*2.1. The Context of This Work*

The TASCMAR project (Tools And Strategies to access to original bioactive compounds from Cultivation of MARine invertebrates and associated symbionts), funded by the European Union in the frame of the Horizon 2020 framework program, offered the opportunity to investigate the molecules produced by marine invertebrates and their symbionts from mesophotic coral ecosystems (MCEs) (30 to 150 m depth). Among the invertebrates investigated, the sponge *Acanthella cavernosa* was collected on the upper mesophotic reef of Eilat at Dekel Beach (51 m depth), in the Gulf of Aqaba (Israel, 2 April 2017, 29◦32- 12.48"N; 34◦56- 55.656"E). The area is characterized by a moderate slope covered with dense patches of hard substrate, mostly calcareous, and is also inhabited by other invertebrates such as octocorals, stony corals, black corals, and sea anemones (Figure 1).

The strain *Chrysosporium lobatum* TM-237-S5 (Figure 2) was among the strains isolated and identified based on its ITS rDNA sequence (Nuclear ribosomal internal transcribed spacer).

**Figure 1.** *Acanthella cavernosa* was collected at 51 m depth in Eilat, Gulf of Aqaba (Israel). (**A**) The sponge in its natural environment, (**B**) The sponge was collected by the remote operating vehicle (ROV) arm, introduced in-situ to the collection basket, and brought to the boat for immediate processing. Two representative pieces were recovered, one for taxonomic identification and the other for symbiont isolation. Both samples were immediately frozen on the boat and shipped in dry ice.

**Figure 2.** Maximum-likelihood tree obtained from ITS rDNA sequence alignment of the strain TM237-S5 and *Chrysosporium* spp. Reliability of the internal branch is represented in red. *Candida albicans* was used as the outgroup. Numbers are Genbank accessions. Th estrain in bold font is the one described in this study. Scale represents substitutions per site.

The strain was cultivated on potato dextrose broth (PDB), potato dextrose agar (PDA), marine broth (MB), and marine agar (MA). Solid phase extraction (SPE) with XAD resin (AMBERLITE™ *XAD*™*16HP* N) was applied in-situ to both liquid (LSF/SPE) and agar-supported cultures (solid-state fermentation and solid-state extraction (SSF/SSE)). It has been previously reported that in-situ XAD extraction coupled to agar-supported cultivation prevents the diffusion of target compounds to the agar layer and traps the target compounds on the resin beads [25].

On day four of incubation (Figure 3E), the resin beads became colored, but were not yet covered by the mycelium (white filaments). On day seven (Figure 3F), the resin beads became darker and the mycelium surface increased. On day 10 (Figure 3G), the recovery time, the resin beads were totally covered by the mycelium. We previously reported that such phenomenon is probably due to the lack of oxygen in the viscous resin layer, which pushes the mycelium to reach the surface to access more oxygen. However, the mycelia remained in contact with the agar to access nutriments, as shown on the agar layer, following the recovery of the resin beads (Figure 3H).

**Figure 3.** 10 days culture of *Chrysosporium lobatum* TM-237-S5 on potato dextrose agar (PDA) (**A**,**B**) and marine agar (MA) (**C**,**D**) coupled to solid-solid extraction (SSE) with XAD resin (AMBERLITE™ *XAD*™*16HP* N). The resin beads remained white to light beige on the marine broth (MB) (**D**), while they turned dark brown on the PDA (**B**), showing that the resin beads trapped the colored compounds secreted by the strain. (**E**–**G**) present the coverage of the resin beads by the mycelium at four, seven, and 10 days. (**H**) depicts the easy recovery of the resin biofilm layers; the mycelium is not incrusted and no compounds flow to the agar. The resin beads, as revealed by the dark brown color, trapped all the produced compounds.

After 10 days of incubation, the resin beads were recovered by filtration from the liquid cultures (10 L), and by scraping the surface of the agar cultures (10 <sup>×</sup> 625 cm<sup>2</sup> petri plates), and washed extensively with water to remove medium residues and any compounds not trapped by the XAD. Resin beads from the PDB, MB, and MA cultures had a light beige color, while the PDA culture was dark brown; most of the color being trapped by the resin beads (Figure 3A–D and Figure 4).

**Figure 4.** 10 days culture of *Chrysosporium lobatum* TM-237-S5 on PDA coupled to SSE with *XAD*™*16HP* N (**a**) and the control culture on PDA without resin (**b**). Without the resin (**b**), the colored compounds were spread in the agar, and their extraction was difficult. With the resin (**a**), the agar remained clear as the resin beads trapped all the colored target compounds.

The compounds trapped in the XAD were eluted with ethyl acetate and analyzed by HPLC coupled to photodiode array PDA, light-scattering LSD, and mass spectrometry MS detectors (Figure 5). According to the recovered quantities of extracts and the diversity of metabolites observed in the chromatograms, the current study focused on the extract from agar-supported cultivation coupled to in-situ solid-state extraction (SSF/SSE, Figure 5B). This SSF/SSE on PDA lead to an overall extract yield of 872 mg/m2 of cultivation surface, corresponding to 2 L of medium (200 mL per plate). HPLC analysis revealed 10 peaks with specific UV absorption spectra (Figure 6), which were totally absent in the liquid culture (LSF/SPE) (Figure 5C).

**Figure 5.** HPLC analysis of the ethyl acetate extract of *C. lobatum* TM-237-S5 cultivated on different media and support.

**Figure 6.** HPLC analysis with LSD and one-dimensional/two-dimensional PDA detections (right). Absorbance spectrum of the compounds investigated.

#### *2.2. Structural Identification of Compounds 1 to 10*

Compounds **1** to **10** in Figure 7 were purified and submitted to one-dimensional and two-dimensional NMR and HRMS analysis. Six compounds were unambiguously identified as peniciphenalenin D (**1**), isolated from *Pebnicillium* sp. ZZ901 [30], coniolactone (**3**), (-)-7,8-Dihydro-3,6-dihydroxy-1,7,7,8-tetramethyl-5H-furo-[2- ,3- :5,6] naphtho[1,8-bc]furan-5-one (**6**), coniosclerodin (**9**), isolated from *Coniothyrium cereale* [30,31], (+)-scleroderolide (**7**), isolated from *Gremmeniella abietina* [32,33], and (+)-sclerodin (**10**), isolated from *Aspergillus silvaticus* [34].

**Figure 7.** Structures of the compounds produced by *Chrysosporium lobatum* TM-237-S5, cultivated for 10 days on PDA medium coupling solid-state fermentation with solid-state extraction (SSF/SSE).

Compounds **2**, **4**, **5**, and **8** were submitted to dereplication based on the Antibase database of microbial compounds (Wiley–VCH) and the natural compounds Reaxys database (Elsevier). Spectroscopic data of these compounds did not match the previously reported compounds or present significant differences, and were submitted to de-novo structural elucidation. Their 1H and 13C NMR data are shown in Tables 1 and 2.


**Table 1.** 13C NMR (125 and 150 MHz) of compounds **2**, **4**, **5**, and **8**.

a,b; the spectra were recorded in MeOD and CD2Cl2, respectively.

The peak at 28 min exhibits a molecular formula of C17H16O5, determined by HRESIMS (*m*/*z* 301.1076 [M + H]+). A careful 1H and 13C NMR analysis of the peak at 28 min revealed a mixture of two compounds with indistinguishable HRMS. De-replication and comparison with published results showed that one of the constituents was unambiguously coniolactone (**3**). As well as **3**, HMBC correlations showed that compound **2** differs only at the ring C configuration. A key HMBC correlation from H-2 (δ<sup>H</sup> 6.70, s) to the carbonyl C-7 (δ<sup>C</sup> 168.1) indicated that, in **2**, the carbonyl at C-7 is connected to C-6 rather than to C-9 in coniolactone (**3**) (Figure 8). So far, all our attempts to separate **2** and **3** by different chromatographic techniques have failed. Compound **2** was named isoconiolactone.


**Table 2.** 1H NMR (500 and 600 MHz) of compounds **2**, **4**, **5** and **8**.

a,b; the spectra were recorded in MeOD and CD2Cl2, respectively.

**Figure 8.** COSY and key HMBC correlations for compounds **2**, **4**, **5**, and **8**.

Compound **4** has a molecular formula C17H16O5, deduced from HRESIMS and NMR data (Tables 1 and 2). According to NMR and MS data, **4** has the same planar structure as the already known fungal metabolite peniciphenalenin F [30]. However, compound **4** and peniciphenalenin F have an opposite optical rotation; negative for **4** (−36.10◦ (c 0.10, MeOH)) and positive for the reported peniciphenalenin F (+16.50◦ (c 0.50, MeOH)). Subsequently, **4** was named (-)-peniciphenalenin F.

Compound **5** had a molecular formula of C18H16O7, deduced from HRESIMS and NMR data (Tables 1 and 2). The NRM data of **5** indicates the presence of two carbonyls, eight aromatic carbons, one oxymethine, one quaternary carbon, and four methyls. NMR comparison with previously reported phenalenone derivatives has shown similarities with the isolated (+)-scleroderolide (**7**) [32,33], except in the C-2 position. Indeed, the aromatic proton, H-2, of scleroderolide is substituted in **5** by a hydroxyl group in C-2 (δ<sup>C</sup> 144.9). This finding is also supported by key HMBC correlations from H-18 (2.61, 3H, s) to C-2 (δ<sup>C</sup> 144.9), C-3 (δ<sup>C</sup> 120.5), and C-4 (δ<sup>C</sup> 109.3). Accordingly, compound **5** has been identified as a new phenalenone derivative and was named (+)-8-hydroxyscleroderolide.

Compound **8** has a molecular formula C18H16O7, deduced from the HRESIMS and NMR data (Tables 1 and 2). Here again, the NRM data of **8** showed the presence of two carbonyl, eight aromatic carbons, one oxymethine, one quaternary carbon, and four methyls. The NMR data of **8** closely resemble those of the previously described and isolated (+)-sclerodin (**10**) [34], except in the C-2 position. The aromatic proton, H-2, of the sclerodin structure is substituted in **8** by a hydroxyl group in C-2 (δ<sup>C</sup> 140.4). This finding is supported by key HMBC correlations from H-18 (2.79, 3H, s) to C-2 (δ<sup>C</sup> 140.4), C-3 (δ<sup>C</sup> 130.3), and C-4 (δ<sup>C</sup> 108.9).

The structure of **8**, including the absolute configuration (13*R*), is secured by a single crystal X-ray crystallographic analyses using anomalous scattering of Mo then CuK α radiation through Bijvoet analysis [35], combining maximum likelihood estimation and Bayesian statistics (Figure 9). Therefore, compound **8** was named (+)-8-hydroxyslerodin.

**Figure 9.** The Oak Ridge Thermal Ellipsoid Plot ORTEP diagram of **8**.

The current study presents the breakthrough advantage of solid-state fermentation coupled with in-situ solid phase extraction (SSF/SSE). Indeed, SSF/SSE leads to a large chemical diversity and higher yields. In addition, the concentration of the target compounds on the resin beads represents an economic and ecofriendly recovery process involving limited quantities of water for medium preparation, limited use of power for static incubation, a limited amount of solvents for the elution of compounds, and reduced waste. We also solved the issue of scale-up on the Platotex technology, offering a reusable 2 m<sup>2</sup> cultivation surface, and more recently on the Unifertex technology.

#### **3. Materials and Methods**

#### *3.1. General Experimental Procedures*

Optical rotations, [α]D, were measured using an Anton Paar MCP-300 polarimeter at 589 nm (Anton Paar, Les Ulis, France). NMR experiments were performed using a Bruker Avance III 600 MHz spectrometer equipped with a TCi cryo-probe head for compounds **8**, **9**, and **10**, and a Bruker Avance 500 MHz spectrometer for compounds **1** to **7** (Bruker, Vienna, Austria). The spectra were acquired in CD3OD (δ<sup>H</sup> 3.31 ppm and δ<sup>C</sup> 49.15 ppm), in CD2Cl2 (δ<sup>H</sup> 5.32 ppm and δ<sup>C</sup> 53.10 ppm) or in Acetone-*d*<sup>6</sup> (δ<sup>H</sup> 2.04 ppm and δ<sup>C</sup> 29.8 ppm and 206.5 ppm) at 300K. High-resolution mass spectra were obtained on a Waters LCT Premier XE spectrometer equipped with an electrospray-time of flight (ESI-TOF) by direct infusion of the purified compounds (Waters SAS, Saint-Quentin-en-Yvelines, France). Pre-packed silica gel Redisep columns were used for flash chromatography using a Combiflash-Companion chromatogram (Serlabo, Entraigues-sur-la-Sorgue, France). All other chemicals and solvents were purchased from SDS (SDS, Peypen, France).

The analytical HPLC system consisted of an Alliance Waters 2695 controller coupled with a PhotoDiode Array Waters 2996, an evaporative light-scattering detector ELSD Waters 2424 detector and a mass detector Waters QDa (Waters SAS, Saint-Quentin-en-Yvelines, France). A Sunfire C18 column (4.6 × 150 mm, 3.5 μm) was used with a flow rate of 0.7 mL/min. The elution gradient consisted of a linear gradient from 100% solvent A to 100% solvent B in 40 min, then 10 min at 100% B (Solvent A: H2O + 0,1 HCOOH, Solvent B: ACN + 0,1% HCOOH). Preparative HPLC was performed on a semi-preparative Sunfire C18 column (10 × 250 mm, 5 μm) using a Waters autosampler 717, a pump 600, a photodiode array detector 2996, and an ELSD detector 2420 (Waters SAS, Saint-Quentin-en-Yvelines, France). XAD resin (AMBERLITE™ *XAD*™*16HP* N) was purchased from DOW (DOW France SAS, Saint-Denis, France).
