Phospholipid fatty acids from Colombian Caribbean sea sponges Ocean and Coastal

Five demosponges belonging to the order Axinellida: Axinella corrugata, Dragmacidon alvarezae, Dragmacidon reticulatum, Ptilocaulis walpersi, Myrmekioderma rea and one sponge belonging to the order Scopalinida: Scopalina ruetzleri were analyzed to assess their fatty acid composition in the phospholipid fraction. Additionally, the seasonal and intraspecific variation in fatty acids composition was assessed in M. rea and D. alvarezae . Fatty acid identification was conducted using an HRGC-MS with an ECL value of methyl ester derivatives, and analyzing their mass spectra. To confirm double bound location, N-acylpyrrolidide derivatives were used studying their mass spectra. In total, 83 fatty acids were identified with chain lengths ranging from C14 to C32. Interestingly, brominated fatty acids were identified, previously suggested for sponges. Polybranched fatty acids such as 4,8,12-trimethyltridecanoic acid (4,8,12-TMTD) and 3,7,11,15-tetramethylhexadecanoic (phytanic acid) were found, without a clear distribution pattern. A predominance of iso-acids (i-15:0 and i-17:0) on anteiso acids were observed. Some seasonal variations in fatty acid (FA) compositions for M. rea and D. alvarezae were observed. The hierarchical Clusters Analysis (HCA) showed that the FA composition was species-specific but not informative at the family or order level.

Marine lipidology searches for new FA structures; new sources of polyunsaturated fatty acids (PUFA) of biological interest; FA roles in cell membranes; biosynthetic pathways, and developing trophic biomarkers in ecosystems (Bergé and Barnathan, 2005); and also their chemotaxonomic significance. Usually, FA isolated from marine sponges have long chains (>C24), but there are few studies on fatty acid composition significance in chemotaxonomy (Rod'kina, 2005). Examples of the utility include the proposed use of the branched FA 4,8,12-trimethylltridecanoic acid as a chemical marker for the Clionaidae and Spirastrellidae families (Erpenbeck and Van Soest, 2007). A review of the distribution of long-chain FA (LCFA) in the three classes of sponges revealed more proximity between Hexactinellida (LCFA>C 28 ) and Demospongiae (LCFA, C 24 -C 28 ); and less relationship with Calcarea, which lacks long-chain fatty acids (Erpenbeck and van Soest, 2005). Another study showed that FA profiles of deep-sea sponges can be used as chemotaxonomic markers and support the concept that sponges acquire building blocks from their endosymbiotic bacteria (de Kluijver et al., 2021). In contrast to shallow-water sponges that are dominated by PUFA with a high percentage of LCFAs (C24-C30), the deep-sea sponges were found to be rich in saturated fatty acids (C14-C20) (Zivanovic et al., 2011).
Continuing our research on fatty acids of marine sponges from the Colombian Caribbean Sea (Castellanos and Duque, 2008;Rodríguez et al., 2010), the aim of this study was to determine phospholipid fatty acid composition of the sponges Axinella corrugata, Dragmacidon alvarezae, Dragmaciodn reticulatum, Ptilocaulis walpersi, Myrmekioderma rea and Scopalina ruetzlei in order to detect regularities or novelties, study the seasonal and intraspecific variation for M. rea and D. alvarezae, and explore if the FA composition in the studied sponges can be used as a chemotaxonomic marker at the species and higher taxonomic levels.

MetHODs cOllectION Of spONges
Samples of sponges were collected through SCUBA diving in several places around Santa Marta, Colombian Caribbean (Table 1) (Vega-Sequeda et al., 2008)bands of 10 × 1 m and 10 × 2 m were used, respectively. A total of 33 species of hard corals were registered. Algae and corals were the components with major cover, nonetheless, averages varied widely in the study area (coral= 4.0 ± 0.9 %-62.0 ± 7.2 %; algae = 30.6 ± 11.0 %-78.3 ± 3.0 %. The sponges were washed with sterile sea water, carefully cleaned of epibiotic organisms and finally frozen (-20°C). The sponges were identified and classified by Dr. Sven Zea, and reference specimens are available at the Makuriwa Colombian Marine Natural History Museum, Marine and Coastal Research Institute -INVEMAR in Santa Marta. Species D. alvarezae and D. reticulatum were described by (Zea and Pulido, 2016). The frozen material was transported to Bogota by air preserving the cold chain.

extrActION AND IsOlAtION Of pHOspHOlIpID frActION
The cleaned sponges were cut into small pieces and extracted with CHCl 3 -MeOH (1:1) to obtain the total lipid fraction. Neutral lipids, glycolipids and phospholipids were separated by column chromatography on silica gel (60-200 mesh) from the total lipidic fraction using a previously reported procedure Rodríguez et al. Fatty acids in Caribbean sea sponges  (Carballeira et al., 2007). The phospholipidic fraction (methanolic fraction) was kept in nitrogen at -10ºC in CHCl 3 : MeOH (1:1), with 0.002% of 2,6-di-tertbutyl-4-methylphenol (BHT) as an antioxidant until their derivatization for GC-MS analysis. For intraspecific variation, the same procedure was done (Table 2). For the data in Table 3, the sponge specimens were pooled to have a sufficient quantity of sample for chemical analyses.

fAMes AND N-AcylpyrrOlIDIDes DerIvAtIves
The fatty acids of the phospholipids were transformed into their methyl esters (FAME) by reaction, under reflux, of the phospholipid fraction (100 mg) with methanolic KOH 2N (3 ml, 4 h), followed by addition 3 mL of methanolic boron trifluoride 14% (Merck TM ). This reaction mixture was then stirred (360 rpm) for 2h at 60°C. The reaction mixture was monitored with thin layer chromatography using benzene/ethyl acetate (10:2 v/v) as the mobile phase and methyl stearate as standard FAME. The organic fractions were obtained by extraction with benzene (5 mL, 3 times), and dried over anhydrous Na 2 SO 4 . Finally, the FAME were purified by column chromatography on silica gel eluting with hexane/diethyl ether 10:1. The purified FAME fraction was stored at -20°C.
To determine the double-bonds and methyl-branching positions in fatty acid chain, N-acylpyrrolidide derivatives were obtained as has been suggested by (Carballeira et al., 2007). These amide derivatives were prepared by direct treatment of FAME fractions (20 mg) with pyrrolidine/ acetic acid (10:1) (2 ml) for 2 h at 100°C. Pyrrolidine (Merck TM ) reagent was distilled before use. The reaction was monitored with thin layer chromatography using benzene/ethyl acetate (10:2 v/v) as the mobile phase, and N-stearoyl pyrrolidide as the reference compound. The reaction mixture was purified by CC (silica gel, hexane/diethyl ether 1:2). The resulting methyl esters and N-acylpyrrolidide fractions were analyzed with high resolution gas chromatography coupled mass spectrometry (HRGC-MS).

HIgH resOlUtION gAs cHrOMAtOgrApHy-MAss spec-trOMetry (Hrgc-Ms) ANAlyses
For the HRGC-MS analysis, a Shimadzu GC-17A chromatograph (Kyoto, Japan) equipped with a crosslinked HP-1 fused silica capillary column (25 m x 0.25 mm i.d.) operating at 70 eV was employed. For FAME analysis, 1 μL of a 1 mg/mL solution in acetone was injected in split mode (1:30), while the temperature of the injector was kept at 300°C. Helium, 5.0 grade, was used as carrier gas, maintaining a 1 mL/min flux. Temperature was kept at 135°C for 4 min, followed by a temperature ramp of 2°C/min until 280°C, maintaining this temperature for 15 min. Detector temperature was set at 300°C.
The fatty acid methyl esters were identified by its retention times in capillary GC, using equivalent chain length (ECL) (Zhang et al., 2010)a method was developed to search all saturated fatty acid methyl esters in the sample, subsequently calculate the ECL value of each fatty acid, and finally identify the molecular structure for each component by comparing the ECL of the interest and that in the customised database. Our method was applied to analyse the fatty acid composition of Eucommia ulmoides seed oil. The results show that major polyunsaturated fatty acids are α-linolenic acid (56.5093% of total fatty acids, TFAs, and by their FAME mass spectral analysis (Mjøs, 2004), with further comparison with standards Rodríguez et al. Fatty acids in Caribbean sea sponges  (C12-C24) and with those of published data (LIPID MAPS; Rodríguez et al., 2010). The abundance relative of each FAME based peak area/total area ratio is reported. For pyrrolidide derivatives, the temperature program started at 200°C for 4 min, followed by a temperature ramp of 2°C/min until 295°C, then maintaining this temperature for 30 min. Others HRGC-MS conditions were the same as for FAME capillary chromatographic analysis. The branches and double bond positions in N-acylpyrrolidide fatty acids was determined by comparison with those of published data (Rodríguez et al., 2010).

AcIDs
The effect of the coefficient of seasonal variation (CV) on the content of phospholipid fatty acids in the M. rea sponge ranged from 12.2% to 90.3%. In the D. alvarezae sponge, the seasonal CV ranged between 6.0% and 106.1%. The intraspecific CV (different specimens A, B, and C collected on the same date) in D. alvarezae ranged from 0% to 132.4%, showing high variability in FA content among the same species (Table 4).

HIerArcHIcAl clUsters ANAlysIs
To evaluate if sponge FA contents are species specific even when there is seasonal and intraspecific variation, a multivariate data analysis (MVDA) was utilized. The analysis of the FA composition in the samples studied was performed using a Hierarchical Cluster Analysis (HCA-Ward distances) (Figure 1), generating a model with two principal components (PCs), representing 55.5% of the variance of the included data, with a correlation value of 0.654 (R 2 ) and a predictive power of 0.498 (Q 2 ). In general terms, all species stood out separately in their overall FA composition. Also, there is a good clustering for sponges collected on the same date (D. alvarezae A, B, and C replicates) and for samples of same species collected at different times (D. alvarezae, and M. rea). However, the FA composition does not serve to discriminate at higher taxonomic levels in the current classification of sponges, as the profile of S. ruetzleri´s, a sponge from the order Scopalinida, is intermingled among the species of the order Axinellida. Aside from this, however, families Axinellida (A. corrugata, D. alvarezae, D. reticulatum and P. walpersi), and Heteroxyidae (M. rea) were separated by their FA composition.
The high content of polyunsaturated FA 22:6 in D. alvarezae suggests the presence of dinoflagellates and microalga as the predominant symbionts (Yang et al., 2017). The presence of short chain FA (14:0, 16:0 and 18:0) is associated to diatoms and dinoflagellates, widespread in all studied sponges, showing that these microorganisms are common symbionts. This implies that FA precursors released from chemo-as well as heterotrophic microbes in sponges contributed to the synthesis of very long chain fatty acids (VLCFA's), identifying sponge-associated bacteria as symbionts of the sponge (van Duyl et al., 2020). It was reported that sponges acquire building blocks from their endosymbiotic bacteria (de Kluijver et al., 2021). The presence of LCFA, known as demospongic acids (≥C 23 ), have been used as sponge biomarkers in different sponge taxa. For instance Haliclona species are characterized by the presence of FA 28:3, 25:2, and 24:1; Aplysina aerophoba by the presence of 22Me-28:2, 30:3, and 28:3, while Dysidea avara included FA such as 26:2, 25:3, 24:1, and ai-23:0. These compounds have been used to determine metabolic rates in sponges of Demospongiae (Koopmans et al., 2015).
With respect to the FA composition found in the present study (Table 3), we can see the sequence 14:0-16:0-18:0-20:0-22:0-24:0-26:0 and 16:1-18:1-20:1-24:1-26:1 occurred, indicating that our results are agree with currently accepted pathways for the main very long-chain non-methylene-interrupted FAs (>C24) (Kornprobst and Barnathan, 2010). However, for sponges of class Calcarea, the high Rodríguez et al. Fatty acids in Caribbean sea sponges  occurrence of iso-and anteiso-FA (>40%) suggests that the sponge cells are the principal producers of these lipids (Schreiber et al., 2006). In this study we reported low contents of these FA (<23%), indicating a probable symbiont-sponge origin as was mentioned above (Table 3). It has been seen that high quantities of the 4,8,12-TMTD acid occur in calcareous sponges (Schreiber et al., 2006). Our results show that there was a predominance or specificity of 4,8,12-TMTD on phytanic acid or vice-versa. The presence of branched acids 4,8,12-TMTD and phytanic acid have been suggested as chemical markers for sponges of the families Spirastrellidae and Clionaidae; but also for Calcarea (<3%) (Schreiber et al., 2006). A review on chemosystematics in sponges showed 4, 8, 12-TMTD acid was regarded unsuitable (Erpenbeck and van Soest, 2007). In our case, the overall FA composition, although species-specific to the point that it could perhaps be used as chemotaxonomic marker, did not serve to distinguish between the studied orders (Scopalinida and Axinellida).
A fatty acid named 18-bromooctadeca-9E,17Ediene-7,15-diynoic acid and others similar FAs were identified in Xestospongia sp. with the atom Bromine position always at the end of hydrocarbon chain (Taniguchi et al., 2008) while in Haliclona sp, an FA was recorded with a structure of 6-bromo-icosa-3Z,5E,8Z,13E,15E pentaene-11,19-diynoic acid, showing than more than twenty fatty acids with bromination at C-6 have been reported in sponges (Aratake et al., 2009). All those results indicate that brominated fatty acids at C-6 with system ∆ 5,9 in marine sponges seem to be restricted to the orders Agelasida, Haplosclerida and Tetractinellida. Interestingly, the FA content is lower for specimens collected in tropical waters compared to those collected in cold northern seas (Blumenberg and Michaelis, 2007). There is strong evidence of bromoperoxidase type enzyme activity in sponges belonging to the order Suberitida; these results should direct future work to experimentally characterize the specific enzymatic activity in marine sponges of this order (Blumenberg and Michaelis, 2007).
The C-2 methoxylated FA content reported here was lower than that reported in other marine sponges (Rod'kina, 2005) (Rodríguez et al., 2010). In general, the methoxy group was between chain lengths of C 15 to C 28 , regardless of whether there were double bonds in the chain. Reported bioassays indicate that these compounds have potential as antimicrobial agents, but the full potential of the alpha methoxylated fatty acids as a therapeutic candidate remains to be determined (Carballeira, 2002). Recently very long-chain α-methoxylated Δ5,9 fatty acids from the Sponge Asteropus niger were shown to be effective inhibitors of topoisomerases IB (Carballeira et al., 2016). On the other hand, the C-2 hydroxylated FA content registered here was lower than that reported in other marine sponges (Rod'kina, 2005).
With respect to the seasonal and intraspecific variation of fatty acids presented here, as the food source determines FA content in sponges, the sponge FA composition will change when the food source changes (Koopmans et al., 2014). In Colombia, a year has two seasons: rainy and dry seasons. The rainy season runs from April to May and October to November, and the dry season is usually December to January and July to August, although this can vary considerably. During the rainy season, more sediment reaches the sea, and more food is available for the sponges. To assess the effect of the seasons, the FA composition of the phospholipid fraction was determined for M. rea and D. alvarezae (Table 4).
Studies on the effect of seasonal and geographical variation in the composition of marine and freshwater sponge fatty acids are scarce. A study on seasonal variation in the contents of demospongic acids (C24-C28) in sponges from the Sea of Japan, collected between January and July, revealed variability ranging from 24% to 30% for the C24:0 fatty acid contents for the sponge Semisuberites cribrosa. The fatty acids with a chain length of C25:1 in Semisuberites cribrosa, C26:2 in Haliclona (Reniera) cinerea and C28:0 in Suberites domuncula presented a CV<15 %, while other fatty acids had variations ranging from 30% to 100% Rodríguez et al. Fatty acids in Caribbean sea sponges (Lüskow et al., 2019). However, the seasonal fluctuation of fatty acid composition is also common in marine sponges and is an adaptive response of the sponge to optimize membrane flexibility at different environmental temperatures (Lüskow et al., 2019) which probably partly explains the high FA content variations encountered in our study (Table 4).

cONclUsIONs
We identified FA with chain length from C 14 to C 32 ; there was a predominance of iso-acids (i-15:0 and i-17: 0) on anteiso acids within sponge studied suggesting a high predominance of bacteria as symbionts in these sponges. Our results show that there was a predominance or specificity of 4,8,12-TMTD on phytanic acid or vice-versa, regardless of the studied taxon level between studied sponges. In this study, the Caribbean sponges showed a lower content of FA bearing bromine compared with those from the Nordic sea. The seasonal variations in the phospholipid fatty acid contents ranged from CV 6.0-106.1%. The intraspecific variation on the phospholipid fatty acid contents ranged from CV 0-132.4%. FA composition could be used as a chemo-taxonomic marker for each of the studied species, even across temporal and individual variation, the latter for the sponges Dragamacidon alvarezae and Myrmekioderma rea. However, at the family and order level, FA composition appears not to be informative.