Isolating and Incorporating Light-Harvesting Antennas from Diatom Cyclotella Meneghiniana in Liposomes with Thylakoid Lipids

Summary

Here, we present a protocol to isolate fucoxanthin chlorophyll a/c binding proteins (FCP) from diatoms and incorporate them into liposomes with natural lipid compositions to study excitation energy transfer upon ion composition changes.

Abstract

The photosynthetic performance of plants, algae and diatoms strongly depends on the fast and efficient regulation of the light harvesting and energy transfer processes in the thylakoid membrane of chloroplasts. The light harvesting antenna of diatoms, the so called fucoxanthin chlorophyll a/c binding proteins (FCP), are required for the light absorption and efficient transfer to the photosynthetic reaction centers as well as for photo-protection from excessive light. The switch between these two functions is a long-standing matter of research. Many of these studies have been carried out with FCP in detergent micelles. For interaction studies, the detergents have been removed, which led to an unspecific aggregation of FCP complexes. In this approach, it is hard to discriminate between artifacts and physiologically relevant data. Hence, more valuable information about FCP and other membrane bound light harvesting complexes can be obtained by studying protein-protein interactions, energy transfer and other spectroscopic features if they are embedded in their native lipid environment. The main advantage is that liposomes have a defined size and a defined lipid/protein ratio by which the extent of FCP clustering is controlled. Further, changes in the pH and ion composition that regulate light harvesting in vivo can easily be simulated. In comparison to the thylakoid membrane, the liposomes are more homogenous and less complex, which makes it easier to obtain and understand spectroscopic data. The protocol describes the procedure of FCP isolation and purification, liposome preparation, and incorporation of FCP into liposomes with natural lipid composition. Results from a typical application are given and discussed.

Introduction

Photosynthetic organisms such as diatoms must cope with ever-changing light conditions and respond with sophisticated acclimation mechanisms that sustain high photosynthetic efficiency and protect from photo-oxidative damage caused by the excessive light. A major light-protective process in photosynthetic eukaryotes is the high energy quenching (q_E) of absorbed light that occurs as the main contribution to the non-photochemical quenching (NPQ) under light stress conditions^1,2,3. The light harvesting antenna complexes (LHC) are involved in the regulation of excitation energy transfer pathways. In response to high light induced low pH in the chloroplast lumen, the antenna system switches from the light harvesting state to the quenching state. This energy dissipative state protects photosystems (PS) and other complexes in the thylakoid membrane from photo-oxidation. In photosynthetic eukaryotes, the q_E is usually induced by two factors^1,2,3. One factor is the specialized light harvesting protein that responds to the low pH. The PsbS protein induces the q_E in higher plants⁴. LhcSRs⁵, modulated by PsbS activity, induce the q_E in green algae⁶. Diatoms possess Lhcx-like proteins which structurally related to LHCSRs^7,8,9,10.

The second factor of q_E is the xanthophyll cycle where carotenoids of the antenna are converted into a photo-protective form by de-epoxidation and reverted by epoxidation. In plants and green algae, violaxanthin is converted to zeaxanthin. In diatoms, diadinoxanthin is converted to diatoxanthin, which then correlates with the extent of NPQ¹¹. The diatom light harvesting antenna possesses some peculiarities although it is evolutionary related to plant and algal LHCs. The switch from light harvesting to photo-protection is enormously fast and the NPQ capacity is higher compared to plants¹². This might be one reason why diatoms are very successful in different ecological niches in a way that they are responsible for up to 45% of the oceanic net primary production¹³. Therefore, diatom light harvesting systems are an interesting object of photosynthesis research.

Diatoms, like the centric species Cyclotella meneghiniana, possess thylakoid intrinsic light harvesting systems named after the pigments they bind - fucoxanthin, chlorophyll (chl) a and c, hence FCP. Light harvesting proteins, such as FCPs, are embedded in the thylakoid membrane system comprising several membrane layers. Diatoms form bands of three thylakoids. This complex situation makes it difficult to study them on the molecular level in the thylakoid membrane. In addition, many components contribute to the regulation of light harvesting (see above). Therefore, in many approaches, the complexes were isolated from the membrane using mild detergents, such as n-Dodecyl-β-D-maltopyranoside (β-DDM), which solubilize the membrane but keep the FCP complexes intact. Many spectroscopic studies were performed using solubilized FCP to investigate intramolecular energy transfer^14,15,16,17. However, this former approach was limited since the regulation of energy transfer needs excitonic interaction with other antenna complexes or photosystems. Hence, these kinds of studies cannot be carried out with solubilized complexes because the interaction between complexes is lost.

An important feature in antenna regulation is the "molecular crowding" of the antenna and photosystems in the thylakoid membrane¹⁸. Formerly, a simple approach was carried out to simulate this effect in vitro. The detergent was removed, which leads to random aggregation of antenna complexes. Although some reasonable data was obtained by this approach^17,19, the detergent removal does not reflect the situation in vivo and has some limitations since the complexes are not interacting in their regular tertiary structure.

The use of liposomes overcomes several of the former limitations. The tertiary structure is still fully intact. The liposome membrane provides a quasi-native environment for the antenna complexes. The membrane separates the inside of the liposome from the outside environment. By these means, liposomes provide two reaction compartments for studies of ion and pH gradients as well as for transport processes. Further, the parameters of the experimental system can be controlled more easily for studies in the thylakoid membrane. Liposomes were already shown to be an excellent tool to study photosynthetic complexes. A major focus in the past was on plant LHC where the effect of altered lipid composition was tested on LHC II²⁰. In other approaches, protein-protein interaction between different LHC II were investigated²¹. Also, some studies in green algae were carried out that describe spontaneous clustering between LHC²². Considering the importance of diatoms for aquatic ecosystems, relatively few studies were performed with antenna complexes of diatoms. Two studies investigated the antenna complexes of the centric Cyclotella meneghiniana, where the clustering of the FCP antenna²³ and responsiveness of FCP to electrochemical gradients²⁴ were shown. Thus, liposomes are an excellent tool to study diatom antennas and their interaction and regulation in nearly native conditions. The liposomes are versatile since many conditions such as lipid composition, liposome size, protein density and the surrounding aqueous phase can be controlled. Furthermore, the method requires low amounts of samples. The experimental system is robust and highly reproducible. The compartmentalization of liposomes allows for studying pH and ion gradients, which are important factors in the regulation of antenna complexes.

Here, we describe the isolation of FCP antenna complexes from C. meneghiniana and their incorporation in liposomes with natural thylakoid lipid composition. Also, we provide exemplary data for the spectroscopic characterization of solubilized FCP and compare them with FCP in liposomes. The method summarizes knowledge and standardized protocols obtained from the improvements of Gundermann and Büchel 2012²³, Natali et al. 2016²², and Ahmad and Dietzel 2017²⁴.

Figure 1: Schematic representation of the workflow. (1) Refers to paragraph 1 which describes cell growth, disruption and thylakoid isolation with following FCP separation on sucrose density gradients; C. m.-Cyclotella meneghiniana cells. (2) Preparation of natural thylakoid lipid mixture (MGDG, DGDG and SQDG) described in paragraph 2 and creation of lipid-detergent micelles with octylglycoside (OG). A defined lipid-micelle size is achieved by extrusion using membranes of a defined pore diameter. FCP and lipid-micelles are unified at a predefined lipid: protein ratio and the detergents OG and β-DDM are removed via controlled dialysis forming FCP proteoliposomes. Please click here to view a larger version of this figure.

Protocol

Note: Photosynthetic complexes such as FCPs are highly vulnerable to light and heat. Always work on ice and under a very dim light.

1. Isolation of FCP from Cells

1. Thylakoid isolation from C. meneghiniana cells
   1. Grow C. meneghiniana in five 500 mL flasks each filled with 300 mL of ASP-medium^23,25 and 50 million cells. Plug the flasks with a cotton stopper and allow the cells to grow to the exponential growth phase for approximately one week on a shaker at 120 rpm with a 16 h light and 8 h dark phase at 40 µmol photons/(m²s) white light and a temperature between 15-18 °C. Check that the cell number is between 1.5-2 million cells/mL with a cell counter chamber.
   2. Centrifuge the cells at 4,000 x g in a precooled rotor with 500 mL centrifuge vials (4 °C) for 15 min in a high-speed centrifuge. Re-suspend the cell pellets in 12 mL of homogenization buffer (HB, Table 1) by pipetting.
      1. Transfer the suspension to a single 50 mL plastic tube. Store the samples at -80 °C or proceed to step 1.1.3.
   3. Pre-cool the bead mill and equipment. Fill the 50 mL beaker of the bead mill to 75% with a glass bead mixture and add the cell suspension. For cell disruption, use 7 x 45 s pulses at full speed with 30 s of cooling between each pulse. Take 20 µL of disrupted cells for quality checks in step 1.1.6.
   4. Filter the disrupted cells over a glass filter funnel and wash by pouring the HB over the glass beads until they appear clear. Pool the wash fraction with the filtrate. Keep the final volume lower than 150 mL.
   5. Centrifuge the sample for 15 min at 140 x g using three 50 mL plastic tubes to pellet the cell debris. Carefully transfer the supernatant to 20 mL polycarbonate ultracentrifugation vials and discard the pellet.
      1. Fill the vials with HB, equilibrate the weight, and centrifuge in a suitable rotor for 1 h at 300,000 x g and 4 °C to pellet the thylakoid membranes.
   6. Use the centrifugation time to check the proportion of disrupted cells by microscopy at 400X magnification with the 20 µL sample taken in 1.1.3. Calculate the ratio between chloroplast free and chloroplast containing frustules (silica shells).
      Note: Diatom cell walls are made of silica, which is visible as highly diffracting substance in the microscope. Chloroplasts occurring as green dots should have been released from the cells if the cell disruption worked.
   7. Re-suspend the membrane pellet with as little washing buffer as possible (0.5-1 mL) using a small painter’s brush. Fill the polycarbonate ultracentrifugation vials with washing buffer (Table 1), equilibrate their weight and centrifuge for 20 min at 200,000 x g and 4 °C.
      1. Re-suspend the washed membranes with a painter’s brush. Add washing buffer only if needed to keep the thylakoid concentration as high as possible. Pool all thylakoids in one sample vial (15 mL).
   8. Pre-dilute the samples using 10 µL of the sample with 90 µL of 100% acetone. Centrifuge it at 12,000 x g for 5 min to pellet precipitated protein. Take 10 µL of the predilution and mix it with 990 µL of 90% acetone.
      1. Measure the absorbance (ABS) of chlorophyll a and c at 664 nm and 630 nm in 90% acetone. Subtract the ABS_750nm from both values. Determine the total chlorophyll content using the following formula:²⁴
         1)       
         2)       
   9. Aliquot thylakoids in portions of 0.5 mg of total chlorophyll in a 1.5 mL reaction tubes, freeze them in liquid nitrogen and store them at -80 °C until further use.
2. Separation and concentration of FCP complexes
   1. Prepare a sucrose gradient solution and fill the ultracentrifuge tubes until the top minus the loading volume (300 – 500 µL). Freeze the tubes at -20 °C until they are completely frozen. Allow the tubes to thaw at +4 °C, which takes 3-4 h for a 17 mL tube.
      1. Repeat the freeze-thaw cycle twice to refine the gradient for better resolution.
   2. Use the samples obtained in 1.1.9 corresponding to 0.5 mg of chlorophyll and adjust with buffer B1 (Table 1) to a final volume of 2 mL. For solubilization, add n-Dodecyl-β-D-maltopyranoside (b-DDM) to a final concentration of 20 mM.
      1. Invert the tube 3 times and place it on ice for 20 min with gentle shaking to avoid foam. Centrifuge for 5 min, at 12,000 x g in a pre-cooled table top centrifuge at +4 °C.
   3. Load the supernatant on the gradient. Do not load more than 125 µg of total chlorophyll per gradient if 17 mL vials are used. Centrifuge for 22 h at 100,000 x g and +4 °C.
      1. Recover the desired brown FCP fractions from the gradient using a syringe (Figure 2A). Take a 5 µL aliquot and dilute it with 995 µL of B1a.
   4. Measure the absorbance (ABS) spectrum between 370 - 750 nm in a UV-VIS-spectrophotometer. Use semi-micro optical glass cuvettes.
      1. Open the spectra manager software, go to spectrum mode and record a baseline from 370 to 750 nm followed by sample measurements. Use the following settings: scan speed, 200 nm/min; data pitch, 0.5 nm; response, medium.
   5. Measure the volume of the recovered sample with a micropipette. Wash the FCP complexes by adding twice the recovered volume with B1 buffer (Table 1). Concentrate in a membrane concentrator with a 30 kDa cutoff at 1,000 x g and +4 °C to an ABS_672nm of at least 20.
      Note: The b-DDM concentration might rise due to micelle enrichment in the sample compartment. This could lead to further solubilization of the FCP complexes! Avoid over-solubilization by adding detergent free buffer B1 if further washing steps are required to remove residual sucrose.
   6. Take a 20 µL aliquot for the controls. Shock freeze the samples in liquid nitrogen and store them at -80 °C.

Figure 2: Purification of FCP, spectroscopic controls and purity check. (A) Typical appearance of a sucrose density gradient after overnight centrifugation. All brown bands contain the FCP pool consisting of FCPa and FCPb. pigm.- unbound pigments, PS - photosystems (B) Absorbance spectra of FCP before (blue line) and after (orange-dashed line) concentration using centrifugal filter devices with 30 kDa cutoff. Particularly, carotenoids are prone to loss from the FCP, which would result in lower absorbance in the region between 500-550 nm. Graphs are normalized to the chlorophyll Q_y maximum at ~670 nm. (C) Chlorophyll a emission spectra with excitation of chl c (465 nm) for testing the functional excitation energy transfer. If the energy transfer of chl c to chl a is hampered, an additional fluorescence band at ~640 nm (chl c) would occur. Graphs are normalized to the emission maximum. (D) Excitation spectra recorded at 675 nm (chl a fluorescence maximum) for testing the energy transfer to chl a from all pigments absorbing between 370 nm and 600 nm. If the energy transfer to chl a is less efficient, the fluorescence yield would decrease especially between 465 and 550 nm. The graphs are normalized to the maximum around 440 nm. The spectra in (B), (C) and (D) are nearly identical if the concentration worked well. (E) Check for purity of the isolated FCP using a Tris-tricine gel²⁸. FCPa and FCPb have subunits between 18-19 kDa. All visible silver-stained proteins larger than 20 kDa are contaminants. Thyl. - Thylakoids Please click here to view a larger version of this figure.

3. Spectroscopic and gel-based controls
   1. Record the absorbance between 370-750 nm of 5 µL of FCP in 995 µL of B1a after step 1.2.5. Compare it with the spectrum obtained in step 1.2.4. Use the same instrumentation and settings as described in step 1.2.4.
      1. Export data as *.csv and import the data into a spreadsheet. Normalize both spectra to the maximum of chl a in the Q_y band of around 672 nm as depicted in Figure 2B.
   2. Calculate the dilution factor of the samples from step 1.2.4 and 1.3.1 by dividing the measured ABS_672nm by the desired ABS_672nm of 0.03. Dilute the samples accordingly with B1a and transfer them to special glass fluorescence cuvettes.
   3. Record chl a fluorescence emission spectrum with a spectrofluorometer to reveal intact light energy transfer from chl c to chl a (Figure 2C).
      1. Use the spectrometer software and go to the spectrum measurement mode with the following settings: mode, emission; slit width excitation and emission, 3 nm; sensitivity, medium; scan speed, 100 nm/min; data pitch, 0.5 nm; excitation wavelength, 465 nm; emission, 600-800 nm. Perform autozero and measure.
   4. Record excitation spectra with the same sample and equipment to reveal intact energy transfer from all pigments to chl a (Figure 2D). Change settings to: mode, excitation; emission wavelength, 675 nm; excitation, 370-600 nm. Record the spectra.
      1. Correct with a rhodamine spectrum in the same range for spectral properties of the lamp – cf. instructions in the user manual.
   5. Mix FCP samples corresponding to 1 µg of chl a with 10 µL of SDS-loading buffer. Incubate for 10 min at 25 °C. Centrifuge for 5 min at 12000 x g in a table top centrifuge.
      1. Load the supernatant on a Tris-tricine gel²⁸. Separate it for 2 h at 150 V and silver-stain it after separation⁴⁰.
         Note: The FCP subunits separate into two prominent protein bands between 18-19 kDa, which are constituents of FCPa and FCPb²⁹ (Figure 2E).

2. Preparation of Liposomes and Incorporation of FCP

1. Preparation of the lipid mixture and lipid detergent micelles
   Note: Lipids are susceptible to warm temperatures combined with oxidative conditions. Try to keep the lipids chilled and under an N₂ atmosphere.
   1. Calculate the desired thylakoid lipid ratios for C. meneghiniana according to Vieler et al. 2007³⁰. Refer to the example given in Supplemental Table 1. Prepare the lipid stock solutions recommended by the manufacturer in a solvent proof container.
   2. Pipette the desired amount of lipids in a 2 mL reaction tube and evaporate the chloroform using a gentle nitrogen flow and try to spread the lipids over the whole area of the tube base. Let the N₂ flow until all solvent is evaporated.
   3. Solubilize the lipid mixture in 29 µL of n-octyl β-D-glucopyranoside solution (OG) at 4 °C for 4 h. Incubate the lipid mixture for 10 min at 30 °C. Incubate the lipids in a sonicator bath for 3 x 3 min at 25 °C interrupted by 30 s on ice.
      1. Add 221 µL of tricine buffer and 250 µL of 4x dialysis buffer.
   4. Use an extruder with 0.1 µm polycarbonate membranes for a defined liposome diameter of 50-70 nm. Assemble the extruder with the membrane and filter support. Avoid air bubbles and tighten the assembly thoroughly.
   5. Fill one syringe with 4x dialysis buffer and pre-wet the extruder until no bubbles can be seen in the second syringe.
   6. Apply the lipid-detergent-micelles to the extruder and press the solution from one syringe to the other forth and back. Repeat this step 5 times until the solution appears homogenous.
      Note: This solution can be stored at 4 °C for several days. Do not freeze!
2. Incorporation of FCP complexes and removal of detergents and aggregates
   Note: In this example we use a lipid/chl a ratio of 12:1, which corresponds to a lipid/protein ratio of roughly 100:1.
   1. Add FCP equal to 20 µg of chl a in a total volume of 500 µL of B1a buffer to 250 μL of the extruded lipid-micelles and 250 µL of 4x dialysis buffer. Incubate the samples for 3 x 3 min at 25 °C in a thermomixer at 1,500 - 3,000 rpm interrupted by a 30 s pause on ice.
   2. Cut the lid of four 1.5 mL reaction tube just under the top giving a ring which still fits on the lid. Prepare 1.5 cm x 1.5 cm pieces of dialysis membrane and wash them in 20 mL of 1x dialysis buffer
   3. Fill 250 µL of the sample to each lid. Carefully, lay the membrane on the lid so that the compartment is completely filled with the sample and no air bubbles occur. Tighten the reaction tube ring on the assembly in order to have a closed compartment.
   4. Dialyze the samples in 50 mL of 1x dialysis buffer overnight (12-16 h) on ice on a tumbling shaker. Replace the used dialysis buffer with fresh one and add 7 mg of adsorbent beads to remove the remaining detergents for at least 6 h.
   5. Replace the dialysis buffer again and dialyze for another 12 h. Recover the liposomes by piercing the dialysis membrane with a 200 µL micropipette tip and aspirate all liposomes from the reaction tube lid.
   6. Optional step: If high purity (>95%) is needed. Prepare a discontinuous density gradient in 17 mL ultracentrifugation vials with steps containing 6%, 10%, 15% and 20% sucrose epichlorhydrin copolymer in dialysis buffer. Load the liposomes on top and ultracentrifuge at 100,000 x g for 4 h in a swinging bucket rotor.
      1. Recover the upper brown band with a syringe, dilute the sample 1:5 with DP and proceed to the next step.
   7. Centrifuge the FCP liposomes in at least 2 mL of 1x dialysis buffer for 1.5 h at 100,000 x g and 4 °C. Recover the liposomes by turning the centrifuge tube at an angle of 45°. Allow the liposomes to move down for 1 min (Figure 3A).
   8. Recover the FCP liposomes in a final volume of 25-50 µL. Avoid disturbing the precipitate.
3. Controls 1: absorbance, fluorescence spectroscopy
   1. Add a 3 µL aliquot of the FCP-liposomes to a 1 mL final volume of 1x dialysis buffer and centrifuge for 5 min at 12,000 x g.
      1. Record the absorbance between 370 and 750 nm of FCP-liposomes with the same equipment as described in 1.2.4. Normalize the spectrum to the maximum in the Q_y region of chl a (670-680 nm) and compare it to the normalized spectrum of solubilized FCP (Figure 3C).
   2. Prepare FCP-liposomes in 1 mL of DP with an absorbance (ABS)=0.03 with respect to the maximum between 670-680 nm. Adjust the solubilized FCP in detergent from step 1.2.6 to the same ABS diluting with B1a.
   3. Record absorbance spectra of both samples as described in 1.2.4. Record chl a fluorescence emission spectrum of both samples as described in 1.3.3.
      Note: The fluorescence yield is decreased in the FCP-liposome sample (Figure 3D and cf. discussion.)

Figure 3: Isolation of FCP proteoliposomes followed by spectroscopic controls and confocal imaging. (A) Recovery of FCP liposomes after centrifugation. Turn the centrifugation tube to 45° and wait approximately 1 min - the liposomes will move down whereas the FCP aggregates which are not incorporated into liposomes stick to the tube wall. (B) Comparison of absorbance spectra of solubilized FCP in detergent (blue) and FCP in liposomes (orange) (C) The same spectra as in (B) normalized to chl a maximum in the red region (~670 nm - Qy peak); solubilized FCP in detergent (blue) and FCP in liposomes (orange). Potentially, there could be a pigment loss mainly of carotenoids visible in the 500-550 nm region. The clustering of FCP in the liposomes may lead to a peak broadening and a slight shift of the chl a maximum (~670 nm) to the red. (D) Emission spectra of solubilized FCP in detergent and FCP in liposome. Clustering of FCP in the liposome enhances energetic interactions of FCP complexes which lowers the fluorescence yield (orange curve) and shifts the emission maxima slightly to the red. Please click here to view a larger version of this figure.

Representative Results

The protocol describes the isolation of total FCP fraction from Cyclotella meneghiniana and incorporation into liposomes with native lipid composition. The thylakoid isolation is highly reproducible, but the thylakoid yield may change. The result is acceptable if more than 50% of all pigments are recovered in step 1.1.4. More than 80% is optimal.

The solubilization of the thylakoids is a critical step. Well-solubilized membranes are obtained if the supernatant after step 1.2.2 contains most of the pigments. The separation of FCP from photosystems is good if a clear brown (FCP) and a green (PS) band are distinguishable (Figure 2A). A prominent yellowish fraction on the top of the gradient is an indication for pigment loss due to over-solubilization. Pigment loss occurs in every preparation. It is acceptable to a certain degree as long as the functionality of the energy transfer to chl a is not affected (Figure 2C, 2D). Also, FCP enrichment step (1.2.4) might induce further solubilization and pigment loss due to concomitant detergent enrichment. Differences in the normalized absorbance spectra are an indication for pigment loss (Figure 2B), which often occurs in the spectral region between 500-550 nm where carotenoids absorb. The pigment loss is acceptable if the complexes are still functional. This means that no major differences occur in emission spectra and excitation spectra (Figure 2C, 2D). For the emission spectra, chl c is preferentially excited at 465 nm and transfers all the excitation energy to chl a, which emits at ~675 nm. If the chl c to chl a energy transfer is disturbed, chl c emits at ~640 nm. In the excitation spectra, the energy transfer to chl a is monitored by changing the excitation wavelength from 370 - 600 nm. In the spectral range from 370 - 440 nm mainly chl a is excited. Excitation around 465 nm favors chl c. In the range > 500 nm carotenoids, mainly fucoxanthins, are excited. The energy transmission is affected, if chl c and carotenoids contribute less to chl a fluorescence. Hence, the fluorescence yield decreases in these spectral regions in respect to chl a excitation between 370-440 nm.

The desired purity of the sample depends on the intended purpose of the study. For example, it is essential to separate photosystems from FCP for spectroscopic measurements while non-pigmented proteins interfere less with the measurements. Here, a purity of 95% is sufficiently high (Figure 2E). For mass spectrometric approaches exclusively, FCP subunits should be visible on a silver-stained gel.

The quality control of the liposome preparation is hard to monitor and can only be seen at the very end of the protocol (2.2.5, Figure 3A). The FCP complexes are destroyed if their color turns from brown to olive green. A good result is if a brown liposome fraction is collected. The percentage of liposomes recovered vs. aggregated pellet is optimally >70%. The incorporation of FCP in the liposomes may again lead to minor pigment loss. A comparison of normalized absorbance spectra before and after incorporation is seen if pigments were lost (Figure 3B, 3C). Some absorbance peaks of FCP-liposomes appear broader and the absorption in the red spectral region (~670 nm - the so called Q_y band) of chl a might be slightly red-shifted. This occurs due to clustering of FCP in the liposome. The chl a fluorescence yield arising from FCP in liposomes is largely reduced due to excitonic interaction in FCP clusters (Figure 3D, cf. discussion).

A typical experiment with liposomes makes use of the liposome's two reaction compartments separated by a lipid bilayer that allows pH or electrochemical gradients. In an example, the impact of a potassium ion gradient on FCP fluorescence yield was tested (Figure 4A). The fluorescence yield was measured in the presence of standard buffer. The fluorescence drops by 30% if KCl was added. The K⁺ gradient was released by the potassium specific ionophore valinomycin, which restores the FCP fluorescence to the initial level (Figure 4B). These changes in fluorescence corroborate the hypothesis that FCP complexes respond to potassium ion concentration gradients²⁴ in vivo.

Figure 4: Typical experiment with FCP liposomes showing the change of FCP fluorescence in response to an electrochemical gradient induced by KCl. (A) Experimental scheme: 1. FCP chl fluorescence in dialysis buffer. 2. Addition of 20 mM KCl inducing an electrochemical gradient. 3. Relaxation of the gradient by the K⁺ selective ionophore valinomycin (4 µM). (B) Drop of FCP fluorescence (675 nm) in response to a K⁺ gradient (blue) and restoration of the same fluorescence yield as before (orange) by relaxation of the gradient (grey). Please click here to view a larger version of this figure.

	buffers / mixtures	composition
cell growth &	Artificial seawater according to	86 mM NaCl
thylakoid preparation	Provasoli, 1957 (ASP)25	21 mM KCl
		4 mM Tris
		8.1 mM MgSO4
		11.8 mM NaNO3
		0.58 mM K3HPO4
		0.16 mM H3BO3
		2 mM silica
		pH 7.7 (H2SO4)
		autoclave
		add sterile: 2.72 mM CaCl2
		add sterile: 0.012 mM FeCl3
		add sterile: 0.092 mM EDTA
		add sterile: 0.02 mM MnCl2
		add sterile: 0.002 mM ZnCl2
		add sterile: 50 pM CoCl2
		add sterile: 25 pM Na2MoO4
		add sterile: 22.5 pM CuCl2
	homogenisation buffer	10 mM MES
		2 mM KCl
		5 mM EDTA
		1 M sorbitol
		pH 6.5 (KOH)
	washing buffer	10 mM MES
		2 mM KCl
		5 mM EDTA
		pH 6.5 (KOH)
	glass bead mixture	75% 1 mm beads
		25% 0.4 mm beads
	acetone	90% (v/v) with H2O
	liquid nitrogen
FCP purification	n-Dodecyl-β-D-Maltopyranoside	10 % (w/v) in H2O
	buffer B1	25 mM Tris
		2 mM KCl
		pH 7.4 (HCl)
	buffer B1a=B1 + b-DDM	0.03 % (w/v) final
	sucrose gradient solution	19 % sucrose (w/v) in B1a
liposome preparation	lipids see supplemental table 1
	N2 gas
	n-octyl β-D-glucopyranoside (OG)	10 % (w/v) in B1a
	tricine buffer	20 mM pH 7.5
	4x dialysis buffer (DB)	80 mM tricine pH7.5
		20 mM MgCl2
		0.04 % (w/v) NaN3

Table 1: List of buffers, stock solutions and mixtures for FCP preparation and incorporation into liposomes.

Supplemental Table 1: Example for lipid and chl a ratio calculation. A lipid/chl a ratio of 12 is used in the example given in Figure 1 and Figure 3. Please click here to download this file.

Discussion

FCP liposomes with natural lipid composition provide a handy, simple and reproducible tool to investigate spectroscopic properties in vitro. The lipid environment in FCP liposomes resembles the situation within the thylakoid membrane, giving rise to experimental results that are closer to natural conditions.

There are several advantages of using C. meneghiniana as a model system for FCP antenna. It grows relatively fast and is more robust in comparison to other diatom model species, e.g., Thalassiosira pseudonana. The cells break relatively easy in the bead mill, which allows for a high yield of intact thylakoid complexes. The major point for C. meneghiniana is the wealth of biochemical and spectroscopic data obtained in the last years^{16,17,23,27,29,31,32,33,34}. The only disadvantage is that an annotated genome is not available yet although genetic manipulations are possible³⁵.

The FCPs were prepared from cultures grown in optimal conditions. Hence, the yield, subunit composition and pigmentation of FCP complexes may change if they are prepared from stress conditions, e.g., high light²⁷ or nutrient limitation. The cell disruption in the bead mill (1.1.3) is crucial for the FCP yield and should be monitored after every change of growth condition by a microscope. The plastids should have been released from the cells so that empty frustules (silica shells) are prevailing. The next critical step is thylakoid solubilization. The efficiency of membrane solubilization interferes with the amount of storage lipids that accumulate under certain stress conditions³⁶. Therefore, the amount of detergent should be adjusted before processing samples obtained under stress conditions to avoid over-solubilization. This issue appears as increased chl c fluorescence emission at ~640 nm. Further, the energy transfer from carotenoids (500-550 nm) to chl a is reduced. This can be seen in the excitation spectrum (Figure 2C, 2D). Certain approaches might be useful to separate the FCPa from FCPb antenna complexes by ion exchange chromatography²⁷.

The transfer of FCP into liposomes is the next critical step. The transfer efficiency, the liposome homogeneity and intactness depend on several parameters. The lipid-micelle-preparation and extrusion provides homogenously sized liposome precursors. In this example (2.1.3-2.1.4), a polycarbonate membrane of 100 nm pore size was used to obtain a proteo-liposome size of ~50 nm. The amount of FCP added to the lipid-detergent micelles defines the protein: lipid ratio in the liposome. Hence, a defined number of FCP per liposome is obtained. In a similar approach plant and green algal LHC complexes clustered spontaneously. This fact was revealed by time-resolved fluorescence and single-molecule spectroscopy²² and provides a sound base for further experiments.

The controlled dialysis is also crucial for the incorporation of FCP into the liposomes. Although the incorporation of the hydrophobic FCP complex is energetically favored, the kinetics of insertion might compete with the kinetics of aggregation. This fact becomes more prominent the more hydrophobic and the larger the complexes are. In this example the smaller FCPa trimer is embedded faster than the larger FCPb nonamer. Proper FCP incorporation requires that the lipid micelles are well-mixed with solubilized FCP complexes before detergent removal via dialysis. Here, the mixing time in 2.2.1 can be varied. It is not recommended to rise the mixing temperature because this might lead to FCP disintegration. Further, the "speed" of detergent removal is important and can be adjusted by the amount of the adsorbent (2.2.4). The final size of the liposomes might vary. Dynamic light scattering, analytical ultracentrifugation and electron microscopy are suitable methods to determine the liposome size range^37,38. For this protocol with natural thylakoid lipid composition and lipid micelle extrusion, we expect the aforementioned diameter of 50 - 80 nm and a nearly sphere like shape revealed by cryo-electron microscopy. These results were obtained with the very same method incorporating green algal LHC²². Liposomes of such small diameter have a strong surface curvature of ~10° per 5 nm which may bend larger antenna complexes and force them to monomerize. This situation mimics the natural curvature in grana margins of algae and plants²². Compared to grana margins, the major part of the thylakoid appears rather flat. Therefore, if larger complexes such as photosystems shall be studied it is recommended to use liposomes of a larger diameter. Another question which arises is: What is the orientation of the complexes? In case of FCPa, we conclude from measurements in response to low pH²⁴ that the orientation is about 50% right-side out and 50% inside out. In the future, methods are required to force the complexes in one direction. In other studies with the bacteriorhodopsin efflux pump, nearly all proteins are spontaneously oriented in the same way (right-side out)³⁹.

There is the question of what induces the fluorescence change when K⁺ is added to the liposomes. Several answers are imaginable. If the K⁺ concentration alone is responsible, the experiment in Figure 4 would result in a further decrease of fluorescence if valinomycin is added after KCl. In a recent study, we provided support for the notion that the K⁺ concentration gradient is responsible for the fluorescence change of FCPa²⁴. Several other factors might influence the fluorescence yield as well. These could be osmotic or surface charge dynamics effects²⁴.

To sum up, the advantage to investigate light harvesting antenna complexes such as FCP in liposomes is that it provides a functional system close to in vivo conditions. But the number of components is still small, which reduces complexity in the experimental results. It allows for simulations of light stress conditions and induction of photo-protective mechanisms. Further, the role of certain lipid species on the function of FCP complexes can be studied. The potential of this method increases with the recently made progresses in cryo-electron microscopy, single molecule spectroscopy and other time resolved spectroscopic methods. This will enable researchers to improve studies of energy transfer in light harvesting systems. Combined with these techniques, the knowledge about FCP in the solubilized and aggregated state can be complemented with data about the contribution of the lipid environment to FCP function.

Materials

Name	Company	Catalog Number	Comments
500 ml centrifuge vials
high speed centrifuge	Heraeus
Bead Mill VI 2	Edmund-Bühler (edmund-buehler.de)		newer version: Vibrogen-Zellmühle Vl 6
Silibeads S 400 µm	Sigmund-Lindner.com	5223-7
Silibeads S 1,-1,3 mm	Sigmund-Lindner.com	4504
VitraPOR filter funnel - por1	ROBU GmbH	21121
polycarbonate ultracentrifuagtion vials (30 mL) for T-865	Beranek Laborgeräte (Laborgeraete-beranek.de)	314348
Ultracentrifuge Discovery 90SE	Sorvall	n.a.
rotor T 865	ThermoFisher Scientific (thermofisher.com)	51411
Neubauer Cell Counter Chamber (improved)	Carl Roth Laborbedarf (Carlroth.com)	T729.1
Zeiss Mikroskop Primostar (7)	Optik-Pro (optik-pro.de)	51428
optical glass cuvettes (6040-OG)	Hellma Analytics (hellma-analytics.com)	"6040-10-10"
V-630 UV-VIS Spectrophotometer (incl. software)	Jasco (jasco.de)	V-630
n-Dodecyl-β-D-Maltopyranoside	ANATRACE (anatrace.com)	D310LA
Ultra-Clear tubes 17 ml for AH629	Beranek Laborgeräte (Laborgeraete-beranek.de)	344061
rotor AH629-17-mL	ThermoFisher Scientific (thermofisher.com)	54285
Membrane concentrator_Centriprep 30 kDa cutoff	Millipore (merckmillipore.com)	4307
Biometra Minigel-Twin	Analytik Jena AG (analytik-jena.de)	846-010-100
Silver Stain Plus Kit	Bio-Rad (bio-rad.com)	1610449
libre office spread sheet	The document foundation	https://de.libreoffice.org/download/libreoffice-still/
special glass cuvettes for fluorescence (101-0S)	Hellma Analytics (hellma-analytics.com)	101-10-20
Spectrofluorometer FP-6500 (incl. Software)	Jasco (jasco.de)	FP-6500
SDS-loading buffer Roti-Load	ROTH (carlroth.com)	K929.1
n-octyl β-D-glucopyranoside	ANATRACE (anatrace.com)	O311
Monogalactosyl Diaclyglycerol (MGDG)	Larodan AB (larodan.com)	59-1300	make stock solution in chloroform
Digalactosyl Diacylglycerol (DGDG)	Larodan AB (larodan.com)	59-1310	make stock solution in chloroform
Sulphoquinovosyl Diacylglycerol (SQDG)	Larodan AB (larodan.com)	59-1230	make stock solution in chloroform
L-alpha-Phosphatidylglycerol (PG)	Larodan AB (larodan.com)	37-0150	make stock solution in chloroform
L-α-Phosphatidylcholine	Sigma-Aldrich (sigmaaldrich.com)	P3782 SIGMA	make stock solution in chloroform
sonicator bath S-50TH	Sonicor (getmedonline.com	SONICOR-S-50TH
mini-Extruder	Avanti Polar Lipids (Avanti.com)	610000
Nuleopore polycarbonate membrane	Avanti Polar Lipids (Avanti.com)	610005
dialysis membrane Visking 14 kDa cutoff	ROTH (carlroth.com)	0653.1	boil in destilled water before use
Biobeads SM2 Adsorbent	Biorad (Bio-rad.com)	152-3920
sucrose epichlorhydrin copolymer - Ficoll 400	Sigma-Aldrich (sigmaaldrich.com)	F4375
Polycarbonate ultracentrifuagtion vials (2.7 mL) for TFT 80.4	Beranek Laborgeräte (Laborgeraete-beranek.de)	252150
rotor TFT 80.4	Millipore (merckmillipore.com)	54356
material listed in order of appearance			For specific safety instructions please refer to material safety sheets and repective manuals. Standard lab material and substances are not listed.