The non-bilayer lipid MGDG and the major light-harvesting complex

1 day ago - The thylakoid membrane of algae and land plants is characterized by its intricate architecture, comprising tightly appressed membrane stac...
0 downloads 6 Views 2MB Size
Subscriber access provided by Kaohsiung Medical University

Article

The non-bilayer lipid MGDG and the major light-harvesting complex (LHCII) promote membrane stacking in supported lipid bilayers Dennis Seiwert, Hannes Witt, Sandra Ritz, Andreas Janshoff, and Harald Paulsen Biochemistry, Just Accepted Manuscript • DOI: 10.1021/acs.biochem.8b00118 • Publication Date (Web): 26 Mar 2018 Downloaded from http://pubs.acs.org on March 26, 2018

Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.

Biochemistry is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.

Page 1 of 25 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Biochemistry

The non-bilayer lipid MGDG and the major light-harvesting complex (LHCII) promote membrane stacking in supported lipid bilayers a

b

c

b

a,*

Dennis Seiwert , Hannes Witt , Sandra Ritz , Andreas Janshoff , Harald Paulsen

a

Institute of Molecular Physiology, Johannes Gutenberg University Mainz, Johannes-von-Müller-Weg 6, 55128 Mainz, Germany

b

Institute of Physical Chemistry, University of Goettingen, Tammannstrasse 6, 37077 Goettingen, Germany

c

Microscopy Core Facility, Institute of Molecular Biology, Ackermannweg 4, 55128 Mainz, Germany

*

Corresponding author

Telephone: 49-6131-3924633, E-mail: paulsen@uni-mainz.de

Abstract The thylakoid membrane of algae and land plants is characterized by its intricate architecture, comprising tightly appressed membrane stacks termed grana. The contributions of individual components to grana stack formation are not fully elucidated yet. As an in vitro model we use supported lipid bilayers made of thylakoid lipid mixtures to study the effect of LHCII, different lipids and ions on membrane stacking, seen as elevated structures forming on top of the planar membrane surface in the presence of LHCII protein. These structures were examined by confocal laser scanning microscopy (CLSM), atomic force microscopy (AFM), and fluorescence recovery after photobleaching (FRAP), revealing multilamellar LHCII-membrane stacks composed of connected lipid bilayers. Both native-like and non-native interactions between the LHCII complexes may contribute to membrane appression in the supported bilayers. However, applying in vivo-like salt conditions to uncharged glycolipid membranes drastically increased stack formation due to enforced LHCII-LHCII interactions, which is in line with recent crystallographic and cryo-electron microscopic data [Wan, T. et al. (2014) Mol Plant 7, 916–919; Albanese, P. et al. (2017) Sci Rep 7, 10067–10083.]. Furthermore, we observed the non-bilayer lipid MGDG to strongly promote membrane stacking, pointing to the longterm proposed function of MGDG in stabilizing the inner membrane leaflet of highly curved margins in the periphery of each grana disc due to its negative intrinsic curvature [Murphy, D. J. (1982) FEBS Lett 150, 19–26].

1 ACS Paragon Plus Environment

Biochemistry 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 2 of 25

Introduction In cyanobacteria, algae, and higher plants the light-reaction of photosynthesis takes place in the thylakoid, making it the most abundant biological membrane system on earth (1). The thylakoid membrane of chloroplasts exhibits a very complex three-dimensional architecture: cylindrical stacks of flattened membrane discs with a diameter of 300 - 600 nm, the so-called grana structures, are interconnected by non-stacked membrane sections termed “stroma lamellae”. Stacking is accompanied by highly curved margins at the periphery of each grana disc (2–4). While the lipid composition, that is 50% non-bilayer lipid monogalactosyl diacylglycerol (MGDG), 30% digalactosyl

diacylglycerol

(DGDG)

and

20%

negatively

charged

lipids

(10%

each

of

phosphatidyldiacylglycerol (PG) and sulfoquinovosyl diacylglycerol (SQDG)), is more or less uniform across the continuous membrane system (5,6), the distribution of proteins is not. Photosystem II (PSII) and the major light-harvesting complex (LHCII) are concentrated in the grana, whereas photosystem I (PSI) and ATP synthase reside in the stroma-facing regions (2–4,7). This so-called lateral heterogeneity is associated with the primary grana functions: the prevention of spillover of excitation energy via spatial separation of PSI and PSII, fine-tuning of photosynthesis, facilitation of state transitions of LHCII, balance between linear and cyclic electron flow, and, most importantly, enhancement of light-harvesting efficiency under low light conditions through the formation of large arrays of PSII-LHCII supercomplexes (2–4,7). Grana stacking is established by the balance of various interactions between adjacent membrane layers comprising electrostatic attraction and repulsion, attractive van der Waals forces, entropy-driven attraction as well as hydration repulsion (8–10). High concentrations of cations are needed to screen the net negative charges on the membrane surface originating from the fraction of negatively charged lipids and acidic extra-membranous domains of PSII and LHCII (8–16). Disclosure of highly resolved LHCII crystal structures around 10 years ago (17) has given rise to the Velcro model, proposing electrostatic interactions between the stromal surfaces of opposing LHCII trimers to be essential for grana formation (17,18). Independently thereof, recent structural investigations have uncovered distinct salt-mediated interactions being formed between the stromal surfaces that may serve as a key determinant for the stacking of thylakoids (19,20). Moreover, destacking of grana membranes was suggested to be induced by phosphorylation of LHCII, sequestering in the latter into stromal lamellae where LHCII can associate with PSI (2,3,21). Although the glycolipids DGDG and MGDG together account for 80% of the lipid matrix and therefore provide the framework of the grana structure, their exact contribution to grana stacking cannot be examined in vivo since mutational studies always target the thylakoid architecture as a whole (for detailed review, see ref. (22)). In vitro studies of membrane stacking have included either the adhesion of preformed bilayers (liposomes) or the formation of multilamellar membrane structures upon hydration of lipids or their integration into protein aggregates (23–29). These approaches did not specifically address the membrane structure present in the non-appressed grana margins, pointing to the need for an in vitro-system that also takes into account the relevance of membrane curvature. The latter has recently been revealed to be induced by CURVATURE THYLAKOID1 (CURT1) proteins

2 ACS Paragon Plus Environment

Page 3 of 25 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Biochemistry

(30); however, little or nothing is known about an involvement of the different thylakoid lipids in this context. A promising way to mimic grana-related membrane stacking in vitro is the formation of multilamellar structures in a flat continuous membrane system deposited on a surface (31). Spreading of small unilamellar vesicles (SUVs) is a straightforward method to obtain such supported lipid bilayers (SLBs) (32–37) since it allows for easy variation of the lipid composition both with and without proteins like LHCII reconstituted into the SUVs (38,39). In order to investigate the impact of LHCII, salt conditions and different thylakoid lipid properties on the degree of membrane stacking, we established the formation of a planar SLB-system starting out from SUVs consisting of either phospholipids or uncharged glycolipids only, while the latter optionally contained a distinct proportion of the non-bilayer lipid MGDG. This enabled us to study the effects of the individual components in a predefined system, monitored via atomic force microscopy (AFM) and confocal laser scanning microscopy (CLSM). Our results indicate that LHCII and MGDG provide major contributions to stabilization of the grana structure.

Materials and Methods Isolation of native LHCII complexes LHCII was prepared from 2-week old pea seedlings according to ref. (40) with the modifications described in ref. (38). The final purification step by ultracentrifugation through a sucrose density gradient was omitted. Instead, aliquots containing precipitated LHCII were frozen in distilled H2O and stored at -20 °C until further use.

Preparation of small unilamellar vesicles (SUVs) Lipid aliquots of 1-palmitoyl-2-oleoyl-sn-glycero-3-phosphoglycerol (POPG; purchased from Avanti Polar Lipids), digalactosyl diacylglycerol (DGDG) and monogalactosyl diacylglycerol (MGDG; each purchased from Lipid Products) dissolved in chloroform were mixed at the desired molar ratios. In case of supported bilayer formation without LHCII protein, C11TopFluor galactosyl ceramide (TopFluor-GC; Avanti Polar Lipids) was added at molar ratios between 1:100 and 1:1000. A thin lipid film was formed on the inner wall of a glass flask by slowly removing the chloroform in a rotary evaporator. After complete removal of the solvent at 40 °C under high vacuum for 1 h, the lipid film was hydrated by rigorous mixing (3 x 30 s at 60 °C; MS2 minishaker, IKA) in ultrapure H2O (resistance R > 18 MΩ) or 50 mM NaCl (for supported bilayer formation on mica with POPG liposomes) at a total lipid concentration of 1.5 mM. For MGDG-containing samples the mixing step was repeated several times until no more lipid traces were detected by eye on the bottom of the flask. Sonication in a tip sonicator (Vibra cell, Sonics & Materials) for 4 min, followed by 3 freeze-thaw cycles, yielded large unilamellar vesicles. The vesicle suspension was then extruded 21x (LipoFast-Basic, Avestin) through a polycarbonate membrane (pore diameter: 100 nm) to obtain SUVs.

3 ACS Paragon Plus Environment

Biochemistry 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Reconstitution of LHCII into SUVs For reconstitution into liposomes, aliquots containing precipitated LHCII were thawed and dissolved in 0.05% (w/v) Triton X-100 / 10 mM TRIS-HCl pH 7.5. Preformed SUVs were mixed with Triton X-100 and TRIS-HCl pH 7.5 to final concentrations of 0.05% (w/v) and 10 mM, respectively. The reconstitution step was performed according to established protocols (38) with some modifications: the protein suspension was added dropwise under continuous mixing at 4 °C in the dark to a molar lipid/protein ratio of 500 – 1000 with respect to LHCII-monomers, and subsequently polystyrene beads (Bio-Beads SM-2, Bio-RAD) at 30 mg/ml were added. In order to remove Triton X-100, the mixture was incubated overnight under constant rotating at 4 °C in the dark, and the supernatant was removed to a new tube containing fresh polystyrene beads, followed by incubation for 1 h. The last step was repeated, and finally the supernatant was collected. For bilayer formation on glass coverslips, all buffers additionally contained 0.1 mM EDTA.

Surface preparation and functionalization Glass coverslips (22 x 22 mm2 1.5H, Marienfeld) were chemically cleaned via Piranha treatment (immersion in a stirred solution of H2SO4 (98%) and H2O2 (33%) at a ratio of 7:3 (v/v)) for 5 min, thoroughly rinsed with ultrapure H2O (10x), dried at 100 °C for 15 min and then kept in a desiccator under vacuum until further use. For bilayer formation on glass, the coverslips were treated with 1% (v/v) (3-aminopropyl)triethoxysilane (APTES; Sigma-Aldrich) in dry toluene for 4 h at 80 °C, rinsed with methanol and chloroform (3x each) and dried in a steam of nitrogen (41). For bilayer formation on 2

mica, thin mica slices (14 x 10 mm ; Glimmer “V5”, Plano) were fixed onto cleaned coverslips with UV curing glue (31). Before adding SUV suspension, the upper mica layers were freshly cleaved (at least 3x) with adhesive film.

Formation of supported lipid bilayers (SLBs) Preformed SUVs or SUVs containing reconstituted LHCII were freshly extruded (21x; LipoFast-Basic, Avestin) through a polycarbonate membrane (pore diameter: 100 nm). For SLB-formation on APTES coated glass, the vesicle suspension (20 µl) was incubated for 5 min on the functionalized coverslips and then rinsed with 0.1 mM EDTA / 10 mM TRIS-HCl pH 7.5 (4 x 1 ml). Prior to SLB-formation on mica, CaCl2 solution (100 mM) was added to the vesicle suspension to a final concentration of 5 mM and the suspension (60 µl) was incubated for 10 min on the freshly cleaved surface, followed by intense rinsing with 10 mM TRIS-HCl pH 7.5 (4 x 1 ml, including 50 mM NaCl in case of SLB-formation with POPG).

4 ACS Paragon Plus Environment

Page 4 of 25

Page 5 of 25 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Biochemistry

Confocal laser scanning microscopy and fluorescence recovery after photobleaching (FRAP) Confocal laser scanning microscopy and FRAP experiments were performed at room temperature on a TCS SP5 laser scanning microscope (Leica) with an 63x oil objective (NA = 1.4; HCX PL APO CS, Leica). A 488 nm argon laser was used to excite LHCII (detection range of photomultiplier: 600 – 700 nm) and TopFluor-GC (detection range: 520 – 600 nm). FRAP measurements were conducted in a circular region of interest (ROI) at constant framerate (1.037 s / frame) with 5 (Figure S2) or 10 frames of pre-bleaching, 2 (Figure S2) or 3 (Figure 1) or 10 frames (Figure 2, Figure S1) of bleaching and 100 frames of post-bleaching. The radius r of the ROI varied between 0.75 and 1.5 µm but was kept constant for each series of experiments (for a given figure). In order to compensate for photobleaching through imaging, the actual fluorescence intensity Ffrap was normalized by the intensity Fref of an unbleached area for every time point as follows: F(t) = Ffrap(t) / Fref(t). The mobile fractions M given by M = F∞ / F0, with F∞ the fluorescence intensity after recovery and F0 the average fluorescence intensity before bleaching, were directly estimated from the plots.

Quantification of membrane surface coverage Estimation of the surface areas covered by supported bilayers and membrane stacks was performed with the Fiji image software (42). For a given fluorescence micrograph, the fluorescence intensity distribution was monitored in circular ROI (in the same order of magnitude as for the FRAP measurements) for i) the background signal of the surface support (membrane-free area) and ii) the signal corresponding to the planar bilayer membrane (membrane area; see purple intensity in Figure 4). The maximal signal of i) was applied as an intensity threshold to create a binary image, which yielded the proportion of pixels related to lipid bilayers including the membrane stacks, i.e. the relative surface area occupied by membranes. Applying the maximal signal of ii) as an intensity threshold gave rise to a binary image which only contained pixels pertaining to the fluorescence signal of the membrane stacks, i.e. yielding the relative surface area covered only by stacked domains. The percentage value from ii) was normalized by i) in order to obtain the proportion of stacked domains on the total membrane area. For the analysis, 7 - 17 micrographs were taken into account per SLB lipid composition obtained from one (POPG), three (DGDG) and four (DGDG/MGDG = 2:1) different sample preparations, respectively.

5 ACS Paragon Plus Environment

Biochemistry 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Atomic force microscopy (AFM) Topographic imaging was performed at room temperature in intermitted contact mode using a commercial AFM (MFP-3D Infinity, Oxford Instruments Asylum Research) with MSCT-F cantilevers (nominal resonant frequency (air): 90 – 160 kHz, nominal spring constant: 0.3 – 1.2 N/m; Bruker).

Quantification of stack heights Estimation of the heights of the membrane stacks presented in Figure 4 was performed with Gwyddion (43). The average height of the planar lipid bilayer was subtracted from the AFM image. A height threshold of 4 nm was applied to identify stacks on the membrane. Only stacks with an area larger than 8 Pixels (~750 nm2) were considered.

Results Formation of surface supported membranes from LHCII-liposomes The formation of SLBs comprises adsorption at the surface followed by crowding, potentially fusion and rupture of the SUVs, a process that is predominantly governed by electrostatic interactions (32– 37). Therefore, glass coverslips were coated with (3-aminopropyl)triethoxysilane (APTES) to generate positive charges at the surface attracting the negatively charged phosphate head groups of the SUVs (diameter: 100 – 150 nm) (36,41,44). In order to facilitate vesicle spreading, SUVs made of pure 1palmitoyl-2-oleoyl-sn-glycero-3-phosphoglycerol (POPG) lipid were first applied for SLB formation. The molar lipid-to-protein ratios (1500 - 3000 with respect to LHCII trimers) used here correspond to protein densities one order of magnitude lower than in the grana (45) with the aim to avoid significant aggregation (46) or even crystallization (17) of trimeric LHCII within the SUVs, which might hamper the spreading process. However, the fluorescence intensity of the chlorophyll molecules bound to LHCII was still sufficient to record CLSM images. The corresponding fluorescence micrographs display a homogeneously distributed fluorescence signal on the surface support (Figure 1a), indicating an extended SLB and a homogeneous distribution of LHCII in the lipid bilayer on the length scale of microscope resolution. Moreover, bright fluorescent domains (diameter ≥ ~1 µm) are also apparent within the SLB that point to multilamellar structures or protein crowding, while smaller dots (diameter < ~1 µm) may also represent non-ruptured vesicles (see magnification in Figure 1a). FRAP measurements in between the bright spots (white circle in Figure 1a) demonstrate the POPG membranes to be fluid as the fluorescence intensity recovers by ~90% after bleaching (Figure 1b); this concomitantly shows that LHCII is fully mobile in solid supported membranes.

6 ACS Paragon Plus Environment

Page 6 of 25

Page 7 of 25 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Biochemistry

Figure 1 Fluorescence micrograph and FRAP curve of an SLB formed by spreading LHCII-containing POPG-SUVs on APTES-coated coverglass. (a) The lipid vesicles form homogenous bilayer membranes, which also contain patches with high fluorescence intensities; scale bars: left = 10 µm; right = 2 µm. (b) Representative FRAP experiment on the lipid bilayer. The time-resolved relative fluorescence intensity F/F0 within the region of interest (white circle in the magnification of the dotted section in (a)) reveals almost full recovery (~90%) after bleaching, indicating a fluid and continuous membrane system.

In order to inspect whether the bright areas mentioned above (see Figure 1a) represent either protein accumulation or, alternatively, alterations in the planar membrane structure, SLBs were stained with the membrane probe Bodipy (Figure 2). The CLSM images show an increased fluorescence signal of both LHCII and Bodipy emission within those domains; hence, the bright spots correlate to membrane patches containing stacked LHCII-membrane structures rather than (lateral) protein aggregates.

7 ACS Paragon Plus Environment

Biochemistry 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Figure 2 Fluorescence micrograph of a Bodipy-stained SLB formed by spreading of LHCII containing POPG-SUVs on APTES-coated coverglass. The patches characterized by high LHCII-fluorescence intensities (left; see Fig. 1a) also exhibit an increased fluorescence signal of the membrane probe Bodipy (middle). The merged image of both LHCII and Bodipy emission (right) demonstrates colocalization of protein and lipid within those patches, pointing to stacked LHCII-membrane structures; scale bars: 10 µm.

This finding is further supported by bleaching experiments conducted on entire patches: FRAP measurements show full recovery of the high fluorescence intensities after bleaching (Figure 3; lipid composition here: POPG/DGDG = 1:1), indicating diffusion processes related to LHCII complexes that would not be observable for LHCII-aggregates. The stacked membrane areas are thus connected to the surrounding planar bilayer (31) from where intact LHCII-proteins can diffuse into the stacks.

8 ACS Paragon Plus Environment

Page 8 of 25

Page 9 of 25 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Biochemistry

Figure 3 FRAP experiment on a patch containing stacked LHCII membranes (see Fig. 2) within an SLB formed by spreading LHCII-containing SUVs at POPG/DGDG = 1:1 on APTES-coated coverglass. A series of fluorescence micrographs (top) before (10 s), immediately after (20 s), and 2 min after bleaching (120 s) shows almost complete bleaching of the stacked domain (arrow). Full recovery within the region of interest (white circle) as monitored by the time-resolved relative fluorescence intensity F/F0 (bottom) reveals that the stack is connected with the surrounding bilayer. Scale bars: 5 µm.

However, since the fluorescence recovery might also arise from reversible bleaching of the pigments coordinated by LHCII (47,48), FRAP experiments were likewise performed on isolated membrane patches as a control (Figure S1). In this case, no significant recovery of the LHCII-fluorescence was observed, which rules out the possibility of reversible pigment-bleaching. To further illuminate the structure of the stacked domains and to investigate whether stack formation is affected by the different thylakoid lipid species, SLB formation was next performed on mica supports while varying the lipid composition of the SUVs.

Investigations on the structure of LHCII-membrane stacks at different lipid compositions In order to assess the impact of negative charges in the lipid head groups (present in POPG but absent in DGDG and MGDG) and of the tendency to promote spontaneous membrane curvature (present in MGDG/DGDG mixtures but absent in DGDG membranes (49)) on stack formation, we used three different thylakoid lipid mixtures: pure POPG, pure DGDG and DGDG/MGDG = 2:1. Higher proportions of MGDG were not examined as they prevent the formation of lamellar vesicles due to the hexagonal phase behavior of the lipid (50). Spreading of LHCII-SUVs containing these lipid compositions was performed on mica, driven by calcium ions, since mica has been shown to enable vesicle spreading for various lipid compositions in combination with calcium (35–37,51–54). Moreover, the flat mica surface preferably allows to discriminate between the solid supported bilayer and elevated membrane stacks via topographic (AFM) imaging as compared to the rough surface structure of glass. Figure 4 displays atomic force microscopy images of the LHCII-SLBs and membrane stacks formed on mica as a function of thylakoid lipid composition. The AFM topographs demonstrate that the lipid bilayer (brown color) can clearly be distinguished from the dark mica surface (Figure 4, top). According to height profile analysis the bilayer has a thickness of ~2.5 nm (Figure 4, bottom; see lane 2 in (a) explicitly), which is somewhat reduced compared to the typical values reported in the literature ranging from 3.5 to 4 nm; we attribute the disparities to compression of soft material such as membranes that can occur during AFM imaging via tapping mode (55). By contrast, rigid LHCII protein is not expected to be compressed by the AFM-tip. Therefore the height of membrane stacks is governed by the height of LHCII rather than by the thickness of the bilayer. POPG-SLBs possess a high surface coverage (Figure 4a), while the membrane density for DGDG- and DGDG/MGDG-SLBs is notably reduced representing partially connected bilayer patches (Figure 4b and c, respectively).

9 ACS Paragon Plus Environment

Biochemistry 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Line profile analysis of the elevated structures (white patches in Figure 4, top) resembling stacked LHCII-membranes as examined in Figures 1-3 allows for assigning their heights. Many stacks were found to exhibit heights of ~6, ~12 and ~18 nm on top of the lipid bilayer surface (Figure 4, bottom).

Figure 4 AFM topographs (top) and height profiles (bottom) of SLBs formed by spreading LHCIIcontaining SUVs at different thylakoid lipid compositions on mica. (a) POPG vesicles form a continuous bilayer membrane system (brown intensity). However, in some areas the bilayer contains bigger gaps (up to a few µm) visible as dark spots within the topographic image. (b) DGDG vesicles lead to the formation of partially connected bilayer patches due to a lower surface coverage. (c) The proportion of 30% MGDG in DGDG membranes barely alters the bilayer structure. For all lipid mixtures, elevated structures can be detected (visible as white patches in a-c top) that resemble the stacked LHCII-membrane domains described in Fig. 1-3. Line profile analysis (cyan lines in a-c top) reveals that many stacks adopt heights of ~6, ~12 and ~18 nm (lanes 1-3 in a-c bottom) on top of the planar membrane surface (height level: 0 nm). Scale bars: 1 µm.

The latter observation is reflected in the corresponding histograms of all stack heights obtained (Figure 5), which were determined via grain analysis of the AFM images: for each lipid composition a distinct maximum, i.e. the most frequent stack height counted, is apparent at ~6 nm (Figure 5a-c). Although less obvious than the latter feature, two additional local maxima can be observed in the probability distribution around ~12 nm and ~18 nm, respectively, for DGDG- as well as DGDG/MGDG -SLBs (Figure 5b and c). These results suggest that the stacks may be composed of multiple layers with a height of ~6 nm each. Since the height of LHCII complexes is about 6 nm according to crystallographic data (17)), each layer would thus consist of single LHCII containing-membrane, of

10 ACS Paragon Plus Environment

Page 10 of 25

Page 11 of 25 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Biochemistry

which the thickness is defined by the height of the protein causing a gap between adjacent lipid bilayers.

Figure 5 Histogram of the stack heights acquired from AFM topographs depending on the thylakoid lipid composition. For each lipid mixture, grain analysis was performed of a 10 x 10 µm2 section from the corresponding AFM image in Fig. 4 (where 5 x 5 µm2 sections are presented). For all lipid compositions, i.e. POPG (a), DGDG (b) and DGDG/MGDG = 2:1 (c), the stack heights show a clear maximum at ~6 nm within a Gaussian-like distribution ranging from 4 - 10 nm. Beyond that border the histograms in (b) and (c) point to two additional maxima around ~12 nm and ~18 nm.

The assumption that the stacks are composed of LHCII-membrane sheets was verified via CLSM imaging of the SLBs at the given lipid compositions. In order to better assess the heights of elevated membrane structures, the images are presented in an intensity-based color code of the LHCIIfluorescence (Figure 6, left). In line with the corresponding AFM topographs, the fluorescence micrographs indicate a high surface coverage of the supported bilayer membranes (purple intensity) for POPG (Figure 6a), in contrast to a rather low surface coverage for DGDG and DGDG/MGDG = 2:1 (Figure 6b and c, respectively). At each lipid composition the LHCII-fluorescence intensity of the

11 ACS Paragon Plus Environment

Biochemistry 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

planar lipid bilayer was quantified via line profile analysis (Figure 6, middle). Intensity profiles across different stacked domains (cyan, yellow and red intensities in Figure 6, left) were then normalized by the bilayer intensity in order to obtain the number of signal repeats corresponding to one LHCIImembrane layer (Figure 6, right). For all lipid mixtures, the results reveal that the stacks give rise to multiples of the LHCII-bilayer signal in whole numbers, thereby validating the proposed structure derived from topographic measurements.

Figure 6 Fluorescence micrographs and intensity profiles of SLBs formed by spreading LHCIIcontaining SUVs at different thylakoid lipid compositions on mica. As also apparent in the AFM topographs (see Fig. 4), the corresponding fluorescence images (left) reveal POPG vesicles to create a continuous lipid bilayer system (purple intensity) on the mica surface (a), whereas DGDG and DGDG/MGDG = 2:1 vesicles form partially connected bilayer patches ((b) and (c), respectively). Line profile analysis (white lines 1, left) was performed to first determine the signal intensities of the LHCII fluorescence corresponding to the planar lipid bilayers (lanes 1, middle). These intensities were then

12 ACS Paragon Plus Environment

Page 12 of 25

Page 13 of 25 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Biochemistry

used to normalize the line profiles (white lines 2-4, left) across different stacked domains (cyan/yellow/red intensities; see also white patches in Fig. 4), giving rise to the multiple signals of one LHCII-bilayer (lanes 2-4, right), i.e. the number of layers of a membrane stack. Note that in (c) line 5 (left) corresponds to a membrane stack of which the fluorescence signal (red intensity) exceeds the detection limit (lane 5, middle). Scale bars: 10 µm.

A feature characteristic for MGDG-containing SLBs is the appearance of high-intensity structures (red intensities in Figure 6c, left), of which the fluorescence emission (at similar microscopy settings with respect to the other lipid mixtures) exceeds the CLSM detection limit (lane 5 in Figure 6c, middle). However, representative FRAP-experiments, similar to those performed on membrane stacks obtained from SLB formation on glass supports (see Figure 3), suggest that these structures still represent lamellar membrane stacks as they are connected with the surrounding bilayer (Figure S2) (31). In summary, the preparation of SLBs from LHCII-SUVs composed of different thylakoid lipids - either POPG, DGDG or a mixture of DGDG and MGDG - is accompanied by the formation of membrane stacks with conserved structure, i.e. stacks containing a variety of single LHCII-membrane layers that are connected with one another and with the surrounding SLB as indicated by full fluorescence recovery after photobleaching (see Figure 3 and Figure S2) (31).

Impact of POPG/DGDG/MGDG composition on the degree of membrane stacking Having verified that the structure of LHCII-membrane stacks is not altered upon variation of the thylakoid lipid composition, we wished to know whether the aforementioned lipid properties, i.e. head group charge and lipid curvature (see above), have an impact on the degree of stack formation in the SLBs. To address this question, the relative surface areas occupied by the supported membranes and the proportions of stacked domains in the overall membrane were determined (Figure 7). The procedure was performed on several CLSM micrographs for each of the established lipid mixtures (see representative images in Figure. 6, left).

13 ACS Paragon Plus Environment

Biochemistry 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Figure 7 Statistical analysis of the relative membrane areas occupied by stacked domains acquired from different fluorescence micrographs depending on the thylakoid lipid composition. POPG-SLBs exhibit a ~2-fold higher surface coverage than SLBs containing DGDG or DGDG/MGDG = 2:1, as indicated by the relative membrane areas. By contrast, only 3% of the POPG membrane area is covered by stacked domains, whereas for DGDG SLBs the stacked areas amount to 9% of the membrane area. The presence of MGDG in DGDG membranes (ratio = 1:2) raises the proportion of stacked areas to 25% of the total membrane area. The error bars represent the standard deviations obtained from n (POPG) = 7, n (DGDG) = 17 and n (DGDG/MGDG) = 12 measurements.

As already indicated by the AFM and CLSM images presented in the last section (see Figure 4 and Figure 6), quantitative analysis confirms a relatively high surface coverage for POPG-SLBs (~65%) and a low surface coverage for DGDG- and DGDG/MGDG = 2:1-SLBs respectively (~35%). However, the stacked domains account for only 3% of the membrane area in the POPG-SLBs in contrast to 9% stacked areas in DGDG membranes. When the latter membranes additionally contain 1/3 of the nonbilayer lipid MGDG, 25% of the total membrane area is occupied by LHCII-membrane stacks. Moreover, besides capturing more membrane area, the membrane stacks in DGDG/MGDG samples in principle seem to be higher compared to the other samples. This notion is based on the observation that only for MGDG containing samples, the fluorescence intensity level of membrane stacks often reaches the CLSM detection limit as illustrated in Figure 6c (see also last section). Those stacks were found to represent up to ~20% of the overall stacked membrane (see exemplary image in Figure S3).

Role of LHCII and influence of monovalent cations To look into the questions whether and to which extent the embedded LHCII complexes may drive membrane stacking within the SLBs, spreading of liposomes lacking LHCII was performed under identical conditions, i.e. driven by calcium ions with POPG-, DGDG and DGDG/MGDG = 2:1-SUVs on mica supports. With respect to bilayer formation, the membrane surface coverage is only slightly higher than in the presence of LHCII as can be seen from membrane staining with the fluorescent labeled lipid C11TopFluor galactosyl ceramide (TopFluor-GC) at the corresponding thylakoid lipid compositions (Figure S4a-c). Notably, all lipid mixtures generate SLBs lacking highly fluorescent domains (Figure S4a-c) or elevated structures (AFM-image: Figure S4d), meaning that no membrane stacking takes place in the absence of LHCII under these conditions. The relevance of cations for grana stacking has been studied intensely in intact chloroplasts or by isolation of thylakoids under low salt conditions. The definite concentration of monovalent cations present in grana is unclear; however, several studies emphasize a strong dependency of the grana structure on high KCl or NaCl concentrations (100 – 200 mM) (8,9,11,14–16). In order to examine if the latter also affect generation of the stacked membrane structures presented in this work, SLBs were formed with SUVs containing DGDG and LHCII. Since we wished to focus only on membrane stacking due to LHCII related interactions, vesicle spreading was performed with uncharged SUVs containing no POPG which rules out effects originating from lipid charge screening by cations. Compared to the standard conditions (Figure S5a; see also Figure 6b), adding 200 mM NaCl prior to

14 ACS Paragon Plus Environment

Page 14 of 25

Page 15 of 25 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Biochemistry

vesicle spreading drastically provokes the formation of membrane stacks leading to sizes up to 10 µm in diameter (increased by a factor of ~10; Figure S5b). Even though bleaching experiments could not be conducted here since the stacks originate from unconnected membrane patches, the CLSM images suggest that those structures still represent continuous membranes.

Discussion LHCII drives membrane stacking within SLBs in a salt dependent manner In the current work, we monitored the formation of membrane stacks in flat lipid bilayers deposited on a surface support. Membrane staining (see Figure 2) in combination with quantitative analysis of the heights (see Figure 4 and Figure 5) and LHCII-signal intensities (see Figure 6) results in a picture of stacked lamellar structures seen with all thylakoid lipid compositions applied. According to FRAP measurements (see Figure 3 and Figure S2) each stack embodies a continuous membrane system, connected to the surrounding bilayer, allowing for free diffusion of the inherent LHCII proteins as can be achieved by different membrane architectures (Figure 8a-b). According to the data in Figure 5 and Figure 6, the stacks are composed of multiples of single membranes. Figure 8a shows a model of such stacked membranes containing junctions between adjacent layers. These junctions (Figure 8c) resemble stalked lipid bilayer structures, which have been observed in membrane fusion events (5659). It should be noted that diffusion of LHCII through such membrane junctions would be energetically unfavorable since hydrophilic LHCII regions would have to pass through hydrophobic regions of the lipid stalk. Additionally, the high local curvature in the stalk region might pose an additional energetic barrier against diffusion of LHCII. Furthermore, the free membrane edges of the membrane sheets impose an additional free energetic cost to form these structures in the first place. Diffusion of LHCII is more easily envisioned in an alternative model of membrane stacks (Figure 8b). This model is reminiscent of the thylakoid grana structure where the thylakoid membrane folds up into multiple layers and then is appressed to form stacks. This membrane folding would lead to multiples of two membrane layers in each stack. However, this is not consistent with the AFM and fluorescence micrograph data presented here, which indicate multiples of single membrane layers in the stacks (Figure 5 and Figure 6); therefore, we prefer the model shown in Figure 8a. Notably, similar structure have been reported to form in the context of membrane stack formation triggered by the presence of polyamides (31). There are two feasible mechanisms for stack formation: i) repetitive rupture of lipid vesicles within a certain area on top of the SLB giving rise to separate bilayer sheets that fuse right after or ii) membrane protrusions shaped out of the planar lipid bilayer (31).

15 ACS Paragon Plus Environment

Biochemistry 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Figure 8 Different models how LHCII and MGDG can induce membrane stacking. (a) Membrane stacks comprised of single lipid bilayers connected by membrane junctions leading to single membrane repeats as found in Figure 5 and 6. (b) Membrane folds resembling the structure in the thylakoid would lead to multiples of two membrane layers in each stack, which is not consistent with Figure 5 and Figure 6. Therefore, we prefer the first model. (c) The stalk structure in (a) is characterized by negative lipid curvature, which is presumably stabilized by an accumulation of MGDG, which possesses high negative intrinsic curvature.

Membrane stacking in grana predominantly relies on LHCII protein providing attractive forces for membrane appression in terms of van der Waals interactions between the complexes (8,9,60). Moreover, specific interactions of the positively charged N-termini (stromal surface) of one membrane with negatively charged surface patches on the stromal side of LHCII trimers in the opposing membrane have been proposed as the Velcro model (17,61,18). Alternatively, monovalent and divalent cations have been shown by new crystallographic data to generate salt bridges between the negatively charged stromal surface areas of opposing LHCII trimers, enhancing attractive forces between the complexes (19). This finding is reflected in recent cryo-electron microscopy data of isolated PSII-LHCII supercomplexes, where specific overlaps between opposing LHCII-trimers were discovered which closely resemble the salt-mediated interactions (salt-bridges) found in the new LHCII

16 ACS Paragon Plus Environment

Page 16 of 25

Page 17 of 25 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Biochemistry

crystal structure (20). In addition, hydrogen bonds between the polar loops of opposing LHCII polypeptides may contribute to the stability of grana structures (21). Regardless of the formation process (see above), the structures obtained in the present in vitro study confirm LHCII-LHCII interactions to be a major driving force towards membrane stacking (as illustrated in Figure 8) since no stacks were detected in the absence of the protein (compare Figure 4 and Figure 6 with Figure S4). However, straightforward conclusions addressing the type of LHCII interactions stabilizing appressed membranes in the SLBs cannot be drawn from our results as the complexes insert into artificial membranes at random orientation. Therefore, the opposing LHCII trimers of adjacent membrane layers can face each other with either the stromal or the lumenal surface exposed, enabling non-native as well as native-like interactions to be established between the trimers. Besides attractive van der Waals forces that are expected to be present in both cases, nonnative interaction may comprise, for instance, specific interactions between hydrophobic amino acids along with stromal-lumenal contacts (which have been observed in LHCII crystals (17)). Since we obtained stacked domains in POPG membranes even at low salt conditions (buffer: 10 mM TRIS-HCl pH 7.5 for SLB formation) (see Figure 1), at which lipid-mediated membrane adhesion is quite unlikely, due to repulsive electrostatic forces between the head groups, it is possible that also LHCII-LHCII interactions according to the Velcro model may contribute to membrane stacking in SLBs (17,61,18). By contrast, the formation of salt bridges between opposing LHCII trimers in the presence +

of cations ((19,20); see above) clearly plays a role: the addition of 200 mM Na drastically enhances stacking within SLBs consisting of uncharged glycolipids only, which rules out a possible impact of +

lipid charge screening by cations (see Figure S5) (3,8,9,26). Nevertheless, Na may additionally contribute to the apposition of DGDG membranes due to disruption of structured water, i.e. oriented water molecules associated with the sugar head groups of DGDG (8,9,25,26).

MGDG stabilizes the highly curved membrane junctions between adjacent bilayers With regard to the impact of thylakoid lipids on grana stacking, much attention has been paid so far to the chemistry of the lipid head groups. PG, represented by POPG, has been shown to induce lateral disaggregation of LHCII trimers within lamellar membranes due to repulsive interactions between the phosphate moiety of the lipid and negatively charged domains of the protein. By contrast, both glycolipids DGDG and MGDG tend to facilitate lateral aggregation of the trimers, presumably representing a beneficial condition for stack formation in thylakoids (62). Moreover, a direct contribution to grana stacking by MGDG and in particular DGDG has been attributed to the galactose rings in the head group region, capable of forming intermolecular hydrogen bonds between adjacent membranes (23,24,26,63). In agreement with the literature, a minor but significant effect was observed for uncharged glycolipid DGDG being beneficial for stack formation in SLBs in comparison to negatively charged phospholipid POPG (9% vs. 3% stacked areas; see Figure 7). Surprisingly, when a proportion of ~30% MGDG is present in DGDG membranes, as much as 25% of the total membrane area is covered by stacked

17 ACS Paragon Plus Environment

Biochemistry 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

domains (see Figure 7). Furthermore, many stacks in DGDG/MGDG-SLBs were found to adopt enormous heights – corresponding to more than 12 bilayer repeats according to lane 5 in Figure 4c – a feature that was not observed for the other samples (see Figure 4 and Figure S3). Since one galactose moiety is missing in MGDG compared to the lamellar lipid DGDG, MGDG adopts a conical shape and hence prefers to form a hexagonal phase (49). Therefore, MGDG tends to induce negative intrinsic curvature within lamellar membranes (64,65) which appears to facilitate membrane stacking in SLBs to a great extent. Unlike the aforementioned studies addressing membrane stacks composed of separate bilayer sheets (23–29), we obtained stacks which presumably contain stalked membrane structures (see Figure 8a) as inferred from the FRAP data. We assume that stalk formation is a key determinant for membrane stacking in all types of SLBs presented in this work (see last section); furthermore, a multitude of stalks may exist between two adjacent membrane layers. Hence, stacking is accompanied by rigorous bending of the lipid bilayer: the stalk regions bridging adjacent sheets are characterized by high curvature stress being exerted on the membrane, i.e. negative curvature pressure in both leaflets of the bilayer. This effect presumably renders stacking of the membranes energetically unfavorable unless the curvature stress is eased by the shape of lipids present in these domains. Our results show that MGDG can fulfil such a function, as stack formation is strongly enhanced in SLBs containing this lipid. We propose that MGDG is enriched in the stalked domains (Figure 8c) where it is able to reduce the stress by its negative intrinsic curvature (49,64,65). As mentioned above, we cannot rigorously exclude the scenario that the membrane stacks analyzed here contain folded-up membranes similar to thylakoid grana. Here, the margin regions are characterized by strong bending of the lipid bilayer, by which lipids are forced to adopt a negative curvature in the inner and a positive curvature in the outer leaflet. This can be eased by an enrichment of MGDG in the inner leaflet of the grana margins and its depletion in the positively curved outer leaflet. This scenario is consistent with the observation of an unchanged MGDG content in the overall margin fraction (5,6). Since the stalked membrane domains proposed in Fig. 8a possibly emerge from subsequent fusion of approaching bilayers on top of the SLBs (as discussed in the last section), our data may further stimulate the ongoing discussion about the role of MGDG on facilitating membrane fusion, coincident with recent studies on IM30 protein being essential for fusion events in the thylakoid membrane (66). Moreover, an involvement of MGDG in edge formation (grana margins) in thylakoids has been proposed (67,68), but experimental support for the latter hypothesis has not yet been provided. The supported bilayer system presented here is well suited to model grana-related membrane stacking since the in vitro stacking process is correlated with the formation of curved membrane domains. We propose that one of the functions of MGDG, contributing some 50% of the total lipid content of thylakoids, is to promote membrane stacking by easing the membrane curvature stress in the margin domains. This effect may be enhanced by CURT1 proteins specifically binding to thylakoid margins (30).

18 ACS Paragon Plus Environment

Page 18 of 25

Page 19 of 25 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Biochemistry

Author contributions D.S., H.W., S.R, A.J. and H.P. conceived the experiments. D.S. performed and analyzed the CLSM experiments. H.W. and D.S. performed and analyzed the AFM experiments. D.S. wrote the manuscript. H.W., S.R., A.J. and H.P. edited the manuscript.

Competing financial interests The authors declare no competing financial interests.

Funding sources The project was funded by the Deutsche Forschungsgemeinschaft through the Goettingen Graduate School for Neurosciences, Biophysics, and Molecular Biosciences Grants GSC 226/2 (to H.W.) and SFB 803 B08 (to A.J.).

Acknowledgement We thank Mária Hanulová for encouraging and constructive comments during the CLSM experiments. Support by the IMB Microscopy Core Facility is gratefully acknowledged.

Abbreviations AFM, atomic force microscopy;

APTES, (3-aminopropyl)triethoxysilane; CLSM, confocal laser

scanning microscopy; CURT, CURVATURE THYLAKOID1; DGDG, digalactosyl diacylglycerol; LHCII, major light-harvesting complex; MGDG, monogalactosyl diacylglycerol; POPG, 1-palmitoyl-2-oleoyl-snglycero-3-phosphoglycerol; SLBs, supported lipid bilayers; SUVs, small unilamellar vesicles.

Supporting information FRAP experiment on an isolated membrane patch formed by spreading LHCII-containing SUVs at POPG/DGDG = 1:1 on APTES-coated coverglass (Figure S1), FRAP experiment on a membrane stack formed by spreading LHCII-containing SUVs at DGDG/MGDG = 2:1 on mica (Figure S2), fluorescence micrograph of an SLB with several high-intensity stacks formed by spreading LHCIIcontaining DGDG/MGDG = 2:1-SUVs on mica (Figure S3), fluorescence micrographs and AFM topograph of SLBs formed by spreading SUVs containing TopFluor-GC at different thylakoid lipid compositions on mica (Figure S4), fluorescence micrographs of SLBs formed by spreading LHCIIcontaining DGDG-SUVs on mica affected by NaCl (Figure S5).

19 ACS Paragon Plus Environment

Biochemistry 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

References 1. Gounaris, K., and Barber, J. (1983) Monogalactosyldiacylglycerol: the most abundant polar lipid in nature, Trends Biochem Sci 8, 378–381. 2. Pribil, M., Labs, M., and Leister, D. (2014) Structure and dynamics of thylakoids in land plants, J Exp Bot 65, 1955–1972. 3. Ruban, A. V., and Johnson, M. P. (2015) Visualizing the dynamic structure of the plant photosynthetic membrane, Nat Plants 1, 1–9. 4. Kirchhoff, H. (2013) Architectural switches in plant thylakoid membranes, Photosynth Res 116, 481–487. 5. Douce, R., and Joyard, J. (1990) Biochemistry and function of the plastid envelope, Annu Rev Cell Biol 6, 173–216. 6. Duchêne, S., and Siegenthaler, P.-A. (2000) Do glycerolipids display lateral heterogeneity in the thylakoid membrane?, Lipids, 739–744. 7. Dekker, J. P., and Boekema, E. J. (2005) Supramolecular organization of thylakoid membrane proteins in green plants, Biochim Biophys Acta 1706, 12–39. 8. Chow, W. S., Kim, E.-H., Horton, P., and Anderson, J. M. (2005) Granal stacking of thylakoid membranes in higher plant chloroplasts: the physicochemical forces at work and the functional consequences that ensue, Photochem Photobiol Sci 4, 1081–1090. 9. Puthiyaveetil, S., van Oort, B., and Kirchhoff, H. (2017) Surface charge dynamics in photosynthetic membranes and the structural consequences, Nat Plants 3, 1–9. 10. Jia, H., Liggins, J. R., and Chow, W. S. (2014) Entropy and biological systems: Experimentallyinvestigated entropy-driven stacking of plant photosynthetic membranes, Sci Rep 4. 11. Izawa, S., and Good, N. E. (1966) Effect of salts and electron transport on the conformation of isolated chloroplasts., Plant Physiol 41, 544–552. 12. Rumak, I., Gieczewska, K., Kierdaszuk, B., Gruszecki, W. I., Mostowska, A., Mazur, R., and Garstka, M. (2010) 3-D modelling of chloroplast structure under (Mg2+) magnesium ion treatment. Relationship between thylakoid membrane arrangement and stacking, Biochim Biophys Acta 1797, 1736–1748. 13. Carter, D. P., and Staehelin, A. (1980) Proteolysis of chloroplast thylakoid membranes. II. Evidence for the involvement of the light-harvesting chlorophyll a/b-protein complex in thylakoid stacking and for effects of magnesium ions on photosystem II-light-harvesting complex aggregates in the absence of membrane stacking, Arch Biochem Biophys 200, 374–386. 14. Rubin, B. T., Chow, W. S., and Barber, J. (1981) Experimental and theoretical considerations of mechanisms controlling cation effects on thylakoid membrane stacking and chlorophyll fluorescence, Biochim Biophys Acta 634, 174–190. 15. Murakami, S., and Packer, L. (1971) The role of cations in the organization of choroplast membranes, Arch Biochem Biophys 146, 337–347. 16. Staehelin, A. (1976) Reversible particle movements associated with unstacking and restacking of chloroplast membranes in vitro, J Cell Biol 71, 136–158. 17. Standfuss, J., van Scheltinga, A. C. T., Lamborghini, M., and Kühlbrandt, W. (2005) Mechanisms of photoprotection and nonphotochemical quenching in pea light-harvesting complex at 2.5 A° resolution, EMBO J 24, 919–928.

20 ACS Paragon Plus Environment

Page 20 of 25

Page 21 of 25 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Biochemistry

18. Daum, B., Nicastro, D., Austin, J., McIntosh, J. R., and Kühlbrandt, W. (2010) Arrangement of photosystem II and ATP synthase in chloroplast membranes of spinach and pea, Plant Cell 22, 1299– 1312. 19. Wan, T., Li, M., Zhao, X., Zhang, J., Liu, Z., and Chang, W. (2014) Crystal structure of a multilayer packed major light-harvesting complex: implications for grana stacking in higher plants, Mol Plant 7, 916–919. 20. Albanese, P., Melero, R., Engel, B. D., Grinzato, A., Berto, P., Manfredi, M., Chiodoni, A., Vargas, J., Sorzano, C. Ó. S., Marengo, E., Saracco, G., Zanotti, G., Carazo, J.-M., and Pagliano, C. (2017) Pea PSIILHCII supercomplexes form pairs by making connections across the stromal gap, Sci Rep 7, 10067– 10083. 21. Janik, E., Bednarska, J., Zubik, M., Puzio, M., Luchowski, R., Grudzinski, W., Mazur, R., Garstka, M., Maksymiec, W., Kulik, A., Dietler, G., and Gruszecki, W. I. (2013) Molecular architecture of plant thylakoids under physiological and light stress conditions: a study of lipid-light-harvesting complex II model membranes, Plant Cell 25, 2155–2170. 22. Kobayashi, K. (2016) Role of membrane glycerolipids in photosynthesis, thylakoid biogenesis and chloroplast development, J Plant Res. 129, 565–580. 23. Deme, B., Cataye, C., Block, M. A., Marechal, E., and Jouhet, J. (2014) Contribution of galactoglycerolipids to the 3-dimensional architecture of thylakoids, FASEB J 28, 3373–3383. 24. Simidjiev, I., Barzda, V., Mustárdy, L., and Garab, G. (1998) Role of thylakoid lipids in the structural flexibility of lamellar aggregates of the isolated light-harvesting chlorophyll a/b complex of photosystem II, Biochemistry 37, 4169–4173. 25. Webb, M. S., Tilcock, C. P. S., and Green, B. R. (1988) Salt-mediated interactions between vesicles of the thylakoid lipid digalactosyldiacylglycerol, Biochim Biophys Acta 938, 323–333. 26. Webb, M. S., and Green, B. R. (1990) Effects of neutral and anionic lipids on digalactosyldiacylglycerol vesicle aggregation, Biochim Biophys Acta 1030, 231–237. 27. Ryrie, I., Anderson, J. M., and Goodchild, D. J. (1980) The role of the light-harvesting chlorophyll a/b -protein complex in chloroplast membrane stacking. Cation-induced aggregation of reconstituted proteoliposmes, Eur J Biochem 107, 345–354. 28. Sprague, S. G., Camm, E. L., Green, B. R., and Staehelin, A. (1985) Reconstitution of lightharvesting complexes and photosystem II cores into galactolipid and phospholipid liposomes, J Cell Biol 100, 552–557. 29. Mullet, J. E., and Arntzen, C. J. (1980) Simulation of grana stacking in a model membrane system. Mediation by a purified light-harvesting pigment-protein complex from chloroplasts, Biochim Biophys Acta 589, 100–117. 30. Armbruster, U., Labs, M., Pribil, M., Viola, S., Xu, W., Scharfenberg, M., Hertle, A. P., Rojahn, U., Jensen, P. E., Rappaport, F., Joliot, P., Dormann, P., Wanner, G., and Leister, D. (2013) Arabidopsis CURVATURE THYLAKOID1 proteins modify thylakoid architecture by inducing membrane curvature, Plant Cell 25, 2661–2678. 31. Gräb, O., Abacilar, M., Daus, F., Geyer, A., and Steinem, C. (2016) 3D-membrane stacks on supported membranes composed of diatom lipids induced by long-chain polyamines, Langmuir 32, 10144–10152. 32. Goksu, E. I., Vanegas, J. M., Blanchette, C. D., Lin, W.-C., and Longo, M. L. (2009) AFM for structure and dynamics of biomembranes, Biochim Biophys Acta 1788, 254–266. 21 ACS Paragon Plus Environment

Biochemistry 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

33. Anderson, T. H., Min, Y., Weirich, K. L., Zeng, H., Fygenson, D., and Israelachvili, J. N. (2009) Formation of supported bilayers on silica substrates, Langmuir 25, 6997–7005. 34. Hardy, G. J., Nayak, R., and Zauscher, S. (2013) Model cell membranes: techniques to form complex biomimetic supported lipid bilayers via vesicle fusion, Curr Opin Colloid Interface Sci 18, 448–458. 35. Richter, R. P., Berat, R., and Brisson, A. R. (2006) Formation of solid-supported lipid bilayers: an integrated view, Langmuir 22, 3497–3505. 36. Kim, Y.-H., Rahman, M. M., Zhang, Z.-L., Misawa, N., Tero, R., and Urisu, T. (2006) Supported lipid bilayer formation by the giant vesicle fusion induced by vesicle–surface electrostatic attractive interaction, Chem Phys Lett 420, 569–573. 37. Seantier, B., and Kasemo, B. (2009) Influence of mono- and divalent ions on the formation of supported phospholipid bilayers via vesicle adsorption, Langmuir 25, 5767–5772. 38. Yang, C., Boggasch, S., Haase, W., and Paulsen, H. (2006) Thermal stability of trimeric lightharvesting chlorophyll a/b complex (LHCIIb) in liposomes of thylakoid lipids, Biochim Biophys Acta 1757, 1642–1648. 39. Zhou, F., Liu, S., Hu, Z., Kuang, T., Paulsen, H., and Yang, C. (2009) Effect of monogalactosyldiacylglycerol on the interaction between photosystem II core complex and its antenna complexes in liposomes of thylakoid lipids, Photosynth Res 99, 185–193. 40. Krupa, Z., Huner, N. P. A., Williams, J. P., Maissan. E., and James. D. R. (1987) Development at cold-hardening temperatures. The structure and composition of purified rye light harvesting complex II, Plant Physiol 84, 19–24. 41. Dewa, T., Sugiura, R., Suemori, Y., Sugimoto, M., Takeuchi, T., Hiro, A., Iida, K., Gardiner, A. T., Cogdell, R. J., and Nango, M. (2006) Lateral organization of a membrane protein in a supported binary lipid domain: direct observation of the organization of bacterical light-harvesting complex 2 by total internal reflection fluorescence microscopy, Langmuir 22, 5412–5418. 42. Schindelin, J., Arganda-Carreras, I., Frise, E., Kaynig, V., Longair, M., Pietzsch, T., Preibisch, S., Rueden, C., Saalfeld, S., Schmid, B., Tinevez, J.-Y., White, D. J., Hartenstein, V., Eliceiri, K., Tomancak, P., and Cardona, A. (2012) Fiji: an open-source platform for biological-image analysis, Nat Meth 9, 676–682. 43. Nečas, D., and Klapetek, P. (2012) Gwyddion. An open-source software for SPM data analysis, Open Phys 10, 181–188. 44. Sumino, A., Dewa, T., Kondo, M., Morii, T., Hashimoto, H., Gardiner, A. T., Cogdell, R. J., and Nango, M. (2011) Selective assembly of photosynthetic antenna proteins into a domain-structured lipid bilayer for the construction of artificial photosynthetic antenna systems: structural analysis of the assembly using surface plasmon resonance and atomic force microscopy, Langmuir 27, 1092– 1099. 45. Kirchhoff, H., Haferkamp, S., Allen, J. F., Epstein, D. B.A., and Mullineaux, C. W. (2008) Protein diffusion and macromolecular crowding in thylakoid membranes, Plant Physiol 146, 1571–1578. 46. Natali, A., Gruber, J. M., Dietzel, L., Stuart, Marc C. A., van Grondelle, R., and Croce, R. (2016) Light-harvesting complexes (LHC) cluster spontaneously in membrane environment leading to shortening of their excited state lifetimes, J Biol Chem, 730101. 47. Jennings, R. C., Garlaschi, F. M., and Zucchelli, G. (1991) Light-induced fluorescence quenching in the light-harvesting chlorophyll a/b protein complex, Photosynth Res 27, 57–64. 22 ACS Paragon Plus Environment

Page 22 of 25

Page 23 of 25 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Biochemistry

48. Barzda, V., Jennings, R. C., Zucchelli, G., and Garab, G. (1999) Kinetic analysis of the light-induced fluorescence quenching in light-harvesting chlorophyll a/b pigment–protein complex of photosystem II, Photchem Photbio 70, 751–759. 49. van den Brink-van der Laan, Els, Antoinette Killian, J., and Kruijff, B. de (2004) Nonbilayer lipids affect peripheral and integral membrane proteins via changes in the lateral pressure profile, Biochim Biophys Acta 1666, 275–288. 50. Sprague, S. G., and Staehelin, A. (1984) Effects of reconstitution method on the structural organization of isolated chloroplast membrane lipids, Biochim Biophys Acta 777, 306–322. 51. Reviakine, I., and Brisson, A. (2000) Formation of supported phospholipid bilayers from unilamellar vesicles investigated by atomic force microscopy, Langmuir 16, 1806–1815. 52. Richter, R. P., and Brisson, A. R. (2005) Following the formation of supported lipid bilayers on mica: a study combining AFM, QCM-D, and ellipsometry, Biophys J 88, 3422–3433. 53. Jass, J., Tjärnhage, T., and Puu, G. (2000) From liposomes to supported, planar bilayer structures on hydrophilic and hydrophobic surfaces: an atomic force microscopy study, Biophys J, 3153–3163. 54. Benes, M., Billy, D., Benda, A., Speijer, H., Hof, M., and Hermens, W. T. (2004) Surface-dependent transitions during self-assembly of phospholipid membranes on mica, silica, and glass, Langmuir 20, 10129–10137. 55. Rädler, J., Radmacher, M., and Gaub, H. E. (1994) Velocity-dependent forces in AFM imaging of lipid films, Langmuir 10, 3111–3115. 56. Aeffner, S., Reusch, T., Weinhausen, B., and Salditt, T. (2012) Energetics of stalk intermediates in membrane fusion are controlled by lipid composition, Proc Natl Acad Sci U S A 109, E1609–E1618. 57. Chernomordik, L. V., and Kozlov, M. M. (2008) Mechanics of membrane fusion, Nat Struct Mol Biol 15, 675–683. 58. Khattari, Z., Köhler, S., Xu, Y., Aeffner, S., and Salditt, T. (2015) Stalk formation as a function of lipid composition studied by X-ray reflectivity, Biochim Biophys Acta 1848, 41–50. 59. Yang, L., and Huan, H. W. (2002) Observation of a membrane fusion intermediate structure, Science 297, 1877-1879. 60. McDonnel, A., and Staehelin, A. (1980) Adhesion between lipsosomes mediated by the chlorophyll a/b light-harvesting complex isolated from chloroplast membranes, J Cell Biol, 40–56. 61. Fehr, N., Dietz, C., Polyhach, Y., Hagens, T. von, Jeschke, G., and Paulsen, H. (2015) Modelling of the N-terminal section and the lumenal loop of trimeric light harvesting complex II, J Biol Chem 43, 26007–26020. 62. Schaller, S., Latowski, D., Jemioła-Rzemińska, M., Dawood, A., Wilhelm, C., Strzałka, K., and Goss, R. (2011) Regulation of LHCII aggregation by different thylakoid membrane lipids, Biochim Biophys Acta 1807, 326–335. 63. Lee, A. G. (2000) Membrane lipids: it’s only a phase, Curr Biol 10, R377-R380. 64. Kollmitzer, B., Heftberger, P., Rappolt, M., and Pabst, G. (2013) Monolayer spontaneous curvature of raft-forming membrane lipids, Soft Matter 9, 10877–10884. 65. Booth, P. J., and Curnow, P. (2009) Folding scene investigation: membrane proteins, Curr Opin Struct Biol 19, 8–13. 66. Heidrich, J., Thurotte, A., and Schneider, D. (2017) Specific interaction of IM30/Vipp1 with cyanobacterial and chloroplast membranes results in membrane remodeling and eventually in membrane fusion, Biochim Biophys Acta 1859, 537–549. 23 ACS Paragon Plus Environment

Biochemistry 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

67. Murphy, D. J. (1982) The importance of non-planar bilayer regions in photosynthetic membranes and their stabilisation by galactolipids, FEBS Lett 150, 19–26. 68. Boudiere, L., Michaud, M., Petroutsos, D., Rebeille, F., Falconet, D., Bastien, O., Roy, S., Finazzi, G., Rolland, N., Jouhet, J., Block, M. A., and Marechal, E. (2014) Glycerolipids in photosynthesis: composition, synthesis and trafficking, Biochim Biophys Acta 1837, 470–480.

24 ACS Paragon Plus Environment

Page 24 of 25

Page 25 of 25 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Biochemistry

-For Table of Contents Use OnlyThe non-bilayer lipid MGDG and the major light-harvesting complex (LHCII) promote membrane stacking in supported lipid bilayers a

b

c

b

a,*

Dennis Seiwert , Hannes Witt , Sandra Ritz , Andreas Janshoff , Harald Paulsen

a

Institute of Molecular Physiology, Johannes Gutenberg University Mainz, Johannes-von-Müller-Weg 6, 55128 Mainz, Germany

b

Institute of Physical Chemistry, University of Goettingen, Tammannstrasse 6, 37077 Göttingen, Germany

c

Microscopy Core Facility, Institute of Molecular Biology, Ackermannweg 4, 55128 Mainz, Germany

*

Corresponding author

Telephone: 49-6131-3924633, E-mail: paulsen@uni-mainz.de

TOC

25 ACS Paragon Plus Environment