Differential Membrane Lipid Profiles and Vibrational Spectra of Three Edaphic Algae and One Cyanobacterium

The membrane glycerolipids of four phototrophs that were isolated from an edaphic assemblage were determined by UPLC–MS after cultivation in a laboratory growth chamber. Identification was carried out by 18S and 16S rDNA sequencing. The algal species were Klebsormidium flaccidum (Charophyta), Oocystis sp. (Chlorophyta), and Haslea spicula (Bacillariophyta), and the cyanobacterium was Microcoleus vaginatus (Cyanobacteria). The glycerolipid profile of Oocystis sp. was dominated by monogalactosyldiacylglycerol (MGDG) species, with MGDG(18:3/16:4) accounting for 68.6%, whereas MGDG(18:3/16:3) was the most abundant glycerolipid in K. flaccidum (50.1%). A ratio of digalactosyldiacylglycerol (DGDG) species to MGDG species (DGDG/MGDG) was shown to be higher in K. flaccidum (0.26) than in Oocystis sp. (0.14). This ratio increased under high light (HL) as compared to low light (LL) in all the organisms, with its highest value being shown in cyanobacterium (0.38–0.58, LL−HL). High contents of eicosapentaenoic acid (EPA, C20:5) and hexadecenoic acid were observed in the glycerolipids of H. spicula. Similar Fourier transform infrared (FTIR) and Raman spectra were found for K. flaccidum and Oocystis sp. Specific bands at 1629.06 and 1582.78 cm−1 were shown by M. vaginatus in the Raman spectra. Conversely, specific bands in the FTIR spectrum were observed for H. spicula at 1143 and 1744 cm−1. The results of this study point out differences in the membrane lipid composition between species, which likely reflects their different morphology and evolutionary patterns.


Introduction
Particular biochemical composition, mainly photosynthetic pigments and membrane lipids of algal species and cyanobacteria, may help to better understand the biodiversity and functioning of photosynthetic organisms. Indeed, pigments and fatty acids (FAs) have a long history of use as taxonomic biomarkers [1,2]. Nowadays, the profiling of the fatty acid composition has become relevant for biotechnological applications such as biofuels and nutritional supplements [3][4][5]. However, the thylakoid architecture and optimal photosynthetic function are dependent on the specific glycerolipid class to which molecular species the fatty acids are associated [6][7][8]. Four classes of glycerolipids make up the thylakoid membrane lipids, namely, monogalactosydiacylglycerol (MGDG), digalactosyldiacylglycerol (DGDG), sulfoquinovosyldiacylglycerol (SQDG), and diacylglycererylphosphoglycerol (PG), of which MGDG is the most abundant.
While the lipid composition of thylakoid membranes of a number of model organisms is well known, the specific composition of the photosynthetic and cell membranes The FTIR and Raman spectra are also reported in order to gain insights into the molecular spectroscopy of diverse photosynthetic organisms.

Species Identification
Five species of algae and three species of cyanobacteria could be distinguished in the initial filtered BG11 suspension, namely, Klebsormidium flaccidum (Kützing; Charophyta), Oocystis sp. and Chlorella sorokiniana (Chlorophyta), two naviculaceae (Bacillariophyceae), Haslea spicula and Navicula pulchripora, M. vaginatus (Cyanophyceae), and two unicellular cyanobacteria, one of them likely related to Synechococcus and the other to Aphanocapsa salina. The three algae and the cyanobacteria used in this study were chosen because there have been no or scarce data regarding their lipid profile; in addition, we were interested in comparing the lipid profiles of algae and cyanobacteria with different morphologies but dwelling under the same environmental conditions. As well, vibrational spectra were expected to show the differences in their morphology and cell organization. Photographs of the selected species for this study are shown in Figure 1. Data on similarity (%) and the sequence total score for the primers used are depicted in Table S1.

Photosynthetic Pigments
The content of the representative photosynthetic pigments for the algae and cyanobacterium under the low light condition (c.a. 15 µmol photons m −2 s −1 ) are reported in Table 1. The relative contents of chlorophyll b, carotenes, and violaxanthin to chlorophyll a were shown to be higher in K. flaccidum (0.82, 0.27, and 0.15, respectively) than in Oocystis sp. (0.63, 0.03, and 0.10, respectively), but the opposite was shown for the ratio of lutein and neoxanthin to chlorophyll a (0.13 and 0.01, 0.40 and 0.12, respectively, for K. flaccidum and Oocystis sp.). In H. spicula, the ratio of fucoxanthin to chlorophyll c1 was higher than that of fucoxanthin to chlorophyll a. The typical xanthophylls zeaxanthin and myxoxanthophyll of M. vaginatus were detected along with chlorophyll a. Representative chromatograms are depicted in Figure S1 for each organism.

Glycerolipids
The base peak chromatogram (BPI) resulting from the UPLC-ESI-QToF-MS analysis pointed out the prevailing glycerolipids of each alga and cyanobacteria. Representative chromatograms of the lipid extract are shown in Figure Table 2). Nonetheless, other MGDG and DGDG species also stood out in M. vaginatus. The glycerolipid profile of H. spicula was abundant in highly unsaturated fatty acids, mainly eicosapentaenoic acid (C20:5, EPA), with MGDG(16:4/20:5), m/z 813.48, standing out (Figure 4a). It should be mentioned that even though the fatty acid C20:5 is placed at the sn-2 position of the glycerol backbone for the sake of simplicity, it is known that the long-chain fatty acid is currently esterified at the sn-1 position in diatoms [30]. This alga exhibited SQDG(16:0/16:1) as the almost unique SQDG species; it accounted for 91.8% of this glycerolipid class (Figure 4b). Few species of PG were detected in the extracts as a whole, with C18:3, C18:2, C18:1, C16:1, and C16:0 being the fatty acyls that were esterified, and no particular PG species stood out. A few PC and PE species were also detected in the algae (Figures 3e,f and 4c). Monoacylated species were only detected for the monogalactosyl class (MGMG) (Figure 3a). Under high light (c.a. 45 µmol photons m −2 s −1 ), the relative content of MGDG in all the algae and the cyanobacterium was diminished as compared to low light (c.a. 15 µmol photons m −2 s −1 ), whereas the relative content of DGDG was enhanced, in particular in M. vaginatus (Table 2). SQDG content was also increased in all the algae and the cyanobacte-rium under HL compared to LL, except for Oocystis sp. ( Table 2). The PG content rose in K. flaccidum and M. vaginatus under HL but decreased in Oocystis sp. and H. spicula (Table 2). Regarding changes in each glycerolipid species, exposure to the moderate high light level brought about a significantly (p < 0.01, t-test) substantial decrease in K. flaccidum of the predominant MGDG(18:3/16:3), as well as in almost all the MGDG species ( Figure 3a); however, with regard to the total content of MGDG species, MGDG(18:3/16:2) and MGDG(18:2/16:2) showed a moderate, not significant, increase. Conversely, the related DGDG(18:3/16:3) and DGDG(18:3/16:2) species increased their content per chlorophyll and percentage of the glycerolipid class ( Figure 3a and Table 2). In this alga, a notable decrease in the most abundant PG(18:3/16:1) was compensated for by an equivalent increase in PG(18:2/16:1) (p < 0.01, t-test) (Figure 3c and Table 2), and the anionic SQDG(18:2/16:0) also increased under HL (p < 0.05, t-test). Conversely, in Oocystis sp., higher light level exposure elicited increased concentrations of almost all the MGDG, DGDG, and SQDG species compounds per chlorophyll a content, including the standing species MGDG(18:3/16:4) (p < 0.05, t-test) (Figure 3b,d and Table 2). However, the increase in this species did not compensate for the rise in all the MGDG species and its percentage was lower under HL than under LL (Table 2). PG(18:1/16:0) and PG(18:3/16:1) also increased under higher light exposure in Oocystis sp. The response of the glycerolipid species to the several light levels did not show a clear pattern in H. spicula; however, augmented content of MGDG(16:4/20:5) was accompanied by a concurrent decrease in contents of the remaining C20:5-containing MGDG species; an increase in the content of the C20:5-containing DGDG species seems to occur, but with statistically significant changes only being accounted for by low abundance species (Figure 4a and Table 2). A statistically significant (p < 0.05, t-test) increase in the contents of the minor species SQDG(16:1/14:1) and SQDG(16:1/16:1) was found in this alga, and the prevailing species SQDG(16:0/16:1) also rose in a per Chla basis but was not significant (p > 0.05, t-test). PG species did not vary significantly in H. spicula. Exposure of M. vaginatus to high light elicited a rise in the content of the most abundant MGDG(18:3/16:2) and DGDG(18:3/16:2) species (p < 0.01, t-test), but a concomitant decrease in the content of MGDG(18:2/16:0) was also observed. A statistically significant increase was also found for SQDG(18:3/16:0), SQDG(18:2/16:0), and PG(18:3/16:0).
The unsaturation index of the membrane (UI), this being calculated as the sum of the number of unsaturations in the two acyl chains in a given species multiplied by the content of such species, was reduced in K. flaccidum as a result of high light exposure (14.9 in HL versus 18.5 in LL); this reduction was likely motivated by the lower content of the MGDG species under HL than under LL (Table 3). Conversely, the other green alga, i.e., Oocystis sp., accounted for an increase in UI of about 45.8% under HL with respect to LL, with relevant increases of UI taking place in DGDG and anionic glycerolipid species. UI also underpinned a notable rise in M. vaginatus from 18.3 under LL to 24.5 under HL (33.9%), to which all the glycerolipid classes contributed. A moderate increase in UI of 8.1% took place in the diatom H. spicula, being mainly due to the contribution of DGDG and anionic glycerolipid species, as in Oocystis sp. The highest UI was counted in H. spicu-la and the lowest one in Oocystis sp. under LL, with similar values being calculated for K. flaccidum and M. vaginatus; however, under HL, K. flaccidum became the organism with the lowest UI.     Principal component analysis (PCA) clearly distinguished the four organisms according to their lipid profile (Figure 5a), and, in particular, H. spicula was separated from the other organisms by Component 1. Conversely, K. flaccidum, Oocystis sp., and M. vaginatus were mainly separated by Component 2. Regression parameters were R 2 X(cum) = 0.38 and Q 2 (cum) = 0.32 for Component 1, and R 2 X(cum) = 0.66 and Q 2 (cum) = 0.43 for Component 2.

FTIR and Raman Spectroscopy
K. flaccidum and Oocystis sp. showed similar spectra between them for both Raman and FTIR spectroscopy ( Figures S3 and S4), whereas H. spicula and especially M. vaginatus had characteristic peaks in the FTIR and Raman spectra, respectively. Main observable bands from the crude spectra, that is, without deconvolution and showing averaged values of frequencies from the four organisms, are illustrated in Tables 4 and 5 for Raman and FTIR, respectively. The characteristic bands of carotenoids at~1004,~1157, and 1525 cm −1 [27,28,[31][32][33] in the Raman spectrum were detected for the four organisms. The band at the higher frequency (1521 cm −1 ), which has currently been assigned to double bond stretching, exhibited the highest intensity in the four spectra, though with comparable intensity to that of the band at 1157 cm −1 in M. vaginatus. It should be noted that a slight red shift was observed in H. spicula for these bands with respect to the other organisms (1013.9, 1158.5, and 1526.8 cm −1 ). Standing bands at 1629.1 and 1582.8 cm −1 were evidenced exclusively in the M. vaginatus spectrum ( Figure S3) and are assigned to C=O stretching and amide I β-sheet from proteins [24,31,34], probably of the phycobilisomes (C-N stretching plus the -N-H bend of amide II [34], but these vibrations could also have a contribution from the tretrapyrrol ring of the phycobilins, which are specific components of the cyanobacteria. A weak band was shown at 1605.9 cm −1 in the algae, which may be indicative of the band corresponding to the vibration of the protein α-sheet [28,29,34]. Bands at 1653 cm −1 have been assigned to ν(C=C) stretching from unsaturated fatty acids either in pure compounds [35] or algae [29,32] and, consequently, the bands at 1654.1 cm −1 in H. spicula and 1681.1 cm −1 in K. flaccidum and Oocystis sp. are likely to stem from the lipid and carotenoid double bonds. A band at 1284 cm −1 in M. vaginatus, with a slight blue shift and weaker intensity in the algae (1268-1270 cm −1 ), is assigned to (C-N) stretching plus the (-N-H) bend of amide III [28,34]. Otherwise, these bands have been attributed to (=C-H) deformations in unsaturated lipids [29]. Two bands at 1327 and 1212 cm −1 may be assigned to chlorophyll a (Chla) [27,33], though these bands may arise from the overlap of several vibration modes (Table 4). Two bands were also specifically detected in M. vaginatus at 667 and 815 cm −1 ; these bands are currently ascribed to ring breathing from nucleotides of DNA or aromatic amino acids [28,31,36]; nonetheless, in the case of cyanobacteria, they could also arise from the tetrapyrrol ring of chlorophyll a or the phycobilins [24]. A weak band at 728 cm −1 was specifically shown in the spectrum of H. spicula, and this band has been ascribed to different modes of vibration, e.g., -C-S-trans [34], the tetrapyrrol ring of Chla [28], or to δ(CNH), β(NH), and δ(ring) from proteins [36]. The band at 2930 cm −1 has been ascribed to the C-H stretching of CH 3 and CH 2 groups of proteins, mainly from aromatic and aliphatic amino acids [33,37,38], but it is also representative of pectin-based carbohydrates [39]; therefore, this band could be an overlapping of C-H vibration modes coming from galactosyl groups of MGDG and DGDG plus the polysaccharides of the cell wall in addition to acyl substituents in amino acids, which would explain the band widening. All the organisms showed different intensities between HL and LL spectra in the C-H stretching band due to methyl and methylene groups at about 2930 cm −1 , but differences in other bands between HL and LL spectra were only evident in those at 1582.8 and 1629.1 cm −1 in M. vaginatus, and in the band at 1604.0 cm −1 of Oocystis sp., whose intensity was decreased under HL ( Figure S3).  FTIR spectra are shown in Figure S4, and the band assignment is depicted in Table 5. The amide I and amide II bands were detected in all the algae and the cyanobacterium at 1638 and~1537 cm −1 , respectively [27,28,34]. Symmetric and asymmetric stretching (ν), as well as the CH 2 and CH 3 group deformations (δ) of the C-H bonds of lipids, were also evident in all the organisms. In the FTIR spectrum, specific bands were observed for H. spicula that correspond to the C-O and C=O vibrations from fats and carbohydrates [27][28][29]34], whereas these bands were not so evident in the remaining organisms. The algae showed relatively lower absorbance under LL than under HL, whereas the opposite happened for M. vaginatus, apart from the band at the highest wavenumber (c.a. 3300 cm −1 ). Decreased relative absorbance of the amide II (1537 cm −1 ) and amide I (1638 cm −1 ) bands was particularly shown in H. spicula and K. flaccidum.

Discussion
In addition to nutrient availability, other environmental factors such as UV resistance or survival mechanisms may elicit species selection or even strains within a given species [45]. Diatoms, together with green algae and cyanobacteria, are usually described as the main phototrophs in edaphic communities [45][46][47][48][49]. In this study, the lipid composition of four phototrophs from an edaphic assemblage with different architectures of the photosynthetic apparatus or morphology is reported. They include two green algae, one unicellular and one filamentous, a diatom, and a filamentous cyanobacterium.
The two green algae showed a similar lipid profile between them, but specific variations in species within each lipid class suggest a somewhat different organization of the photosynthetic apparatus. These subtle differences may account for variations in photosynthetic efficiency and capability to respond to inhibitory conditions (e.g., xanthophyll cycle and non-photochemical quenching). Indeed, principal component analysis (PCA) placed each alga separately ( Figure 5). The photosynthetic apparatus of Oocystis sp. seems to be highly dependent on DGDG species, which had higher diversity than in K. flaccidum, with the highly unsaturated MGDG(18:3/16:4) being the major building block of the photosynthetic membranes and a substantial decrease of the MGDG/DGDG ratio under HL (4.41) compared to LL (7.14). In K. flaccidum, two slightly less unsaturated MGDG species were abundant, namely, MGDG(18:3/16:3) and MGDG(18:3/16:2), with the content of the less unsaturated species increasing under HL. In this latter alga, the ratio of MGDG to DGDG also dropped under HL (3.82) compared to LL (2.87), but to a lesser extent than in Oocystis sp. Given that DGDG is tightly associated with the LHCII complex and, in particular, with the water-splitting complex [7,15,50], the rise in the content of DGDG species under HL suggests changes in the organization of the photosynthetic apparatus regarding higher LHCII aggregation and hydrophobic mismatch [14]. However, this result is somewhat contradictory with the expected involvement of non-bilayer lipids, i.e., MGDG, in the membrane energization state [15,22]. The high decrease in the MGDG/DGDG ratio of Oocystis sp. under HL compared to LL may be related to a different dominant lipid phase, with an inverted hexagonal (H II ) form under LL but increasing the bilayer surface under HL, which is noteworthy given that the HL condition used in this study may not be considered stressing for photosynthetic performance. Conversely, in K. flaccidum, the increase in the lipid bilayer surface was not so pronounced, a fact that may suggest that the compensation point of photosynthesis (E k in the P-I curve) is attained at a higher irradiance in K. flaccidum than in Oocystis sp. Furthermore, the reduction in the unsaturation index of K. flaccidum under HL compared to LL is surprising, which might be related to a more stable organization of the thylakoids that facilitates membrane fluidity. K. flaccidum is a filamentous alga, whereas Oocystis sp. is unicellular ( Figure 1); thus, it may be hypothesized that the chloroplasts of the cells of a given filament are interconnected and respond as a whole, which implies that the measured composition is averaged between more exposed cells to light and the cells undergoing self-shading, whereas all Oocystis sp. cells are exposed and self-dependent, which, consequently, likely renders a more homogeneous response to the different light intensities. These results point out that within the general architecture of the photosynthetic apparatus, subtle differences may exist between organisms, which are likely related to their specific physiological and morphological characteristics and ultimately become determined by the particular requirements of each organism for adaptation to the conditions of the ecosystem where they currently dwell.
In Oocystis sp., the relative content (about 1% of total glycerolipids) of PG species did not change from LL to HL conditions. However, the most abundant PG under LL, namely, PG(18:3/16:0), underwent a drastic reduction of its content under HL (from 33.89% to 23.55% of total PGs), whereas the opposite happened with the less unsaturated species PG(18:1/16:0), whose content rose from 23.57% to 42.93%. A similar trend was observed in K. flaccidum but with the PG species PG(18:3/16:1) and PG(18:2/16:1) instead. Since PG has been related to the functional integrity of PSI and to the PSI/PSII coupling [51,52], these changes might be due to a reorganization of the electron transfer between both photosystems. It is noteworthy that fewer PG species were detected in K. flaccidum than in Oocystis sp., and C16:0 was dominant in this latter alga at the sn-2 position of the glycerol backbone whilst it was C16:1 in K. flaccidum. This differential composition of C16 fatty acyls between both algae suggests again differences in the thylakoid membrane structure and organization that might be related to morphological and eco-physiological adaptation. A somewhat similar pattern to that of PG species was observed in SQDG species in Oocystis sp., but no relevant changes were observed in K. flaccidum for the SQDG species between LL and HL. The rise in less saturated acyl chains, with the concomitant drop of more unsaturated acyl chains, in the negatively charged glycerolipids (that is, PG and SQDG) is likely to be related to the structural organization of the PSII complex and the assembly of the extrinsic proteins in the oxygen-evolving complex as well as in the electron transport between the plastoquinones Q A and Q B [7,53].
The glycerolipid profile of M. vaginatus resembled that of cyanobacteria [24,[54][55][56], thus suggesting a similar photosynthetic membrane architecture. However, more DGDG species were detected in M. vaginatus in the present study than those previously reported for a marine strain of Synechococcus sp. [24], which may suggest a higher dependence on plasticity for the required changes of the oxygen-evolving complex (OEC) and LHCII aggregation to respond to environmental changes [53]. The filamentous character of M. vaginatus may also be a reason for that. The same acyl changes were shown in the most abundant species of both MGDG and DGDG, a fact that could be related to a capacity for fast acclimation to changes in environmental conditions associated with variations in the membrane phase. The conversion of MGDG into DGDG relies only on the addition of an additional galactosyl group, but this fact induces a relevant physicochemical change in the membrane, which is the conversion of non-bilayer-forming MGDG to bilayer-forming DGDG and, consequently, a lipid phase transformation [10,15]. However, the relative proportion of these most abundant DGDG species did not change with the light intensity. The MGDG/DGDG ratio was 2.64 and 2.15 under LL and HL, respectively; provided that this ratio is independent of the chlorophyll content or cell number, it seems evident that high light conditions imply a higher DGDG content, a fact that is likely related to an increase in the PSI/PSII ratio and enhanced thermotolerance [7,57]. Swelling of the thylakoid lumen under illumination in cyanobacteria has been reported [58], and this feature implies a reduction of the membrane curvature, a fact that is likely to be associated with the increase in the content of the bilayer-forming DGDG species.
As is well known, the composition and structure of the thylakoid membrane of diatoms differ substantially from those of green and other algal classes [12,19,[59][60][61]. In diatoms, fucoxanthin plays a more active role than that played by the xanthophylls in the green algae [11], with fucoxanthin being the major contributor to the fucoxanthin protein complexes (FPC), along with chlorophyll a. The high ratio of fucoxanthin:Chla, measured spectrophotometrically in this study (0.60) under quite a low irradiance (see Table 1), is in agreement with such a notion. Typical features regarding the membrane lipid composition of diatoms were also shown for H. spicula in this study. They include dominance of MGDG and DGDG species with long chain highly unsaturated fatty acyls, mainly C20:5 (EPA) and C20:4 (AA), a near complete absence of C18 fatty acyls, and abundant C16 fatty acyl chains with up to four double bonds in addition to the prevalence of SQDG with C16 acyl chains instead of PG [10,19]. Indeed, SQDG(16:0/16:1) accounted for 91.8% and 73.9% of SQDGs and 78.8% and 89.1% of SQDG + PG, under LL and HL, respectively. As a general pattern, SQDG species rose, whereas PG species remained unchanged under HL compared with LL, which is in agreement with the relevant role played by SQDG in diatoms regarding the inhibition of diadinoxanthin de-epoxidation as opposed to the shield of MGDG surrounding the FPCs that favors diadinoxanthin and diatoxanthin solubilization [19,62]. Nonetheless, MGDG and DGDG species with a C18/C16 composition were reported to be abundant in Haslea ostrearia, whereas C20:5/C16:n species were a minority [30].
The Raman spectrum is more stable and sensitive to non-polar compounds, whereas FTIR spectrum is more reliable and sensitive to polar compounds (i.e., proteins) [27]. Therefore, major signals in FTIR are due to proteins, which account for about 70% of the thylakoid membrane area [8], and saccharides, whereas major bands in Raman are due to lipids and especially to double bonds from carotenoids and fatty acyl chains of glycerolipids (Tables 4 and 5). Nonetheless, aliphatic chains of amino acids may contribute to Raman signals from proteins as well [33]. The subtle differences in lipid composition and, possibly, thylakoid membrane organization between K. flaccidum and Oocystis sp. were not reflected in the vibrational spectra. Thus, both algae showed overlying FTIR and Raman spectra. Nonetheless, two features could help distinguish between the algae, given that they have a differential Raman relative intensity in the high-frequency range around 2930 cm −1 , which could be related to a high content of galactolipids in Oocystis sp. compared to K. flaccidum and a ratio of the band at 1638 cm −1 to the band at 1040 or 1045 cm −1 in the FTIR spectrum, which is higher than 1 for Oocystis sp. and lower than 1 for K. flaccidum ( Figure S4), though this feature might depend on the nutritional status [29,43,63].
The major Raman bands at 1157 and 1524.9 cm −1 , observed in the two green algae, were also shown in H. spicula and M. vaginatus (Table 4), a fact that shows the relevance of carotenoids in Raman spectra. However, the Raman spectrum of M. vaginatus exhibited four specific bands, which may likely arise from the phycobilisomes (PBSs), a distinctive feature of cyanobacteria. Indeed, the reduction of the relative intensity of the bands at 1582.8 and 1629.1 cm −1 under HL compared to LL supports such a conclusion. The typically reported FTIR bands of cyanobacteria were also shown for M. vaginatus in this study [42,44]. Conversely, specific bands could be ascribed to H. spicula in the FTIR spectrum, which can be derived from the fucoxanthin-protein complexes (FPCs), being, in turn, specific features of diatoms [11,19,62,64]. Using resonance Raman spectroscopy with different excitation wavelengths, the organization of FPCs in the diatom Cyclotella meneghiniana has been studied with regard to both the carotenoid binding and oligomerization of the FPCs [11,65]. These studies have shown fucoxanthin molecules specifically binding to the FPCa or the FPCb complexes, with fucoxanthin molecules in three different excitation stages that mainly absorb in the blue, red, or green wavelengths. The prevalence of diadinoxanthin binding to FPCb could also be discerned [65]. Two ring-breathing modes of chlorophyll c2 were also shown to rely on the trimeric or the oligomeric forms of the FPCs. The different organization of the thylakoids between green algae and diatoms [62] gives rise to specific bands for the same chemical groups, depending on the particular environment, as this fact may shift the absorption maxima. In this regard, specific bands were observed for H. spicula in the FTIR spectrum, which may be due to chlorophyll c and fucoxanthin, two molecules that are not present in the green algae and the cyanobacteria.

Experimental Setup
Algae were collected from an agricultural field in Burgos (Spain) with geographical coordinates 42 • 29 44.2" N (latitude) and 3 • 48 17.5" W (longitude). The soil is classified as calcaric cambisol (CMc) according to the FAO [66]. The main chemical properties are: pH (water 1:2.5 w/v) 7.9; electrical conductivity (water 1:5 w/v, 25 • C) 0.498 dS m −1 ; soil organic matter 3.92%; total nitrogen 2.28 g kg −1 . For algae collection, a small piece of geotextile polypropylene fiber (120 g m −2 ) was placed in a run-off zone of the cultivated field and left for three weeks. After this period, the textile fiber was removed, immersed in BG-11 culture medium, and transported to the laboratory for isolation of the algae and microscope inspection. Five aliquots of the liquid suspension were obtained from the fiber-derived culture and inoculated in agar plates with BG-11 medium. Subsequent sub-samplings of diverse colonies in the agar cultivations obtained in the initial agar plate cultures were carried out until non-axenic isolates of each algal species were obtained, visually confirmed by microscopic inspection.
The diverse algal species were identified by morphological inspection using a microscope and 16S or 18S rDNA sequencing, along with the phylogenetic tree position for the forward + reverse primer similarity, and the determination of the photosynthetic pigment profile by means of high-pressure liquid chromatography with diode-array detection (HPLC-DAD,Thermo Scientific, Madrid, Spain).
Samples used in this study for biochemical analysis were obtained from BG-11 liquid medium cultures for all the algae and from BG-11 solid agar medium in the case of M. vaginatus. All cultures of algae and cyanobacteria were kept in a culture chamber with an illumination of~15 (low light, LL condition) or~45 µmol photon m −2 s −1 (high light, HL condition), and 20 • C. The algal and cyanobacterial cells were allowed to grow for at least three growth cycles under each illumination before they were used for the experiments.

Photosynthetic Pigment Measurement
Photosynthetic pigments in a methanolic extract were analyzed by high-performance liquid chromatography with photodiode array detection (HPLC-DAD) using the same chromatographic method as in Montero et al. [24]. A FINNIGAN SURVEYOR PLUS chromatography system (Thermo Scientific) equipped with a Quaternary LC pump, autosampler, and PDA detector was used for the HPLC-DAD measurements. Pigments were identified according to the retention time and UV-Vis spectrum (350-700 nm). Chlorophyll a (Chla), zeaxanthin (Z), and β-carotene (bC) were quantified after regression curves were drawn using commercial standards from SIGMA-ALDRICH, Merck KGaA, Darmstadt, Germany (references are C5753 for Chla, 1733122 (USP) for Z, and 1065480 (USP) for bC). Other chlorophylls were also quantified using the Chla regression curve. The regression curve of Z was used for the quantification of other xanthophylls as well. Each pigment was quantified at its own maximum absorption wavelength (Max-Plot).

UPLC-QToF-MS Measurements
A dichloromethane:methanol (2:1, v/v) extract was analyzed by ultra-performance liquid chromatography coupled to quadrupole time-of-flight mass spectrometry (UPLC-QToF-MS), as in Montero et al. [24]. After lipid extraction, the supernatant was evaporated to dryness and the pellet resuspended in methanol:water (9:1, v/v). An Acquity™ UPLC system (WATERS, Manchester, UK) equipped with an automatic injector (Sample Manager) and a binary solvent pump (Binary Solvent Manager) was used for liquid chromatography. The output of the liquid chromatographer was connected to a SYNAPT G2 HDMS mass spectrometer (WATERS, Manchester, UK), with a time-of-flight analyzer (QToF) and an electrospray ionization source (ESI). The chromatographic column was an Acquity UPLC BEH HSS T3 100 × 2.1 mm, 1.7 µm p.s., with a 10 × 2.1 mm precolumn (vanguard column) of the same stationary phase. Solvents were (A) methanol:water:formic acid (50:50:0.5, v/v/v) and (B) methanol:acetonitrile:formic acid (59:40:0.5, v/v/v), both with 5 mM ammonium formate. Compounds were eluted at a flow rate of 0.35 mL/min using the gradient that follows: initial, 100% A; 1 min, 100% A; 2.5 min, 20% A; 4 min, 20% A; 5.5 min, 0.1% A; 8.0 min, 0.1% A; 10 min, 100% A; this was kept isocratic for 2 min to recover initial pressure before the next injection. Samples were analyzed with negative ionization using an MS E method [67]. Compounds were identified by the m/z value, the elemental composition compatible with the isotopic distribution and relative retention time, and specific fragments from the MS E function. Samples from three independent cultures were analyzed for each alga and the cyanobacterium. For glycerolipid quantification, the commercial standards that follow were used: from Avanti Polar Lipids Inc 59752-57-7 and reference 37-1220-9). Fucoxanthin was quantified using the zeaxanthin standard, and chlorophylls were quantified using the Chla standard from SIGMA-ALDRICH, indicated above.

Fourier Transform Raman (FT-Raman) and ATR-Infrared (ATR-FT-IR) Spectroscopies
Attenuated total reflectance Fourier transform infrared (ATR-FT-IR) and FT-Raman spectra were done from both liquid medium and solid medium cultures for the algae. Aliquots of 5 mL volume of the algal suspension were filtered through PTF membranes, and the cell biomass was immediately placed on special devices used for the spectroscopic measurements of FT-Raman and ATR-FT-IR. For the solid medium cultures, an amount of M. vaginatus biomass or algae biomass was taken from the agar plates using a disposable microstreaker and placed on the special device for the spectroscopic measurements. Samples from two independent cultures were analyzed for every illumination. A Perkin Elmer Spectrum 100 FT-IR spectrometer with a Universal ATR sampling accessory was used for the ATR-FT-IR measurements. Spectral resolution was 4 cm −1 , and 16 cumulative spectra were acquired for each measurement. FT-Raman spectra were acquired with an FT-Raman Bruker RFS100/S instrument equipped with a LASER Klastech, Senza series (1064 nm, 500 mW), and a CCD Bruker D418-T (range 851-1695 nm),(Bruker Española S.A., Rivasvaciamadrid, Madrid, Spain).. The diameter of the LASER spot over the sample was 1000 µm. Other parameters were 10 mm slit, a scan rate of 1.6 kHz, and number of accumulations of 1024 [24]. All samples were harvested just a few minutes before the measurement was carried out.

Statistical Analysis
Pairwise comparisons of the diverse lipid contents between low (LL) and high (HL) light were done using the Student t-test, and significant differences were accepted for p < 0.05. Samples from three independent cultures (n = 3) were measured for each illumination.
Principal component analysis (PCA) of the UPLC-QToF data was done using the Extended Statistic (XS) application included in MarkerLynx ® software (V4.1 SCN 803, WATERS, Manchester, UK). This application includes the statistical tools of the SIMCA-P+ software package (Umetrics EZ info 2.0; Umea, Sweden). The statistical parameters R 2 X(cum) and Q 2 (cum), which explain the variability of X-variables and indicate model predictive capability, respectively, were determined [68]. Previously, a three-dimensional data array (Pareto-scaled) comprising the variables sample (including the blanks), retention time_m/z values (molecular features), and normalized (scaled to Pareto variance) signal intensity of the m/z value was generated using MarkerLynx ® software (WATERS, Manchester, UK).

Conclusions
The membrane lipid profile of two green algae, one filamentous and one unicellular, a diatom, and a cyanobacterium from an edaphic assemblage, is shown. They differ in the relative proportion of the main glycerolipid classes and the response to low and high light exposure. The differential lipid composition between them allowed their clear separation in principal component analysis (PCA). The high abundance of MGDG(18:3/16:4) and the high variety of DGDG species with a high DGDG/MGDG ratio in Oocystis sp. compared to K. flaccidum suggest subtle differences in the architecture of the photosynthetic apparatus between the two green algae. Nonetheless, these subtle differences in lipid composition were not clearly reflected in the Raman and FTIR spectra. The typical lipid composition of diatoms, with an abundant content of C20:5 and C20:4 acyls in MGDG and DGDG species, was also found for H. spicula in this study. This alga showed specific bands in the FTIR spectrum, whereas M. vaginatus showed specific bands in the Raman spectrum.