Seasonal acquisition of chill tolerance and restructuring of membrane glycerophospholipids in an overwintering insect: triggering by low temperature, desiccation and diapause progression Free

Poikilotherm organisms, from bacteria, through plants to animals, show capacity for remodelling of the lipid composition of their biological membranes in response to changing ambient temperature(Sinensky, 1974; Steponkus,1990; Cossins, 1994; Hazel, 1995; Hazel, 1997; Wallis and Browse, 2002). Cold acclimation characteristically leads to one or a combination of the following adjustments as summarized by Hazel (Hazel,1989; Hazel,1995): (a) increased proportion of unsaturated fatty acids; (b)restructuring of glycerophospholipid (GPL) molecular species; (c) elevated relative proportions of glycerophosphoethanolamines (GPEtns) to glycerophosphocholines (GPChols); (d) reduced proportions of plasmalogens compared to diacyl GPLs; (e) adjustment of cholesterol content. Such changes has been postulated to prevent perturbation of membrane structure and function when the body temperature changes. Direct genetic proof for the role of increased unsaturation in enhancing survival at decreasing ambient temperatures has been obtained in cyanobacteria and plants(Browse et al., 1994; Murata, 1994; Allakhverdiev et al., 1999; Inaba et al., 2003). Also in insects, acclimatory responses to cold has been shown to involve restructuring of GPL composition in biological membranes (Bennet et al., 1997; Ohtsu et al., 1998; Kostál and Simek, 1998; Hodková et al., 1999; Hodková et al., 2002; Bayley et al., 2001; Holmstrup et al., 2002; Slachta et al., 2002a; Kostál et al., 2003; Overgaard et al., 2005).

Decreasing ambient temperature is most often considered to be the triggering factor for acclimatory changes in GPL composition. In insects,however, the triggering factors and transduction cascades for GPL compositional changes have not been analyzed in detail. Based on earlier studies, at least three different triggering mechanisms might be involved:ambient temperature, hydration state of the organism (mediated viaosmolarity/concentration of specific solutes) and endogenous factors such as developmental hormones. In the field situation, all three factors may operate in parallel, which makes it difficult to distinguish between their effects. Changes in temperature intervene both daily and seasonally. Naturally occurring diurnal variations in temperature were sufficient to induce the nocturnal increase of cold tolerance in adults of Drosophila melanogaster through the process of `rapid cold hardening'(Lee et al., 1987) and this was accompanied by small but significant changes in the GPL composition(Overgaard et al., 2005). Many insects that overwinter in temperate and polar habitats enhance their cold hardiness during the acclimatization process, which starts upon entering to diapause and continues during autumnal decrease of ambient temperatures(Denlinger, 1991; Slachta et al., 2002b). Changes in GPL composition were observed both in the field and in the laboratory studies during such winter cold hardening (Bennet et al., 1997; Hodková et al., 1999; Hodková et al., 2002; Kostál et al., 2003). Partial dehydration often represents an inseparable part of the complex acclimatory change related to cold acclimation(Ring and Danks, 1994; Block, 1996; Danks, 2000; Williams et al., 2004). It raises the question of whether the GPL changes observed during cold acclimation may be triggered by changes in hydration state of the organism. Indeed, typical membrane alterations resulting in a higher degree of unsaturation were induced by mild desiccation stress applied to the collembolan Folsomia candida(Bayley et al., 2001; Holmstrup et al., 2002). Many insects change their hormonal milieu seasonally, in response to specific token stimuli such as photoperiod, and enter diapause(Denlinger, 1985; Tauber et al., 1986; Danks, 1987; Kostál, 2006). Similar remodelling of GPL composition was observed in response to cold exposure and to transition from direct development to diapause in two insects, adult P. apterus (Hodková et al.,2002), and larvae of the fly Chymomyza costata(Kostál et al.,2003).

The main objective of this study was to describe the nature and triggering factors of seasonal changes in GPL composition in the bug Pyrrhocoris apterus L. (Insecta: Heteroptera). The changes observed during autumnal acclimatization of wild bugs were compared with the acclimation changes induced under controlled laboratory conditions,where three potential triggering/modulating factors were examined separately:low temperatures, desiccation and progression of diapause. Survival at sub-zero temperatures, and several other physiological parameters related to it, were measured in parallel and linked to the changes in GPL composition

In the present study, wild and laboratory-reared insects were compared. Wild adults were collected in Ceské Budejovice (Czech Republic) on two occasions during 2005, on 23rd October and 27th December. All insects were in reproductive diapause and they passed through the process of winter cold hardening between October and December(Hodek, 1968; Kostál and Simek,2000). Fig. 1Ashows the air temperatures as recorded by the weather station of the Czech Institute of Hydrometeorology in Ceské Budejovice. The insect samples were transported to the laboratory (10 min) without changing their temperature and were then processed immediately.

Laboratory-reared insects came from a culture that originated from adults collected in Ceské Budejovice during spring 2005. F₂ embryos and nymphs were reared under conditions inducing reproductive diapause in the adults, i.e. constant temperature of 25°C and a short-day photoperiod of 12 h:12 h, L:D (Hodek, 1968). Dry seeds of the linden tree (Tilia parviflora Ehrh.) and water were provided ad libitum. Adult F₂ insects were kept at a constant 20°C, with a short-day photoperiod and with access to food and water, for 4 weeks. Such adults were considered to be in deep reproductive diapause and their physiological state was considered as the initial state. Three different treatments were then applied (depicted in Fig. 1B). (1) Progression of diapause: the insects were kept under constant conditions (access to food and water) for another 8 weeks. (2) Desiccation: the insects were kept at 20°C, with a short-day photoperiod, and denied access to liquid water and food for 14 days. Relative humidity of the air fluctuated between 40–60%. Such a treatment results in approximately 50% mortality(Tollarová and V.K., unpublished data), and in the present study only survivors were taken for analyses. (3) Cold acclimation: the insects were exposed to gradually decreasing temperatures: 20°C (day)/10°C (night)during the first week, followed by 15°C/5°C and 10°C/0°C during the second and third weeks, respectively (all 3 weeks in short days),followed by 5 weeks at constant 0°C and continuous darkness. Such an acclimation protocol simulates the natural drop of temperatures during autumn(see Fig. 1A) and the bugs achieve higher chill tolerance than if constant temperatures (in contrast to thermoperiodic regime) are used for acclimation(Kostál et al.,2001).

The fresh mass (FM) was measured in five males and five females(individually) using a Sartorius balance with sensitivity of 0.1 mg. Dry mass(DM) was measured after drying the specimens at 65°C for 3 days. Hydration(in mg H₂O mg^–1 DM) was calculated from the gravimetric data. Haemolymph samples for osmolality measurements were collected from 10 insects (five males, five females) by cutting off one of the antennae and allowing it to bleed into a calibrated capillary tube. Osmolalities were measured in 10–15-nl droplets using the Clifton Nanoliter Osmometer (Clifton Technical Physics, USA).

The concentrations of the two most abundant `winter' polyols (ribitol and sorbitol) that typically accumulate in diapausing cold-acclimated P. apterus adults (Kostál and Simek, 2000; Kostál et al., 2001), were measured in 70% ethanol extracts of whole body samples (five males, five females, taken individually). The extraction,derivatization and analytical procedures (gas chromatography coupled to mass spectrometry) were the same as described by Kostál and Simek(Kostál and Simek,1995). All chemicals used for analysis were purchased from Sigma-Aldrich Co. (St Louis, MO, USA).

The temperature of crystallization of body fluids (supercooling point, SCP)was measured in individual insects (16 males, 16 females) at a cooling rate of 0.3° min^–1 using the method described(Nedved et al., 1995). Chill tolerance was assessed by exposing groups of 20 insects (10 males, 10 females)to a temperature of –15°C (maximum, –14.0°C; minimum,–15.3°C) for various periods (1, 4, 7, 10, 13, 20 or 28 days). The capacity for coordinated crawling at 25°C, 2 days after the exposure, was used as a criterion for survival. Survival rate was described using the parameter LT50, which gives the time of exposure that is subsequently survived by 50% of the population sample. The parameter Lt50 was calculated from nonlinear logistic regression with the sigmoid curve as described(Kostál et al.,2001).

Thoracic muscle and fat body tissues were rapidly (3 min) dissected from the adults. Tissues from three individuals were pooled to obtain one sample,and each sample was replicated six times (three times with males, three times with females). Total lipids were extracted in ice-cold chloroform:methanol(2:1) solution using the method of Folch et al.(Folch et al., 1957) slightly modified according to Kostál and Simek(Kostál and Simek,1998). After the extraction, the solvents were evaporated under a stream of nitrogen, and lipids were stored at –80°C until analysis.

High performance liquid chromatography (HPLC) combined with electrospray ionisation mass spectrometry (ESI-MS) (Han and Gross, 1995; Hsu and Turk,2000; Brooks et al.,2002; Wang et al.,2005) was performed on a quadrupole ion LCQ mass spectrometer(Thermo, San Jose, CA, USA) coupled to a Rheos 2000 ternary HPLC system (Flux,Basel, Switzerland), equipped with a FAMOS autosampler and Thermos thermostat(both LC Packings-Dionex, Amsterdam, The Netherlands). The stored dry samples were dissolved in 1 ml of methanol and 5 μl aliquots were injected into a 150×2 mm i.d. 3.5 μm Synergi Polar HPLC column (Phenomenex, Torrance,CA, USA). The mobile phase was composed of (A) 10 mmol l^–1ammonium acetate (NH₄Ac) in methanol, (B): 10 mmol l^–1 NH₄Ac in water and (C) isopropanol. A linear gradient of A:B:C changing from 90:10:0 to 70:0:30 within 14 min was used with a flow rate of 300 μl min^–1. The column temperature was maintained at 30°C. The mass spectrometer was operated either in positive or negative ion detection modes at +4 kV or –3.6 kV, respectively. Capillary temperature was 240°C; nitrogen was used as both the sheath and the auxiliary gas. For MS2 and MS3 fragmentations, ion isolation windows were 5 and 2 Da, respectively. Maximum ion injection time was 100 ms, collision energies were 30% (MS2) or 35% (MS3), mass range of 600–800 Da was scanned each 0.5 s. All chemicals used for extraction and analysis were purchased from Sigma-Aldrich Co.

As there were no significant differences between males and females, the data sets for both sexes were pooled. Owing to the complexity of instrument response to particular GPL molecular species, which is dependent not only on features of the chemical structure but also on experimental conditions,results of GPL analysis were preferably expressed in relative values, i.e. relative molar proportion of each lipid species from the total of 100%. Our analysis was focused on major GPEtn and GPChol molecular species, which were present at or above 1%. It is known that GPEtns plus GPChols represent more than 80% of all glycerophospholipids in P. apterus tissues(Hodková et al., 1999). The ratio of total GPEtns to total GPChols (GPEtn/GPChol) was calculated for each tissue/treatment and the relative proportions of individual fatty acyls were calculated based on GPL data. Unsaturation ratio (UFA/SFA) was calculated from the proportional composition of fatty acyls.

Unpaired two-tailed t-tests (parametric) were used to compare means of two variables with normal Gaussian distribution (F-tests were used to verify equality of variances). Mann–Whitney U-tests (non-parametric) were used to compare the medians of SCP variables (with skewed distribution of data). To investigate if there were any general trends in the changes of GPL composition caused by different treatments, all data were analysed together using a multivariate statistical analysis, Principal component analysis (PCA; Simca-P software, Umetrics, Umea,Sweden). Using this method, the data are represented in K-dimensional space (where K is equal to the number of variables) and reduced to a few principal components (two principal components, PC1 and PC2, in this study) that describe the maximum variation within the data. The principal components are then displayed as a set of scores, which highlight clustering. Differences between the mean scores for different treatments along the axes PC1 and PC2 were tested using one-way ANOVA followed by a post-hocStudent–Newman–Keuls test (SNK).

Wild adults of P. apterus collected in the field during October and December had similar dry masses (Table 1). Although their hydration decreased from October to December,this change was statistically insignificant (P=0.0601; t-test). Osmolality of haemolymph, however, increased almost threefold (from 326 to 904 mosmol kg^–1) and high amounts of winter polyols accumulated: 27.0 and 3.8 μg mg^–1 DM of ribitol and sorbitol, respectively. The temperature of crystallization of body fluids (SCP) decreased from –11.0 in October to –19.2°C in December and the chill tolerance (measured as a Lt50 at –15°C)increased from <1 to 18.9 days.

There were no significant differences in the measured physiological parameters between the laboratory-reared insects during their initial state and those collected in the field in October (compare Tables 1 and 2; P>0.05 for all parameters, t-tests or U-test). When laboratory adults were subjected to various treatments, their dry mass did not change significantly(Table 2). Hydration remained stable during the progression of diapause at 20°C, but it significantly decreased during the desiccation and cold-acclimation treatments. Desiccating conditions caused the loss of 51% of initial water content and an approximate twofold increase in haemolymph osmolality. Cold acclimation was accompanied by the loss of 34% of initial water content and more than a threefold increase in osmolality. No winter polyols accumulated during the progression of diapause. Small amounts of polyols (slightly above the analytical thresholds: 2.5 or 1.0μg of ribitol or sorbitol, respectively, in the sample) were found in desiccated insects and high amounts of polyols, i.e. 15.9 and 11.4 μg mg^–1 DM of ribitol and sorbitol, respectively, were found in the cold-acclimated specimens. Supercooling capacity and chill tolerance at–15°C remained relatively low during diapause progression and desiccation while cold acclimation caused the SCP to decrease from –13.9 to –18.9°C and the Lt50 at –15°C to increase from <1 to 14.6 days.

Gradient HPLC efficiently separated GPLs from other lipid classes,particularly from triacylglycerols, which were dominant in the sample extracts. In this way, highly reproducible full-scan positive ESI mass spectra of the GPChol and GPEtn fractions were obtained with minimal outward peak overlapping and ion suppression. As shown in Fig. 2, [M+H]⁺ ions dominated in the full-scan ESI spectra of GPEtn and GPChol fractions. Twelve to 13 major GPEtn and GPChol molecular species were consistently detected in the tissue samples of P. apterus (Figs 2, 3). Fig. 3 shows a typical HPLC chromatogram (upper trace), and extracted chromatograms of principal GPEtn and GPChol species. The identity of each major ion was verified by fragmenting the ion and studying the MS2 and MS3 spectra containing preferential losses of headgroups and fatty acyls from the sn-1 or sn-2 positions of the glycerol backbone. All GPL species were composed of only five different fatty acyls (FAs): palmitoyl (16:0), heptadecanoyl (17:0), stearoyl (18:0),oleyl (18:1) and linoleyl (18:2). The GPL compositions were qualitatively similar in the wild (October) and laboratory-reared (initial state) insects in both tissues (compare Fig. 4Aand Fig. 5A with Fig. 6A and Fig. 7A, respectively). There were, however, quantitative differences at the level of individual GPL species. The single largest difference was found in the proportion of 18:2/18:2-GPChol in the muscle tissue, which was 28.2% in the wild insects and only 19.4% in the laboratory-reared insects (P<0.0001, t-test). Wild specimens showed similar GPEtn/GPChol ratios in both tissues (muscle, 1.0±0.03; fat body, 0.9±0.2). By contrast,laboratory-reared insects had a ratio of 1.5±0.3 in the muscles and 0.5±0.1 in the fat body.

The complex change in GPL composition during seasonal acclimatization in the field is shown in Figs 4and 5. Two most noticeable alterations were: (1) the increase of 16:0/18:2-GPEtn by 6.6 or 5.3% (muscle or fat, respectively), and (2) the decrease of 18:2/18:2-GPChol by 4.7% (both tissues). The GPEtn/GPChol ratios significantly increased from October to December: to 1.4±0.2 in the muscles (P=0.0011, t-test), and to 1.2±0.2 in the fat body (P=0.0293, t-test). FA composition changed relatively little during autumnal acclimatization (Table 3). Statistically significant increases in the proportion of palmitoyl (16:0) were registered in both tissues, and the proportions of linoleyl (18:2) tended to decrease. Such changes resulted in a slight, but statistically significant,decrease in UFA/SFA ratio in both tissues.

The diapause progression and desiccation treatments were accompanied by relatively small changes in GPL composition of the muscle(Fig. 6B,C) and fat body(Fig. 7B,C). The relative proportions of GPEtn/GPChol remained unchanged (P>0.05, t-test): muscle after diapause progression, 1.4±0.3; muscle after desiccation, 1.6±0.4; fat body after diapause progression,0.6±0.1; and fat body after desiccation 0.6±0.1. Relatively small changes were observed in FA composition, and the UFA/SFA ratios did not change significantly (Table 4).

The cold-acclimation treatment resulted in considerable restructuring of GPL composition. The general patterns of changes were similar in both tissues(compare Fig. 6D with Fig. 7D). The proportion of 16:0/18:2-GPEtn increased by 8.1% and 3.3% in the muscle and fat body tissue,respectively; 18:0/18:2-GPEtn was the second most increased species (by 5.9%and 3.5%). Also, 18:2/18:2-GPChol decreased by 8.8% and 6.5%, and 16:0/18:2-GPChol decreased by 3.5% and 2.5%. The GPEtn/GPChol ratios significantly increased in both tissues, to 3.6±1.1 in the muscle(P=0.0010, t-test), and to 0.7±0.1 in the fat body(P=0.0037, t-test). Changes were observed also at the level of FA composition in both tissues; the UFA/SFA ratios significantly decreased(Table 4).

PCA identified two principal components (PC1 and PC2), which together accounted for 82.8% of the total variation seen among different treatments(field and laboratory data were analyzed in a single run). Fig. 8A presents the scores of the two principal components for individual treatments. In the two-dimensional space described by PC1 and PC2 scores, vectors may by drawn, which connect an initial state with a state after the application of a treatment. Similar patterns in changes of GPL composition appear as similar orientations of the vectors. The length of each vector is proportionally related to the significance of the change. The vectors displaying the autumnal acclimatization in the field and the cold acclimation in the laboratory were relatively long and of similar orientation in both tissues. The differences between the relevant scores of principal components were significant atα=0.05 (ANOVA followed by SNK test)(Fig. 8B,C). In contrast, the vectors related to diapause progression and desiccation (these were not drawn in Fig. 8A) were relatively short and the differences between the scores were mostly insignificant.

Insect cold acclimation is a complex physiological change consisting of many components (Lee, 1991; Danks, 1996; Renault et al., 2002; Sinclair et al., 2003) of which the membrane restructuring is just one. Hence, it is difficult to estimate the effect of any single component in isolation. In P.apterus, depression of SCP temperature down to between –15 and–20°C is necessary to extend the range of survival temperatures. This phenomenon was observed repeatedly in the field(Kostál and Simek,2000) (Table 1) and in the laboratory cold-acclimation experiments (Hodková and Hodek,1997; Kostál et al.,2001) (Table 2). The low SCP alone, however, does not guarantee high levels of chill tolerance(Bale, 1987). About 25% of diapausing P. apterus adults showed SCP lower than –15°C already prior to cold acclimation (both in the field and in the laboratory,see the values of 75 percentiles in Tables 1 and 2). The minimum SCP values in such adults was –20°C. Despite that, none of those adults survived longer than 1 day at –15°C (Tables 1, 2). This shows clearly that additional adjustments are required to reach the maximum chill tolerance. Among them, accumulation of ribitol and sorbitol play an important role(Kostál and Simek,2000; Kostál et al.,2001; Kostál et al.,2004a). Both compounds were also found in cold-acclimated P. apterus in relatively high concentrations in this study (Tables 1, 2). Enhancing the capacity to maintain ion gradients across cell membranes and epithelia at sub-zero temperatures is another adjustment that probably contributes to increasing chill tolerance of cold-acclimated P. apterus(Kostál et al., 2004b). The potential contribution of membrane restructuring to the increasing chill tolerance of this insect was hypothesized first by Hodková et al.(Hodková et al., 2002). They described the differences in the GPL composition of membranes between laboratory, warm-acclimated insects and those collected in the field during February [fig. 4 in Hodková et al.(Hodková et al.,2002)]. We verified and further extended their results using more advanced analytical techniques and comparing the responses in the field and variously treated laboratory animals.

The increase in the relative proportion of 1-palmitoyl-2-linoleyl-sn-GPEtn at the expense of 1,2-dilinoleyl-sn-GPChol was the most prominent specific change observed in the wild adults of P. apterus during their autumnal acclimatization in the field. It was one component of a more general trend,i.e. the increase in the total proportion of GPEtn species counteracted by the decrease in the total proportion of GPChol species(Fig. 3B, Fig. 4B), which also can be expressed as an increase in the GPEtn/GPChol ratio. Similar changes were registered in two tissues, thoracic muscle and fat body. It should be noted here that whereas the locomotory function of muscle does not change seasonally, the physiology of fat body changes profoundly. The fat body is the principal tissue for intermediary metabolism in insects(Keeley, 1985). Its major function in reproductively active females is the synthesis and release of vitellogenic proteins for yolk formation during oocyte maturation(Telfer, 2003). At early phases of diapause, the fat body synthesizes large amounts of energy reserves in the form of diapause proteins, triacylglycerols and glycogen. During the later phases of diapause, the synthetic activity of the fat body declines and the lipidic and glycogen reserves are slowly utilized(Tauber et al., 1986). Upon stimulation by low temperature, the fat body glycogen becomes the main source for the rapid biosynthesis of polyol cryoprotectants(Storey and Storey, 1991).

The extensive seasonal restructuring of GPL composition in P. apterus occurred without major changes in the proportions of the FAs. A moderate increase (by 2.4%) in the proportion of palmitoyl (16:0) was found in both tissues (Table 3). The most common response of ectotherm membranes to cold is to increase the proportion of unsaturated FAs and/or increase the number of double bonds per FA. This response was not seen in P. apterus. In fact, autumnal acclimatization led to a slight decrease in the UFA/SFA ratio in both tissues (Table 3). Our FA data,however, were not obtained by direct analysis of FAs but were derived from original analytical data of GPL species. Nevertheless, direct GC/MS analysis of FAs in P. apterus [performed by Hodková et al.(Hodková et al., 1999)]confirmed a positive correlation between the unsaturation ratio and ambient temperature, and a negative correlation between the proportion of palmitoyl and temperature.

In the field, various factors triggering/stimulating GPL restructuring may not be easily distinguished. Our laboratory experiments allowed us to separate three potential factors: low temperatures, desiccation, and endogenous factors related to diapause progression. Whereas the low temperatures stimulated significant changes (Fig. 5D, Fig. 6D), the later two factors appeared to have relatively small effects on GPL composition in P. apterus membranes.

The changes in GPL composition triggered by low temperatures in the laboratory were similar to the acclimatization changes observed under the field situation. The GPEtn/GPChol ratio increased, as a result of the increases in 16:0/18:2-GPEtn and 18:0/18:2-GPEtn countered by the decrease in 18:2/18:2-GPChol. Results of multivariate statistical analysis (PCA) further supported the overall similarity between the restructurings seen in the field and laboratory and in both tissues (Fig. 8). This was so despite the fact that the initial states widely differed between the two situations and between the two tissues as well. Such uniformity in the response suggests a common adaptive value of the underlying GPL changes in relation to cold.

The ratio of GPEtn/GPChol generally tends to be higher in cold-than in warm-acclimated ectotherms (Hazel,1989; Hazel, 1995; Hazel, 1997). Molecules of the two lipid classes differ in their geometries: GPChols assume a cylindrical shape and pack efficiently in a lamellar gel phase; GPEtns adopt a more conical form, which is readily accommodated in the reverse hexagonal phase. Consequently, less efficient packing of GPEtns in the lamellar phase disrupts the organization of the membrane and thus counters the effects of low temperatures, which tend to organize the membrane in a gel phase(Hazel, 1989). The increasing GPEtn/GPChol ratio in a membrane may thus help to prevent the formation of a gel phase in membranes at low ambient temperatures. In our earlier work(Slachta et al., 2002a), we detected a slight increase in the GPEtn/GPChol ratio in response to low temperatures in the non-diapausing adults of P. apterus. It increased from 1.0 to 1.1 (NS, t-test) in the muscles, and from 0.7 to 1.0(P=0.0032, t-test) in the fat body. Non-diapausing adults were not able, however, to accomplish the other cold-acclimation adjustments seen in diapausing adults (SCP depression, polyol accumulation and ion gradients regulation) and no, or only a weak, increase in their chill tolerance was observed (Slachta et al.,2002a).

It is more difficult to search for an adaptive explanation for the observed changes in relative proportions of specific molecular species, such as the 16:0/18:2-GPEtn. The range of temperatures at which membranes composed of GPEtns remain fluid, i.e. T_m–T_h(where T_m is a temperature of transition from lamellar gel to liquid-crystalline phase and T_h is a temperature of transition from liquid-crystalline to reverse hexagonal phase) markedly increases with a decrease in acyl chain length(Lewis et al., 1989). Hodková et al. (Hodková et al., 2002) suggested that a relative increase of 16:0 FA and its specific pairing with 18:2 FA in the GPEtn molecule may help to maintain the ambient temperature within a suitable interval above T_myet below T_h, and protect membranes from unregulated phase transitions caused by either a drop of ambient temperature or partial dehydration, or both occurring at the same time, as is normal in overwintering insects. Generally, the phase behaviour, fluidity and other properties of membranes are finely tuned to serve various purposes in different environmental situations. Changing the lipidic composition is the main tool of such tuning. Some changes may represent just a side effect without any explicit adaptive value. However, increasing the relative proportion of 16:0/18:2-GPEtn, as a specific response to cold, might be of some adaptive value as it was found in two tissues with different physiological functions,and it was observed also in the fly, Chymomyza costata(Kostál et al., 2003),and water strider (M. Hodková, unpublished results), which are ecologically and phylogenetically distant to P. apterus.

It has been proposed that some of the physiological adaptations to cold might be derived from more ancestral adaptations to desiccation(Ring and Danks, 1994; Block, 1996; Danks, 2000; Williams et al., 2004). Populations of P. apterus probably colonized central Europe from Asia Minor and the Near East relatively recently, during the climate warming after the end of the last glacial period. The ancestral habitats were prone to drying, which could drive the evolution of physiological adaptations to desiccation in this species. Moreover, P. apterus adults typically lose up to one third of body water during cold acclimation and overwintering(Kostál and Simek,2000; Kostál et al.,2004b) (Tables 1, 2). In this study, we attempted to test whether desiccation triggers some changes in GPL composition. In the collembolan, Folsomia candida, a mild desiccation stress was sufficient to elicit the restructuring of membrane GPLs(Bayley et al., 2001; Holmstrup et al., 2002). In addition to alteration of membrane lipids, Folsomia candida responded to desiccating conditions by accumulating glucose and myo-inositol, which led to the increase in osmolality, and subsequently influenced the rates of water uptake/loss (Bayley and Holmstrup,1999). Our experiments showed that the loss of one half of body water, associated with doubling the haemolymph osmolality, induced relatively small changes in GPL composition in P. apterus and that the overall pattern of these changes differed from the changes induced by low temperatures(Fig. 7). For example, a decrease in the proportion of 16:0/18:2-GPEtn was noted in response to desiccation whereas an increase was detected as a result of cold acclimation(see above). No, or negligible, accumulation of polyhydric solutes was observed in response to desiccation in P. apterus(Table 2).

The preparation for winter survival starts long before the autumnal decline of environmental temperatures in P. apterus(Kostál and Simek,2000), and in many other insects as well(Danks, 1987; Kostál, 2006). Hodková et al. (Hodková et al., 2002) found that the relative proportion of 16:0/18:2-GPEtn was higher in short-day reared (diapausing) than in long-day reared adults of P. apterus (non-diapausing, reproducing) that were maintained in the laboratory at 26°C. Moreover, when non-diapausing females of P. apterus had the corpus allatum surgically removed (which prevents juvenile hormone production and causes entry into a diapause-like state), the GPL profiles in their muscle and fat body tissues changed in a similar manner to those of females in which diapause was induced photoperiodically(Hodková et al., 2002). Similar results were obtained with larvae of the fly Chymomyza costata, where relative proportions of 16:0/18:2-GPEtn in total GPLs extracted from the muscle and fat body tissues significantly increased in response to both low temperature stimulus and the transition from direct development to diapause, i.e. without any change in temperature(Kostál et al., 2003). Such results suggest that the seasonal restructuring of GPL composition is photoperiodically activated already during the entrance into diapause, which typically happens (long) before the seasonal decline in temperatures, and that the changes are hormonally mediated. By contrast, the changes in GPL composition during the diapause progression, as described in this paper, were relatively small. Moreover, the orientations of the vectors of diapause-related changes(Fig. 8A) were opposite in the muscle and fat body tissues. Such a clear difference in the vector orientations might reflect specific (changes of) physiological roles in two different tissues during diapause, which may require synthesizing/maintaining specific GPL species. In this light, relatively deep and uniform restructuring of GPL composition of both tissues in response to cold implies an adaptive mechanism assisting in survival at low temperatures.

The study was supported by the Czech Science Foundation grant no. 206/03/0099.