Mass spectrometry imaging reveals the spatial distribution of essential lipids in Daphnia magna – potential implications for trophic ecology

ABSTRACT Lipids and fatty acids are key dietary components for the nutrition of organisms at all trophic levels. They are required to build cellular structures such as cell membranes, serve as energy storage, and take part in signal transduction cascades. For decades, ecological research investigated how dietary fatty acid availability contributes to the fitness of individuals and their populations. The omega-3 polyunsaturated fatty acid eicosapentaenoic acid (EPA) is of particular interest because its dietary availability determines the fitness of many aquatic consumers. Because of the small body size of zooplankton, only bulk tissue fatty acid analysis was previously performed, and thus the tissue-specific importance of EPA for zooplankton remained elusive. We used matrix-assisted laser desorption/ionization–mass spectrometry imaging (MALDI-MSI) to reveal the tissue-specific distribution of functional phospholipids in the herbivorous zooplankton Daphnia magna. We demonstrate several lipid species for heart, egg, gut, gonad, somatic, and neurological tissues of D. magna, including the compound eye as well as the optical and cerebral ganglion. The compound eye revealed a large diversity in lipid species containing EPA, which were also found in other neurological tissues and eggs. Such knowledge of tissue-specific fatty acid requirements is essential to investigate how selective allocation of dietary fatty acids within this key grazer affects processes on a functional and molecular level from the individual to food web scales. This methodological advancement will facilitate investigations on how invertebrate physiology and behaviour adjust to changing environmental conditions and potentially affect food web structures, including the trophic transfer of dietary fatty acids.


Introduction
The success of an individual depends on its ability to survive, grow, and reproduce in its environment, which has a direct bearing on overall ecological fitness. Fitness optimization typically is faced with trade-offs; for example, assimilated dietary energy can be either allocated to an improved adaptation to an individual's current environment or to future generations via investment into reproductive tissues. Lipids and their fatty acids (FAs) contain more energy than carbohydrates or proteins and are functionally important for the survival of organisms as constituent parts of cell membranes (e.g., phospholipids), as hormone precursors, and as energy storage (e.g., triacylglycerols). The lipid composition can vary substantially among species, and lipids are transferred to consumers at higher trophic levels of aquatic food webs (Strandberg et al. 2015, Jardine et al. 2020. For example, aquatic primary producers differ in their biosynthetic ability to produce specific FAs (Lang et al. 2011, Taipale et al. 2013) and, consequently, consumers retain or bioconvert dietary FAs to physiologically required FAs ("trophic upgrading"; e.g., Murray et al. 2014).
Herbivorous zooplankton are a major trophic link between primary producers and higher consumers in aquatic food webs (Sommer et al. 2012) and are also key for the trophic transfer of LC-PUFA Vrede 2006, Sperfeld andWacker 2012). Dietary EPA is particularly important for somatic growth and reproduction of Daphnia sp. Boersma 2003, Martin-Creuzburg andvon Elert 2009) because endogenous synthesis of EPA from ALA or DHA does not fully compensate its limited uptake via food to sustain optimal growth conditions at ambient temperature (Weers et al. 1997, von Elert 2002 and decreases the survival of cladocerans (Müller-Navarra et al. 2000) and copepods ). In addition to taxonomic differences in PUFAs among algal species, decreasing LC-PUFAs has been mainly linked to eutrophication and browning of freshwater ecosystems (Taipale et al. 2018).
Detailed understanding of the physiological mechanisms driving specific dietary PUFA demands requires knowledge of tissue-specific lipid composition and metabolism. To date, analyzing lipids at the tissuespecific levels on aquatic invertebrates has been hampered by their relatively small body size, which makes tissue-specific dissection difficult to perform in the quantities needed. In this study, we applied matrixassisted laser desorption/ionization-mass spectrometry imaging (MALDI-MSI) to investigate the tissue-specific variability in lipid composition of D. magna at a small spatial scale (∼25 μm). In contrast to optical methods that provide spatial information about the absorption of electromagnetic waves with different wavelengths, MALDI-MSI displays the relative distribution of different mass to charge ratios, corresponding to the masses of ions generated from different compounds, across a sample with a spatial resolution of ∼5-50 µm (Sparvero et al. 2012, Thomas et al. 2012, Ràfols et al. 2018. This method has already been used to identify different organ-specific phospholipid species containing LC-PUFAs in terrestrial invertebrates (Bhandari et al. 2015) as well as in individual layers of the human retina (Zemski Berry et al. 2014). Because experiments in many species have shown EPA and DHA to contribute to behavioral and cognitive traits, we tested the hypothesis that lipids containing EPA are selectively retained in neuronal structures of D. magna, as shown experimentally for other larger species (Pilecky et al. 2021). Introducing this method to aquatic ecology has the potential to increase our understanding about lipid metabolism of physiologically important FAs, particularly PUFAs, within zooplankton and will further provide new insights into tissue-specific retention of essential lipids in smallsized organisms across the planktonic food web.

Culture of Daphnia magna, algae, and sample preparation
The unicellular green alga Acutodesmus obliquus (Kützing, 1833) and the cryptophyte Cryptomonas ozolinii (Skuja, 1939) were cultured in Wright's cryptophyte (WC) medium at 20°C (Guillard and Lorenzen 1972). Both algae are easily ingested by Daphnia but differ in their nutritional quality with respect to LC-PUFA composition (Taipale et al. 2013). A. obliquus lacks PUFAs with more than 18 C-atoms, whereas C. ozolinii is rich in EPA and DHA (Ahlgren et al. 1992, Masclaux et al. 2009; Table 1, Fig. 1). D. magna (obtained from the University of Clermont Auvergne, France) were kept in filtered (<0.7 μm) lake water (20°C) supplemented with 10% AdaM medium and fed ad libitum with either alga (Klüttgen et al. 1994). Daphnia were continuously fed over generations with either diet and collected for analysis 2 days after the onset of egg formation (primiparity).

Separation and quantification of phospholipid fatty acids
We filtered 100 mL of algal culture through precombusted GF/C filters (VWR, Leuven, Belgium), and the filters were frozen immediately at −80°C. D. magna were washed twice in 0.9% NaCl to remove attached algae and were subsequently frozen. All measurements were performed in quadruplet. Lipids were extracted from all samples according to Heissenberger et al. (2010). Briefly, freeze-dried samples were homogenized and mixed with chloroform:methanol (2:1 v/v) following the addition of 0.9% NaCl, sonicated for 10 min, vortexed for 1 min and centrifuged (1500 g) (all 3 times) to remove non-lipid materials. Extracted lipids were evaporated to a final volume (1.5 mL) under a gentle N 2 gas flow. Prior to FA methylation, zooplankton phospholipids (PL) and triacylglycerols (TAG) were separated by one-dimensional thin-layer chromatography (TLC) on 10 cm × 10 cm silica gel plates (TLC silica gel 60, Merck, Darmstadt, Germany) using hexane:diethyl ether:methanol:formic acid (90:20:3:2, v/v/v/v) as mobile phase. The TLC method has been previously described elsewhere (Böhm et al. 2014). Before the solvent reached the end of the plate, the development was stopped and the plates were dried, sprayed with 0.05% (w/v) 8-anilino-4-naphthosulfonic acid in methanol, and viewed under UV-light to identify zones of TAG and PL that were subsequently scraped off the plate. For fatty acid methyl ester (FAME) formation, samples were incubated with sulfuric acid: methanol (1:100 v/v) for 16 h at 50°C, following the addition of KHCO 3 and hexane. Samples were shaken, vortexed for 1 min and centrifuged (1500 g), and the upper organic layers were collected, pooled, and concentrated under a N 2 gas flow.
FAMEs were quantified using a gas chromatograph (TRACE GC ThermoFisher Scientific, Waltham, MA; Flame Ionisation Detector: 260°C, carrier gas: 1 mL/min He, detector gases: 40 mL/min H 2 , 45 mL/min N 2 , 450 mL/min air, 140°C temperature [hold: 5 min], temperature ramp 4°C/min to 240°C [hold: 20 min] = 50 min total run time). An SP-2560 column (100 m, 0.25 mm internal diameter, 0.2 µm film thickness; Supelco, Bellefonte, USA) was used for FAME separation and quantification. Chromeleon 7 software (Thermo-Fisher) was used for peak integration. FAMEs were identified by comparison of retention times to known reference standards (37-component FAME mix, 47885-U, Supelco). FA mass fractions were quantified using calibration curves based on different known standard concentrations. FAs were reported as relative values (weight %) compared to the total mass fraction of identified FAMEs. All values are given as mean (standard deviation).

Matrix-assisted laser desorption/ionization (MALDI)mass spectrometry imaging (MSI)
Adult D. magna were embedded into 5% carboxymethylcellulose (Sigma-Aldrich) because the optimal cutting temperature compound has been shown to suppress ion generation in MALDI-MSI (Nelson et al. 2013). Histological slices (20 µm) were prepared by cryocutting on a cryostat Leica CM1850 (Leica, Wetzlar, Germany) at −18°C and mounted onto ITO-coated glass slides (Bruker Daltonics, Bremen, Germany). Slides were washed in 50 mM ammonium formate (pH 6.4) for 10 s to remove remaining embedding medium, followed by freeze-drying and storage in sealed tubes with silica beads at −20°C until imaging. 1,5-diamino naphthalene was applied to the slides by sublimation using a selfmade sublimation chamber (Rettberg, Göttingen, Germany) at 140°C and 50 mTorr (Caughlin et al. 2017) for 3 min. MALDI imaging was performed on an Autoflex Speed MALDI-TOF Instrument (Bruker Daltonics) using only negative mode. Mass spectra were recorded in reflectron mode in the range of m/z 500-1200. Lipids were identified by database comparison (lipidmaps.org) and by interpretation of the electrospray ionization (ESI)-MS n spectra obtained by collisioninduced dissociation (CID) from lipid extracts, obtained as described earlier, on an amaZon speed ETD ion trap instrument (Bruker Daltonics, Billerica, MA, USA). Tentative identification of lipid species was based on data from CID where available, or by selecting the  Fig. 1).  The ion at m/z 687.5 was only detected in neural tissue and could not be unequivocally identified but might correspond to PA(diSDA). Another noteworthy signal was obtained at m/z 901.5, which was identified as [PI(diEPA)-H] − . This lipid is mainly located in the compound eye, as well as the cerebral ganglion, although in the latter the ion at m/z 903.  (Fig. 2 and 3).

Discussion
This study demonstrates that dietary PUFAs are selectively allocated and retained in D. magna visual and reproductive tissues in the form of polar lipids containing EPA. As hypothesized, lipids containing EPA were most abundant in neuronal tissue, particularly the compound eye and the cerebral ganglion, but to a lesser extent also in gonads and eggs. The tissue-specific EPA retention (Kainz et al. 2004, Mariash et al. 2011, Hartwich et al. 2013 indicates that EPA in D. magna, and perhaps also in other cladocerans, is required for light detection and/or movement (e.g., diel vertical migration) and reproduction, and thus to optimize fitness. This finding aligns with experimental studies linking long-chain n-3 PUFA deprivation to impaired neurological functions in other invertebrates and vertebrates (Pilecky et al. 2021).
Natural populations of herbivorous zooplankton like D. magna are subjected to feeding pressure by both vertebrate (fish) and invertebrate predators (Kerfoot and Sih 1989). Only their extremely high population growth rates (enabled by their parthenogenetic reproduction) and adaptive behavioural responses allow them to compensate the high mortality rates caused by predation and sustain viable populations despite massive predation losses (Boersma et al. 1998). To perform adaptive behavioral responses, daphnids require an accurate perception of environmental cues that reflect their individual predation risk. Such cues include light (required for the prey perception of visually oriented predators; Ringelberg 1999) and chemical cues (so-called kairomones that indicate the presence of predators; Weiss et al. 2018, Hahn et al. 2019. Both light and kairomone perception require sensory and neuronal structures for an adequate and adaptive processing of environmental cues. The functioning of neuronal and sensory tissues of aquatic consumers depends on the selective allocation of essential FAs to those tissues (Tocher and Harvie 1988). Consequently, the ability of prey organisms, such as D. magna, to selectively allocate essential dietary FAs to their neuronal tissue is probably a key determinant of their individual predation and mortality risk and ultimately of their population's survival probability.
To understand the mechanisms that determine this adaptive allocation of essential dietary compounds as well as the potential physiological constraints for those allocation processes, it is thus crucial to not only measure the bulk FA content of Daphnia, but to explore the patterns of allocation of these compounds into their tissues. This step was not possible with the analytical methods used by aquatic ecologists to date. MALDI-MSI allows us to gain this information on a high spatial resolution and thus will allow ecophysiologists and population ecologists to investigate these key processes in more detail than previously possible. Ultimately, this knowledge will help us understand the physiological (and behavioural) constraints of zooplankton to sustain viable populations despite massive and continuous predation losses. Despite large differences between the FA profiles of dietary A. obliquus and C. ozolonii, these diets did not impact the PL composition of D. magna to a similar extent. The only significant difference in PL was observed in higher relative ALA content compared to EPA when dietary EPA was lacking, which suggests a high resilience and buffering capacity of Daphnia achieved by adapting gene expression to the individual requirements (Windisch and Fink 2018. Daphnids convert ALA to EPA at a rather high rate depending on the EPA content of the diet (Pilecky et al. 2022). However, it remains to be tested which physiological traits were affected by the observed differences in ALA and EPA, and whether these differences were due to a lack of dietary EPA or dietary energy, which had to be invested into n-3 PUFA conversion (e.g., ALA to EPA).
We hypothesized that low availability of dietary EPA can limit predator avoidance in Daphnia by reduced diel vertical migration, although whether due to reduced perception of kairomones or altered behavior as a response to kairomone perception is unclear (Brzeziński and von Elert 2015). Diel vertical migration can also be blocked by histamine antagonists (i.e., by blocking a G-protein coupled receptor), whose activity is highly associated with n-3 PUFAs in neuronal membranes (McCoole et al. 2011, Pilecky et al. 2021. Unknown is how these thresholds for light and kairomone perception shift as a response to altered biochemical composition of the neuronal tissue, induced by different dietary qualities. In addition to predator perception, the ability to evade the actual predation event is under high selective pressure in zooplankton. Thus the assumption that zooplankton with a limitation of n-3 LC-PUFA in their diet can develop worse neurophysiological traits that account for a lower predator evasion and thus decrease survival is plausible. These hypotheses can now be tested for zooplankton and other small animals using MALDI-MSI.
Furthermore, the investment of energy into the conversion of n-3 PUFAs reduces somatic growth and reproduction of Daphnia (Becker and Boersma 2005, Abrusán et al. 2007, Sperfeld and Wacker 2012, Ravet et al. 2012. The physiological EPA requirements were found to be temperature dependent because Daphnia integrate more EPA into their cell membranes at lower temperature (Sperfeld and Wacker 2012). This finding is particularly interesting in conjunction with global warming, which is predicted to reduce the n-3 LC-PUFA production by phytoplankton (Hixson andArts 2016, von Elert and. Thus, the presented identification of tissue-specific PUFA retention may be key to answering questions about effects of altered dietary FA supply on tissue-specific FA allocation and zooplankton performance. The use of MALDI-MSI has some limitations. Signal intensity cannot be simply used for quantification because individual compounds give different responses based on their different ionization efficiency (Porta et al. 2015), and intensities between different sample spots can vary because of inhomogeneities in MALDI matrix deposition and ion suppression effects due to other components in the tissue. Thus, only relative abundances of the same lipid class can be considered and compared. Here, we focused only on ions detectable in the negative mode, which excludes PC. However, functional n-3 PUFAs have been described to be particularly linked to PE, PI, and PS (Chen et al. 2009) because their tendency to form negative membrane curvature benefits the activity of receptors involved in neurological signaling, particularly in conjunction with the flexibility unique to n-3 PUFAs (Seddon 1990, Albert et al. 1998). In addition, lipid species that are difficult to ionize, such as TAGs (i.e., storage lipids), cannot be routinely analyzed using MALDI-MSI. Despite these limitations, MALDI-MSI provides as yet unseen details about the somatic FA distribution of small zooplankton or other aquatic consumers (e.g., <1 mm). Such detailed Figure 3. Overlay of selected MALDI-MS-generated images with a brightfield microscopic image of Daphnia magna containing one egg with representative mass spectra. Phospholipids (PLs) containing EPA were most abundant in neuronal tissue, particularly the compound eye, as well as the egg. The signal at m/z 901.5 (dark blue) was characteristic for the compound eye and tentatively assigned to the PI containing 2 EPAs while m/z 903.5 (light blue) was more prevalent in the cerebral ganglion and represents mainly the PIs containing EPA and ARA. The image of m/z 881.5 (red) shows one of the EPA-containing PLs that is particularly abundant in the egg, but it was also found in the optical ganglion and most likely constitutes PI(EPA/18:1). The heart also contained some discriminating lipids such as the one detected at m/z 831.5 (green), which is either PI(ALA/16:0) or PI(LIN/16:1). PLs characteristic of the gut consisted of rather short FAs, such as PE(14:0/14:1) indicated by the ions at m/z 632.5 (purple). Gonads were little different to gut and somatic tissue for most m/z values, but 721.5 (yellow) was most likely associated with them and constitutes PA(EPA/18:0). PE = phosphatidylethanolamine, PA = phosphatic acid, and PI = phosphatidylinositol. lipid analyses of aquatic invertebrates will provide the basis to elucidate how organs (e.g., eyes, brains, or gonads) modify their FA composition in response to dietary or environmental changes, which are expected to increase with climate change (Maazouzi et al. 2008).

Conclusions
This study shows that EPA is abundant in neural and reproductive tissues of D. magna regardless of dietary EPA supply, suggesting that this LC-PUFA is truly essential for these tissues in D. magna, and likely other cladocerans. Thus, they must either allocate sufficient EPA or convert dietary precursors to satisfy the needs of the neuronal tissues and subsequently enable reproduction. Organ-specific differences in the EPA content of polar lipids could be identified, and further research is warranted to provide more detailed information about changes in lipid composition and metabolism. Testing how this composition is altered depending on changes in environmental conditions (i.e., water temperature, browning), as well as food sources, will be particularly interesting. Revealing if and how zooplankton are able to compensate for a lack of dietary LC-PUFAs and how environmental changes affect their neurological structures with subsequent effects on environmental fitness will open new research avenues in plankton ecology.