Preparation and Characterization of Membrane Fractions Enriched in Outer and Inner Envelope Membranes from Spinach Chloroplasts

J. Biol. Chem. we have described a method for the separation of membrane fractions enriched in outer and inner envelope membranes from spinach chloro- plasts. The two envelope membranes have a different weight ratio of acyl lipid to protein (2.5-3 for the outer envelope membrane and 0.8-1 for the inner envelope membrane). The two membranes also differ in their polar lipid composition. However, in order to prevent the functioning of the galacto1ipid:galactolipid galac- tosyltransferase during the course of envelope membrane separation, we have analyzed the polar lipid composition of each envelope membrane after ther- molysin treatment of the intact chloroplasts. The outer envelope membrane is characterized by the presence of high amounts of phosphatidylcholine and digalacto-syldiacylglycerol


Preparation and Characterization of Membrane Fractions Enriched in
Outer and Inner Envelope Membranes from Spinach Chloroplasts 11. BIOCHEMICAL CHARACTERIZATION* (Received for publication, May 26, 1983) Maryse Anne Block, Albert-Jean Dorne, Jacques Joyard, and Roland Douce 13273-13280), we have described a method for the separation of membrane fractions enriched in outer and inner envelope membranes from spinach chloroplasts. The two envelope membranes have a different weight ratio of acyl lipid to protein (2.5-3 for the outer envelope membrane and 0.8-1 for the inner envelope membrane). The two membranes also differ in their polar lipid composition. However, in order to prevent the functioning of the galacto1ipid:galactolipid galactosyltransferase during the course of envelope membrane separation, we have analyzed the polar lipid composition of each envelope membrane after thermolysin treatment of the intact chloroplasts. The outer envelope membrane is characterized by the presence of high amounts of phosphatidylcholine and digalactosyldiacylglycerol whereas the inner envelope membrane has a polar lipid composition almost identical with that of the thykaloids.
No phosphatidylethanolamine or cardiolipin could be detected in either envelope membranes, thus demonstrating that the envelope membranes, and especially the outer membrane, do not resemble extrachloroplastic membranes. No striking differences were found in the fatty acid composition of the polar lipids from either the outer or the inner envelope membrane. The two envelope membranes also differ in their carotenoid composition.
Among the different enzymatic activities associated with the chloroplast envelope, we have shown that the Mg"+-dependent ATPase, the UDP-Ga1:diacylglycerol galactosyltransferase, the phosphatidic acid phosphatase, and the acyl-CoA thioesterase are associated with the inner envelope from spinach chloroplasts whereas the acyl-CoA synthetase is located on the outer envelope membrane.
In the preceding paper (l), we have described a method for the separation of two membrane fractions enriched in outer and inner envelope membranes from spinach chloroplasts. The identification of the two membrane fractions was based on electrophoretic and immunochemical studies and was made possible by the recent characterization of two envelope polypeptides, El0 and E24, as specific markers for the outer envelope membrane ( 2 ) .
* The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore he hereby marked "adurrtisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

______
This article is dedicated to Professor Andrew A. Benson.
To characterize further the two membrane fractions obtained, we analyzed the polar lipid composition of each membrane fraction. However, we have demonstrated that a galac-to1ipid:galactolipid galactosyltransferase is associated with the outer envelope membrane (3). This enzyme, which catalyzes an interlipid exchange of galactose (4), becomes active during the course of envelope membranes purification (3) and catalyzes, by its action on MGDG' molecules, the rapid formation of diacylglycerol and unnatural galactolipids such as tri-GDG and tetra-GDG (3,5 ) . Therefore, in order to obtain a polar lipid composition which could represent the in uiuo situation, we analyzed the polar lipid composition of' each envelope membrane after thermolysin treatment of intact chloroplasts. We also analyzed the fatty acid and pigment composition or each envelope membranes. Finally, we have determined the localization within envelope membranes of enzymatic activities known to be associated with envelope membranes (for reviews, see Refs. 6 and 7). The results of these experiments are reported in this article.

Outer and Inner Chloroplast Envelope Membranes I1
Envelope merrbranes were p u r i f i e d f r o m t h e l y s a t e s ( s w o l l e n c h l o r o p l a r t s ) by centrifugation through a step sucrose gradient as described by D O U C~ e t a l . (8.4).
PPepaPation Of

Polar Lipid Composition of Outer and Inner Envelope Membranes-
We have previously demonstrated that the galactolipidgalactolipid galactosyltransferase, localized on the outer envelope membrane, becomes active during the course of envelope membrane preparation (3). This enzyme catalyzes the interlipid exchange of galactose between galactolipid molecules and induces the formation of unnatural galactolipids such as tri-GDG and tetra-GDG and diacylglycerol at the expenses of MGDG (4). Consequently, in order to obtain a polar lipid composition of each envelope membrane which could represent the in vivo situation, the galactolipid:galactolipid galactosyltransferase must be destroyed prior to fractionation of isolated intact chloroplasts. This was achieved by thermolysin treatment of intact chloroplasts, as demonstrated by Dorne et al. (3,5).
The polar lipid composition of outer and inner envelope membranes is given in Table I and is compared with that of whole envelope and thylakoids prepared according to Douce et al. (8). The outer envelope membrane has an unusually high lipid to protein ratio of 2.5-3 when compared with that of the inner membrane (0.8-1) or of thylakoids (0.4-0.5) ( Table I). The lipid to protein ratio value measured in whole envelope fraction is an average of those obtained with each envelope membrane: 1.5-2. Quantitatively, the lipids present in each fraction are identical, but the proportions in which they are found are different (Table I). In the outer envelope membrane, glycolipids, (galactolipids and sulfolipid) ac-  counted for 52% of the total lipids, a value which is much lower than in the inner membrane (84%) or in the thylakoids (84.5%). Again, the value obtained in the whole envelope is an average value (75%) of that in the outer and the inner envelope membrane. The two galactolipids (MGDG and DGDG) are present in the ratio 0.6:l in the outer membrane, a much lower value than those measured in the inner membrane (1.6:1), whole envelope (1.2:1), or thylakoids (2:l). No tri-GDG, tetra-GDG or diacylglycerol could be detected in the different membrane fractions, except when non-thermolysintreated chloroplasts were used. In this case, the MGDG to DGDG ratio in outer envelope membrane was 0.08:l instead of 0.6, thus demonstrating the necessity to destroy the galactolipidgalactolipid galactosyltransferase to prevent the interlipid exchange of galactose during the experiment. In agreement with previous observations (5,18), Table I demonstrates that the outer membrane contains large amounts of PC, which is the major phospholipid in this membrane. There is much less phospholipid in the inner envelope membrane (16% of the total lipids, compared to 48% in the outer membrane) where phosphatidylglycerol is the major phospholipid ( Table  I). Table I also demonstrates that the inner envelope membrane has a polar lipid composition close to that found in thylakoids (see also Refs. 5 and 18). The absence of PE in chloroplasts has been a somewhat controversial issue. Because this phospholipid is the main component of extrachloroplast membranes, it is sometimes difficult to be certain that its presence in chloroplast membrane fractions is not merely due to contamination. However, no PE and no cardiolipin could be detected in either outer or inner envelope membranes, thus indicating the absence of contaminating extrachloroplastic membranes.
The fatty acid composition in polar lipids has been determined for each membrane fraction (Table 11). The total fatty Outer and Inner Chloroplast Envelope Membranes II

Polar lipid composition of the different membranes isolated from spinach chloroplasts
Thermolysin-treated intact, purified spinach chloroplasts (see "Materials and Methods") were used for these experiments. This treatment destroys the galacto1ipid:galactolipid galactosyltransferase localized on the outer surface of the outer envelope membrane (3,5) and prevents the formation of unnatural galactolipids such as tri-GDG and tetra-GDG; therefore, this treatment allows the determination of the polar lipid distribution in chloroplast membranes which represents the in uiuo situation. The different membrane fractions were prepared according to Douce et al. (8,9) for whole envelope and thylakoids and according to Block et al. (1) (see Fig. 1) for the membrane fractions enriched in outer and inner envelope membranes. Polar lipids were extracted and analyzed as described under "Analyses of Envelope Lipids." The values obtained were corrected for cross-contamination of fractions 2 and 3, as estimated from data presented in Fig. 5 from Block et al. (1). This composition is from a representative experiment and has been reproduced at least three times. The values are expressed as weight per cent fatty acids.    acids are more saturated in the outer envelope membrane that in the other chloroplast membranes. The major differences observed between the fractions enriched in outer and inner envelope membranes are the decrease in the proportions of linolenic acid and the increase in the proportions of linoleic and palmitic acids in the fraction enriched in the outer envelope membrane. However, when each polar lipid is taken separately, the increase in the proportions of linoleic acid appears to be due to the increase in the amount of PC in the outer envelope membrane. Likewise, the decrease in the proportions of linolenic acid appears to be due to the decrease in the amount of MGDG in the outer envelope membrane. As a matt,er of fact, the fatty acids were more saturated in PC than in galactolipids (especially in MGDG) and no striking differences could be detected in each individual polar lipids from both envelope membranes (Table 11).

TABLE I11
Carotenoid composition of membrane fractions enriched in outer and inner envelope membranes isolated from spinach chloroplasts The different membrane fractions analyzed were prepared according to the procedure described in Fig. 1 Carotenoid Analyses of Outer and Inner Envelope Membranes-The carotenoid composition of each membrane fraction is shown in Table 111. In both fractions, violaxanthin accounted for more than 50% of the total carotenoids. However, the proportions in which the carotenoids were present in each fraction are different (Table 111). For instance, the fraction enriched in outer envelope membrane (fraction 2 or light fraction) was characterized by a higher proportion of neoxanthin than the fraction enriched in inner envelope membrane (fraction 3 or heavy fraction) (Table 111). Finally, the amount of carotenoids was 3 times lower in the outer membrane fraction than in the inner one, on a protein basis, and this quantitative difference was even more pronounced on a lipid basis.
Distribution of Enzymatic Activities Associated with Envelope Membranes within Outer and Inner Envelope Membranes-We have previously characterized up to 20 different enzymatic activities associated with chloroplast envelope membranes (6, 7 ) . Therefore, we analyzed some of them in the fractions enriched in outer and inner envelope membranes. However, since the outer envelope membrane represents a small percentage of the total envelope protein (about 30 to 35%), even small amount of contamination by inner envelope membrane may be of major significance, especially in experiments designed for enzyme localization.

TABLE IV Distribution of enzymatic activities associated u i t h t h e chloroplast envelope in membrane fractions enriched in outer and inner envelope
membranes Membrane fractions enriched in outer (fraction 2 ) and inner (fraction 3 ) envelope membranes were prepared as described in fig. 1 according to Block et al. (1). The different enzymatic activities were measured as described under "Enzymatic Assays." Kinetic experiments were used to determine the initial velocity of the reactions. The values shown in this table are from representative experiments and have been reproduced a t least three times. OM, outer membrane; IM, inner membrane. It is interesting to compare the values of the ratio "Activity in Fraction Z/Activity in Fraction 3" (A2/A3) with the values ofthe ratio area of the rockets in fraction 2/area of the rocket in fraction 3" (S2/S3j obtained by Block et al. (1) in the immunoelectrophoresis experiments presented in Fig. 5 from Ref. 1. When the immunoelectrophoresis was done using antibE24, an antibody raised against outer envelope polypeptide E24, the ratio S2/S3 was 4.57, a value which can be compared with that of the ratio A2/A3 for acyl-CoA synthetase. When the immunoelectrophoresis was done using antikE37, an antibody raised against inner envelope polypeptide E37 ( l ) , the ratio S2/S3 was 0.13, a value which can be compared with those of the ratios A2/A3 with UDP-Gal:diacylglycerol galactosyltransferase, acyl-CoA thioesterase, phosphatidic acid phosphatase, and, to a lesser extent, with Mg"-dependent ATPase.

4.33
Mg''-dependent ATPase which has been characterized by Douce et al. (8) and Joyard and Douce (19). This activity is not attributable to coupling factor because it is insensitive to dicyclohexylcarbodiimide and antisera raised against coupling factor 1 (8). As shown in Table IV, this activity seems to be associated with the membrane fraction enriched in inner envelope membrane, with a specific activity 2.5-fold over that of the membrane fraction enriched in outer membrane, on a protein basis.
We analyzed also the activity of several enzymes involved in the synthesis of galactolipids, the major chloroplast polar lipids. Among these enzymes, phosphatidic acid phosphatase is a key enzyme: this alkaline membrane-bound enzyme (14, 20) is responsible for the synthesis of the diacylglycerol backbone of the main chloroplast polar lipids (21). The diacylglycerol thus formed is then galactosylated by a UDP-Gal& acylglycerol galactosyltransferase, which is probably the best enzymatic marker for envelope membranes (22). Table IV clearly indicates that these two enzymes are both clearly associated with fraction 3 and therefore are probably localized on the inner envelope membrane.
Two other envelope enzymatic activities have also been measured in the two membrane fractions: the acyl-CoA synthetase (15,20,23) and the acyl-CoA thioesterase (16), which could play an important role in the transport of oleic acid and palmitic acid from their site of synthesis, the chloroplast stroma (24), to the cytosol. Indeed, this hypothesis is strengthened by our observation that the acyl-CoA synthetase and the acyl-CoA thioesterase are localized in the membrane fractions respectively enriched in the outer and in the inner envelope membranes (Table IV). DISCUSSION This study is the first demonstration that the outer envelope membrane has a very high lipid to protein ratio (2.5-3), a value which is almost 3 times higher than that of the inner envelope membrane (0.8-1) and 6-8 times higher than that of the thylakoids (0.4-0.5). These observations are consistent with freeze-cleavage studies of outer and inner envelope membranes: the density of the particles displayed on the outer envelope membrane is much lower than on the inner membrane (25,26), thus suggesting that the inner envelope membrane contains more protein that the outer. The differences in the lipid to protein ratio observed in our two fractions explain also the differences in the densities of each membrane fraction (1.08 and 1.13 g/crn'? for fractions 2 and 3, respectively).
Both envelope membranes contain galactolipids and sulfolipid; our results are in agreement with those of Cline et al. (18). All these observations provide further support for early experiments by Billecocq et al. (27) and Billecocq (28,29) which have demonstrated with specific antibodies that the outer leaflet of the outer envelope membrane contains both galactolipids and sulfolipid. The same result was obtained by cytochemical analyses of thin sections of spinach leaves by Carde et al. (30). A characteristic feature of the outer envelope is the presence of large amounts of PC (5, 18). However, we always found that the outer envelope membrane contains more glycolipids (galactolipids and sulfolipid) than phospholipids whereas it is the reverse in the study by Cline et al. (18). This difference can be easily explained by the presence, on the outer envelope membrane, of the galactolipidgalactolipid galactosyltransferase as demonstrated by Dorne et al. (3,5). As shown by Van Besouw and Wintermans (4), this enzyme catalyzes the interlipid exchange of galactose between galactolipids. The very low level of MGDG, the presence of tri-GDG and tetra-GDG in all the membrane fractions from pea reported by Cline et al. (18), as well as in spinach envelopes prepared from non-thermolysin-treated spinach chloroplasts (3) reflect the functioning of the galactolipidgalactolipid galactosyltransferase during membrane preparation. Therefore, this enzyme activity should be destroyed before the rupture of the chloroplast envelope membranes in order to obtain a polar lipid composition close to the situation in vivo (3): the total amount of glycolipids in the study by Cline et al. (18) was underestimated because diacylglycerol (which is formed from MGDG) was not analyzed. Indeed, it was shown by Cline and Keegstra (31) that their two membrane fractions contained between 5 and 6% diacylglycerol; therefore, one can suggest that in vivo the two membrane fractions obtained by Cline et al. (18) have a polar lipid composition where the amount of glycolipids is higher than the amount of phospholipids, close to that presented in this study where the interlipid exchange of galactose between galactolipids was prevented by protease treatment of intact chloroplasts by thermolysin.
Analyses of the outer and inner envelope membranes from thermolysin-treated spinach chloroplasts demonstrate that, qualitatively, the polar lipids of both types of membranes are identical but the proportions in which they are present are different (Table I) (5,18). In the outer envelope membrane, Outer and Inner Chloroplast Envelope Membranes 11 13285 the major galactolipid is DGDG, whereas it is MGDG in the inner membrane. In the outer envelope membrane, the major phospholipid is PC, whereas it is PG in the inner membrane.
In addition, membranes show asymmetry in the distribution of the polar lipids; therefore, the problem is made more complicated since we deal with four bilayer halves which probably have their own lipid composition. For instance, it is possible that, on the outer envelope membrane, MGDG could be located solely on the inner leaflet, in a position which is not accessible to the galacto1ipid:galactolipid galactosyltransferase which is located on the outer surface of this membrane (3). In agreement with Cline et al. (181, our data suggest that the inner envelope membrane has a polar lipid composition highly comparable to thylakoids. Interestingly, in the bluegreen alga Anacystis nidulans, thylakoids and the cell envelope have also the same polar lipid composition (32). In chloroplasts, such a situation will strongly favor the fusion of vesicles produced by the inner envelope membrane with growing thylakoids; see, for instance, electron microscopic studies by Carde et d . (30). However, our demonstration that envelope polypeptides are completely different from those of thylakoids (33) strongly suggests that considerable modification must take place following the init.ia1 steps of invagination.
Except for PC molecules, the polar lipid composition of envelope membranes almost resembles that of blue-green algae (32). The endosymbiotic theory proposes that a primitive eukaryote took up blue-green algae or Prochloron to yield chloroplasts. Then, the evolution of organelles from endosymbiotic precursors would involve their integration with the host cell structurally, numerically, and biochemically. The plasma membrane of the prokaryotic symbiont is shown as homologous with the inner membrane of the chloroplast envelope, whereas the vacuolar membrane of the host cell is shown as homologous with the outer chloroplast envelope membrane (34,35). However, the chemical composition of the membranes surrounding the chloroplasts causes difficulty in the interpretation of their phylogeny. In fact, it is difficult to understand why the outer membrane of the chloroplast envelope contains large amounts of typical blue-green algae lipids such as galactolipids and sulfolipid and does not resemble in its chemical composition the extrachloroplastic membranes, which contain large amounts of phosphatidylethanolamine, a phospholipid absent from all chloroplast membranes. However, we must keep in mind that Gram-negative bacteria as well as blue-green algae also contain a membrane external to the cytoplasmic membrane which, although morphologically similar to the cytoplasmic membrane, contains fewer and different proteins (such as the pore protein that allows the diffusion of low molecular weight molecules; Ref. 36). Consequently, it is possible that the outer envelope membrane is homologous with the outer membrane of the prokaryotic symbiont and not with the vacuolar membrane of the host cell.
The distribution of envelope enzymatic activities on each envelope membrane is interesting. The values of the ratio of the activities obtained in fraction 2 (enriched in outer envelope) to the activities obtained in fraction 3 (enriched in inner envelope) can be compared with data presented by Block et 01. (1) using quantitative immunoelectrophoresis (see Table   IV). So far, we have demonstrated that two enzymatic activities, it.. the galacto1ipid:galactolipid galactosyltransferase and the acyl-CoA synthetase, are located in the membrane fraction enriched in outer envelope membrane whereas the UDP-Ga1:diacylglycerol galactosyltransferase and the acyl-CoA synthetase (and probably the Mg2"dependent ATPase) are concentrated in the membrane fraction enriched in inner envelope membrane (3,5,37). However, there is a striking difference between our data (Table IV) and those of Keegstra's group (18, 31): these authors concluded from their experiments that the UDP-Gakdiacylglycerol galactosyltransferase is located on the outer envelope membrane of pea chloroplasts. Such a localization is surprising since: 1) fatty acids used for galactolipid synthesis are formed in the stroma of chloroplasts (24); 2) acyl carrier proteins are probably the in vivo acyl donor for galactolipids (38, 39); 3) phosphatidic acid phosphatase, which synthesizes diacylglycerol, the in vivo substrate for galactosylation enzymes, is located on the inner envelope membrane (Table IV). Since pea is an "18:3" plant and spinach a "16:3" plant (see, for instance, Heinz (40) and Douce and Joyard (21)) it is possible that the difference observed could be due to species specificities. For instance, Heinz and Roughan (41) and Gardiner and Roughan (42) have shown that the enzymes catalyzing the conversion of phosphatidic to diacylglycerol and diacylglycerol to MGDG are significantly less active in chloroplasts from 18:3 plants than from 16:3 plants. Therefore, it is possible that the biosynthetic pathways for galactolipids are quite different in these two types of plants (41)(42)(43). Under these conditions, spinach should have two pathways for galactolipid synthesis (one for 16:3 galactolipids and one for 18:3 galactolipids), whereas pea should have only one (for 18:3 galactolipids, since this species lacks 16:3 galactolipids). Unfortunately, Keegstra" has obtained the same data with pea and spinach; therefore, it is clear that further work has to be done in order to solve this problem.
The localization of acyl-CoA synthetase and acyl-CoA thioesterase on the outer and inner envelope membrane, respectively, of spinach chloroplasts has major implications in glycerolipid metabolism in leaf cells. As discussed by Block et al. (37), the localization of acyl-CoA synthetase on the outer envelope membrane together with that of the phosphatidic acid phosphatase on the inner membrane preclude the use of acyl-CoA as in vivo fatty acid donor for galactolipid synthesis.
In fact, recently, Frentzen et al. (38, 39) have clearly demonstrated that phosphatidic acid is synthesized within the chloroplast envelope by acylation of sn-glycerol3-phosphate with acyl-primed acyl carrier protein. The presence of acyl-CoA synthetase on the outer envelope membrane supports the view that this enzyme could be involved in fatty acid export outside the chloroplast (Refs. 15 and 44; see also 37).
Finally, extensive genetic studies on the envelope proteins will provide invaluable information for elucidating their functions, the mechanisms of their biosynthesis and assembly, the mechanisms of regulation of their production, and the structural-functional relationship between them.

Outer and Inner
Chloroplast Envelope Membranes II