Acetyl-acyl carrier protein is not a major intermediate in fatty acid biosynthesis in spinach

Transcription

1 Eur. J. Biochem. 213, (1993) 0 FEBS 1993 Acetylacyl carrier protein is not a major intermediate in fatty acid biosynthesis in spinach Jan G. JAWORSKI', Dusty POSTBEITTENMILLER* and John B. OHLROGGE2 ' Chemistry Department, Miami University, Oxford OH, USA * Department of Botany and Plant Pathology, Michigan State University, East Lansing MI, USA (Received November 24, 1992/January 25, 1993) EJB /6 The extent to which acetylacyl carrier protein (acetylacp) is an intermediate in fatty acid biosynthesis was examined. AcetylACP was the least effective primer of fatty acid synthesis by spinach extracts when compared to acetylcoa, butyrylacp or hexanoylacp. Furthermore, the rate of acetylacpprimed fatty acid synthesis was inhibited significantly by cerulenin, indicating that the slow utilization of acetylacp was predominantly by 3oxoacylACP synthase I. In lightincubated isolated chloroplasts with high rates of fatty acid synthesis (greater than 800 nmol h' 'mg chlorophyll'), the rate of acetylacp metabolism was at least 1030fold slower than the rate of butyrylacp metabolism. The relatively slow metabolism of acetylacp provided in situ evidence that (a) butyrylacp was formed principally from condensation of malonylacp with acetyl CoA and (b) acetylacp was a minor participant in fatty acid biosynthesis. What is the sequence of reactions that constitutes the predominant fatty acid biosynthetic pathway in plants? As first delineated over two decades ago (Volpe and Vagelos, 1973 ; Stumpf, 1987) the fatty acid biosynthetic pathway in plants and bacteria included the transfer of an acetyl moiety from CoA to acyl carrier protein (ACP), catalyzed by acetylco A: ACP transacylase (acetyl transacylase). The spinach acetyl transacylase has been characterized and in vitro data suggested that it was potentially a ratelimiting step in this pathway (Shimakata and Stumpf, 1983b). The acetylacp was presumably then used along with malonylacp by 3 oxoacylacp synthase I (KAS I) in the condensation reaction that initiates synthesis of the acyl group. KAS I catalyzed each of the subsequent condensation reactions until a C,,,, fatty acid was synthesized, at which point, KAS I1 catalyzed the final condensation reactions. Recently, however, another condensing enzyme, KAS 111, has been discovered in plants (Jaworski et al., 1989; Walsh et al., 1990) and Escherichia coli (Jackowski and Rock, 1987; Jackowski et al., 1989). KAS 111 is distinguished from the previously studied KAS I by its ability to use acetylcoa directly, rather than acetylacp. The in vitro activity of KAS 111 in spinach is at least fivefold higher than that of acetyl transacylase, and both enzymes have a comparable K, for acetylcoa. Therefore, the in vitro data suggest that acetylcoa could normally bypass the acetyl transacylase reaction and enter fatty acid synthesis directly by way of the KAS 111 reaction. Correspondence to J. G. Jaworski, Chemistry Department, Miami University, Oxford, OH 45056, USA Fax: f Abbreviations. ACP, acyl carrier protein ; KAS, 3oxoacylACP synthase (abbreviated from the common name 3ketoacylACP synthase). Enzymes. 3Oxoacyl[acylcarrierprotein] synthase, 3oxoacyl ACP synthase (EC ) ;[acylcamerprotein] acetyltransferase, acetylcoa:acp transacylase (EC ). However, direct evidence for the presence of acetylacp in spinach leaf was recently provided (PostBeittenmiller et al., 1991). Furthermore, a dynamic role for acetylacp in lipid metabolism in vivo was implicated by the changing levels of acetylacp that were coincident with changes in the overall rate of fatty acid synthesis. The objectives of the present study were to (a) determine the relative contribution of acetylcoa and acetylacp in the initial reactions of fatty acid synthesis in spinach in vivo and (b) explain the changing levels of acetylacp in response to changes in the rates of fatty acid synthesis. EXPERIMENTAL PROCEDURES Materials [ 1 '4C]AcetylCoA (56 Ci/mol) was prepared from sodium [I4C]acetate by the method of Clough et d. (1989) and [214C]malonylCoA (50 Ci/mol) was prepared from [14C]malonic acid (Rutkoski and Jaworski, 1978). Acetyl, butyryl, and hexanoylacps were prepared chemically from the respective acylimidazole (Cronan and Klages, 1981) and recombinant spinach ACPI (Clough et al., 1992). All other chemicals were reagent grade or better. Enzyme assays A crude spinach leaf extract was prepared from a % ammonium sulfate precipitate (Shimakata and Stumpf, 1982), which was then dissolved and dialyzed against 50 mm potassium phosphate ph 8.0, 20% (by vol.) glycerol, and 2 mm dithiothreitol. For KAS assays, the reactions were carried out in two steps, initially producing the unstable 3 oxoacylacps, and then in the second step, enzymatically reducing the KAS product to the stable saturated acylacp. The initial reaction, carried out under initial velocity con

2 982 ditions for 5 min at 32"C, contained 100 mm Tris ph 8.0, 800 pm dithiothreitol, 10pM spinach ACP, 10 pm [2 14C]malonylCoA (50 Ci/mol), 3 pl crude spinach leaf extract (57 pg protein),with 10 pm acetylcoa or acylacp and 100 pm cerulenin as indicated in a total volume of 50 pl. Reactions were stopped by transferring an aliquot of the reaction mix at the designated interval to an equal volume of 10% (masdvol.) trichloroacetic acid, mixing, and storing on ice for 515 min. All samples were diluted with 1 ml 5% trichloroacetic acid, centrifuged, and the supernatant discarded. In the second step, the 3oxoacylACPs were solubilized in 100 mm Tris ph 8.0 and reduced to the saturated acylacp by reaction with 1 mm NADH, 2 mm NADPH and 1 pl crude spinach leaf extract (19 pg protein) treated with 10 mm Nethylmaleimide to inhibit irreversibly KAS activity. For assay of acetyl transacylase, the reaction mixture contained 100 mm Tris ph 8.0, 800 pm dithiothreitol, 10 pm spinach ACP, 10 pm [l'4c]acetylcoa (56 CUmol), and 15 pl crude spinach leaf extract (285 pg protein) in a total volume of 150 pl. The acetylacp was recovered for analysis as described above except that the reduction reaction was omitted. The acylacps were then analyzed as described below. Chloroplast assays Chloroplasts from hydroponically grown spinach (Spinacia olerucea) were prepared using a 40% Percoll cushion as previously described (Roughan, 1987). Chloroplast incubations (500 pl) with [14C]acetate were performed according to Roughan (1987) with additions as described in the text and with the following modifications. Reactions were stopped with an equal volume of 5% trichloroacetic acid rather than chlorofodmethanol (1 : 1). The intactness and activity of the isolated chloroplasts was verified by their ability to incorporate acetate into longchain fatty acids at high rates. Aliquots (50 PI) were removed from the trichloroacetic acid suspension, transferred to 2 ml chloroform/ methanol (1 : 1) and partitioned against 0.9 ml water. The chloroform layer was recovered, dried and the radioactivity in lipid residue determined by scintillation counting. Chloroplast preparations routinely exhibited acetate incorporation rates greater than 800 nmol. h'. mg chlorophyll'. ReversephaseHPLC analysis of acetylcoa The acetylcoa was analyzed essentially as previously described (PostBeittenmiller et al., 1992). The trichloroacetic acid suspensions from the chloroplast incubations were incubated on ice for 515 min, centrifuged, and the supernatant was recovered. The trichloroacetic acid was removed by extracting four times with diethyl ether and the aqueous phase was dried under vacuum. The dried precipitate was resuspended in 0.1 M potassium phosphate ph 5 to a final concentration equivalent to 1 mg chlorophywm1, and the CoA esters separated by HPLC using a reversephase C,, column (4.6 X 250 mm; the Nest Group) and a linear phosphate buffer/acetonibiile gradient (Corkey, 1988). The mass of acetylcoa was determined by absorbance at 254 nm. The radioactivity into the acetylcoa was determined by collecting 1ml fractions (1 min) from the HPLC analysis and measuring incorporation using liquid scintillation counting. To confirm that these methods resulted in quantitative recovery of the CoA esters, [14C]acetylCoA was added to some of the preparations before trichloroacetic acid precipitation ACP \ AcetylACP AcetylCoA.L = malonylcoa CoAcp4t \ malonylacp acetoacetyi ACP Fig. 1. Scheme for the initial reactions of fatty acid biosynthesis in plants. Bold arrows illustrate the reactions with major activity. and processing for HPLC analysis. In these preparations, greater than 90% of the radioactivity was recovered in a single peak corresponding to the acetylcoa standard. Analysis of acylacps Tissue sampling and extraction of acylacps from spinach leaves were as described previously (PostBeittenmiller et al., 1991). For analysis of acylacps, the trichloroaceticacidprecipitated proteins from leaf extracts, chloroplast incubations, or enzyme assays were pelleted in a microfuge at 4 C. The protein precipitates were redissolved in 50 mm Mops ph 7.6, 10 mm Nethylmaleimide, separated by urea/ PAGE, and transferred to nitrocellulose. The acylacps were visualized using either immunoblotting techniques with antibody directed against spinach ACP or by direct autoradiography as previously described (PostBeittenmiller et al., 1991). Relative 14C levels in each acylacp were obtained by a densitometric scan of the autoradiogram. RESULTS AND DISCUSSION In vitro utilization of acetylacp Both acetylcoa and acetylacp are potential primers for the first condensation reaction of fatty acid synthesis. Previous in vitro data (Jaworski et al., 1989) suggested that direct incorporation of acetylcoa into butyrylacp may be the main pathway for fatty acid synthesis, largely bypassing the acetyl transacylase reaction (Fig. 1). To investigate further the relative importance of the KAS I11 and acetyl transacylase reactions in fatty acid synthesis, the preference of the fatty acid synthase enzymes for different potential primers was examined. An ammonium sulfate precipitate of a spinach leaf homogenate containing all the enzymes of fatty acid synthase was prepared. Its ability to incorporate ['4C]malonylCoA into fatty acids was determined using the primers acetylcoa, the preferred substrate of KAS I11 (Jaworski et al., 1989), acetylacp, butyrylacp, or hexanoyl ACP, all substrates of KAS I (Shimakata and Stumpf, 1982). In these experiments, spinach ACP was used and the acyl ACPs were prepared from the respective acylimidazole (Cronan and Klages, 1981), a chemical procedure that maintains the native conformation of the acylacp. The reaction rates were linear under the conditions utilized, and the concentration of the acetylcoa, and the acylacps were all 10 pm. In order to distinguish between the utilization of each primer by KAS I and KAS 111, reactions were carried out in the presence or absence of 100pM cerulenin. This potent antibiotic inhibitor of fatty acid synthesis is known to inhibit this process irreversibly by specifically alkylating the active site of KAS I (Kauppinen et al., 1988>, resulting in 100%

3 983. t 12 0 with cerulenin c 0 CI minus cerulenin h 10 F C E a O6 E.. r O4. > L u 02 v) U Y COA 2 0ACP 4 0ACP 6 0ACP Primer Fig. 2. Priming specificity for fatty acid synthesis in the presence and absence of cerulenin. 3OxoacylACP synthase reactions were catalyzed by a 4080% saturated ammonium sulfate precipitate from spinach leaf. All primers as indicated were 10 pm and, when present, cerulenin was 100 pm. Products were reduced and dehydrated and then were separated by native PAGE and blotted to nitrocellulose as described in Experimental Procedures. Relative 14C in acylacp products was obtained from a densitometric scan of the autoradiogram. To calculate the rates of each reaction, the relative 14C incorporated into products in 5 min with each primer was compared with the I4C incorporated into acylacp with acetylcoa plus cerulenin. The rate of 3oxoacylACP synthase was determined independently with 10 pm [ 4C]acetylCoA and 100 pm cerulenin as described previously (Jaworski et al., 1989). Each data point is the average of three experiments 2 SD. 2 : 0, acetyl; 4 : 0, butyryl; 6 : 0, hexanoyl Time (s) Fig. 3. Incorporation of [14C]acetate into acetylcoa and acyl ACPs in spinach chloroplasts incubated in the light. Chloroplasts were preincubated in the light for 2 min before addition of [ Clacetate and incubation in the light for the indicated time. Products were separated by native PAGE and blotted to nitrocellulose as described in Experimental Procedures. The I4C in acylacps were obtained from a densitometric scan of the autoradiogram and expressed as a percentage of the maximum label in butyrylacp (30 and 120s). Label in acetylcoa was obtained directly from liquid scintillation counting of HPLC fractions as described in Experimental Procedures. These labeling experiments were repeated four times and the data for the figure is taken from a single representative experiment. In these experiments, maximum label was pmol [ ClacetylCoA, pmol [14C]acetylACPI and 8 pmol [ C] butyrylacpumg chlorophyll. 2 : 0, acetyl; 4: 0, butyryl. inhibition of KAS I at concentrations less than 10 pm (Shimakata and Stumpf, 1982). In contrast, 100pM cerulenin had no effect on KAS Ill (Jaworski et al., 1989). Thus, reactions carried out in the presence of cerulenin were used to measure the substrate specificity of only KAS 111, while those carried out in the absence of cerulenin measure the combined activities of KAS I and KAS 111. The assays were carried out under conditions that each primer was extended by only two carbons and the reaction products were analyzed by native PAGE and autoradiography (Jaworski et al., 1989). The results obtained from a densitometric scan of the autoradiogram (Fig. 2) illustrate that acetylacp was the least effective primer tested. In the presence of cerulenin, with inactive KAS I, acetylcoa and butyryl ACP were the preferred substrates, hexanoylacp was less effective, and acetylacp was used poorly to prime fatty acid synthesis. In the absence of cerulenin, KAS I was active, and the priming activity of both butyrylacp and hexanoyl ACP increased and was comparable to acetylcoa. The ability to use acetylacp improved, but even in the absence of cerulenin, it was still a relatively poor substrate. These data indicated that neither KAS I nor KAS 111 effectively used acetyl ACP as a primer under in vitro assay conditions and, furthermore, that acetylacp was used primarily by KAS I rather than KAS 111. Under in vivo conditions, it would be expected that the difference in the rates of utilization of acetylcoa and acetylacp would be even greater, since the level of acetylcoa in spinach chloroplast was estimated to be ym (PostBeittenmiller et al., 1992), while the level of acetylacp was estimated to be pm (PostBeittenmiller et al., 1991). Shimakata and Stumpf (1983a) previously reported a relative activity of spinach KAS I with acetylacp closer to the activity with butyrylacp or hexanoylacp than shown in Fig. 2. However, the earlier study used E. coli ACP and the butyrylacp and hexanoylacp substrates were prepared from the respective mixed anhydrides. It has been reported that the rates of some reactions which utilize ACP may be influenced by the source of ACP (Guerra et a]., 1986; Clough et al., 1992). Furthermore, acylacp preparations from mixed anhydrides are more variable than acylacps prepared enzymatically (Jaworski and Stumpf, 1974; Rock and Cronan, 1979) or from acylimidazoles (Cronan and Klages, 1981). AcetylACP metabolism in chloroplasts The low rate of fatty acid synthesis when acetylacp was used in vitro as a primer led to an examination of its role in in situ fatty acid synthesis. The utilization of acetyl ACP as an intermediate was examined using lightincubated isolated chloroplasts undergoing rates of fatty acid synthesis comparable to in vivo rates, i.e. greater than 800 nmol. min. mg chlorophyll. If acetylacp is a major intermediate in fatty acid synthesis reactions, its rate of radiolabeling due to the flux of carbon through the acetyl group should mimic that of the other intermediate acyl groups. The incorporation of [I4C]acetate into acetylcoa and intermediate acylacps in chloroplasts was determined over a 5min period. At each time, the trichloroaceticacidsoluble fraction of the chloroplasts was used for analysis of acetylcoa by HPLC and the trichloroaceticacidinsoluble fraction was used for analysis of the acylacps by autoradiography and immunoblot techniques. The synthesis of acetylcoa is catalyzed by acetylcoa synthetase whose activity is 35fold faster than the in vivo rates of fatty acid synthesis (Roughan et al., 1979). Thus, in less than los, the acetylcoa pool was maximally labeled (Fig. 3). ButyrylACP was also fully

4 984 labeled within 10 s, while hexanoylacp and octanoylacp were near maximum labeling after 30 s (data not shown). In sharp contrast to the rapid labeling of these acyl groups, the acetylacp was labeled much more slowly and had not reached its maximum after 5 min. Thus the time required for acetylacp to reach maximum labeling was at least 30fold greater than for acetylcoa or butyrylacp. During the course of the incubation, the levels of acetylcoa and acyl ACPs were essentially unchanged as determined by HPLC and imniunoblots (data not shown). The above results suggested that the flux of carbon from acetylcoa through acetylacp was very slow compared to the flux through butyrylacp during fatty acid synthesis. However, it should be noted that the size of the respective acylacp pools would also influence the kinetics of their labeling. Furthermore, as fatty acid synthesis proceeds, 14C in butyrylacp would arise from entry of labeled carbons from two different molecules, malonylacp and either acetylcoa or acetylacp. The malonylcoa pool was at least 2050fold smaller than the acetylcoa pool, based on the HPLC analysis of the acylcoas, and was approximately the same size as the malonylacp pool (PostBeittenmiller et al., 1991). These small pool sizes, coupled to its use in each cycle of fatty acid synthesis, would result in very rapid turnover of the malonylthioester pools and thus they would have essentially the same specific radioactivity as the precursor, acetylcoa. The acetylacp pool was at least twofold greater than the butyrylacp (PostBeittenmiller et al., 1991) and, as a result, its specific radioactivity might be expected to change more slowly than butyrylacp. An algebraic model for the labeling of the intermediates of the initial reactions of fatty acid synthesis was developed in order to analyze this question more carefully (see Supplement to this paper). This model predicted that if butyrylacp was synthesized primarily from the condensation of acetylcoa and malonylacp, catalyzed by KAS 111, its specific radioactivity should change at essentially the same rate as the acetylcoa pool. On the other hand, it could be predicted that if butyryl ACP arose primarily from KAS I activity, condensing acetyl ACP and malonylacp, the specific radioactivity of butyryl ACP would change at a rate that would be nearly the same as that of acetylacp. Rates of butyrylacp labeling intermediate between acetylcoa and acetylacp would be proportional to the relative contribution of KAS 111 and KAS I. Furthermore, the model predicted that the rate of change in the specific radioactivity of butyrylacp relative to the rates for acetylcoa and acetylacp was primarily influenced by the relative rates of KAS I11 and JSAS I. Since the specific radioactivity of butyrylacp changed at essentially the same rate as for acetylcoa, the data support the proposal that fatty acid synthesis is largely primed directly by acetylcoa and that acetylacp is a comparatively minor intermediate. The relative metabolism of acetylcoa and acylacp could also be demonstrated with a pulsekhase of [I4C]acetate in lightincubated chloroplasts. Isolated chloroplasts were preincubated in the light with [I4C]acetate to label extensively the pools of acetylcoa and the acylacps before adding a 50fold excess of unlabeled acetate. The disappearance of 14C from the intermediate pools during the next 45 s was then measured. As shown in Fig. 4, 14C was lost at a much slower rate from the acetylacp pool than from all the other intermediates examined, and moreover, the loss of label from butyryl, hexanoyl, and octanoylacp closely followed the loss of 14C from acetylcoa. Estimates of the initial rates for loss of isotope indicated that the pools of all I. a Chase time (seconds) Fig. 4. Nonradioactive acetate chase of 14C from acetylcoa and acylacps in spinach chloroplasts incubated in the light. Chloroplasts were preincubated with [I4C]acetate in the light for 2 min and then in the dark for 2 min before addition of a SOfold excess of unlabeled acetate followed by incubation in the light for the indicated time. Products were separated by native PAGE and blotted to nitrocellulose as described in Experimental Procedures. The I4C levels in (0) acetyl (2: 0), (W) butyryl (4 : 0), (0) hexanoyl (6: 0) and (A) octanoyl (8:O) ACPs were obtained from densitometric scans of an autoradiogram. Inset: autoradiogram of nitrocellulose blot of acylacps. Label in acetylcoa (A) was obtained directly from liquid scintillation counting of HPLC fractions as described in Experimental Procedures. These chase experiments were repeated three times and the data for the figure is taken from a single representative experiment. other intermediates were diluted with the unlabeled acetate at least 1 0fold faster than the acetylacp. As in the previous experiment, it could be calculated that if butyrylacp resulted primarily from the condensation of malonylacp and acetylcoa by KAS 111, its specific radioactivity would decrease at approximately the same rate as the acetylcoa. If, however, the butyrylacp resulted from the condensation of malonylacp and acetylacp by KAS I, the specific radioactivity of the butyrylacp would decrease at a rate similar to the acetylacp specific radioactivity. Because the rate of decrease in the I4C in butyrylacp was essentially the same as the rate of decrease of acetylcoa label and much more rapid than that of acetylacp, these data provide further in situ evidence that the butyrylacp was primarily produced from malonylacp and acetylcoa and that the initial condensation of fatty acid synthesis is catalyzed by KAS I11 rather than KAS I. Why does acetylacp accumulate in the dark? The data presented above indicate that acetylacp is used very slowly relative to acetylcoa as a primer for fatty acid synthesis. However, we recently reported that the pool of acetylacp changes markedly when spinach plants are

5 985 Fig. 5. Decrease in acetylacp levels in spinach leaf when shifted from dark to light. Analysis of spinach acylacps from leaves of plants that were in the dark overnight before turning on the lights at zero time. Samples were obtained at the indicated time and acyl ACPs separated by 1.3 M urealpage, and blotted to nitrocellulose for immunoblot development. Data are presented for isoform I1 instead of isoform I because, when using 1.3 M uredpage, a better separation of acetylacp (2 : OACP) from ACP was obtained for isoform 11 than for isoform I. The relative levels of the individual acylacp isoforms I and I1 in both spinach (PostBeittenmiller et al., 1991) and spinach chloroplasts (PostBeittenmiller et al., 1992) are the same. shifted from light to dark (PostBeittenmiller et al., 1991). AcetylACP levels increase 45fold in the dark and become the major form of acylacp. When the spinach plants are transferred again to light, the acetylacp pool decreases after 3 5 min to levels found in lightgrown leaves (Fig. 5). During this period, the rate of fatty acid synthesis increases approximately sixfold (Browse et al., 1981). Although these data suggest that acetylacp pools are substantial and are dynamic, it should also be noted that several minutes are required to reduce the acetylacp pool (Fig. 5), compared to the much more rapid metabolism observed for the other acyl ACPs and acetylcoa (Figs 3 and 4). What metabolic changes lead to, first, the accumulation of acetylacp in the dark and then its decrease in the light? Recently it was reported that the concentrations of acetyl CoA do not change substantially between light and darkincubated chloroplasts (PostBeittenmiller et al., 1992). Thus, the observed lightdependent changes in the acetylacp levels are probably not related simply to a mass action response to fluctuations in the levels of acetylcoa. Other metabolic changes that could lead to an increase in acetyl ACP levels in the dark are either (a) an increased rate of synthesis of acetylacp by acetyl transacylase or decarboxylation of malonylacp or (b) a decreased rate of utilization of acetylacp by KAS I. The possibility that acetylacp levels may have risen in the dark because of increased rates of decarboxylation of malonylacp was considered because synthesis of acetyl ACP from malonylacp has been suggested to occur in E. coli (Tsay et al., 1992b). However, as shown in Fig. 6, when chloroplasts are transferred to the dark, malonylacp concentrations drop to very low levels in less than 10 s, whereas acetylacp levels require several minutes to rise. Thus it is unlikely that malonylacp decarboxylation could be serving as a source of acetylacp in the dark. Could an increase in the acetyl transacylase activity in the dark contribute to increases in acetylacp levels. It has been suggested that the acetyl transacylase reaction is actually catalyzed by KAS I11 in vivo (Jackowski et al., 1989). Furthermore, analysis of the purified enzyme from E. coei (Tsay et al., 1992a) and spinach (Clough et al., 1992) have demonstrated that KAS 111 can catalyze the acetyl transacylase reaction. If the majority of the acetyl transacylase reaction was catalyzed in vivo by KAS 111, it might be expected that this reaction would be more favored in the absence of Fig. 6. Disappearance of malonylacp from chloroplasts shifted from the light to the dark. Chloroplasts were preincubated with r4c]acetate in the light for 2 min before shifting to the dark for the indicated time. Products were separated by native PAGE and blotted to nitrocellulose as described in Experimental Procedures. (A) Autoradiogram of blot. (B) Immunoblot development of a duplicate of the gel used to obtain autoradiogram. The dark band on the immunoblot between malonylacp and acetylacp is unesterified ACP, i.e. ACPSH. It does not appear in the autoradiogram because it lacks a radiolabeled acyl group. 2 : 0, acetyl; 4 : 0, butyryl. malonylacp. However, when we compared the in vivo rate of acetylacp production from [14C]acetate in light and darkincubated chloroplasts, the initial rates of incorporation of label were not significantly different under the two conditions (data not shown). Considering that there is a sixfold decrease in the rate of fatty acid synthesis in dark versus light, we conclude that there was a comparatively minor change in the in vivo acetyl transacylase activity and it was insufficient to explain the large increase in acetylacp levels. The level of acetylacp would also have risen if there was a decrease in the rate of its utilization by KAS I. We had proposed previously (PostBeittenmiller et al., 1991) that the decrease in the rate of fatty acid synthesis in the dark was due to a decrease in the activity of acetylcoa carboxylase, and the very rapid loss of malonylacp observed in chloroplasts incubated in the dark further supports that proposal (Fig. 6). In less than 10 s after the shift to the dark, the malonylacp went from being maximally labeled to being essentially undetectable. Furthermore, immunoblot analysis of these samples verified a large decrease in the concentration of malonylacp. Although acetylacp is a poor substrate for KAS I (Fig. 2), its utilization is probably through this reaction. Since malonylacp is also required for KAS I activity, the decrease in malonylacp level would lead to a comparable decrease in the activity of KAS I. Thus, a decrease in the activity of KAS I in the dark, leading to a decrease in the utilization of acetylacp, provides the most likely explanation for the increase in the steadystate levels of acetylacp in the dark. In summary, we have evaluated in vitro and in isolated chloroplasts the role of acetylacp in the initial reactions of fatty acid synthesis. We have concluded: (a) the major primer of fatty acid biosynthesis in spinach is acetylcoa, with acetylacp playing only a minor role as a substrate for fatty acid biosynthesis (Fig. 1) ; (b) the changes in the levels of acetyl ACP which occur in the light and dark are due to changes in the slow rate of utilization of acetylacp by KASI which in turn is modulated by the levels of malonylacp.

6 986 We thank Kunimitsu Nakahira, Paul Roessler, and Grattan Roughan for critical discussions during the preparation of the manuscript. This work was supported by National Science Foundation Grant DCB (to JGJ) and National Science Foundation Grant DCB (to JBO and DPB). REFERENCES Browse, J., Roughan, P. & Slack, C. (1981) Biochem. J. 196, Clough, R. C., Barnum, S. R. & Jaworski, J. G. (1989) Anal. Biochem. 176, Clough, R. C., Matthis, A. L., Barnum, S. R. & Jaworski, J. G. (1992) J. Biol. Chem. 267, Corkey, B. A. (1988) Methods Enzymof. 166, Cronan, J. & Klages, A. (1981) Proc. Nud Acad. Sci. USA 78, Guerra, D. J., Ohlrogge, J. B. & Frentzen, M. (1986) Plant Physiol. 82, Jackowski, S., Murphy, C. M., Cronan, J. E. & Rock, C. 0. (1989) J. Biol. Chem. 264, Jackowski, S. & Rock, C. 0. (1987) J. Biol. Chem. 262, Jaworski, J. & Stumpf, P. (1974) Arch. Biochem. Biophys. 162, Jaworski, J. G., Clough, R. C. & Barnum, S. R. (1989) Plant Physiol. 90, Kauppinen, S., SiggaardAnderson, M. & von WettsteinKnowles, P. (1988) Carlsberg Res. Commun. 53, PostBeittenmiller, D., Jaworski, J. G. & Ohlrogge, J. B. (1991) J. Biol. Chem. 266, PostBeittenmiller, D., Roughan, G. & Ohlrogge, J. B. (1992) Plant Physiol 100, Rock, C. & Cronan, J. (1979) J. Biol. Chem. 254, Roughan, P. (1987) Methods Eniymol. 148, Roughan, P. G., Holland, R. & Slack, C. R. (1979) Biochem. J. 184, Rutkoski, A. & Jaworski, J. (1978) Anal. Biochem. 91, Shimakata, T. & Stumpf, P. K. (1982) Proc. Nut1 Acad. Sci. USA 79, Shimakata, T. & Sturnpf, P. K. (1983~) Arch. Biochem. Biophys. 220,3945. Shimakata, T. & Stumpf, P. K. (1983b) J. Biol. Chem. 258, Stumpf, P. K. (1987) in The biochemistry ofplmts (Stumpf, P. K. & Conn, E. E., eds) vol. 9, pp , Academic Press, New York. Tsay, J., Oh, W., Larson, T. J., Jackowski, S. & Rock, C. 0. (1992a) J. Biol. Chem. 267, Tsay, J. T., Rock, C. 0. & Jackowski, S. (1992b) J. Bucteriol. 174, Volpe, J. J. & Vagelos, P. R. (1973) Annu. Rev. Biochem. 42, Walsh, M. C., Klopfenstein, W. E. & Harwood, J. L. (1990) Phytochemistry 29, Supplement material to: Acetylacyl carrier protein is not a major intermediate in fatty acid biosynthesis in plants Jan G. JAWORSKI', Dusty POSTBEITTENMILLER' and John B. OHLROGGE' * Chemistry Department, Miami University, Oxford OH, USA Department of Botany and Plant Pathology, Michigan State University, East Lansing MI, USA rl I Vrnct*m I Algebraic model of the initial reactions of fatty acid synthesis An algebraic model was developed that could predict the change in the specific activity of the intermediates of fatty acid synthesis, depending on the relative rates of KAS 111, KAS I and the forward and reverse reaction of acetylcoa: ACP transacylase. Equations used to calculate the changing specific radioactivity of pools of acetylcoa, acetylacp, malonylacp and butyrylacp are given below and are consistent with the corresponding scheme for fatty acid synthesis (Fig. 7). SR, and SR, refer to the specific radioactivity of the indicated intermediate at the beginning and end of each calculation. b', V, V,,,, and V, refer to in situ rates of acetylcoa carboxylase, acetylcoa:acp transacylase, 3oxoacylACP synthase, and malonylcoa:acp transacylase, expressed as pmov 0.01 s. Calculations were executed and graphed using a spreadsheet (Excel v3.0, Microsoft Corp.). With iterations of these calculations, graphs mimicking Figs 3 and 4 were obtained. The initial specific activities used were zero except for acetate, which was 56 Cum01 and the values of the reaction rates used are given below. 2 : 0, acetyl ; 4 : 0, butyryl ; mal, malonyl. Equations %(z o cc,a) [mob o c<,a) X SR, (2 o CoA) + V,, X sr(ac,,a,e) + Kc, (rev) x SR(2 o ACP) (Kcc + Kct f Vk,, 11,) x sr(z o C~A)]/ mo4, o Co~) W ( z o ACP) [mol,z o ACP) X SR, (2 o AcP) + V,t X SRv o coa) vact (rev) x sr(z o ALP) vka, I x sr(z o ~ce)l/mob o ACP) U I 6:OACP Fig.7. Schematic flow chart of fatty acid biosynthesis used to develop the algebraic model predicting the level of radioactivity in pools of FAS intermediates. Assumptions It was assumed that all pools of intermediates were at steadystate levels Consequently, the rate of acetylcoa carboxylase was equal to the rate of malonylcoa :ACP transacylase. These rates were sevenfold higher than the rate of butyrylacp synthesis to account for the need of one malonylacp for each cycle of fatty

7 A '"t KASl 1 B KAs.lu1 KASl n x) 40 Tlmo (r) Time (s) Fig. 8. Algebraic model of the labeling experiment illustrated in Fig. 3. Values shown are from the calculations that predicted the changing level of radioactivity in pools of fatty acid synthase intermediates in isolated spinach chloroplasts following addition of [I4C]acetate. (A) The ratio of KAS IIYKAS I activity = 50: 1. (B) The ratio of KAS IIYKAS I activity = 1 : 1. acid synthesis leading to palmitoylacp synthesis. The sum of the rates of KAS I11 plus KAS I was equal to the rate of butyrylacp synthesis and was one seventh the rate of MCT. Since there is a steadystate level of acetylacp, the rate of acetyl transacylase (forward) minus acetyl transacylase (reverse) was set equal to the rate of KAS I. The rate of acetylcoa synthesis from acetate was 2.2 pmol (0.01 s)' mg chlorophyll' (calculated from an average empirical value of 800nmol. h'. mg chlorophyll') and was 8/7 the rate of acetylcoa carboxylase. The endogenous levels of acetylcoa, acetylacp, malonylacp, and butyrylacp used in the model were 380, 16, 8, and 4 pmollmg chlorophyll, respectively. RESULTS Both in the case of the labeling of acetylcoa and acylacp pools with [14C]acetate (Fig. 8) and the nonradioactive chase of the c s s I 8 nme (s) Time (s) Fig.9. Algebraic model of the chase experiment illustrated in Fig.4. Values shown are from the calculations that predicted the changing level of radioactivity in pools of fatty acid synthase intermediates in isolated spinach chloroplasts following unlabeled acetate chase of chloroplasts prelabeled with ['4C]acetate. (A) The ratio of KAS IIYKAS I activity = 50: 1. (B) The ratio of KAS IIYKAS I activity = 1 : 1. c u acetate (Fig. 9), the algebraic model closely approximated the empirical data when using KAS IIVKAS I ratios greater than 40: 1. In contrast, when a KAS III/KAS I ratio equal to 1 : 1 was used, the model predicted that the acetylacp would be labeled nearly as fast as the butyrylacp. Thus, these calculations supported the conclusion that there is only a minor contribution to fatty acid synthesis in spinach by KAS I catalysis of the condensation of acetylacp with malonylacp.