23–26]. Glycine (Gly) was found to be the major amino acid in both ICCs with 257 and 295 residues/1000 residues for *A. cannabina* and *S. carnosus*, respectively. This result is in accordance with the Gly-X-Y amino acid model in which Gly occurs in every third position. Relatively high contents of aspartic acid (Asx; 100 and 94 residues/1000 residues), glutamic acid (Glx; 82 and 84 residues/1000 residues), alanine (Ala; 72 and 89 residues/1000 residues), and proline (Pro; 58 and 56 residues/1000 residues) were observed for *A. cannabina* and *S. carnosus*, respectively. Both ICCs presented the characteristic high threonine (Thr) and serine (Ser) content (approximately 6% each) and low lysine (Lys) and hydroxylysine (Hyl) content, previously reported for *Ircinia* [23]. Low Hyl content (5–6 residues/1000 residues) has also been reported for acid- and pepsin-soluble collagens isolated from shark skin [39]. Moreover, the sum of Thr and Ser of both sponges' ICCs is similar to that of collagens reported for lower vertebrates and invertebrates. Additionally, no differences in the Lys content of the two different ICCs were observed, as already indicated from the same absorption bands in the IR spectra at 2922 cm−<sup>1</sup> attributed to amide B (Table 3) [37].

**Table 4.** Amino acid composition (residue/1000) of intercellular collagen (ICC) isolated from *A. cannabina* and *S. carnosus*.


<sup>1</sup> Asx: Asp + Asn. <sup>2</sup> Glx: Gln + Glu.

Compared to *S. carnosus*, ICC from *A. cannabina* contained higher amounts of methionine (Met), phenylalanine (Phe), leucine (Leu), and isoleucine (Ile), but lower amounts of Gly, Ala, and hydroxyproline (Hyp). The percentage of the remaining amino acids is in relatively good agreement to the above-mentioned studies, especially the amounts of Asx, Glx, Pro, His, and Ala. The number of sulfur-containing Met residues was significantly higher in the ICCs of both sponges (Table 4), as compared to collagen from porcine dermis (6 residues/1000 residues) [43].

Nevertheless, the overall percentages of Hyp were lower than those reported for other sponges [22]. Imino acids are involved in hydrogen bonding, therefore affecting the stability of the collagen triple helix and its thermal behavior [37,39,44]. The imino acid content value is usually lower in marine collagens in comparison to mammalian collagens, resulting in a lower thermal denaturation temperature [33].

The ICCs from both sponges contained approximately 12 tyrosine (Tyr) residues per collagen molecule, indicating that their nonhelical telopeptides, where all of the Tyr residues are located, were intact [45]. The reduced values for Gly, Hyp, and Hyl can also be attributed to the existence of glycoproteinaceous impurities, known to be strongly associated with collagen [46].

#### *2.6. Thermal Behaviour*

It is well established [47] that upon increasing temperature, thermal denaturation of collagen is taking place, during which hydrogen bonds break and helices unfold, leading to the formation of collagen coils. This process is accompanied with appreciable heat absorption and can, therefore, be monitored with DSC. Indeed, the DSC curves of the hydrated collagen samples (Figure 4) clearly indicate two major endothermic peaks. Previous DSC studies also revealed collagen's bimodal transition and concluded that the higher temperature peak was due to the helix-coil transition of collagen (denaturation of collagen), while the lower temperature peak originated from the breaking of the hydrogen bonds between collagen molecules or the defibrillation of the solubilized collagen fibrils [48,49]. This is attributed to the fact that the inter-triple helix hydrogen bonds responsible for the fibrillation are easier to break than the intra-triple hydrogen bonds that are responsible for helix formation [48].

In the present study, the thermal behavior of the ICCs isolated from *A. cannabina* and *S. carnosus* were monitored after removal of the entangled glycoconjugates [46]. The yield of the described procedure was 38% and 46% (*w*/*w*) for *A. cannabina* and *S. carnosus*, respectively. The low endothermic transition had its peak maximum transition temperature (Tmax) at 25.4 ◦C (Δ*H* value 1.27 J g−1) and 32.9 ◦C (Δ*H* value 5.74 J g−1) for the ICCs from *A. cannabina* and *S. carnosus*, respectively (Figure 4). The high temperature endothermic peak had a Tmax of 44.6 ◦C (Δ*H* value 0.37 J g−1) and 51.6 ◦C (Δ*H* value 17.65 J g−1) for the *A. cannabina* and *S. carnosus* ICCs, respectively (Figure 4). As is clearly evident from the examination of both reversing and non-reversing components of the thermograms, the total heat flow for the thermal denaturation of collagen involves a significant non-reversing component, while the reversing component is negligible. This is in line with previous studies that showed that collagen denaturation endotherms in fibers and in basement membranes are governed by an irreversible rate process [50,51] and not by equilibrium thermodynamics, as previously hypothesized. Given the irreversibility of the process within the time frame of temperature modulation (60 s), the transitions are registered as essentially a non-reversing event in temperature-modulated differential scanning calorimetry (TMDSC), although, in general, unfolding of proteins is a complex phenomenon that encompasses both reversible and irreversible steps [52].

**Figure 4.** Temperature modulated DSC data of intercellular collagen (ICC) isolated from *A. cannabina* (**A**) and *S. carnosus* (**B**). The total (- - -), non-reversing (—) and reversing heat (-·-) flows are presented (curves are shifted vertically for clarity).

A rather low Tmax value, as that observed for the ICC from *A. cannabina*, was reported earlier for collagen isolated from edible jellyfish (26.0 ◦C) [53]. On the other hand, Tmax values around 31 to 33 ◦C, as that measured for the ICC from *S. carnosus*, have been observed for an array of collagens isolated from tropical fish [34,54]. Moreover, the ICC from *S. carnosus* exhibited a higher Δ*H* value (5.74 J g−1) than that of *A. cannabina* (1.27 J g−1). It is widely accepted that Tmax is directly correlated with imino acid content, body temperature of the specimen, and environmental temperature [55,56], whereas the enthalpy change (Δ*H*) can be influenced by molecular stability, directly correlated with the amino acid sequence in collagen.

In our case, the ICCs from both sponges contain low amount of iminoacids (Table 4) in comparison to that of terrestrial organisms (approximately 200 residues/1000 residues), with the ICC from *A. cannabina* displaying the lowest content (96 residues/1000 residues). The observed difference between the Tmax of the ICC samples from the two sponges could, therefore, be attributed to the imino acid content difference, and especially to the Hyp content difference. This phenomenon might also be related to the superior stability of the ICC from *S. carnosus*, due to the high content of the Gly-X-Y sequence, as confirmed by the elevated percentage of Gly (17.9% vs. 15.0% *w*/*w* for *S. carnosus* and *A. cannabina*, respectively). This, in agreement with previous reports [33,57], might be an additional justification for the high value of Tmax despite the low amount of imino acids. Additionally, as previously reported [58], the high Asp (pK ≈ 3.9) and Glu (pK ≈ 4.3) content can contribute to ion pair formation with the basic residues at neutral pH, resulting in increased stability, which might partially compensate for the decreased stability deriving from the low Hyp content (Table 4) [59]. Another possible reason might be the intensely-localized sulfur bonding interactions associated with the higher Met content [43].

Finally, an additional low temperature endothermic peak (Tmax = 17.9 ◦C, Δ*H* = 1.65 J g−1) was observed in the case of the ICC from *A. cannabina*. In contrast to the previous transitions discussed, the examination of both the reversing and non-reversing components of this specific transition suggests that this process is, to a great extent, reversible. Taking into consideration that during the denaturation of small proteins (for instance lysozyme) [60] the reversible unfolding has the largest contribution, whereas the irreversible process still remains well detectable, we tentatively ascribe this low temperature transition to the denaturation of small molecular weight collagen species that are present in this sample.

The CD spectra of the ICCs from the two sponges in the region of 190 to 250 nm are depicted in Figure 5A,B. Both samples showed a rotatory maximum at about 221 nm, a minimum at 193–196 nm, and a consistent crossover point (zero rotation) at about 212 nm. These spectral characteristics are typical of a collagen triple-helix structure [61–63]. The corresponding mean residue ellipticities, [*θ*]221, as a function of temperature, are shown in Figure 5C,D. The [*θ*]221 values decreased with temperature due to decomposition of the collagen triple helical structure, and indicated denaturation temperatures of 24.3 ◦C and 28.2 ◦C for the ICCs from *A. cannabina* and *S. carnosus*, respectively, in good agreement with the obtained results from the conducted DSC studies.

It has been earlier shown that thermal denaturation temperature of collagens from different sources correlates directly with the imino acid (Pro and Hyp) content [43,64]. Actually, higher imino acid content facilitates intra- and intermolecular crosslinking resulting in a more stable triple helical structure of the collagen molecule [44]. A good linear correlation was observed earlier when measured denaturation temperatures were plotted against the corresponding numbers of Hyp residues, this effect being less pronounced with respect to the Pro content [43]. The amino acid composition analysis of the investigated sponges (Table 4) confirms the above observations, since *A. cannabina* presents a lower Hyp, but equal Pro, content as compared to *S. carnosus* resulting, therefore, in a concurrently-reduced Td. Interestingly, cold-water fish collagens have low Td since their imino acid contents are very low [65], in contrast to the Td of skin collagen of terrestrial mammals which are 37 ◦C and 40.8 ◦C, respectively [43], both possessing high imino acid content.

**Figure 5.** CD spectra in the region of 190–250 nm (recorded at 20 ◦C) and temperature effect on the CD spectra at 221 nm of intercellular collagen (ICC) isolated from *A. cannabina* ((**A**) and (**C**), respectively) and *S. carnosus* ((**B**) and (**D**), respectively).

#### **3. Discussion**

The presence of collagen in freshwater, as well as marine sponges, was unequivocally established more than 50 years ago by the work of Bronsted and Carlsen [66] and Gross and his coworkers [22]. Since then, many investigations regarding the fine structure and physicochemical properties of marine collagen have been performed. However, to the best of our knowledge, such extensive studies on the collagenous profile of sponge material have been conducted only on *C. reniformis* and *Ircinia* species [23,25–27,29]. In this context, the main purpose of the current study was the morphological characterization of various isolated collagenous materials (InSC, ICC, and SlC) from *A. cannabina* and *S. carnosus*, while further biochemical and biophysical characterization was undertaken only for the ICCs, given their relatively higher solubility and purity.

It has also been proven that in Demospongiae, collagen, constituting exclusively the intercellular organic framework, amounts to approximately 10% of the total organic matter [27,67]. In the present study, collagen content was experimentally calculated to amount for the 12.6% and 5.0% dry weight of *A. cannabina* and *S. carnosus*, respectively. The co-isolation of collagen with spicules is justified by the spicule formation procedure, generally accomplished by specialized cells that supply mineral ions or organic macromolecular particles, primarily consisting of proteins, carbohydrates, lipids, and seldom by nucleic acids [68].

Furthermore, the aforementioned characteristic insolubility has prevented the determination of the thermal behavior of sponge collagens. To our knowledge, only a few efforts have been made to determine their thermal behavior [25,69]. Our results corroborate to the existing knowledge that the thermal stability of marine collagens, which exhibit lower denaturation temperatures due to their lower content of imino acids, is generally lower than that of mammalian collagens. The low denaturation temperature of sponge collagen may also reflect the ambient temperature in which marine organisms live [70]. Moreover, the thermal stability of collagen is also directly correlated with the environmental and body temperatures of organisms [71].

Overall, the low denaturation temperature of sponge collagen observed in the present study enables gelatin extraction at lower temperature compared to mammalian gelatin, therefore providing an economic benefit for using marine sponges as a raw material of gelatin for the food industry [72].

All of our results point out that sponges contain collagen that retains its helical structure throughout the isolation procedure and all of its measured characteristics confirm the less crosslinked form, as verified by the IR, amino acid analysis, DSC, and CD data.
