- Email: [email protected]
Membrane lipids: It’s only a phase
bb10j04.qxd
10/5/00
5:07 pm
Page R377
Dispatch
R377
Membrane lipids: It’s only a phase Anthony G. Lee
All biological membranes contain lipids that prefer to adopt a non-bilayer phase. Recent results suggest that, in the thylakoid membrane, membrane proteins force all the lipids to adopt a bilayer structure, and that the non-bilayer-forming lipids in the thylakoid membrane serve to drive the formation of membrane stacks. Address: Division of Biochemistry and Molecular Biology, School of Biological Sciences, University of Southampton, Southampton SO16 7PX, UK. Current Biology 2000, 10:R377–R380 0960-9822/00/$ – see front matter © 2000 Elsevier Science Ltd. All rights reserved.
It is now 75 years since Gorter and Grendel [1] published their proposal that the lipids in a biological membrane are organised as a bilayer. The beauty of the bilayer structure is that its formation is self-driven; the bilayer forms because the hydrophobic fatty acyl chains of the lipids want to be out of contact with water, but the polar headgroups want to be in water. There can be no doubt that the bilayer model for the membrane is correct, but there remains a major puzzle — many of the lipids found in biological membranes are now known not to form bilayers, at least when on their own. A recent paper by Simidjiev et al. [2] has considered the importance of non-bilayer forming lipids for the proper functioning of one particular membrane, the thylakoid membrane. The internal photosynthetic membrane — the thylakoid membrane — of higher plant chloroplasts is the most abundant membrane found in nature. The membrane has a lipid composition unlike that of any other membrane (Figure 1). About 50% of the total lipid content is monogalactosyldiacylglycerol (MGDG), about 30% is digalactosyldiacylglycerol (DGDG) and about 10% is diacylphosphatidylglycerol (PG) [3]. As well as having unusual headgroups, MGDG and DGDG are unusual in containing two highly unsaturated fatty acyl chains (linoleic acid); most natural lipids contain one saturated and one unsaturated chain. The unusual chemical structures of MGDG and DGDG give them unusual packing properties. The ways that lipids pack together depends on their shape. A simple phospholipid, such as a phosphatidylcholine has a cylindrical shape; the cross-sectional area of the phosphocholine headgroup is about the same as that of two fatty acyl chains (Figure 2). Cylinders are ideally suited to packing side by side in a bilayer; however, not all lipids have a cylindrical shape. The headgroup of MGDG is quite large, and so MGDG
might be expected to be cylindrical and pack as a bilayer and, indeed, this is what is found when the chains are saturated [4]. But the two highly unsaturated chains in native MGDG give the chains a much larger than normal crosssectional area; the shape of native MGDG is therefore that of a cone (Figure 2). Cones do not pack well into planar arrays and prefer to adopt curved structures. In particular, MGDG has a tendency to pack into the so-called hexagonal HII structure shown in Figure 2. The structure gets its name because it consists of rod-like aggregates of infinite length packed into hexagonal arrays. The polar lipid headgroups are located at the centre of the rods along with some water, and the fatty acyl chains point outwards so that the surfaces of the rods are hydrophobic; this arrangement is called the HII hexagonal phase, to distinguish it from an arrangement with the hydrophobic fatty acyl chains pointing in towards the centre of the rods with the polar headgroups on the outside — this is the hexagonal HI phase. It is obvious that a biological membrane cannot be built from lipids that are in the hexagonal HII phase. So why has nature chosen to have a non-bilayer forming lipid as the major lipid in the thylakoid membrane and how is this lipid prevented from forming the hexagonal HII phase that it prefers? The recent work of Simidjiev et al. [2] shows that the answers to both these questions involve interactions Figure 1 O O
CH2OH OH
O O
HO
O
OH
O
Monogalactosyldiacylglycerol (MGDG) O
CH2 O O
O
O O HO
O
OH
CH2OH OH
HO
OH
OH
O
Digalactosyldiacylglycerol (DGDG) O-
O
O P O
O O
O
OH OH
O
Diacylphosphatidylglycerol (PG)
Current Biology
The structures of the three major classes of lipid found in the thylakoid membrane.
bb10j04.qxd
10/5/00
R378
5:07 pm
Page R378
Current Biology Vol 10 No 10
Figure 2
Figure 3 Stacked thylakoid membranes
Cylinder
Unstacked membranes
Bilayer
LHC-II Current Biology
Stacking of thylakoid membranes. LHC-II is concentrated in the stacked region, along with other proteins such as photosystem II and the cytochrome b/f complex. Other proteins such as photosystem I and the CF0CF1 (ATP synthase) complex are concentrated in the unstacked regions. Stacking is mediated by LHC-II molecules. Interactions between LHC-II molecules will bring the surfaces of the appressed membranes close together. This could allow interactions between MGDG headgroups. The diagram is based on Anderson and Andersson [14]. Cone
Hexagonal phase
Current Biology
The concept of lipid shape and its effects on lipid packing. In a cylindrical lipid, the cross-sectional area of the lipid headgroup is about equal to that of the chains; a lipid of this type will pack as a bilayer. In a cone shaped lipid, the cross-sectional area of the lipid headgroup is less than that of the chains; a lipid of this type will pack in the hexagonal HII phase.
between MGDG and a major protein of the thylakoid membrane, the light harvesting complex (LHC-II) of photosystem II, also called the chlorophyll a/b protein. LHC-II is responsible for trapping light and passing the trapped energy to photosystem II. The first observation reported by Simidjiev et al. [2] is that, when purified LHC-II is mixed with MGDG at a molar ratio of about 80 MGDG molecules per LHC-II molecule, the MGDG adopts a bilayer structure. It is known that LHC-II contains three hydrophobic transmembrane α helices, forming a scaffold around which the chlorophylls are arranged [5]. Transmembrane α helices, being roughly cylindrical in shape, are ideally suited to interact with lipids in a bilayer arrangement. The helices presumably force MGDG to adopt a bilayer structure. This is likely to be a general property of membrane proteins; it has long been known, for example, that cytochrome oxidase can force cardiolipin, another lipid that favours non-bilayer structures, into a bilayer [6]. So the presence of LHC-II can stop MGDG adopting its preferred hexagonal HII structure, but why is MGDG present in the first place? The second most abundant lipid of the thylakoid membrane, DGDG, is a lipid that only forms bilayers [4]; the headgroup on DGDG is larger than that on MGDG, so that, even with two unsaturated fatty acyl chains, DGDG has a cylindrical shape. Why then is
the membrane lipid not just DGDG? A possible answer comes from studies of another role of LHC-II in the thylakoid membrane, that of organising the structure of the thylakoids. Thylakoids are arranged as stacks or grana, connected by a system of unstacked (non-appressed) stroma membranes (Figure 3). Stacking leads to a nonrandom distribution of the components of the thylakoid membrane, with photosystem I being highly enriched in the stroma membranes and photosystem II being concentrated in the stacked membranes [7]. This stacking is dependent on the presence of cations. When isolated thylakoids are suspended in a low salt medium, the membrane stacks fall apart and at the same time the membrane proteins become randomised between the stroma and grana membranes [8]. LHC-II is important in the stacking process [9]: when LHC-II, purified with its native lipid, is reconstituted into bilayers of phosphatidylcholine, vesicles are formed which aggregate in the presence of divalent cations with formation of extensive membrane stacks. The LHC-II molecules are found to be clustered at the adhering membrane regions [9]. It has been suggested that aggregation between the LHC-II particles is mediated by hydrophobic interactions between the LHC-II molecules and that the role of the cations is to neutralise surface charges on the proteins [9]. LHC-II is present in the thylakoid membrane in the form of trimers. In the stacks, these trimers are arranged into an orderly structure and the long-range order can be detected from distinctive changes in circular dichroism (CD) spectra. When LHC-II is stripped of its lipids it forms disordered aggregates. If the delipidated LHC-II is then mixed with MGDG, however, stacked lamellar structures are formed showing the CD spectra characteristic of
bb10j04.qxd
10/5/00
5:07 pm
Page R379
Dispatch
R379
ordered LHC-II [2]. Electron micrographs of the reconstituted membranes show that the LHC-II molecules are present in a regular hexagonal array. If LHC-II is mixed with other lipids such as DGDG or PG, grana-like stacks of membrane are not formed: stacking is only observed in the presence of MGDG [10].
[5]. When interactions between LHC-II molecules bring about membrane stacking, the surfaces of the stacked membranes will therefore be brought very close together. Hydrogen bonds could then form between MGDG headgroups in the two surfaces, leading to a strengthening of the interaction between the two surfaces.
These studies, in conjunction with previous studies, show that the three major species of lipid present in the thylakoid membrane each have distinct effects on the packing of LHC-II. LHC-II forms trimers in the membrane, but only in the presence of PG; PG is thought to bind at protein–protein interfaces in the trimer, stabilising trimer formation [11]. The role of DGDG is different. DGDG binds at the periphery of the trimer and allows the trimers to pack into two-dimensional crystalline arrays; DGDG is therefore involved in lipid-mediated contacts between adjacent trimers and MGDG cannot substitute for DGDG in this role [11]. Finally, the role of MGDG is to drive stacking of the bilayers to form grana [2]. Stacking has two advantages for the cell. The first is that it allows efficient packing of the thylakoid membranes within the chloroplast. The second is that it provides a mechanism for separation of the components of photosystem I and II within the thylakoid membrane [9].
A potential problem, however, is that strong interactions could also occur between galactose headgroups in the plane of the membrane. This type of interaction is indeed seen in monolayers of MGDG containing saturated fatty acyl chains, where strong interactions between adjacent headgroups lead to the formation of highly condensed, solid-like monolayers [4]. A way to prevent lateral headgroup interactions would be to keep the headgroups well apart. Having two polyunsaturated fatty acyl chains per lipid molecule would achieve this; the bulky chains, by keeping the headgroups apart, will prevent the formation of lateral hydrogen bonds between the headgroups. The headgroups will then be poised to form transverse hydrogen bonding between two appressed membrane surfaces.
What is still unknown is how the presence of MGDG leads to membrane stacking. All discussions of the role of nonbilayer lipids tend to separate into a debate on the relative importance of physics and of chemistry. The physicists argument is as follows. When a lipid that favours a nonbilayer phase is forced to adopt a bilayer structure it exists in a state of physical ‘frustration’. This will be manifested as an increase in surface free energy (increased surface tension). This increased surface tension could, in some way, modulate the function of a membrane protein or change the way the protein packs into the membrane [12]. The chemists argument is that what is important is the chemical structure of the lipid headgroup, particularly its ability to take part in hydrogen bonding with water molecules or with adjacent lipid headgroups; the fact that the lipid may prefer to adopt a non-bilayer structure is irrelevant, as long as this can be prevented from occurring in the native membrane. The uncharged galactose headgroup of MGDG is difficult to hydrate; rather then hydrogen bonding to water, it is expected that strong hydrogen bonds will be formed between adjacent galactose headgroups. This has been demonstrated in studies of monolayers of MGDG, where stronger interactions are observed between MGDG molecules than between phosphatidylcholines with comparable degrees of chain unsaturation [13]. It is possible that it is this hydrogen bonding potential that leads to the formation of stacked membranes in the presence of MGDG. The LHC-II protein only extends about 5 Å above the surface of the membrane on the outer (stromal) surface
The paper by Simidjiev et al. [2] does not answer the problem of whether MGDG is present in the thylakoid membrane because of its hydrogen bonding potential or because of its potential to form non-bilayer phases. What it does do is define the role of MGDG in the thylakoid membrane and show that the three major species of lipid in the membrane, MGDG, DGDG and DG, all have their own distinctive roles to play. References 1. Gorter E, Grendel F: On bimolecular layers of lipoids on the chromocytes of the blood. J Exp Med 1925, 41:439-443. 2. Simidjiev I, Stoylova S, Amenitsch H, Javorfi T, Mustardy L, Laggner P, Holzenburg A, Garab G: Self-assembly of large, ordered lamellae from non-bilayer lipids and integral membrane proteins in vitro. Proc Natl Acad Sci USA 2000, 97:1473-1476. 3. Harwood JL, Russell NJ: Lipids in Plants and Microbes. London: George Allen & Unwin; 1984. 4. Sen A, Williams WP, Quinn PJ: The structure and thermotropic properties of pure 1,2-diacylgalacosylglycerols in aqueous systems. Biochim Biophys Acta 1981, 663:380-389. 5. Kuhlbrandt W, Wang DN, Fujiyoshi Y: Atomic model of plant lightharvesting complex by electron crystallography. Nature 1994, 367:614-621. 6. Rietveld A, van Kemenade TJJM, Hak T, Verkleij AJ, de Kruijff B: The effect of cytochrome c oxidase on lipid polymorphism of model membranes containing cardiolipin. Eur J Biochem 1987, 164:137-140. 7. Hankamer B, Barber J, Boekema EJ: Structure and membrane organization of photosystem II in green plants. Annu Rev Plant Physiol 1997, 48:641-671. 8. Barber J: Influence of surface charges on thylakoid structure and function. Annu Rev Plant Physiol 1982, 33:261-295. 9. McDonnel A, Staehelin LA: Adhesion between liposomes mediated by the chlorophyll a/b light-harvesting complex isolated from chloroplast membranes. J Cell Biol 1980, 84:40-56. 10. Simidjiev I, Barzda V, Mustardy L, Garab G: Role of thylakoid lipids in the structural flexibility of lamellar aggregates of the isolated light-harvesting chlorophyll a/b complex of photosystem II. Biochemistry 1998, 37:4169-4173. 11. Nussberger S, Dorr K, Wang DN, Kuhlbrandt W: Lipid-protein interactions in crystals of plant light-harvesting complex. J Mol Biol 1993, 234:347-356.
bb10j04.qxd
10/5/00
R380
5:07 pm
Page R380
Current Biology Vol 10 No 10
12. Gruner SM: Intrinsic curvature hypothesis for biomembrane lipid composition: A role for nonbilayer lipids. Proc Natl Acad Sci USA 1985, 82:3665-3669. 13. Bishop DG, Kenrick JR, Bayston JH, MacPherson AS, Johns SR: Monolayer properties of chloroplast lipids. Biochim Biophys Acta 1980, 602:248-259. 14. Anderson JM, Andersson B: The dynamic photosynthetic membrane and regulation of solar energy conversion. Trends Biochem Sci 1988, 13:351-355.
10/5/00
5:07 pm
Page R377
Dispatch
R377
Membrane lipids: It’s only a phase Anthony G. Lee
All biological membranes contain lipids that prefer to adopt a non-bilayer phase. Recent results suggest that, in the thylakoid membrane, membrane proteins force all the lipids to adopt a bilayer structure, and that the non-bilayer-forming lipids in the thylakoid membrane serve to drive the formation of membrane stacks. Address: Division of Biochemistry and Molecular Biology, School of Biological Sciences, University of Southampton, Southampton SO16 7PX, UK. Current Biology 2000, 10:R377–R380 0960-9822/00/$ – see front matter © 2000 Elsevier Science Ltd. All rights reserved.
It is now 75 years since Gorter and Grendel [1] published their proposal that the lipids in a biological membrane are organised as a bilayer. The beauty of the bilayer structure is that its formation is self-driven; the bilayer forms because the hydrophobic fatty acyl chains of the lipids want to be out of contact with water, but the polar headgroups want to be in water. There can be no doubt that the bilayer model for the membrane is correct, but there remains a major puzzle — many of the lipids found in biological membranes are now known not to form bilayers, at least when on their own. A recent paper by Simidjiev et al. [2] has considered the importance of non-bilayer forming lipids for the proper functioning of one particular membrane, the thylakoid membrane. The internal photosynthetic membrane — the thylakoid membrane — of higher plant chloroplasts is the most abundant membrane found in nature. The membrane has a lipid composition unlike that of any other membrane (Figure 1). About 50% of the total lipid content is monogalactosyldiacylglycerol (MGDG), about 30% is digalactosyldiacylglycerol (DGDG) and about 10% is diacylphosphatidylglycerol (PG) [3]. As well as having unusual headgroups, MGDG and DGDG are unusual in containing two highly unsaturated fatty acyl chains (linoleic acid); most natural lipids contain one saturated and one unsaturated chain. The unusual chemical structures of MGDG and DGDG give them unusual packing properties. The ways that lipids pack together depends on their shape. A simple phospholipid, such as a phosphatidylcholine has a cylindrical shape; the cross-sectional area of the phosphocholine headgroup is about the same as that of two fatty acyl chains (Figure 2). Cylinders are ideally suited to packing side by side in a bilayer; however, not all lipids have a cylindrical shape. The headgroup of MGDG is quite large, and so MGDG
might be expected to be cylindrical and pack as a bilayer and, indeed, this is what is found when the chains are saturated [4]. But the two highly unsaturated chains in native MGDG give the chains a much larger than normal crosssectional area; the shape of native MGDG is therefore that of a cone (Figure 2). Cones do not pack well into planar arrays and prefer to adopt curved structures. In particular, MGDG has a tendency to pack into the so-called hexagonal HII structure shown in Figure 2. The structure gets its name because it consists of rod-like aggregates of infinite length packed into hexagonal arrays. The polar lipid headgroups are located at the centre of the rods along with some water, and the fatty acyl chains point outwards so that the surfaces of the rods are hydrophobic; this arrangement is called the HII hexagonal phase, to distinguish it from an arrangement with the hydrophobic fatty acyl chains pointing in towards the centre of the rods with the polar headgroups on the outside — this is the hexagonal HI phase. It is obvious that a biological membrane cannot be built from lipids that are in the hexagonal HII phase. So why has nature chosen to have a non-bilayer forming lipid as the major lipid in the thylakoid membrane and how is this lipid prevented from forming the hexagonal HII phase that it prefers? The recent work of Simidjiev et al. [2] shows that the answers to both these questions involve interactions Figure 1 O O
CH2OH OH
O O
HO
O
OH
O
Monogalactosyldiacylglycerol (MGDG) O
CH2 O O
O
O O HO
O
OH
CH2OH OH
HO
OH
OH
O
Digalactosyldiacylglycerol (DGDG) O-
O
O P O
O O
O
OH OH
O
Diacylphosphatidylglycerol (PG)
Current Biology
The structures of the three major classes of lipid found in the thylakoid membrane.
bb10j04.qxd
10/5/00
R378
5:07 pm
Page R378
Current Biology Vol 10 No 10
Figure 2
Figure 3 Stacked thylakoid membranes
Cylinder
Unstacked membranes
Bilayer
LHC-II Current Biology
Stacking of thylakoid membranes. LHC-II is concentrated in the stacked region, along with other proteins such as photosystem II and the cytochrome b/f complex. Other proteins such as photosystem I and the CF0CF1 (ATP synthase) complex are concentrated in the unstacked regions. Stacking is mediated by LHC-II molecules. Interactions between LHC-II molecules will bring the surfaces of the appressed membranes close together. This could allow interactions between MGDG headgroups. The diagram is based on Anderson and Andersson [14]. Cone
Hexagonal phase
Current Biology
The concept of lipid shape and its effects on lipid packing. In a cylindrical lipid, the cross-sectional area of the lipid headgroup is about equal to that of the chains; a lipid of this type will pack as a bilayer. In a cone shaped lipid, the cross-sectional area of the lipid headgroup is less than that of the chains; a lipid of this type will pack in the hexagonal HII phase.
between MGDG and a major protein of the thylakoid membrane, the light harvesting complex (LHC-II) of photosystem II, also called the chlorophyll a/b protein. LHC-II is responsible for trapping light and passing the trapped energy to photosystem II. The first observation reported by Simidjiev et al. [2] is that, when purified LHC-II is mixed with MGDG at a molar ratio of about 80 MGDG molecules per LHC-II molecule, the MGDG adopts a bilayer structure. It is known that LHC-II contains three hydrophobic transmembrane α helices, forming a scaffold around which the chlorophylls are arranged [5]. Transmembrane α helices, being roughly cylindrical in shape, are ideally suited to interact with lipids in a bilayer arrangement. The helices presumably force MGDG to adopt a bilayer structure. This is likely to be a general property of membrane proteins; it has long been known, for example, that cytochrome oxidase can force cardiolipin, another lipid that favours non-bilayer structures, into a bilayer [6]. So the presence of LHC-II can stop MGDG adopting its preferred hexagonal HII structure, but why is MGDG present in the first place? The second most abundant lipid of the thylakoid membrane, DGDG, is a lipid that only forms bilayers [4]; the headgroup on DGDG is larger than that on MGDG, so that, even with two unsaturated fatty acyl chains, DGDG has a cylindrical shape. Why then is
the membrane lipid not just DGDG? A possible answer comes from studies of another role of LHC-II in the thylakoid membrane, that of organising the structure of the thylakoids. Thylakoids are arranged as stacks or grana, connected by a system of unstacked (non-appressed) stroma membranes (Figure 3). Stacking leads to a nonrandom distribution of the components of the thylakoid membrane, with photosystem I being highly enriched in the stroma membranes and photosystem II being concentrated in the stacked membranes [7]. This stacking is dependent on the presence of cations. When isolated thylakoids are suspended in a low salt medium, the membrane stacks fall apart and at the same time the membrane proteins become randomised between the stroma and grana membranes [8]. LHC-II is important in the stacking process [9]: when LHC-II, purified with its native lipid, is reconstituted into bilayers of phosphatidylcholine, vesicles are formed which aggregate in the presence of divalent cations with formation of extensive membrane stacks. The LHC-II molecules are found to be clustered at the adhering membrane regions [9]. It has been suggested that aggregation between the LHC-II particles is mediated by hydrophobic interactions between the LHC-II molecules and that the role of the cations is to neutralise surface charges on the proteins [9]. LHC-II is present in the thylakoid membrane in the form of trimers. In the stacks, these trimers are arranged into an orderly structure and the long-range order can be detected from distinctive changes in circular dichroism (CD) spectra. When LHC-II is stripped of its lipids it forms disordered aggregates. If the delipidated LHC-II is then mixed with MGDG, however, stacked lamellar structures are formed showing the CD spectra characteristic of
bb10j04.qxd
10/5/00
5:07 pm
Page R379
Dispatch
R379
ordered LHC-II [2]. Electron micrographs of the reconstituted membranes show that the LHC-II molecules are present in a regular hexagonal array. If LHC-II is mixed with other lipids such as DGDG or PG, grana-like stacks of membrane are not formed: stacking is only observed in the presence of MGDG [10].
[5]. When interactions between LHC-II molecules bring about membrane stacking, the surfaces of the stacked membranes will therefore be brought very close together. Hydrogen bonds could then form between MGDG headgroups in the two surfaces, leading to a strengthening of the interaction between the two surfaces.
These studies, in conjunction with previous studies, show that the three major species of lipid present in the thylakoid membrane each have distinct effects on the packing of LHC-II. LHC-II forms trimers in the membrane, but only in the presence of PG; PG is thought to bind at protein–protein interfaces in the trimer, stabilising trimer formation [11]. The role of DGDG is different. DGDG binds at the periphery of the trimer and allows the trimers to pack into two-dimensional crystalline arrays; DGDG is therefore involved in lipid-mediated contacts between adjacent trimers and MGDG cannot substitute for DGDG in this role [11]. Finally, the role of MGDG is to drive stacking of the bilayers to form grana [2]. Stacking has two advantages for the cell. The first is that it allows efficient packing of the thylakoid membranes within the chloroplast. The second is that it provides a mechanism for separation of the components of photosystem I and II within the thylakoid membrane [9].
A potential problem, however, is that strong interactions could also occur between galactose headgroups in the plane of the membrane. This type of interaction is indeed seen in monolayers of MGDG containing saturated fatty acyl chains, where strong interactions between adjacent headgroups lead to the formation of highly condensed, solid-like monolayers [4]. A way to prevent lateral headgroup interactions would be to keep the headgroups well apart. Having two polyunsaturated fatty acyl chains per lipid molecule would achieve this; the bulky chains, by keeping the headgroups apart, will prevent the formation of lateral hydrogen bonds between the headgroups. The headgroups will then be poised to form transverse hydrogen bonding between two appressed membrane surfaces.
What is still unknown is how the presence of MGDG leads to membrane stacking. All discussions of the role of nonbilayer lipids tend to separate into a debate on the relative importance of physics and of chemistry. The physicists argument is as follows. When a lipid that favours a nonbilayer phase is forced to adopt a bilayer structure it exists in a state of physical ‘frustration’. This will be manifested as an increase in surface free energy (increased surface tension). This increased surface tension could, in some way, modulate the function of a membrane protein or change the way the protein packs into the membrane [12]. The chemists argument is that what is important is the chemical structure of the lipid headgroup, particularly its ability to take part in hydrogen bonding with water molecules or with adjacent lipid headgroups; the fact that the lipid may prefer to adopt a non-bilayer structure is irrelevant, as long as this can be prevented from occurring in the native membrane. The uncharged galactose headgroup of MGDG is difficult to hydrate; rather then hydrogen bonding to water, it is expected that strong hydrogen bonds will be formed between adjacent galactose headgroups. This has been demonstrated in studies of monolayers of MGDG, where stronger interactions are observed between MGDG molecules than between phosphatidylcholines with comparable degrees of chain unsaturation [13]. It is possible that it is this hydrogen bonding potential that leads to the formation of stacked membranes in the presence of MGDG. The LHC-II protein only extends about 5 Å above the surface of the membrane on the outer (stromal) surface
The paper by Simidjiev et al. [2] does not answer the problem of whether MGDG is present in the thylakoid membrane because of its hydrogen bonding potential or because of its potential to form non-bilayer phases. What it does do is define the role of MGDG in the thylakoid membrane and show that the three major species of lipid in the membrane, MGDG, DGDG and DG, all have their own distinctive roles to play. References 1. Gorter E, Grendel F: On bimolecular layers of lipoids on the chromocytes of the blood. J Exp Med 1925, 41:439-443. 2. Simidjiev I, Stoylova S, Amenitsch H, Javorfi T, Mustardy L, Laggner P, Holzenburg A, Garab G: Self-assembly of large, ordered lamellae from non-bilayer lipids and integral membrane proteins in vitro. Proc Natl Acad Sci USA 2000, 97:1473-1476. 3. Harwood JL, Russell NJ: Lipids in Plants and Microbes. London: George Allen & Unwin; 1984. 4. Sen A, Williams WP, Quinn PJ: The structure and thermotropic properties of pure 1,2-diacylgalacosylglycerols in aqueous systems. Biochim Biophys Acta 1981, 663:380-389. 5. Kuhlbrandt W, Wang DN, Fujiyoshi Y: Atomic model of plant lightharvesting complex by electron crystallography. Nature 1994, 367:614-621. 6. Rietveld A, van Kemenade TJJM, Hak T, Verkleij AJ, de Kruijff B: The effect of cytochrome c oxidase on lipid polymorphism of model membranes containing cardiolipin. Eur J Biochem 1987, 164:137-140. 7. Hankamer B, Barber J, Boekema EJ: Structure and membrane organization of photosystem II in green plants. Annu Rev Plant Physiol 1997, 48:641-671. 8. Barber J: Influence of surface charges on thylakoid structure and function. Annu Rev Plant Physiol 1982, 33:261-295. 9. McDonnel A, Staehelin LA: Adhesion between liposomes mediated by the chlorophyll a/b light-harvesting complex isolated from chloroplast membranes. J Cell Biol 1980, 84:40-56. 10. Simidjiev I, Barzda V, Mustardy L, Garab G: Role of thylakoid lipids in the structural flexibility of lamellar aggregates of the isolated light-harvesting chlorophyll a/b complex of photosystem II. Biochemistry 1998, 37:4169-4173. 11. Nussberger S, Dorr K, Wang DN, Kuhlbrandt W: Lipid-protein interactions in crystals of plant light-harvesting complex. J Mol Biol 1993, 234:347-356.
bb10j04.qxd
10/5/00
R380
5:07 pm
Page R380
Current Biology Vol 10 No 10
12. Gruner SM: Intrinsic curvature hypothesis for biomembrane lipid composition: A role for nonbilayer lipids. Proc Natl Acad Sci USA 1985, 82:3665-3669. 13. Bishop DG, Kenrick JR, Bayston JH, MacPherson AS, Johns SR: Monolayer properties of chloroplast lipids. Biochim Biophys Acta 1980, 602:248-259. 14. Anderson JM, Andersson B: The dynamic photosynthetic membrane and regulation of solar energy conversion. Trends Biochem Sci 1988, 13:351-355.