Several clear alterations are evident in the membrane fatty acid composition in response to high temperature growth. A decrease in trienoic fatty acids, including strongly diminished 16:3, produced exclusively within the prokaryotic pathway in the chloroplast, is consistent with the general decrease of polyunsaturated fatty acids in response to high temperature. Concurrent with the reduced accumulation of trienoic fatty acids at high temperature is an increase in linoleic acid, 18:2, and an increase in 16:0 (Figs. 1 and 3). The observation that 16-C fatty acids show a pattern distinct from that of 18-C fatty acids suggests that individual fatty acid classes may have specific roles in maintaining optimal membrane function as well as different mechanisms governing their synthesis. This idea of distinct roles for particular fatty acids in the membrane is also supported by the relatively similar degree of membrane unsaturation in most of the lines examined (Table 2), despite the differences found in growth or photosynthetic stability.
It is remarkable that these general alterations are observed even in mutant lines that are deficient in steps early or late in the desaturation pathways. For example, the profile of fad5, which is deficient in 16:0 desaturation in the chloroplast compared to fad7 fad8, which is deficient in the final desaturase step forming trienoic fatty acids, show similar overall trends in response to high temperature. These temperature responses must take into account the two glycerolipid pathways that operate in parallel in Arabidopsis, one in the chloroplastic envelope membranes and one in the endoplasmic reticulum [38, 39].
Previous characterizations of all of the individual mutant background lines used in this study have clearly indicated the primary lipid species acted upon by the specific desaturase (or acyltransferase) activity missing in each mutant (Table 1). The temperature-induced alterations in the membrane composition can be considered in the context of the particular mutant background with its corresponding deficiency in a given step of the glycerolipid pathway. In this case, the mutations examined primarily affect chloroplastic lipids. In addition, the polyunsaturated 16-C fatty acids can be used as reliable markers for the major chloroplastic lipids monogalactosyldiacylglycerol and digalactosyldiacylglycerol, since they are the only lipids that contain these fatty acids. In regard to lipid compositions following growth at elevated temperature, examination of the proportions of individual lipids from the related species, Brassica napus, has shown that significant changes do not occur in leaf membranes from plants grown at 20°C and 30°C . This finding suggests that it is the degree of fatty acid unsaturation that varies most appreciably at these temperatures and not the levels of the major leaf lipids themselves .
Temporal basis for temperature-induced membrane alterations
The time required for the fatty acid composition to adjust to high-temperature growth conditions demonstrated that major alterations in leaf membranes do not occur rapidly in response to elevated temperature. The ~60-h period before changes become evident in the fatty acid profile corresponds with the occurrence of new lipid synthesis and turnover and therefore does not suggest a mechanism by which temperature induces modifications of existing membrane fatty acids. These results are also consistent with labeling studies that show fatty acids produced by the prokaryotic pathway accumulate prior to those synthesized via the eukaryotic pathway. The abrupt changes in 16-C fatty acids beginning 60 h after the shift to high temperature compared to the gradual and more continuous alterations observed for 18-C fatty acids (Fig. 2) also fits with these labeling patterns. It is well established by a number of time-course radiotracer labeling studies, including several conducted on most of the mutant lines used here, that earlier stages in the glycerolipid pathway show alterations prior to those formed through the eukaryotic pathway [27, 33–35, 40]. These include the formation of saturated fatty acids prior to the accumulation of unsaturated species and the earlier production of palmitate-containing species in the prokaryotic pathway. Overall, these time-course observations suggest that the modulation of membrane unsaturation levels plays a role in longer-term acclimation of the plant. Transient fluxes in environmental temperature are therefore not likely to result in pronounced alterations in the composition of leaf membranes.
Membrane composition in response to temperature in mutant lines
The fatty acid composition resulting from high-temperature growth in the different genetic backgrounds reveals a complex regulatory system. Thus, despite deficiencies in several enzymatic steps in the different mutants, the membrane fatty acid composition undergoes adjustments similar to those observed in the wild type in response to elevated temperature. The similar increase in 18:2 levels after high temperature growth in the fad5, fad6, act1, act1 fad6 and wild-type lines suggests that the percentage of 18:2 may be important in membrane acclimatization. These results also suggest that desaturase activities as well as flux from extrachloroplastic membranes might also be controlled to mediate the response to temperature (see below). The accumulation of 18:2 (as opposed to 18:1) in the lines tested implicates it as a preferred species in membranes adapted to high-temperature growth. Although unknown, one speculation for 18:2 for this may be due to the "intermediate" level of disorder represented by diunsaturated acyl chains in the membrane. Fatty acids with one double bond, such as 18:1, impart a greater relative degree of disorder to the membrane than acyl groups containing two double bonds which, in turn, confer only slightly less disorder to the membrane relative to triunsaturated 18:3. The composition of the fad7 fad8 mutant line supports this contention, in which the mol% of 18:2 is almost 2-fold higher than that of wild-type membranes and this line exhibited the best growth at elevated temperature. However, biological membranes are complex, dynamic structures and it is likely that other, unknown factors will be important to maintain optimally functioning membranes.
The fatty acid profiles of several mutants grown at high temperature suggests that the control of flux of fatty acids from the eukaryotic pathway is partly responsible for the changes observed in 18:2 and 18:3 due to temperature. In the fad6 mutant grown at high temperature, the proportions of 18:2 and 18:3 are highly similar to the proportions observed in the wild type, despite that essentially all of the 18:2 must be derived from the eukaryotic pathway. Labeling studies have shown previously that the fad6 line exhibits a decrease in lipid synthesis via the prokaryotic pathway , and this same mechanism, which presumably operates to maintain specific physical properties of the chloroplast membranes, may also be involved in mediating compositional adjustments of the membranes in response to increased temperature. In this case, a possible mechanism might be a reduction in the amount of 18:2 fatty acids transferred to the chloroplast for desaturation by the FAD7 and FAD8 desaturases. The fad6 profile also reveals that activity of the FAD2 enzyme, operating in the endoplasmic reticulum, might also be subject to temperature regulation.
Similarly, in the act1 mutant line, in which entry of fatty acids into the prokaryotic pathway is blocked by the step catalyzed by GPAT, the primary difference is a decrease in the level of 16:0 when grown at elevated temperatures. This presumably reflects an increase into the eukaryotic pathway but with a similar temperature-responsive regulation. Flux of 18:2 from the eukaryotic pathway back into the chloroplast also appears to be modulated, as the proportions of 18:1 and 18:2 are slightly elevated in act1 at both temperatures, while the 18:3 level is highly similar to wild type.
The fatty acid composition of the fad7 fad8 mutant reveals alterations that occur in response to elevated temperature in the absence of all trienoic fatty acid-forming desaturase activity in the chloroplast (Fig. 3E). This profile also suggests that transfer of fatty acids from the eukaryotic pathway may be an important component in temperature regulation of the membrane composition. Such a mechanism is possible considering that the major proportion of 18:3 produced in this line must be synthesized through the eukaryotic pathway via the FAD3 desaturase . The resulting low level of 18:3 detected in plants grown at high temperature is evidence that the FAD3 enzyme in the endoplasmic reticulum also is subject to temperature regulation. Thus it appears that desaturase enzyme activity is inversely regulated by increased temperature, in agreement with previous proposals as a likely mechanism . Analysis of the data presented here suggests that the desaturases that catalyze trienoic fatty acid formation (FAD7, FAD8 and FAD3) and the FAD5 desaturase are the enzymes likely to have the greatest impact if regulated in this way.
Membrane fatty acid composition has a role in enhancing photosynthesis to tolerate high temperature
Studies conducted on the ability of plants to acclimate to elevated temperature have mainly focused on components most likely to affect the stability of photosynthetic electron transport, particularly PSII. In this respect, the composition of the chloroplast thylakoids is expected to be important in the thermal tolerance of photosynthetic electron transport . Moon et al.,  have implicated the unsaturation level of PG as being important in the removal and replacement of damaged D1 proteins in plants. However, the direct relationship pointing to protein-lipid associations being involved in stabilizing the D1 protein at high temperature has only recently been suggested . The results presented here imply a regulatory mechanism that confers a similar overall composition in response to temperature regardless of the initial fatty acid alteration in the membranes due to mutation. A possible reason for such apparently stringent control might be that compositions that are deleterious for membrane function are curtailed, utilizing desaturase pathways that are present in both chloroplastic and extrachloroplastic locations.
Measurement of PS II activity during leaf heating was used here as a sensitive indicator of the thermostability that might be conferred by the different membrane compositions. In this case, the temperature at which the quantum yield of PS II electron transport collapses (TP) was determined in intact leaves. The fluorescence measurements were conducted on plants grown at 17°C, to minimize variation and other potential responses, such as increases in the synthesis of heat shock proteins that might be induced at higher temperatures. In addition, only at the 17°C growth temperature were differences observed in the total membrane unsaturation level among the different lines tested as estimated by the double bond index, which might suggest that the largest influence of the fatty acid composition occurs at lower and more moderate temperatures.
The fad6 mutant exhibited the maximum fluorescence yield enhancement of the lines grown at 17°C. This maximum TP is essentially the same as that obtained from several plants grown at 29°C, including the wild-type line (Table 3B). The 2°C difference apparent in the mutant lines grown at 17°C may be an accurate indication of the magnitude of PSII thermal stability that can be conferred by adjustments in the membrane fatty acid composition and therefore may reflect the extent to which these adjustments can contribute to high-temperature acclimation in Arabidopsis. While it is unknown how specific fatty acids influence thermal stability, analysis of the act1 fad6 mutant described in this study demonstrates that alterations in the relative abundance of 16:1 and 18:1 have an effect. Both act1 and fad6 mutant plants display enhanced stability compared to wild type as determined by fluorescence measurements whereas the act1 fad6 line, which does not exhibit elevated 16:1 but slightly higher levels of 18:1, shows no statistical difference in the TP temperature compared to wild type (Table 3). This difference in fluorescence TP temperatures among these three mutants, as well as a lack of a correlation between the TP values and the double bond index, illustrates that the relative level of a specific lipid class can influence thermal stability. Measurements of diffusion rates of light-harvesting complexes in desaturase mutants of cyanobacteria also point to distinct roles that lipids may have in photosystem stability. In a recent study, the interaction of phycobilisomes with reaction centers was proposed to be stabilized by lipids as opposed to being affected by the level of membrane unsaturation directly . Studies using spectroscopic methods have also suggested that overall lipid acyl chain disorder in cyanobacterial membranes is similar despite differences in growth temperature or unsaturation levels  and that protein-to-lipid interactions in membranes seem to be a key parameter in membrane dynamics [45, 46].
Measurements of a single physical parameter attributable to alterations in the membrane composition have often not correlated with results apparent in whole plant performance tests. Similar to the results presented here, Murakami, et al.,  have shown that chlorophyll fluorescence measurements might not be the ideal indicator to assess whole-plant performance. For example, although the fad7 fad8 line showed the fastest growth rate at high temperature, it did not show a significant difference in thermal stability based on fluorescence yield. The use of other functional measurements, such as CO2 uptake rates or O2 evolution, may provide indicators that more reliably assess potential high-temperature tolerance, although these measurements can also lead to unpredictable results regarding whole-plant physiology. In an Arabidopsis mutant devoid of virtually all digalactosyldiacylglycerol in chloroplast membranes, O2 evolution was found to be unaffected despite large changes in thylakoid membrane organization and in fluorescence energy transfer characteristics . Thus it is likely that additional differences stemming from the relative levels of distinct fatty acids in the membrane affect thermal tolerance and that membrane unsaturation levels will not be the exclusive factor. Other induced changes that impact membrane structure and function can be important. For example, alterations in the length and positions of double bonds within the acyl chain of a lipid molecule can confer widely differing properties and suggest that specific proportions of distinct fatty acid classes may be necessary for optimum membrane function .
A number of processes are called into play in response to stresses such as high temperature. Induction of heat shock proteins serves as one countermeasure to respond to elevated temperatures  and their specific roles in thermal tolerance are becoming clearer [2–4, 49, 50]. Correlations between the antioxidant status of plants and thermotolerance have also been noted recently [51–53]. Additional, perhaps more species-specific, responses are likely to be important in protecting against harmful effects of high temperature growth, including the accumulation of small molecules such as glycinebetaine and through the regulation of carbon fixation via rubisco activase [6, 54].
Although decreases in the amounts of trienoic fatty acids may be an important determinant for plant thermal tolerance, lines possessing elevations in other, distinct fatty acid species, exhibit different characteristics. A triple desaturase mutant of Arabidopsis demonstrated that all trienoic fatty acids in leaf membranes are not essential for growth at low temperature . This fad7 fad8 fad3 mutant also displayed thermostability but actually died after prolonged exposure to 33°C, suggesting that some trienoic fatty acids are indeed essential for high temperature growth . The upper limit of chronic high temperature exposure for all lines tested here was about 34°C, where plants continue to grow and flower but exhibit reduced seed yield (data not shown), consistent with the idea that some trienoic fatty acids are essential. Although not addressed in the present study, it would be of interest to determine if this decreased seed yield from high-temperature-grown plants was related to insufficient levels of linolenic acid, a precursor required for the synthesis of jasmonic acid, a signaling molecule necessary for pollen development .
The results relating to growth performance at high temperature in this study are noteworthy in view of the fact that Arabidopsis is not considered a high temperature tolerant plant . In more heat tolerant species, adjustment of the membrane fatty acid composition may well have greater significance in providing a rational means to control plant high-temperature tolerance. For example, suppression of a FAD7 homolog to concomitantly raise 18:2 and decrease 18:3 in tobacco enabled enhanced growth at elevated temperature . Thus elimination of trienoic fatty acids might be the most critical aspect of altering the membrane composition to favor such enhanced growth. It is unclear, however, if the corresponding increases in monounsaturated and diunsaturated fatty acids that occur in such lines also contribute to the improved tolerance. Several of the mutant lines examined in this study also possessed very low levels of trienoic fatty acids but were accompanied by distinct alterations in other fatty acids. Most of the mutant lines exhibited elevations in the fatty acid species that serves as a precursor to the mutated step. The fad5 line grown at 17°C displayed 16-C fatty acids that were most similar to that of the wild-type line grown at 36°C, and had a membrane fatty acid profile almost identical to the wild type when each was grown at 36°C (Fig. 3A,3B). While such a composition has no deleterious effects at moderate or elevated temperatures, this line, as well as the fad6 line, shows impaired growth and chloroplast development at low temperature . Based on the similar composition to high temperature-grown wild-type plants, the reaction catalyzed by the FAD5 desaturase, which has been identified as an expressed sequence tag , may be an additional target to further manipulate tolerance to high temperature.