Crucian carp (Carassius carassius) is an excellent vertebrate model for studies on temperature adaptation in biological excitable membranes, since the species can tolerate temperatures from 0 to +36°C. To determine how temperature affects the lipid composition of brain, the fish were acclimated for 4 wk at +30, +16, or +4°C in the laboratory, or seasonally acclimatized individuals were captured from the wild throughout the year (temperature = +1 to +23°C), and the brain glycerophospholipid and sphingolipid compositions were analyzed in detail by electrospray-ionization mass spectrometry. Numerous significant temperature-related changes were found in the molecular species composition of the membrane lipids. The most notable and novel finding was a large (∼3-fold) increase of the di-22:6n-3 phosphatidylserine and phosphatidylethanolamine species in the cold. Since the increase of 22:6n-3 in the total fatty acyl pool of the brain was small, the formation of di-22:6n-3 aminophospholipid species appears to be a specific adaptation to low temperature. Such highly unsaturated species could be needed to maintain adequate membrane fluidity in the vicinity of transporters and other integral membrane proteins. Plasmalogens increased somewhat at higher temperatures, possibly to protect membranes against oxidation. The modifications of brain lipidome during the 4-wk laboratory acclimation were, in many respects, similar to those found in the wild, which indicates that the seasonal changes observed in the wild are temperature dependent rather than induced by other environmental factors.
- docosahexaenoic acid
- mass spectrometry
- polyunsaturated fatty acid
the essential functions of cell membranes have to be maintained in all metabolic and environmental conditions, and, therefore, adaptive mechanisms regulating membrane lipid composition and physical properties exist (8, 26, 92). Among ambient factors, temperature has the largest effect on the physical properties of biological membranes, especially in ectothermic animals, which are, therefore, excellent models for studying the temperature-induced restructuring of excitable membranes, including remodeling of lipid compositions. The present view of temperature acclimation is based on the theory of homeoviscous adaptation of membrane lipids (77). The potential biochemical mechanisms of temperature acclimation in vertebrate membranes include modulation of 1) cholesterol content, 2) phospholipid (PL) class composition, and 3) PL acyl chain composition, and 4) restructuring of glycerophospholipid (GPL) molecular species, i.e., “reshuffling” existing acyl chains to form new pairs for the new molecular species without changing the overall fatty acid composition (8, 13, 40, 71, 92).
Some adaptive phenomena modulating lipid composition of vertebrate tissues are very rapid and occur within a few hours, whereas others require a considerably longer time. For example, after a sudden temperature drop, the PL class composition of membranes of ectothermal animals is altered within a few hours, but the full response in acyl chain desaturation takes several weeks (32, 35, 36, 40, 92). On the other hand, seasonal changes in membrane physiology may require other cues from the environment, in addition to temperature. Therefore, changes in membrane lipidomes may differ in animals that are acclimated to different temperatures in the laboratory vs. those seasonally acclimatized in the wild.
Temperature responses in molecular species composition of fish brain have been reported previously only for the two main GPL classes, i.e., phosphatidylcholines (PC) and phosphatidylethanolamines (PE) (12, 16). However, recent developments in membrane biology encourage conducting more detailed analysis of PL molecular species also in the quantitatively smaller classes of lipids. The recent evidence implies that biological membranes are heterogeneous and contain microdomains with a specific lipid composition (50, 76). According to this so-called “raft” model, specific domains enriched in cholesterol and sphingolipids (SL) may assemble in Golgi and the plasma membrane, and they are postulated to play a role in protein and lipid transport and signal transduction. Alternatively, domain formation may occur at lipid-protein interfaces, where certain lipids could be enriched to form a shell or annulus (2). A logical prediction arising from the putative heterogeneous nature of the membranes is that the adaptive adjustments of lipid composition may vary between different lipid domains (96). Consequently, the adaptive changes of lipid composition may be larger in certain lipid classes with specific functions than in the bulk membrane, which warrants detailed analyses of the whole brain lipidome.
Here, we analyzed by electrospray ionization-mass spectrometry (ESI-MS) the changes in the brain lipidome of crucian carp (Carassius carassius) occurring either in laboratory acclimation or seasonally in the wild. Crucian carp is an excellent ectothermal model organism, as it tolerates temperatures ranging from 0 to +36°C and also complete anoxia for over 4 mo at +2°C (39, 66). We show that cold induces several alterations in the lipid composition of brain membranes. The most notable one was the major increase in the content of the highly unsaturated di-22:6n-3 phosphatidylserine (PS) and PE species probably needed to maintain adequate membrane fluidity in the vicinity of transporters/channels or other integral membrane proteins. Significant changes were also noted in the concentrations of PE plasmalogen, which could relate to temperature-dependent changes in the need to protect the membranes against lipid oxidation or to promote membrane fusion.
MATERIALS AND METHODS
Crucian carp were captured from a small lake located near Joensuu, Eastern Finland, every second month from May 2002 to March 2003. The fish traps were kept at the same location and depth throughout the year, and every time five specimens (body mass 21−47 g) were taken for analyses of brain lipids. At the time of sampling, water temperature was measured at the trap. In addition, in August 2004, crucian carp were captured from the same lake and transferred to laboratory tanks, and during September-October acclimated to +30, +16, and +4°C for 4 wk (n = 6–7). Comparable groups were formed from individual fish having body weight of 20–65 g, length of 10–19 cm, and brain weight of 80–160 mg. The fish at +30 and +16°C were fed with TetraMin aquarium fish food, rich in all essential polyunsaturated fatty acids (PUFA) (detailed composition in Supplemental Table S1). (The online version of this article contains supplemental data.) However, at +4°C, crucian carp do not eat and thus were not offered any food. The Ministry of Agriculture and Forestry of Finland has permitted to the University of Joensuu legal rights to maintain experimental animals. All of the experiments were made with consent of the local Committee for Animal Experimentation in the University of Joensuu and follow the national and European legislation on animal experiments.
The whole brain of the fish was removed, washed twice in physiological saline, and stored at −80°C for 3–4 wk. The brain tissues were homogenized and extracted according to Folch et al. (17). Extract aliquots were spiked at the one-phase stage with a cocktail consisting of diunsaturated PL and saturated SL standards for mass spectrometry (detailed below). After evaporation under nitrogen flow, the lipids were dissolved either in the chromatography solvent for liquid chromatography-ESI-MS (LC-MS) (see below) or in chloroform-methanol (1:2 vol/vol) for direct infusion tandem mass spectrometry (MS/MS) experiments. In the latter case, 1% NH4OH was added just before the analysis. In addition, the total PL and cholesterol contents of the brain extracts from the laboratory-acclimated fish were determined by spectrophotometric methods (4, 18).
Synthetic di-14:1, di-20:1, and di-22:1 PC species were purchased from Avanti Polar Lipids (Alabaster, AL). Di-16:1, di-20:1, and di-22:1 PE and PS species were synthesized from the corresponding PC species using phospholipase D-mediated (Streptomyces species, Sigma, St. Louis, MO) transphosphatidylation (45). The 16:0/18:2 and 18:0/18:2 phosphatidylinositol (PI) species were isolated from total yeast PI (Sigma) by C18 reversed-phase chromatography using 1–5% H2O in methanol as the eluent. The 14:0 and 17:0 ceramides were purchased from Avanti Polar Lipids and deuteriated 16:0 sulfatide from Larodan (Malmö, Sweden). The 15:0, 21:0, and 25:0 sphingomyelin (SM) species, and the 15:0 and 25:0 galactosylceramide (GalCer) species were synthesized from the lyso forms (Avanti Polar Lipids) and the respective acid chlorides (Larodan), as detailed previously (60). The lipid standards were dissolved in chloroform-methanol (9:1) and stored at −20°C. The concentrations of PL standards were determined by phosphate analysis (4), and those of SL species according to Naoi et al. (62).
The molecular species compositions of major PL classes PC, PE, PS, PI, and SM were analyzed using LC-MS, essentially as described previously (38, 45). Chromatographic separations were carried out isocratically on Lichrosphere diol-modified silica columns (250 × 1 mm; 5-μm particles) at 35°C by using an Ultimate nano-HPLC system equipped with a Famos autosampler (LC Packings, Amsterdam, the Netherlands) or by using Surveyor plus HPLC system (Thermo Electron, Waltham, MA). The isocratic solvent system consisted of hexane-isopropanol-water-formic acid-triethylamine (628:348:24:2:0.8 vol/vol) and was eluted at 50 μl/min. The eluent was introduced to the electrospray source of a Quattro Micro triple quadrupole mass spectrometer (Micromass, Manchester, UK) or a Finnigan LTQ linear ion trap mass spectrometer (Thermo) operated in the negative-ion mode, which allows the detection of all of the GPL classes and SM (choline lipids as formate adducts). The source temperature was 90°C, and the capillary voltage 3.8 kV. Default values were used for the other parameters. The spectra were scanned from 500 to 1,000 mass-to-charge ratio with a frequency of 1 scan/s.
MS/MS and Data Analysis
To confirm the identity of the lipid molecular species, the crude lipid extracts in 1:2 chloroform-methanol were infused into the source of the quadrupole mass spectrometer at the flow rate of 6 μl/min. The collision energy of the instrument was set to 25–65 eV, and negative and positive ion modes were used. Argon was used as the collision gas. The PC, PE, PS, PI, SM, sulfatide, GalCer, and ceramide species were selectively detected using class-specific MS/MS scanning modes (10, 41, 81) (Fig. 1). The fatty acids present in the molecular species were identified employing acyl chain-specific precursor ion scans and product ion scans (14, 29). After the identification steps, quantification of the molecular species of the LC-MS data was carried out based on several internal standards (45) using the LIMSA software (27). The concentration of the lipid classes was obtained by summing of the concentrations of the molecular species in a class.
Fatty Acid Analysis
The lipid extract (∼15 mg lipid) was dried under nitrogen flow, and the lipid-bound and so-called free fatty acids were converted to methyl esters by heating at 95°C with 1% H2SO4 in methanol under nitrogen atmosphere for 100 min. The fatty acid methyl esters were extracted by hexane in two steps, dried, concentrated, and analyzed by a gas-liquid chromatograph (GC) using both flame ionization and mass detection (6890N network GC with flame ionization and 5973 MSD, Agilent) and DB-wax capillary columns (30 m, ID 0.25 mm, film 0.25 μm, J&W Scientific), as detailed in Käkelä et al. (46). Alkenyl chains of plasmalogen PL species were detected as dimethylacetals (DMA). The fatty acid methyl esters and DMA were identified based on retention time, mass spectrum, and comparisons with authentic (Sigma) and natural standards of known composition and published reference spectra (Christie W.W., www.lipidlibrary.co.uk/ms/masspec.html).
The statistical significance of the differences in the relative amounts of studied lipids and their side chains in the experimental groups were studied by one-way ANOVA, followed by the Newman-Keuls test of the means (P < 0.05). Regression analysis was used to study the temperature dependence of the highly unsaturated molecular species concentrations in the brain PE and PS, and the average double bond content of the sidechains of total lipids.
Fatty Acid Composition of the Total Brain Lipids
The acyl and alkenyl chain compositions of total brain lipids were determined by GC, as detailed in materials and methods. The average number of double bonds per hydrocarbon chain decreased nearly linearly with increasing water temperature, i.e., from 1.64 at 4°C to 1.28 at 30°C (Fig. 2), and this decrease was mainly due to increased proportions of saturated fatty acyl (SFA) and alkenyl chains, as well as a somewhat diminished proportion of the polyunsaturated ones (PUFA) at the higher temperatures (Fig. 3). Notably, the proportion of 22:6n-3 changed only slightly with the temperature (12 mol% at 30°C, 14 mol% at 4°C). Although the total proportion of monounsaturated fatty acids (MUFAs) did not differ significantly between the temperature groups (Fig. 3A), clear changes in the proportions of individual MUFAs were observed (Fig. 3B). In the cold, n-7 MUFA were favored over n-9 MUFA. The average hydrocarbon chain length varied only slightly, i.e., from 18.1 carbons in the warm water to 18.3 carbons in the cold (also in the wild).
The temperature dependence of the average number of double bonds (rising from 1.44 in August to 1.70 in November) was very similar to that of the laboratory fish (Fig. 2). The levels of saturated acyl and alkenyl chains were highest during the summer and decreased with temperature (Fig. 3C). The PUFA content showed an opposite pattern, i.e., increased from August to November.
Lipid Class Composition
The lipid class composition was quite independent of the temperature, with only two significant differences (Supplemental Fig. S1A): the percentage of SM was higher in the brains of the fish kept at 30°C and 16°C (5 and 7 mol%, respectively) vs. those kept at 4°C (3 mol%). Also, the proportion of sulfatides was significantly higher at 30°C (10 mol%) than at 4°C (5 mol%). The cholesterol-to-PL ratio was significantly higher in the brains of the fish kept at 30°C (0.75 ± 0.04) vs. those acclimated to 16°C and 4°C (each 0.62 ± 0.03) (n = 6–7).
No marked seasonal changes were found in the lipid class composition, including SM (Supplemental Fig. S1B).
PC Molecular Species
The molecular species compositions of each major PL class were analyzed by ESI-MS. In diacyl PCs, the relative amount of 18:1/18:1 increased toward cold and 16:0/16:1 peaked at 16°C, but was also higher at 4°C than at 30°C. The increases happened at the expense of 16:0/16:0, 16:0/18:1, and 18:0/18:1 (Fig. 4A). All polyunsaturated PC species increased as the temperature decreased, with the largest increase being observed for 16:0/22:6n-3. The changes in alkenyl-acyl species (plasmalogens) were generally smaller than those found for diacyl species.
The main PC species, i.e., 16:0/18:1, peaked at the highest water temperature in July and then decreased toward the winter. In contrast, two other major PC species, 16:0/16:1 and 16:0/22:6n-3 had their highest levels in May and March (Fig. 5A, Table 1). In addition, several minor PC species varied seasonally (Table 1).
PE Molecular Species
All major diacyl PE species were polyunsaturated, and their relative concentrations varied significantly with temperature (Figs. 4B and 6A). Notably, the di-22:6n-3 species was threefold more abundant at 4 vs. 30°C. Also, several other polyunsaturated species (e.g., 18:0/20:4, 18:1/20:4, and 16:0/22:6n-3) were more abundant in the cold, whereas the reverse was true for the plasmalogen species 18:0a/22:6n-3, having maximum values at 30°C (Fig. 4B).
From the summer to the winter, the proportion of di-PUFA PE species rose from 11 to 17% (Fig. 5, B and D), but the main di-PUFA species di-22:6n-3 increased less than in the laboratory (Fig. 6A). The 18:0/22:6n-3 was the most prominent PE species during the summer, but its levels decreased toward winter and fell below those of the di-22:6n-3 and 18:1/22:6n-3 species (Fig. 5B). The major PE plasmalogen species, 18:0a/22:6n-3 and 18:1a/20:4, varied little with season, but had minimum levels in May (Table 2). In addition, 18:0a/22:5 and several lesser plasmalogen species composed of C16 and C18 chains reached their minimum levels during the period of cold water (November '02 to March '03 or May '02) (Table 2).
PS Molecular Species
The most striking change among all studied lipid classes was the fivefold increase of di-22:6n-3 PS from 30 to 4°C (Fig. 4C). The temperature-dependent variation of the di-22:6n-3 and the di-PUFA PS totals were very pronounced and exponential (Fig. 6B). In contrast, the content of 18:0/22:6n-3, which accounts for one-half of total PS, hardly varied with the water temperature. Among other PS species, 22:1/22:6n-3, 16:0/18:1, and 18:0/18:1 decreased significantly with decreasing temperature (Fig. 4C).
Among the PL classes studied, the di-PUFA PS species were again those that varied most, i.e., from 18% in the summer to 28% in the winter (Fig. 5, C and D). The increase was mostly due to the di-22:6n-3 species, which increased linearly with decreasing temperature (Fig. 6B). In contrast, the three other major PS species, i.e., 18:0/22:6n-3, 18:0/18:1, and 16:0/18:1, diminished toward the winter (Fig. 5C, Table 3).
PI Molecular Species
The 18:0/20:4n-6 species comprised up to 80 mol% of total PI at each acclimation temperature. Statistically significant changes were found only for the minor 18:1/20:4n-6 species, which increased from 2 to 10%, as well as for the 16:0/20:4n-6 and 18:0/22:6n-3 species, which decreased toward cold (from 7 to 4% and 4 to 1%, respectively) (Supplemental Fig. S2A).
The major species 18:0/20:4n-6, which comprised ∼40% most of the year, dropped to 30% in July (Supplemental Fig. S2B). The next most abundant species, 18:1/20:4n-6, increased from summer (7%) to winter (15%), whereas the minor 16:0/20:4n-6 and 18:0/22:6n-3 species decreased slightly toward winter.
SL Molecular Species
Among the brain SL classes, the most prominent changes in molecular species composition were observed for the sulfatides and ceramides. The concentration of long-chain sulfatides, e.g., 24:1 and 26:1, were higher, and those of several (smaller) species with shorter acyl chains were lower at 30 vs. 4°C (Supplemental Fig. S3A). In contrast, in ceramides, there were less 24:1 and more 18:0 at 30 vs. 4°C (Supplemental Fig. S3B). In general, the molecular species compositions of SM and GalCer did not vary with the acclimation temperature (Supplemental Fig. S5, C and D). However, minor SM species 18:0 was more prominent in the fish acclimated to 30°C than in the other groups.
No significant seasonal changes were observed in the molecular species composition of the SL classed studied.
The present results show that, at north-temperate latitudes, the brain lipidome of a eurythermic teleost species, the crucian carp, is strongly modified by seasons. The seasonal changes in lipid composition of the brain are in many respects similar, although not identical, to changes induced by temperature acclimation in the laboratory, suggesting that temperature is the main environmental cue that primes membrane lipidome of the crucian carp brain for winter stresses, characterized by low temperatures and prolonged anoxia. To our knowledge, this is the first study to indicate a direct connection between seasonal acclimatization and thermal acclimation of the brain lipidome in an ectothermic vertebrate. The details of this connection are discussed below.
Bulk Lipid Properties
Fatty acid composition.
The compositions of lipid membranes of temperature-acclimated organisms have been found to change to compensate for the direct effects of the change in temperature and to maintain the proper semifluid state of the membrane, a process termed “homeoviscous adaptation” (33, 77). The physical properties of lipids are largely regulated by their fatty acid composition. The fatty acid composition of the brain of the laboratory and wild fish were very similar and also displayed similar temperature responses (Fig. 3 and Supplemental Fig. S4). At low temperatures, the brains contained lower levels of SFA, higher levels of PUFA, and the average MUFA chain was shorter. All of these changes are compatible with the model of homeoviscous adaptation of membrane fluidity and may originate from 1) selective incorporation of certain dietary fatty acids into membrane lipids, or 2) changes in endogenous fatty acid metabolism, producing highly unsaturated fatty acids (from dietary PUFA precursors) and Δ9-desaturated MUFA at low temperatures (31, 65).
Lipid class composition.
The membranes of ectothermic animals have been reported to respond to changes in temperature initially by rapidly modulating their lipid class composition (34, 49). Later, other mechanisms, such as acyl chain desaturation, become important, and the original lipid class composition is recovered (40, 92). Accordingly, we detected a clear reduction of cholesterol and SM in the brain of cold-acclimated laboratory fish, compared with those acclimated to higher temperatures. However, no major long-term changes could be observed in the lipid class composition of the brain of crucian carp captured from the wild throughout the year.
Cholesterol and SM have mutual affinity (70) and are thought to form “rafts” in biological membranes, supposedly involved in several cellular functions, e.g., lipid and protein transport, as well as signal transduction (1, 76). Since lowering of the temperature increases the size of “rafts” in model membranes (e.g., Ref. 74), the diminished levels of cholesterol and SM in the brain membranes of cold-acclimated fish could be an adaptive response targeted to maintain proper size and viscosity of these domains. Consistently, decreased cholesterol content in hepatocyte plasma membrane of cold-acclimated rainbow trout, particularly in the detergent-insoluble (raft) fraction, has been reported (95).
PL Molecular Species
In contrast to the PL class composition, the molecular species composition of the brain PLs varied markedly with temperature. In most fish tissues, the membrane fluidity is maintained at low temperatures by replacing a SFA with a MUFA in the sn-1 position of the GPL molecule having a PUFA residue in the sn-2 position (e.g., Refs. 8, 54). At low temperature, brain PE of fish has been reported to prefer 18:1 over 18:0 as the acyl pair of 22:6n-3 (12, 16). Our findings, especially for the wild crucian carp, confirmed this, and the phenomenon was seen also in other GPL classes: during winter, numerous MUFA/PUFA species increased, while the corresponding SFA/PUFA species decreased. However, in the laboratory fish, replacing of the sn-1 SFA residue of PE and PS with MUFA played a smaller role, and the adaptive response was based mostly on adjusting the amount of the di-PUFA species.
PS and PE molecular species.
Independent of temperature, most PE and PS species were polyunsaturated and contained 22:6n-3, thought to give the excitable membranes special properties, like enhanced fluidity, permeability, and compressibility (23, 79). A novel, interesting finding is the threefold higher level of the di-22:6n-3 PS and PE species in the laboratory fish acclimated to 4 vs. 30°C. A similar, albeit lesser, change was observed in the wild fish. In summer, greater fluidity of lipid membranes and specific membrane domains may be needed for proper brain function due to large thermal gradients of temperate lakes, e.g., when fish move across the thermocline in catching different food items. Furthermore, seasonal acclimatization may involve other factors, in addition to temperature changes, in particular, hypoxia or anoxia (88). Hypoxic conditions may set additional requirements for brain lipidome or may limit lipid metabolism and, therefore, could attenuate temperature-induced changes in membrane lipid composition. Di-22:6n-3 PE and PS have been detected earlier in the brain of rainbow trout (Salmo gairdneri) (5), and these species are common in vertebrate retina and sperm (6, 52). However, their participation in the thermal acclimation of fish brain in the laboratory or the wild has not been known before.
Plasma membranes of eukaryote cells are highly asymmetric, with most of PE and all of PS residing in the inner leaflet (94). In the red blood cells, PS and PE account for up to 70–80% of the lipids of the inner leaflet (87). Since PS and PE together constitute ∼55% of crucian carp brain PLs (Supplemental Fig. S1), their concentration in the inner leaflet of plasma membranes of neuronal cells is also likely to be very high. Thus the observed temperature-induced changes in PS and PE molecular species compositions should have a significant impact on this membrane compartment in particular. Among the common fatty acids, 22:6n-3 is the most potent enhancer of membrane compressibility and may thus facilitate conformational transitions of integral membrane proteins (48, 78). Studies with probes have shown larger lateral mobility for membranes consisting of 22:6n-3-containing GPL compared with the less unsaturated GPL (73, 80, 83).
PS has a negatively charged head group and, therefore, significant affinity for membrane proteins, often containing positively charged residues at the bilayer boundary (51, 59, 68, 69). Therefore, it is likely that observed increase of di-PUFA PS species occurs to maintain the local environment of membrane proteins optimal for their function. A recent study with crucian carp showed a clear, positive thermal compensation of Na-K-ATPase activity as a reduced Q10 value in winter, resisting the otherwise strong, direct thermal effect inactivating the enzyme catalysis (88). It is likely that the increases of the highly unsaturated membrane lipids observed in the present study contributed to this effect, since 22:6n-3 containing PLs activate Na-K-ATPase of excitable membranes of vertebrates (21, 84). It has also been proposed that 22:6n-3 could intervene transmembrane proteins and weaken interhelical packing, thereby changing protein kinetics (25). Unesterified 22:6n-3 has been reported to modulate sodium, calcium, and potassium currents (47, 67, 89) and inhibit neuron GABA responses (28).
Due to its tendency to adopt bent conformations, 22:6n-3 increases the cross-sectional area of a PL molecule and hence decreases membrane thickness (42, 72). The 22:6n-3 chains also increase membrane permeability to water and small solutes (37, 43). In addition, the 22:6n-3 lipids increase membrane exfoliation and fusion and lipid flip-flop (3, 93). The mentioned properties are all crucial for active transport functions of the membranes.
Increase of PS containing 22:6n-3 may also be needed to maintain signal transduction at low temperatures, since these molecular species activate protein kinase C and enhance G protein-coupled signal transduction (22, 30, 61, 64). Recently, Takamori et al. (82) showed that, due to very high protein concentration, most of the PL in vertebrate synaptic vesicles is in contact with a protein and that most PS species contain a 22:6n-3 residue. Studies in vitro and in vivo suggest that 22:6n-3 and other n-3 PUFAs modulate partitioning of membrane proteins between different lipid microdomains and protein clustering rate, thereby affecting receptor functions and signal transduction (53, 58). In addition, studies with artificial membranes have shown that di-22:6n-3 GPL species are superior to SFA/PUFA species in excluding cholesterol from fluid lipid phases (e.g., Refs. 11, 97). Furthermore, atomic force microscopy studies have revealed the great effectiveness of 22:6n-3-containing lipids to drive lipid phase separation (75).
PE and its plasmalogen derivatives are thought to be important for membrane and cell fusion (55), modulation of membrane curvature (85), and cell division (15). PE has been shown to be important for protein membrane insertion and folding in bacteria (90), and thus its fatty acid composition may be important for such functions in brain membranes as well. However, since PE is much more abundant than PS and not negatively charged, PE is likely to contribute more to the overall properties of the membranes.
The total plasmalogen content (seen best as the totals of DMA, i.e., alkenyl chains) was somewhat higher in laboratory fish acclimated at 30°C or wild fish captured during summer, compared with the cold-acclimated and wild fish captured during winter time (Fig. 3). Exposure to elevated temperatures increases oxidative stress of the fish brain (56, 57), which could be attenuated by antioxidatively active plasmalogens (9). The brain of crucian carp is partially inactivated in cold and anoxic winter conditions (63), which may reduce the need of plasmalogens against oxidative damage in winter. Oxidative stress is expected to increase after winter anoxia in warming and oxygen-rich waters of early summer.
Plasmalogens have been suggested to act as storage forms of PUFA (7, 19). Thus the 22:6n-3-containing plasmalogens of warm-acclimated fish may act as temporary storages of valuable 22:6n-3 in conditions in which there is more 22:6n-3 available. In cultured macrophages, the levels of plasmalogens were coupled to the supply of 22:6n-3 and 22:5n-3, while no such connection between the levels of plasmalogens and 20:4n-6 was found (19). Consistently, only a minor amount of 20:4n-6-containing plasmalogen species was found in this study at the high-acclimation temperatures.
PC molecular species.
In contrast to PS and PE, PC of the crucian carp brain contained only negligible amounts of di-PUFA species, independent of temperature. However, replacement of SFA/MUFA species by MUFA/MUFA in the cold was a more pronounced response in PC than in other GPL classes (Fig. 4A, Table 1). The cylindrical PC molecules are thought to have a stabilizing effect on membrane bilayer, and perhaps this function would have been compromised by incorporating too many destabilizing PUFA residues (increasing the diameter of the acyl chain compartment) (24). All SFA/PUFA PC species increased during the cold acclimation and toward winter. The 16:0/22:6n-3 species, which represents 10% of the PC in the laboratory and wild, is probably also important for the adaptation, as it accounts for 50% of PC in brain synaptic membranes of Antarctic fish Pagothenia borchgrevinki living at subzero temperatures (54).
PI molecular species.
The temperature response of PI differs from the other GPL in many respects. Under all thermal conditions, 1) there were very few di-PUFA PI species, 2) PI preferentially incorporated 20:4n-6, and 3) 18:0 was most often paired with 20:4n-6. These results suggest that PI has a different role from those of PE and PS in membrane function and temperature adaptation. In laboratory acclimation, the molecular species composition of PI was clearly less dependent on environmental temperature than those of the other GPL. Besides being part of the GPI anchors of proteins, PI is phosphorylated to PIPs (phosphoinositides), which are crucial for cell signaling and membrane trafficking (44, 91). Large modulation of PI acyl chains does not seem to be required to allow these crucial phenomena to proceed at the different temperatures. The only obvious temperature-induced change in PI species, the increase of 18:1/20:4n-6 in the cold, occurred in the laboratory and wild and presumably affected membrane fluidity.
SL molecular species.
While the total level of SM in the crucian carp brain seems to respond to temperature, only minor changes in the molecular species composition of SM were observed in laboratory or wild fish. Because SM chain length affects the distribution of GPI-anchored proteins in rafts (20), large acyl chain variations in SM could impair raft-dependent signaling. In the laboratory, modest decrease in sulfatide acyl chain length occurred at the lower temperatures and may contribute to homeoviscous adaptation. Modest changes observed in ceramide species profiles are difficult to interpret, as this lipid is both a precursor and degradation product of more complex SLs (86).
Perspectives and Significance
ESI-MS was used to study fish brain lipidome temperature adaptation for the first time, and novel adaptive responses to temperature variation were found. Most importantly, PS and PE molecules containing two PUFAs, preferentially 22:6n-3, increased markedly at low temperatures in the laboratory, as well as in the wild. These changes may help to maintain proper physical properties of the membranes and thereby allow proper functions of membrane-associated phenomena, including conformational flexibility of proteins, membrane fusion, and signaling. The modifications of membrane lipid composition that occurred during the 4-wk laboratory acclimation were, in many respects, similar to those found seasonally in the wild. This indicates that changes observed in the wild are temperature dependent rather than induced by other environmental factors, such as variation of light and oxygen level or nutritional status. However, the differences observed between the brain lipid compositions of laboratory and wild fish indicate also that the time scale of adaptation (slow in nature, fast in the laboratory) or the factors listed above may also modulate the adaptive responses. In addition to finding a direct connection between seasonal acclimatization and thermal acclimation of the brain lipidome in an ectothermic vertebrate, the present results imply that, besides determining fatty acid and lipid class compositions, future studies on adaptive adjustments of membrane lipids will have to consider changes in the molecular species composition.
This work was supported by the Academy of Finland (fellowship funding no. 111261 to R. Käkelä, and project funding no. 210400 to M. Vornanen).
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
- Copyright © 2008 the American Physiological Society