acyltransferase activity

Transcription

1 Proc. Natl. Acad. Sci. USA Vol. 85, pp , June 1988 Biochemistry Altered regulation of lipid biosynthesis in a mutant of Arabidopsis deficient in chloroplast glycerol-3-phosphate acyltransferase activity LJERKA KUNST*, JOHN BROWSE*t, AND CHRIS SOMERVILLE* *Michigan State University-Department of Energy Plant Research Laboratory, Michigan State University, East Lansing, MI 48824; and tdsir, Plant Physiology Division, Palmerston North, New Zealand Communicated by P. K. Stumpf, January 25, 1988 (received for review November 30, 1987) ABSTRACT The leaf membrane lipids of many plant species, including Arabidopsis thaliana (L.) Heynh., are synthesized by two complementary pathways that are associated with the chloroplast and the endoplasmic reticulum. By screening directly for alterations in lipid acyl-group composition, we have identified several mutants of Arabidopsis that lack the plastid pathway because of a deficiency in activity of the first enzyme in the plastid pathway of glycerolipid synthesis, acyl- ACP:sn-glycerol-3-phosphate acyltransferase (EC ) (where ACP is acyl carrier protein). The lesion results in an increased synthesis of lipids by the cytoplasmic pathway that largely compensates for the loss of the plastid pathway and provides nearly normal amounts of all the lipids required for chloroplast biogenesis. However, the fatty acid composition of the leaf membrane lipids of the mutants is altered because the acyltransferases associated with the two pathways normally exhibit different substrate specificities. The remarkable flexibility of the system provides an insight into the nature of the regulatory mechanisms that allocate lipids for membrane biogenesis. In the present model of glycerolipid metabolism in higher plants (Fig. 1), two pathways contribute to the synthesis of chloroplast glycerolipids in leaf cells (1-6). The chloroplast is the sole site of de novo fatty acid synthesis (7) and the main products of this process are C16:0-ACP and C18:1-ACP (8). These fatty acids either enter the "prokaryotic pathway" of lipid biosynthesis through acylation of GroP within the chloroplast (9) or are exported as CoA thioesters (3, 10) to enter the "eukaryotic pathway" associated with the endoplasmic reticulum (2, 3). Most of the enzymes of the prokaryotic pathway are located in the inner membrane of the chloroplast envelope where they catalyze the synthesis of acyl2gropgro, acyl2grogal, acyl2grogalgal, and SL (in which acyl2gro is linked to a pyranose that bears an -SO3 group on carbon 6), the major glycerolipids of the thylakoid membranes (11-14). The eukaryotic pathway is responsible for the synthesis of the glycerolipids such as acyl2gropcho, acyl2gropetn, and acyl2groplns, which are found primarily in extrachloroplast membranes (4). In addition, however, acyl2gro moieties from acyl2gropcho are transferred from the endoplasmic reticulum to the chloroplast where they are utilized for synthesis of the galactolipids and SL (1, 5, 15-17). In the majority of higher plants acyl2gropgro is the only product of the prokaryotic pathway, and the remaining chloroplast lipids are synthesized entirely by the eukaryotic pathway (3, 4). These species are known as "18:3 plants." However, in a number of species, including Arabidopsis thaliana, both pathways contribute to the synthesis of acyl2grogal, acyl2grogalgal, and SL (17). These species The publication costs of this article were defrayed in part by page charge payment. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C solely to indicate this fact. CYTOPLASM 4 *--PA; PI PC PE *- PA -+ -_, JI-- 18:1-CoA 4 Lyso-PA Lyso-PA T - G3P t 18:1 -CoA t l 16:0-CoA Acet ite_*16:0-acp *o 18:0-ACP -18:1-ACP 0 Lyso-PA.42-I PG4-4-PA - X DAG G3P l.j MGD DGD 14 CHLOROPLAST FIG. 1. An abbreviated scheme for lipid biosynthesis in the leaves of a 16:3 species. The steps identified by numbers are as follows: 1, fatty acid synthesis; 2, elongation; 3, desaturation; 4 and 6, acyl transfer to sn-glycerol 3-phosphate; 5 and 7, acyl transfer to lysophosphatidic acid. The enzymatic defect in the actl mutant identified in this paper is indicated by a break in the pathway at reaction 4. Symbols for various components (and their abbreviations, used throughout the rest of this paper) are as follows: G3P, sn-glycerol 3-phosphate (GroP); DAG, diacylglycerol (acyl2gro); PA, phosphatidic acid (acyl2grop); lyso-pa, 1-acyl-sn-glycerol 3-phosphate (1- acyl-grop); PI, phosphatidylinositol (acyl2gropins); PC, phosphatidylcholine (acyl2gropcho); PE, phosphatidylethanolamine (acyl2gropetn); PG, phosphatidylglycerol (acyl2gropgro); MGD, monogalactosyldiacylglycerol (acyl2grogal); DGD, digalactosyldiacylglycerol (acyl2grogalgal); SL, sulfolipid; 16:0, hexadecanoyl (C16:0); 18:0, octadecanoyl (C18:0); 18:1, cis-9-octadecenoyl (C18:1); CoA, coenzyme A; ACP, acyl carrier protein. characteristically contain substantial amounts of hexadecatrienoic acid (C16:3), which is found only at the sn-2 position of galactolipid molecules produced by the prokaryotic pathway. These species have been termed "16:3 plants" to distinguish them from 18:3 plants, whose galactolipids contain predominantly a-linolenate (3, 18). We previously described the isolation of a number of mutants of Arabidopsis with altered fatty acid composition. The mutants were identified by direct analysis of leaf fatty acid composition of individual mutagenized plants by Abbreviations: C,, x, fatty acid (or acyl group) containing n carbons and x double bonds (cis configuration unless indicated otherwise); other abbreviations are defined in the legend to Fig. 1.

2 4144 Biochemistry: Kunst et al. gas/liquid chromatography. Several of these mutants were characterized as being deficient in specific desaturases (19, 20). Here we describe the biochemical characterization of a class of mutants with greatly reduced levels of C16:3 acyl groups, due to a deficiency of chloroplast acyl-acp:grop acyltransferase (EC ) activity. Because these mutants lack the activity of the first enzyme of the prokaryotic pathway, the mutation effectively converts a 16:3 plant into an 18:3 type. Thus, the mutants offer an opportunity to examine both the effects of this change on the regulation of glycerolipid metabolism and the physiological significance of the 16:3/18:3 dimorphism. MATERIALS AND METHODS Plant Material. The lines of Arabidopsis thaliana (L.) Heynh. described here were descended from the Columbia wild type. The mutant lines JB3, JB25, JB28, and LK8 were isolated after mutagenesis with ethyl methanesulfonate (21). Before use in experiments, the line JB25 was backcrossed to the wild type at least three times, and an individual with the mutant phenotype was reselected from a segregating population. Plants were grown under continuous fluorescent illumination [ ,uE m -2. 1; 1 E (einstein) = 1 mol of photons] at 220C on a perlite/vermiculite/sphagnum (1:1:1 by volume) mixture irrigated with mineral nutrients (21). Lipid Analysis. Plants were frozen in liquid N2, and lipids were extracted and quantitated as described (17). Lipids were separated by thin-layer chromatography on silica gel-coated plates with a solvent system of chloroform/acetone/methanol/acetic acid/water (100:40:20:20:10 by volume). The kinetics of lipid biosynthesis in intact Arabidopsis plants were followed by measuring incorporation of [14C]acetate into the various lipids (17). Chloroplast Labeling. Chloroplasts used in labeling experiments with [14C]glycerol-3-phosphate were prepared by grinding 20 g of leaf tissue in 200 ml of 0.45 M sorbitol/20 mm Tricine-KOH, ph 8.4/10 mm EDTA/10 mm NaHCO3/0.1% (wt/vol) bovine serum albumin. The extract was filtered through Miracloth (Calbiochem), centrifuged at 270 x g for 90 sec, and resuspended in buffer A (0.3 M sorbitol/20 mm Tricine-KOH, ph 7.6/5 mm MgCl2/2.5 mm EDTA). The chloroplast suspensions were then layered on Percoll gradients prepared by centrifuging 50% (vol/vol) Percoll in buffer A at 43,000 x g for 30 min in a Sorvall SS-34 rotor (22). The gradients were centrifuged at 13,000 x g for 6 min in a Sorvall HB-4 rotor. Intact chloroplasts, which formed a band near the bottom of the gradient, were recovered, diluted with buffer-a, pelleted at 3000 x g for 90 sec in a Sorvall HB-4 rotor, and resuspended in buffer A. Intact chloroplasts (400,g of chlorophyll per ml) were incubated with shaking at 25 C in 0.33 M sorbitol/25 mm Hepes-NaOH, ph 7.9/10 mm NaHCO3/2 mm EDTA/1 mm MgCl2/1 mm MnCl2/0.15 mm sodium acetate/0.4 mm [14C]- glycerol-3-phosphate (30 mci/mmol; 1 mci = 37 MBq) for 20 min under illumination (150,E-m2.S 1-) and then in darkness for 20 min. Reactions were stopped by adding chloroform/methanol (1:1, vol/vol), and lipids were recovered from the chloroform layer after partitioning against 0.2 M H3PO4/2 M KCI (23). Enzyme Assays. Whole-cell extracts were prepared by lysing intact protoplasts in 20 mm Tricine-KOH, ph 8.4/10 mm EDTA. Chloroplast extracts were prepared by isolating intact chloroplasts from protoplasts (24) and then rupturing the chloroplasts in 10 mm Tricine-KOH, ph 7.6/1 mm MgCl2. The chloroplast components then were layered on density step gradients composed of 0.93 and 0.6 M sucrose in 10 mm Tricine'NaOH, ph 7.6/4 mm MgCl2 and were centrifuged for 1 hr in a swinging-bucket rotor at 72,000 x g. The envelopes were collected from the interface of the 0.93 Proc. Natl. Acad. Sci. USA 85 (1988) M and 0.6 M sucrose layers. The fraction containing the chloroplast stromal components was recentrifuged at 130,000 x g for 2 hr to remove any remaining membranes and used immediately for enzyme assays. GroP acyltransferase and monoacyl-grop acyltransferase activities were assayed at 220C essentially as described (9, 25), with ["4C]acyl-CoA as the substrate for chloroplast extracts and [(4C]acyl-ACP for whole cell extracts. The 80-,ul GroP acyltransferase assay mixtures contained 250 mm Mops-NaOH (ph 7.4), 50 tug of bovine serum albumin, 5,uM acyl-acp or acyl-coa, 2 mm L-GroP, and tug of chloroplast stromal protein or jig of protein from whole-cell extracts. The same mixture was used for the monoacyl-grop acyltransferase assays except that 1-oleoyl-GroP was used instead of GroP and pug of chloroplast envelope protein was used instead of the stromal extract. Ribulose-bisphosphate carboxylase (26) and phosphoenolpyruvate carboxylase (27) were assayed essentially as described. [14C]C16:0-ACP (55 mci/mmol) and [14C]C18 1-ACP (55 mci/mmol) were gifts from J. Ohlrogge (Michigan State University). RESULTS Genetic Analysis. Approximately 3000 randomly chosen individuals from two independently mutagenized populations (M2) of plants were screened for altered leaf fatty acid composition. Four lines, designated JB3, JB25, JB28, and LK8, were identified as being deficient in C16:3. The reduction in C16:3 was not accompanied by any increase in the precursors C16:0, C16:1, or C16:2. Genetic complementation tests indicated that the four lines have a lesion at the same locus (results not shown). Therefore, we characterized only one of the lines in detail. The representative mutant line, JB25, is normal in appearance and growth characteristics but can be readily distinguished from the wild type by the reduced amount of C16:3 in the leaf lipids (Table 1). To determine the genetic basis for the alteration in lipid fatty acid composition, the mutant line JB25 was crossed with wild type as the maternal parent. The fatty acid composition of the F1 progeny was indistinguishable from the wild type (Table 1), suggesting a recessive mutation. The frequency of individuals with the mutant phenotype in the F2 population resulting from self-fertilization of F1 plants was also measured by gas chromatography of leaf samples. Of 271 plants analyzed, 64 lacked C16:3 whereas the remaining individuals had wild-type levels of C16:3. This pattern of segregation is a good fit (2 = 0.36; P > 0.5) to the 3:1 hypothesis and indicates that the difference is due to a single recessive nuclear mutation at a locus we designate act]. Biochemical Characterization. Isolated, intact chloroplasts from the wild type incorporated label from [14C]GroP into acyl-grop, acyl2grop, acyl2gropgro, and acyl2grogal (Fig. 2). By contrast, chloroplasts from the mutant were unable to Table 1. Fatty acid composition of total leaf lipids of wild-type (WT) and mutant Arabidopsis grown at 22 C Fatty mol % (mean + SD, n = 10) acid WT (WT x JB25)F1 JB25 C16: ± ± ± 0.3 C16: ± ± 0.4 C16:lt* 1.6 ± ± ± 0.4 C16:2 0.5 ± ± ± 0.3 C16: ± ± 0.3 C18:0 1.7 ± ± 0.2 C18:1 3.0 ± ± ± 0.8 C18: ± ± ± 0.8 C18: ± ± ± 1.5 *Trans isomer.

3 Biochemistry: Kunst et al. WT JB25 - acyl2grogal - acyl2grop amm*- acy12gropgro #6 Aft- - acyl-grop FIG. 2. Distribution of radioactivity among the polar lipids of wild-type (WT) and mutant (JB25) Arabidopsis after [(4C]GroP labeling of isolated chloroplasts. The same amount of radioactivity was applied to each lane. synthesize acyl-grop, acyl2grop, or acyl2grogal but accumulated label in acyl2gropgro. The lack of acyl-grop and acyl2grop accumulation in the mutant suggested that the mutant was deficient in GroP acyltransferase, the first enzyme of the prokaryotic pathway (Fig. 1). On the basis of the chloroplast labeling studies, we assayed the activities of the chloroplast enzymes involved in acyl2grop synthesis. Because of the presence in crude extracts of both chloroplast and microsomal acyltransferases, we first purified chloroplasts and then assayed stromal extracts and the envelope fraction for activity. Stromal extracts of the mutant exhibited only 3.8% of the wild-type activity of the plastid isozyme of GroP acyltransferase (Table 2). Since the chloroplasts were slightly contaminated with protoplasts (1.9% of total cellular phosphoenolpyruvate carboxylase activity was detected in JB25 stromal extract), some of the residual activity in the mutant was due to contamination of the chloroplast fraction by cytoplasmic enzymes. The Table 2. Enzyme activities in chloroplast or whole-cell extracts of mutant and wild-type Arabidopsis Activity, pmol/min per mg of protein Enzyme Chloroplast Whole cell GroP acyltransferase* Wild type JB Wild type + JB25t ND LK8 ND 3.9 Monoacyl-GroP acyltransferaset Wild type JB LK8 ND 3.4 RbuP2 carboxylase Wild type JB P-enolpyruvate carboxylase Wild type JB Values are the means of 3-5 assays. ND, not determined. *Chloroplast and whole-cell extracts were assayed with C18:1-CoA or C18i1-ACP, respectively. tequal volumes of mutant and wild-type extract were mixed before assay. SChloroplast and whole-cell extracts were assayed with C16:0-CoA or C16:0-ACP, respectively. Ribulose-bisphosphate (RbuP2) carboxylase and phosphoenolpyruvate carboxylase activities are given as nmol/min per mg of protein. Proc. Natl. Acad. Sci. USA 85 (1988) 4145 mutant had wild-type levels of the chloroplast enzymes ribulose-bisphosphate carboxylase and monoacyl-grop acyltransferase (Table 2). Therefore, we believe that the acti locus specifically controls the activity of the plastid isozyme of GroP acyltransferase. Chloroplast GroP acyltransferase can use either acyl-acp or acyl-coa for the acylation reaction, but when both are present, ACP thioesters are exclusively used as substrate (9). Since ACP thioesters are confined to the chloroplasts, it was of interest to determine whether they could be used as acyl donors by the other leaf acyltransferases. Therefore, we assayed crude leaf extracts by measuring the incorporation of [14C]C18:1-ACP into lipids. Under these conditions, leaf extracts of the mutants JB25 and LK8 exhibited <5% the GroP acyltransferase activity of the wild type (Table 2). On the other hand, when [14C]C18:1-CoA was used as the substrate in assays of crude leaf extracts, the total GroP acyltransferase activity in the mutant JB25 was 82% of the level in wild-type plants (results not shown). This indicates that acyl-acp thioesters are not good substrates for extraplastid acyltransferases. Labeling of Leaves. To investigate the consequences of the enzyme deficiency on lipid biosynthesis in the mutant, we applied [14C]acetate to leaves of mutant and wild-type plants and followed the redistribution of radioactivity in polar lipids during the next 142 hr. The mutation caused dramatic differences in the pattern of [14C]acetate incorporation in JB25 when compared to the wild type (Fig. 3). As we discussed previously (17), the labeling kinetics for wild-type plants demonstrate the parallel operation of the two pathways of lipid synthesis. Flux through the prokaryotic pathway leads to the substantial labeling of acyl2grogal at early times, whereas the subsequent transfer of 14C from acyl2gropcho to acyl2grogal and acyl2grogalgal occurs via the eukaryotic pathway. In contrast, the mutant contained the label primarily in acyl2gropcho at short times, whereas acyl2grogal contained <3% ofthe total radioactivity. During the course of the experiment there was a steady and substantial decline of radioactivity in acyl2gropcho, which was accompanied by increased labeling of galactolipids and SL, so that by the end of the experiment the distribution of 14C among the various polar lipids was similar to the wild type. These kinetics are consistent with a precursor-product relationship between acyl2gropcho and the chloroplast glycolipids and indicate that acyl2grogal in JB25 is made almost entirely by the eukaryotic pathway. In these respects the a60 A 606 V~~~~~~~ (D U..T40 0 E ~0 310 ~~~~ ae co ~ ~ ie() lgsae FIG. 3. Distribution of radioactivity in leaf lipids of (A) wild-type and (B) mutant JB25 after labeling with [14C]acetate. Data in A were taken from ref. 17 and are included to facilitate comparison of mutant and wild type. e, acyl2gropcho; m, acyl2grogal; Eo, acyl2grogal- Gal; *, SL; A, acyl2gropgro; o, acyl2gropetn; *, acyl2groplns.

4 4146 Biochemistry: Kunst et al. Proc. Natl. Acad. Sci. USA 85 (1988) labeling kinetics are extremely similar to those observed in analogous experiments with 18:3 plants, in which all the chloroplast glycerolipids except acyl2gropcho are derived from the eukaryotic pathway (3-5). The deficiency in the chloroplast acyl-acp:grop acyltransferase found in JB25 would be expected to block acyl2gropgro synthesis by the prokaryotic pathway. However, acyl2gropgro does become labeled in the mutant, although the extent of '4C incorporation into this lipid is only about half of that found in the wild type (Fig. 3). Lipid Composition. In wild-type Arabidopsis the prokaryotic pathway is responsible for producing approximately 70% of the total leaf acyl2grogal, 12% of the acyl2grogalgal, 63% of the SL, and 85% of the acyl2gropgro (17). The total lipid content of leaves from the mutant was the same as the wild type on both fresh-weight and chlorophyll bases (data not shown). Analyses of several different batches of plants consistently revealed that the mutant had 10-25% less acyl2gropgro than the wild type. There was also a 9% decrease in acyl2grogal, a 12% increase in acyl2grogalgal, a 30% decrease in SL, a 12% increase in acyl2gropcho, and a 10% increase in acyl2gropetn (Table 3). The similar proportions ofeach lipid in the mutant and wild type, together with the data from the labeling experiment (Fig. 3), indicate that the lack of synthesis of galactolipids and SL by the prokaryotic pathway in the mutant is largely compensated for by increased production of these lipids via the eukaryotic pathway. The differential effect of the mutation on the amounts of the various chloroplast-specific lipids reflects the various degrees to which these lipids are normally produced by the prokaryotic pathway (17). The increased amounts of acyl2gropcho and -Etn are consistent with (but not proportional to) the increased flux through the eukaryotic pathway. To determine whether acyl2gropgro in the mutant had the characteristic structure of a product of the prokaryotic pathway, the purified lipid was digested with Rhizopus lipase and the fatty acid compositions of the 1-acyl-GroPGro and released fatty acid were determined (17). This analysis indicated that in the mutant, 75% of the fatty acid at the sn-2 position of acyl2gropgro was C16. In the wild type the sn-2 position was 83% C16 (17). By contrast, other polar lipids in the mutant contained >90% C18 fatty acids at sn-2 of the glycerol, indicating that they were produced by the eukaryotic pathway. The chloroplast-specific acyl group C16:3, which is characteristic of prokaryotic acyl2grogal, is virtually absent from the lipids of the mutant and is replaced by C18 acyl groups (primarily C18:3). The 95% reduction in the amount of unsaturated C16 fatty acids in acyl2grogal is not accompanied by an increase in C16 fatty acids in any other lipid. Thus, the molecules of C16:0 that would normally be utilized by the prokaryotic pathway appear to be elongated and desaturated Table 3. to C18:1 before being exported from the chloroplast to enter the eukaryotic pathway. The implication is that, in the wild type, the amount of C16:0 exported to the cytoplasm is not simply regulated by the availability of C16:0-ACP. In a detailed analysis of wild-type Arabidopsis, we showed (17) that for every 1000 fatty acid molecules made in the chloroplast, 615 enter the eukaryotic pathway (117 as C16 and 498 as C18). A similar analysis of the mutant showed that the increase in flux through the eukaryotic pathway (to 950 per 1000) is made up almost entirely of C18 fatty acid chains (126 C16 plus 824 C18). However, the C16/Cj8 ratio in acyl2grop- Cho, -Etn, and -Ins are the same as in the corresponding lipids of the wild type (Table 3). In contrast, the C16/Cj8 ratio in galactolipids and SL of the mutant is in each case less than the ratio calculated for these lipids synthesized by the eukaryotic pathway in wild-type Arabidopsis (table 4 of ref. 17). Thus, the additional C18 fatty acids entering the eukaryotic pathway in the mutant are found specifically in the additional quantities of chloroplast lipids (galactolipids and SL) that are produced by the eukaryotic pathway in response to the loss of the prokaryotic pathway. The mutation causes an increase in the amount of C18:1 and a decrease in the amount of C18:3 in all of the extrachloroplast (acyl2gropcho, -Etn, and -Ins) lipids of the mutant. There was little or no effect on the amount of C18:2 in these lipids (Table 3). The data indicate a 10-20% reduction in the extent of C18:1 desaturation in these lipids in the mutant relative to the wild type. It seems likely that this is caused by the inability of the endoplasmic reticulum C18.1 desaturase to completely metabolize the increased flux of lipid through the eukaryotic pathway in the mutant. DISCUSSION In the mutant line JB25, a single recessive nuclear mutation at the act] locus causes a specific deficiency in the activity of the GroP acyltransferase. Three other allelic mutants show all of the changes in fatty acid composition that have been described here for JB25, indicating that all the changes are direct consequences of the deficiency in acyltransferase activity. Analysis of these mutants provides clear evidence that an extensive network of controls exists to regulate leaf lipid metabolism and maintain suitable glycerolipid and fatty acid composition of cellular membranes. Regulation of Glycerolipid Synthesis. The labeling data (Fig. 3) and the lipid analysis (Table 3) indicate that loss of the acyltransferase activity does not result in the accumulation of precursors (C18:1- and C16:0-ACP) upstream of the enzyme deficiency but causes a redirection of lipid metabolism so that the eukaryotic pathway predominates in the mutant. Somewhat unexpectedly, this redirection has little effect on the amount of each lipid that accumulates or on the total Fatty acid composition of leaf lipids from wild-type (WT) and mutant Arabidopsis grown at 220C mol % Fatty acyl2grogal acyl2grogalgal SL acyl2gropgro acyl2gropcho acyl2gropetn acyl2groplns acid WT JB25 WT JB25 WT JB25 WT JB25 WT JB25 WT JB25 WT JB25 C16: C16:It C16: C16: C18: C18: C C18: (%) , Not detected.

5 Biochemistry: Kunst et al. glycerolipid content of the tissue. The enhanced flux through the eukaryotic pathway in the mutant is accompanied by a change in the proportion of individual lipids synthesized by this pathway. Thus, the amount of lipids such as acyl2grop- Etn and acyl2gropins in the extrachloroplast membranes of the mutant is similar to that in the wild type. However, to compensate for the loss of the prokaryotic pathway, the amount of acyl2gropcho synthesized must be increased about 3-fold to provide the acyl2gro moieties required for normal levels of glycerolipid synthesis by the chloroplast. However, the amount of acyl2gropcho that accumulates is apparently only that required for synthesis of the extrachloroplast membranes, since the level of this lipid in the mutant is only slightly higher than in the wild type. Clearly, the synthesis of acyl2gropcho and export from the microsomal membranes are regulated in concert to meet the requirements for lipid synthesis by chloroplasts. Although it is not currently possible to infer the mechanisms, it is apparent that membrane biosynthesis by the chloroplast and the microsomal membranes is closely coordinated. Altered C18/C16 Ratio. In wild-type Arabidopsis, C16 fatty acids represent half of the acyl chains found in lipids made by the prokaryotic pathway. However, loss of the prokaryotic pathway in the mutant does not result in redirection of the C16 chains into the eukaryotic pathway. Instead the C16:0-ACP is apparently elongated and desaturated to C18:1-ACP before export from the chloroplast. Thus, the overall ratio of C16 to C18 chains is reduced from 0.3 in the wild type to 0.18 in the mutant (Table 1). It is noteworthy that the primarily extrachloroplast lipids (acyl2gropcho, -Etn, and -Ins) have levels of C16:0 similar to that of the wild type. The implication is that the amount of export of C16:0 is not regulated simply by the availability of C16:0. However, the increased C18/C16 ratio suggests that elongation activity is regulated by availability of substrate (C16:0) and that this is determined by competition between alternative pathways of C16:0 metabolism. Synthesis of acyl2gropgro. Evidence from other studies (13, 28) indicates that isolated chloroplasts are able to synthesize acyl2gropgro at rates sufficient to meet the requirements for chloroplast membrane biogenesis and that the pathway for synthesis of this lipid involves the same pools of acyl- and acyl2grop used for the synthesis of prokaryotic galactolipids and SL. It is puzzling, therefore, that the mutant contains at least 75% as much acyl2gropgro as the wild type, even though it exhibits less than 4% of the wild-type level of GroP acyltransferase activity. One possibility is that the amount of acyl2grop made in the mutant with the residual 4% of chloroplast acyltransferase activity is adequate to meet most of the requirement for acyl2gropgro synthesis and is utilized preferentially for synthesis of that lipid. A reduction in the size of the acyl-grop and acyl2grop pools might explain why these compounds did not accumulate radioactivity during the [14C]GroP labeling of isolated chloroplasts (Fig. 2). Since only very small amounts of prokaryotic acyl2grogal are found in the mutant (Table 3), this explanation would require that acyl2gropgro synthesis from acyl2grop be efficiently maintained at the expense of acyl2gro synthesis. The other possibility seems to be that an alternative source of acyl2grop is used for acyl2gropgro synthesis in the mutants. However, the predominance of C16 fatty acids on the sn-2 position of acyl2gropgro in the mutant is most consistent with this lipid being derived from the prokaryotic pathway in the chloroplast rather than from any other source. Evolutionary Implications. A preliminary physiological characterization of the mutants suggests that the loss of the prokaryotic pathway is not deleterious. The general appearance of the mutant plants is similar to the wild type, and they Proc. Natl. Acad. Sci. USA 85 (1988) 4147 are not impaired in growth or development under standard growth conditions (unpublished data). Thus, the question that inevitably arises concerns the dispensability of the prokaryotic pathway for glycerolipid synthesis in the mutant, as well as in naturally occurring 18:3 plants and in the fruits of 16:3 species (29). In the course of evolution, the majority of higher plants have abandoned the prokaryotic pathway, so that it persists to varying degrees only in less advanced genera (18). Since a single mutation can eliminate the prokaryotic pathway, it seems reasonable to suggest that there must be some physiological advantage associated with the presence of the prokaryotic pathway in certain circumstances. We anticipate that further analysis of the mutants described here will provide useful insights into why 16:3 plants and 18:3 plants coexist. We thank S. Hugly, P. McCourt, S. Gardiner, J. Ohlrogge, G. Rouighan, and R. Slack for helpful discussions. This work was supported in part by grants from the U.S. Department of Energy (AC02-76ER01338) and the McKnight Foundation. L.K. was supported in part by a fellowship from Monsanto. 1. Roughan, P. G., Holland, R. & Slack, C. (1980) Biochem. J. 188, Williams, J. P. & Khan, M. (1982) Biochim. Biophys. Acta 713, Roughan, P. G. & Slack, C. (1982) Annu. Rev. Plant. Physiol. 33, Roughan, P. G. & Slack, C. R. (1984) Trends Biochem. Sci. 9, Heinz, E. & Roughan, P. G. (1983) Plant Physiol. 72, Gardiner, S. E. & Roughan, P. G. (1983) Biochem. J. 210, Ohlrogge, J. B., Kuhn, D. N. & Stumpf, P. K. (1979) Proc. Nati. Acad. Sci. USA 76, Soll, J. & Roughan, P. G. (1982) FEBS Lett. 146, Frentzen, M., Heinz, E., McKeon, T. A. & Stumpf, P. K. (1983) Eur. J. Biochem. 129, Andrews, J. & Keegstra, K. (1983) Plant Physiol. 72, Heinz, E. (1977) in Lipids and Lipid Polymers in Higher Plants, eds. Tevini, M. & Lichtenthaler, H. K. (Springer, Heidelberg), pp Joyard, J. & Douce, R. (1977) Biochim. Biophys. Acta 486, Andrews, J. & Mudd, B. (1985) Plant Physiol. 79, Douce, R. & Joyard, J. (1979) Adv. Bot. Res. 7, Murphy, D. J. & Stumpf, P. K. (1980) Plant Physiol. 66, Ohnishi, J. & Yamada, M. (1980) Plant Cell Physiol. 21, Browse, J., Warwick, N., Somerville, C. R. & Slack, C. R. (1986) Biochem. J. 235, Jamieson, G. R. & Reid, E. H. (1971) Phytochemistry 10, Browse, J., McCourt, P. J. & Somerville, C. R. (1985) Science 227, Browse, J., McCourt, P. J. & Somerville, C. R. (1986) Plant Physiol. 81, Haughn, G. W. & Somerville, C. R. (1986) Mol. Gen. Genet. 204, Cline, K., Andrews, J., Mersey, B., Newcomb, E. H. & Keegstra, K. (1981) Proc. Natl. Acad. Sci. USA 78, Browse, J., McCourt, P. J. & Somerville, C. R. (1986) Anal. Biochem. 152, Somerville, C. R., Somerville, S. C. & Ogren, W. L. (1981) Plant Sci. Lett. 21, Bertrams, M. & Heinz, E. (1981) Plant Physiol. 68, Pierce, J. W., McCurry, S. D., Mulligan, R. M. & Tolbert, N. E. (1982) Methods Enzymol. 89, Stitt, M., Bulpin, P. V. & ap Rees, T. (1978) Biochim. Biophys. Acta 544, Roughan, P. G. (1985) Biochim. Biophys. Acta 835, Whitaker, B. D. (1986) Planta 169,