Present address: Institut f. Allgemeine Botanik und Pflanzenphysiologie, Friedrich-Schiller-Universität Jena, Dornburger Str. 159, 07743 Jena, Germany.
Towards functional proteomics of membrane protein complexes: analysis of thylakoid membranes from Chlamydomonas reinhardtii
Article first published online: 29 DEC 2001
The Plant Journal
Volume 28, Issue 5, pages 595–606, December 2001
How to Cite
Hippler, M., Klein, J., Fink, A., Allinger, T. and Hoerth, P. (2001), Towards functional proteomics of membrane protein complexes: analysis of thylakoid membranes from Chlamydomonas reinhardtii. The Plant Journal, 28: 595–606. doi: 10.1046/j.1365-313X.2001.01175.x
- Issue published online: 29 DEC 2001
- Article first published online: 29 DEC 2001
- Received 9 May 2001; revised 13 August 2001; accepted 30 August 2001.
- thylakoid membranes;
- light-harvesting proteins;
- two-dimensional gel electrophoresis;
- mass spectrometry;
- Chlamydomonas reinhardtii
Functional proteomics of membrane proteins is an important tool for the understanding of protein networks in biological membranes but structural studies on this part of the proteome are limited. In this study we undertook such an approach to analyse photosynthetic thylakoid membranes isolated from wild-type and mutant strains of Chlamydomonas reinhardtii. Thylakoid membrane proteins were separated by high-resolution two-dimensional gel electrophoresis (2-DE) and analysed by immuno-blotting and mass spectrometry for the presence of membrane-spanning proteins. Our data show that light-harvesting complex proteins (LHCP), that cross the membrane with three transmembrane domains, can be separated using this method. We have identified more than 30 different LHCP spots on our gels. Mass spectrometric analysis of 2-DE separated Lhcb1 indicates that this major LHCII protein can associate with the thylakoid membrane with part of its putative transit sequence. Separation of isolated photosystem I (PSI) complexes by 2-DE revealed the presence of 18 LHCI protein spots. The use of two peptide-specific antibodies directed against LHCI subunits supports the interpretation that some of these spots represent products arising from differential processing and post-translational modifications. In addition our data indicate that the reaction centre subunit of PSI, PsaA, that possesses 11 transmembrane domains, can be separated by 2-DE. Comparison between 2-DE maps from thylakoid membrane proteins isolated from a PSI-deficient (Δycf4) and a crd1 mutant, which is conditionally reduced in PSI and LHCI under copper-deficiency, showed the presence of most of the LHCI spots in the former but their absence in the latter. Our data demonstrate that (i) hydrophobic membrane proteins like the LHCPs can be faithfully separated by 2-DE, and (ii) that high-resolution 2-DE facilitates the comparative analysis of membrane protein complexes in wild-type and mutants cells.
An estimation of the size of the Arabidopsis chloroplast proteome has indicated that, in addition to the 87 plastid-encoded proteins, probably about 2300 nuclear-encoded proteins are present in the chloroplast (Abdallah et al., 2000). It has been suggested that a large number of these proteins are related to photosynthesis. The thylakoid membranes within the chloroplast are the subcompartment in which the primary reactions of photosynthesis occur. These reactions are performed by about 100 proteins that are organized in the four major multisubunit protein complexes, photosystem I (PSI), PSII, the ATP-synthase complex and cytochrome b6/f complex. In addition to proteins that function in the conversion of light energy to chemical energy, several other proteins are present in the membrane that are required for assembly, maintainance and regulation of the four multiprotein complexes (for reviews see Choquet and Vallon, 2000; Merchant and Dreyfuss, 1998; Snyders and Kohorn, 1999; Vener et al., 1998; Wollman et al., 1999; Adam, 2000). A global analysis of thylakoid membrane proteins is a way to establish number and identity of proteins localized to this subcompartment. The standard technique for quantitative proteome analysis combines protein separation by high-resolution (isoelectric focusing/SDS-PAGE) two-dimensional gel electrophoresis (2-DE) with mass spectrometry (MS) or tandem MS (MS/MS) identification of selected protein spots. Such an approach was taken recently to analyse lumenal and peripheral thylakoid proteins (Peltier et al., 2000). From these studies, it was estimated that at least 200–230 different lumenal and peripheral proteins do exist. For analysis of membrane spanning thylakoid proteins, one-dimensional SDS-PAGE and MS/MS were combined to analyse reversible protein phosphorylation (Michel et al., 1991; Vener et al., 2001). Coupling of liquid chromatography with electrospray-ionization MS and MS/MS succeeded in the characterization of intact intrinsic thylakoid membrane proteins (Whitelegge et al., 1998; Corradini et al., 2000). However, no two-dimensional map of intrinsic thylakoid membrane proteins has been established so far. The separation of hydrophobic intrinsic membrane proteins by 2-DE appears to be difficult. As reviewed by Santoni et al. (2000), in none of the 2-DE-based proteomic approaches using yeast, tabacco or Arabidopsis, transmembrane proteins could be identified by MS or MS/MS. However, a limited number of membrane proteins from animals have been resolved on 2-DE gels and identified (Santoni et al., 2000).
In this study we used a proteomics approach to investigate photosynthetic membranes isolated from the eukaryotic green alga Chlamydomonas reinhardtii. Chlamydomonas reinhardtii is an excellent model system to study bioenergetic processes (reviewed in Hippler et al., 1998; Merchant et al., 1999), as numerous mutants exist that are affected in photosynthesis due to chloroplast or nuclear mutations. In addition methods for both nuclear (Debuchy et al., 1989; Kindle et al., 1989; Mayfield and Kindle, 1990) and chloroplast transformation (Boynton et al., 1988) are well established. The recent generation of an EST database for C. reinhardtii (80 000 records at NCBI, July 2001) makes this organism very suitable for proteomics. Thus, the development of a 2-DE system to analyse thylakoid membranes from C. reinhardtii would be an important tool to comparatively analyse wild-type and mutant strains. Here, we succeeded in separating thylakoid membrane proteins from C. reinhardtii by high-resolution 2-DE. Analysis of the 2-DE maps by immuno-blotting and mass spectrometry demonstrates the presence and separation of light-harvesting proteins, which proves that such membrane-spanning proteins can be reliably analysed by such a high-resolution method. In addition our data indicate that proteins that are more hydrophobic than the light-harvesting complex proteins (LHCPs), such as the PsaA polypeptide, that possesses 11 transmembrane domains, can be separated by 2-DE. We also show 2-DE separation of thlyakoid membrane proteins from two photosynthetic mutants, a PSI-deficient strain (Δycf4) and a strain (crd1) that is conditionally reduced in PSI and LHCI under copper-deficiency. This analysis reveals that most of the LHCI spots are present in the PSI-deficient mutant but absent from thylakoids isolated from the crd1 mutant grown under copper-deficiency. Our data demonstrate that (i) hydrophobic membrane proteins such as the LHCPs can be faithfully resolved in 2-DE, and (ii) that the high-resolution 2-DE technique allows comparative analysis of two-dimensional protein maps of wild-type and mutant thylakoids. The methods described here paves the way to functional proteomics of membrane protein complexes.
LHCPs can be separated by 2-DE
In this study we showed the separation of thylakoid membranes and associated proteins by 2-DE isolated from wild-type cells of C. reinhardtii (Figure 1). For the resolution of membrane proteins on 2-DE the solubilization of proteins as well as the sample fractionation appear to be very critical (reviewed in Molloy, 2000). To prepare protein extracts from thylakoid membrane proteins suitable for 2-DE, the proteins were separated from pigments and lipids by chloroform/methanol precipitation. When the chloroform/methanol and the water/methanol supernatants were tested for the presence of proteins by the bicinchoninic acid method no protein could be determined in the water/methanol fraction whereas the chloroform/methanol supernatant contained less than 4% of the total amount of protein applied to precipitation, indicating that the precipitation procedure does not lead to a significant loss of proteins. This extraction procedure and protein precipitation subsequently allowed the analysis of thylakoid membrane proteins by 2-DE. For high-resolution protein separation of thylakoids on 2-DE we found that the addition of thiourea to the dissolving buffer was essential, as described in Pasquali et al. (1997). The resolution and reproducibility of protein spots obtained by 2-DE improved when the non-ionic detergent β-dodecyl maltoside was added to the dissolving buffer. Separation of wild-type thylakoid membrane proteins by 2-DE resulted in an average resolution of about 350–400 spots on each gel (from about 15 gels), about 80% of which were fully reproducible. From a representative 2-DE map as shown in Figure 1, 370 spots could be evaluated after separation of wild-type thylakoid proteins and silver-staining of the gel. To test for resolution of hydrophobic proteins we focused our analysis on the presence of LHCPs using immuno-blotting and mass spectrometry for protein identification.
LHCP have three transmembrane domains. They belong to the class of proteins that span the membrane with α-helical transmembrane domains only (Kuhlbrandt et al., 1994). How hydrophobic are LHCPs? The calculation of the grand average hydropathy index (GRAVY) score (Kyte and Doolittle, 1982) for Lhcb1 and Lhcb2, the major light-harvesting proteins of PSII, gives values of −0.097 and −0.128 for the precursor protein and the mature form of Lhcb1 (Imbault et al., 1988), respectively, and values of 0.114 and 0.101 for the precursor protein and the mature form of Lhcb2 (deduced from the longest EST clone present in the database AV390087, which encodes 189 amino acids (AS)), respectively. For the most abundant LHCI protein (see below), a Lhca1 [deduced from the longest EST clone found in the database AW661131, which encodes 200 AS and corresponds to p22.1 (Bassi et al., 1992)] GRAVY scores of 0.007 and 0.047 can be determined for the precursor and the mature protein, respectively. Proteins with slightly negative or positive GRAVY scores that possess transmembrane domains are difficult to separate by 2-DE (Santoni et al., 2000). However, immuno-blot analysis of LHCPs using polyclonal and peptide antibodies as well as mass spectrometric analysis demonstrates that LHCPs can be well separated by 2-DE. The immuno-blot analysis of 2-DE separated wild-type thylakoids using anti-LHCII (anti-P25K; Michel et al., 1983) and anti-LHCI antibodies (anti18.1 (LHCa4); Bassi et al., 1992) are shown in Figure 1 (lower panels). Using anti-LHCII antibodies 13 protein spots can be identified (spots 12–24). These spots match with highly abundant protein spots on the silver-stained gel (see numbering in Figure 1). The anti-LHCI antibodies (anti18.1) recognize 12 proteins spots on the immuno-blot (Figure 1, spots 1–11, 25), which also can be attributed to spots on the silver-stained gel. Spots with a higher molecular mass than 27.2 kDa, which are recognized by the anti18.1 antibodies, correspond most likely to the minor LHCII protein CP29 (see boxed region in Figure 1, lower panel) since this LHCII protein has a high homology to p18.1 (Bassi et al., 1992). One of the major LHCII spot (spot 22), with the most basic isoelectric point (IP), was cut out of the gel, digested in-gel by trypsin and analysed by mass spectrometric collision induced fragmentation, which resulted in several peptide masses and sequence tags (Figure 2a; Table 1). Peptide masses were obtained when the mass spectrometer was used in a parent ions scan mode (Wilm et al., 1996a) searching for Y1 ions on 147 Da (Y1-ion of Lys) and for immonium ions on Leu/Ile (86 Da). A sequence tag consists of the following information (Wilm et al., 1996b), a peptide mass and an internal amino acid sequence together with corresponding fragment masses. One sequence tag has a peptide mass of 2242 Da and a short amino acid sequence with the corresponding fragment masses of (439.7)VT(640). Another tag has a peptide mass of 1356.7 Da and an internal amino acid sequence together with the corresponding fragment masses of (983.0)KTG(1269.0) (Figure 2a). These sequence tags together with the information of other peptide masses (Table 1) identify the protein as an lhcb1 gene product of C. reinhardtii (Imbault et al., 1988) using peptide search (http://www.mann.embl-heidelberg.de/Services/PeptideSearch/). Analysis of tryptic peptides from spot 24 by MS/MS using parent ions scan on 147 and 86 Da also identified this spot as Lhcb1 (Table 1). According to sequence alignments, the N-terminal sequence of the mature Lhcb1 starts at position 28 with Ala as first amino acid (Bassi et al., 1992), defining the sequence from position 1–27 as putative transit peptide. Interestingly, some of the tryptic peptides identified by mass spectrometry match with the deduced peptide sequences of lhcb1 from position 9–21.The fact that Lhcb1, containing Lys9, is also found with a more acidic IP (see Table 1; spot24) could be explained by phosphorylation of the protein, which has been reported to occur in the light (Wollman and Delepelaire, 1984).
|Spot 2 M/z||Mass (neutral) (Da)||Delta mass (monoisotopic) (Da)||Identified as||Peptides (Lhca1)|
|368.9 (+ 4)||1471.6||− 0.7||Y1-Scan (K)||166–179|
|390.9 (+ 4)||1559.6||− 1.3||Y1-Scan (K)||170–182|
|637.2 (+ 2)||1272.4||− 1.2||Y1-Scan (K)||74–83|
|774.5 (+ 1)||773.5||− 0.0||Y1-Scan (K)||170–176|
|948.0 (+ 2)||1894||− 0.1||Y1-Scan (K)||146–165|
|1190.1 (+ 1)||1189.1||− 1.7||Y1-Scan (K)||166–176|
|Spot 22 M/z||Mass (neutral) (Da)||Delta mass (monoisotopic) (Da)||Identified as||Peptides (Lhcb1)|
|381.0* (+ 2)||762.0||− 0.6||L/I-Scan||10–16|
|445.6* (+ 2)||889.2||− 0.3||Y1-Scan (K)||9–16|
|478.5 (+ 3)||1432.5||− 0.7||L/I-Scan||9–21|
|679.3* (+ 2)||1356.7||− 1.0||Y1-Scan (K)||109–120|
|892.2* (+ 1)||891.2||− 1.7||L/I-Scan||9–16|
|902.8* (+ 2)||1803.6||− 2.7 (1.4)||L/I Scan||92–108|
|1248* (+ 1)/||1247.0/1248.8/||− 0.1/1.1/0.9||L/I-Scan||9–20|
|625.4 (+ 2)/||1248.6||10–21|
|417.2 (+ 3)|
|1027.8* (+ 3)||3080.4||− 1.2||Y1-Scan (K)||82–108|
|1122.1 (+ 2)||2242.2||− 0.6||Y1-Scan (K)||9–21|
|1177.1 (+ 1)||1176.1||− 0.5||L/I-Scan||10–20|
|Spot 24 M/z||Mass (neutral) (Da)||Delta mass (monoisotopic) (Da)||Identified as||Peptides (Lhcb1)|
|601.1 (+ 3)||1800.3||− 0.6||L/I-Scan||92–108|
|891.2 (+ 1)||890.2||− 0.7||Y1-Scan (K)||9–16|
|941.8 (+ 1)||940.8||− 1.7||L/I-Scan||199–206|
|955.9 (+ 1)||954.9||− 0.7||Y1-Scan (K)||21–30|
|979.7 (+ 1)||978.7||− 0.8||L/I-Scan||84–91|
|1122.6 (+ 2)||2243.2||− 0.4||Y1-Scan (K)||9–21|
Lhcb1 may associate with the thylakoid membrane and assemble into trimeric LHCII with part of its putative transit peptide
To verify the finding that a fraction of the Lhcb1 may contain part of its putative transit sequence when the protein associates with the thylakoid membrane immuno-dot blot analyses were performed. Peptides, that correspond to the first 16 N-terminal amino acids of the mature Lhca3 and 4 proteins (p14.1 and p15.1, respectively) and to the first 25 amino acids of the transit peptide of the Lhcb1 protein, were spotted onto nitrocellulose. The dot blot was then decorated with p11 antibodies which recognize LHCII proteins. These antibodies were raised against the p11 protein that was isolated from SDS-PAGE fractionated thylakoids of C. reinhardtii (Vallon et al., 1986). The antibodies only recognize the putative transit peptide of Lhcb1 but not the LHCI peptides (data not shown). In a next step we raised antibodies against the Lhcb1 peptide, which corresponds to the first 25 amino acids of the putative transit peptide of the Lhcb1 protein. When thylakoid membranes were separated by SDS-PAGE, these peptide antibodies recognize the p11 protein on immuno-blots (data not shown). The Western blot analysis of 2-DE separated thylakoids, using the anti-Lhcb1 peptide antibodies, clearly shows that spots 21, 22, 23 and 24 correspond to the longer form of the Lhcb1 protein (data not shown). This strongly supports the mass spectrometric results that thylakoid membrane-associated Lhcb1 may contain part of its putative transit peptide. To check whether pre-processed Lhcb1 assembles into functional LHCII trimers, we isolated trimers by green gel deriphat electrophoresis from wild-type thylakoid membranes (Hippler et al., 2000; Peter and Thornber, 1991) performed in-gel protein digestion with trypsin and analysed the resulting peptides by mass spectrometry. Parent ions scans on Lys as Y1 (147 Da) and on the immonium-ion of Leu/Ile (86 Da) identified a peptide with a mass of 888.7 and 888.6 Da, respectively, that correspond to the expected peptide mass of the tryptic Lhcb1 peptide 9–16 of 889.5 Da with a mass deviation of −0.8 and −0.9 Da, respectively. Mass spectrometric collision induced fragmentation of this peptide (m/z 445. 6; charge + 2; Figure 2b) and revealed several sequence tags which identify the peptide as the Lhcb1 tryptic peptide 9–16 (Figure 2b).
Characterization of LHCI proteins by 2-DE
For further characterization of LHCI proteins, enriched PSI particles were separated by 2-DE and analysed by silver-staining and immuno-blotting (Figure 3). The immuno-blots were probed with LHCI antibodies anti-18.1, anti-17.2, anti-15.1 and anti-14.1. Antibodies anti-15.1 and anti-14.1 are peptide antibodies and directed against the first 16 N-terminal amino acids of mature p14.1 (Lhca3) and p15.1 (Lhca4). These peptide sequences are unique to these LHCI proteins so that the antibodies should specifically recognize the corresponding proteins but not other members of the LHC protein family. When the anti-14.1 peptide antibodies are used to probe the Western blot of 2-DE separated enriched PSI-particles, four spots can be detected (Figure 3, lower panels, spots 3, 4, 5 and 26), that correspond to spots 3, 4, 5 which are also recognized by the anti-18.1 antibodies (see Figure 1, lower panel). It is of note that these spots do not differ in size but only in their IEP. On the silver-stained 2-DE map of wild-type thylakoids these spots are covered by LHCII proteins (Figure 1) but can be visualized on the silver-stained 2-DE map of enriched PSI particles (Figure 3, upper panel). The second peptide antibody anti-15.1 is less sensitive but recognizes two spots on the immuno-blot of 2-DE separated PSI particles, which correspond to spots 9 and 11 that are also identified with the anti-18.1 antibodies. Using anti-17.2 antibodies (Bassi et al., 1992), that also react against LHCI proteins, three further spots (6, 27 and 28) are detected on the PSI immuno-blot (Figure 3, lower panel). This antibody also strongly cross reacts with LHCII proteins (Hippler et al., 2000). The fact that no spots corresponding to LHCII proteins are found on the PSI immuno-blot indicate that these proteins are absent from the enriched PSI preparation. When this blot is probed with anti-18.1 antibodies, 12 spots, which already had been recognized on the 2-DE immuno-blot of thylakoids, are identifed. In addition three further minor spots (spots 29, 30 and 31) are detected. The 15 spots are enriched in 2-DE of PSI compared to the 2-DE of thylakoids, which indicates that these proteins co-purify with PSI and suggests that they belong to the antenna that is functionally connected to PSI.
The most abundant LHCI protein, spot 2, was further analysed by mass spectrometry. From tryptic in-gel digestion of spot 2 and mass spectrometric collision-induced fragmentation several sequence tags were obtained (data not shown). One tag was identified with a peptide mass of 1191.6 Da and two internal amino acid sequence with its corresponding fragment masses of (147.1)LK(388.6) and (1022.3)GPY(1339.4). For the other tag, a peptide with a mass of 1894 Da, fragmented into an internal sequence with following fragment masses (147.0)AF(365.2). When the mass spectrometer was used in a parent ion scan mode several other peptide masses were obtained (Table 1). Using mascot (http://www.matrixscience.com/), these sequence tags together with other peptide masses identify the protein as lhca1 (p22.1; Bassi et al., 1992) from Chlamydomonas EST sequences (Grossman et al., 2000). In total 96 EST clones representing the lhca1 gene were identified. From the EST clone with the longest open reading frame found, 200 amino acids can be deduced (AW661131).
A total of 18 LHCI spots are enriched in 2-DE of PSI compared to the 2-DE of thylakoids, which indicates that these proteins co-purify with PSI and suggests that they belong to the antenna that is functionally connected to PSI. The fact that several spots with similar molecular masses but different IPs could be separated, shows that an exact determination of LHCI stoichiometry by one-dimensional SDS-PAGE is not possible. To determine the relative stoichiometry of the LHCI subunits, ‘spot volumes’ (using Phoretix 2-D software, Phoretix International, Newcastle-upon-Tyne, UK) of the individual LHCI proteins were evaluated after staining of 2-DE separated PSI particles with colloidal Coomassie. In the Coomassie stain, LHCI spots, 1, 2, 3, 4, 6, 8, 9, 10 and 25 could be visualized (data not shown). This staining procedure has a higher dynamic range in staining of proteins than silver-staining. From this analysis (three independent gels were evaluated) it appears that spot 2, a Lhca1 protein, is the most abundant LHCI subunit. When LHCI spots with the lowest spot volume values (spots 6, 9 and 10), are set to 1, the Lhca1 protein is about 50 times more abundant than these low abundant spots. Spots 3 and 4 that correspond to the Lhca3 proteins are each about 30-fold and spots 1, 8 and 25 are about four-fold enriched compared with spots 6, 9 and 10.
The 2-DE analysis of thylakoid membranes isolated from a PSI-deficient mutant.
The potential to separate thylakoid membrane proteins by 2-DE opens up the possibility of starting a differential analysis of membranes from mutant and wild-type strains. One advantage of C. reinhardtii as a phototrophic model organism is the availability of mutant strains that are affected in photosynthesis. In this study, we have analysed thylakoid membranes isolated from a PSI-deficient mutant, a Δycf4 strain, and a crd1 mutant strain by 2-DE. The Δycf4 strain is a knock-out mutant lacking the chloroplast encoded ycf4 gene, which results in deficiency of PSI (Boudreau et al., 1997). The crd1 mutant is conditionally reduced in PSI and LHCI under copper-deficiency (Moseley et al., 2000). The 2-DE maps are shown from a pH range of 3.5–8 and a molecular mass range from about 45 to 10 kDa to focus on LHC proteins (Figure 4). A comparison between the two maps indicates clear differences. LHCII spots (12–24) can be identified in both 2-DE separations. However, LHCI spots are strongly diminished (spots 2, 7, 8) or absent (spots 1, 3, 6, 9, 10, 11, 25) in the 2-DE map of thylakoids isolated from copper-deficient crd1 cells, whereas these spots are present in the map of thylakoids isolated from the PSI-deficient mutant. In 2-DE maps from crd1 thylakoids isolated from copper-supplemented cells or from thylakoids isolated from copper-deficient Δycf4 cells, all LHCI spots are present (data not shown). It has been observed previously that strains carrying deletion mutations in the core polypeptides of PSI (Girad-Bascou et al., 1987; Takahashi et al., 1991) still contain LHCI proteins. Our 2-DE results are compatible with this finding, in addition our results reveal the presence of spots 1, 2, 6, 7, 8, 9, 10, 11 and 25 (see numbering; Figure 4), thus proving that these LHC proteins accumulate with the thylakoid membranes independently from the PSI core complex. These results also show that the positions of LHCI spots on the 2-DE maps are independent of the presence of the PSI complex and hence are not due to incomplete solubilization of PSI complexes. Immuno-blotting results indicated the absence of LHCI subunits when crd1 mutant cells were grown in copper-deficient medium (Moseley et al., 2000). The 2-DE results (see Figure 4) demonstrate that the crd1 mutation strongly affects all LHCI spots so far identified with our gel system, which strengthens the view that the crd1 mutation affects the entire LHCI proteins under copper deficiency. It is of note that other spots of unknown identity are absent or enriched in the 2-DE map of copper-deficient crd1 thylakoids compared with that of copper-supplemented Δycf4 thylakoids.
In this study, we succeeded in the separation of thylakoid membrane proteins isolated from C. reinhardtii by two-dimensional gel electrophoresis. We have demonstrated that proteins such as the LHCPs, which span the membrane with three transmembrane domains, can be reliably separated by such a high-resolution gel system. In addition our data show that this 2-DE system can be used for differential analysis of wild-type and mutant thylakoids.
To answer the question whether hydrophobic proteins such as the reaction centre proteins of PSI can be separated by using our 2-DE protocol, we tested for the presence of the PsaA polypeptide. The PSI reaction centre subunit PsaA possesses 11 transmembrane domains, which leads to a GRAVY score of about 0.2. In comparison the PSII reaction centre subunit D1 has five transmembrane domains and a GRAVY score of about 0.28. When wild-type thylakoids were separated by 2-DE and analysed in an immuno-blot experiment, we could visualize about four spots, when the blot was probed with anti-PsaA specific antibodies. These spots have isoelectric points ranging from about 5.8–6.6 and a molecular mass of about 70 kDa. They match with spots on the silver-stained gel shown in Figure 1 (see Figure 1, upper panel, boxed region), indicating that proteins which are more hydrophobic than the LHCPs, with up to 11 transmembrane domains can be separated by 2-DE. The fact that Naver et al. (2001) also succeeded in the separation of PsaA by using our 2-DE protocol, clearly supports our finding. However, a high number of membrane proteins exist that have GRAVY scores higher than 0.5. It is not clear whether the 2-DE procedure presented here will also facilitate the separation of these proteins. Nevertheless, the reproducible 2-DE separation of transmembrane proteins such as the LHCPs is a step into functional proteomics of thylakoid membrane proteins from C. reinhardtii.
The separation of LHCPs from PSII and PSI by 2-DE enabled us to differentiate between more than 30 LHCP spots. For the major LHCII proteins of Arabidopsis 5 Lhcb1, 4 Lhcb2 and 1 Lhcb3 genes have been identified, whereas only six genes encoding the LHCI antenna proteins have been found (Jansson, 1999). The 2-DE analysis of thylakoids identified at least 13 spots that correspond to major LHCII proteins. These high number of spots may on one hand, also reflect the high number of genes that encode for LHCII proteins in Chlamydomonas but on the other hand, can be rationalized by differential processing of the transit peptide or mature protein as shown by mass spectrometric analysis of Lhcb1. Such a differential processing may lead to the removal of charged amino acids so that pre-processed and processed proteins differ in their IPs. The pre-Lhcb1 protein has a theoretical IP value of 5.98, whereas the putative mature Lhcb1 (starting from Ala28) has one of 4.8. Spot 22, that was analysed by mass spectrometry has an experimentally determined IP of 5.5, which matches the theoretical IP expected when Lys9 is the first N-terminal amino acid. Differential processing of pre-LHCPs has been reported in in vitro studies for wheat, pea, tobacco, tomato and corn (Clark et al., 1989, 1990; Cline, 1988; Kohorn and Yakir, 1990; Lamppa and Abad, 1987; Pichersky et al., 1987) when pre-LHCPs were imported into isolated chloroplasts from wheat or pea. When the putative transit peptide of Lhcb1 from C. reinhardtii is compared with the processing sites of the pre-Lhcb1 sequence from wheat (see Figure 5) it appears that the secondary processing site of the wheat sequence, which is believed to be within the mature protein (Clark and Lamppa, 1991), is very similar to the site in the C. reinhardtii sequence that was suggested from sequence alignments (Bassi et al., 1992). Furthermore, the primary processing site (Mullet, 1983) of the wheat transit sequence resembles the site in the C. reinhardtii sequence which can be deduced from the mass spectrometric results (Figure 5). Interestingly, in an organelle-free assay enriched for the chloroplast-soluble processing enzyme (Abad et al., 1989) as well as with the recombinant stromal processing peptidase (Richter and Lamppa, 1998), pre-Lhcb1 is processed at the second site. However, Thr-3 (see Figure 5.; starting the N-terminus at the primary processing site) is found to be phosphorylated in thylakoid membranes from spinach and Arabidopsis (Michel et al., 1991; Vener et al., 2001). Thus, processing at the primary site occurs in vivo. Our data also indicate that processing of the pre-Lhcb1 sequence of C. reinhardtii can occur at the putative primary processing site which would leave a rather unusual short transit sequence. However, pre-processed Lhcb1 assembles and integrates into the thylakoid membrane in a stable manner, since spot 22 can not be removed from thylakoid membranes by washing at pH 12 with sodium carbonate (data not shown). This suggests that the second processing event may occur at the thylakoid membrane. The fact that the tryptic peptide 9–16 of Lhcb1 is found by mass spectrometric analysis of tryptically digested trimeric LHCII (Figure 2b) suggests that preprocessed Lhcb1 assembles into trimeric LHCII. Our findings show that the two processing events which have been found by in -vitro studies of higher plant pre-Lhcb1 (see above) also may occur in vivo in C. reinhardtii. However, the functional role of these alternative processing sites in the assembly or regulation of Lhcb1 remains to be determined. For C. reinhardtiiRüfenacht and Boschetti (2000) described four distinct stromal processing proteases, indicating that processing of transit peptides is a rather complex process.
The analysis of LHCI polypeptides by 2-DE of thylakoids and enriched PSI particles identified 18 protein spots. Spot 2, representing a Lhca1 polypeptide and spots 3 and 4 representing Lhca3 polypeptides are the most abundant LHCI subunits. These proteins are more than 10-fold more abundant, according to the evaluation of spot volumes, than Lhca2 or 4 spots identified by immuno-blotting or mass spectrometry. In higher plants, LHCI can be separated into three subcomplexes; the LHCI-730 complex is a heterodimer composed of Lhca1 and Lhca4 and LHCI-680A and LHCI-680B are homodimers of Lhca3 and Lhca2, respectively (Knoetzel et al., 1992; Schmid et al., 1997). This clear differentiation could not be obtained for the LHCI complex isolated from C. reinhardtii (Bassi et al., 1992), where the LHCI subcomplexes having a 77 K fluorescence emission of 705 or 680 nm did not differ in their subunit composition. From our study it also appears that the subunit stoichiometry of the LHCI complex differs from that of higher plants, since several LHCI subunits are of rather low abundance. This heterogeneity may have functional impact and assembly of a rather heterogeneous LHCI complex could be a way to modulate transfer of excitation energy to the PSI reaction center. In C. reinhardtii at least six different genes encoding LHCI subunits exist (Bassi et al., 1992). The fact that 18 LHCI spots could be detected by mass spectroscopy and immuno-blot analysis indicate that post-translational modifications must account for the presence of some of these spots. The use of two peptide-specific antibodies support this interpretation, since the anti-14.1 and anti-15.1 antibodies detect four and two different spots, respectively. Different processing of the transit peptide or the mature protein, as in the case of the Lhcb1 protein, could be one reason for multiple forms of LHC proteins. Interestingly all these LHCI spot are strongly diminished or absent from 2-D maps of crd1 thylakoids isolated from copper-deficient cells. This implicates a regulatory mechanism for the entire LHCI complex which leads, due to the lack of crd1 function, to a down-regulation of LHCI subunits under copper deficiency.
Our study shows that separation of thylakoid membrane proteins of C. reinhardtii by two-dimensional gel electrophoresis is possible, which opens the door for systematic proteomics of thylakoid membrane proteins and other protein-rich membrane systems. In addition the methods described here will facilitate comparative proteomics of wild-type and photosynthetic mutants of C. reinhardtii.
Strains and media
Chlamydomonas reinhardtii wild-type and mutant strains were grown in Tris acetate phosphate medium (TAP) as described (Harris, 1989) at a light intensity of 20 μE m−2 s−1.
Isolation of thylakoid membranes and PSI complexes
The isolation of thylakoid membranes purified by centrifugation through a sucrose step gradient and the isolation of PSI particles were as described (Chua and Bennoun, 1975; Hippler et al., 1997). Chlorophyll concentrations were determined according to (Porra et al., 1989). Protein determination in solution was carried out by the bicinchoninic acid method (Sigma, Taufkirchen, Germany) according to the manufacturer's instructions.
Prior to electrophoresis, thylakoids were solubilized at a 10 : 1 (w/w) ratio of dodecyl maltoside : chlorophyll. Deriphat-PAGE (7.5% T) was performed according to (Peter and Thornber, 1991).
Protein precipitation (chloroform/methanol)
Thylakoid membrane proteins were precipitated using a method similar to that described by Wessel and Fluegge, (1984). Four hundred microlitres of methanol were added to 100 μl of thylakoid membranes (5 mm Hepes, pH 7.6; 10 mm EDTA), having a protein concentration of about 2 μg μl−1, vortexed and centrifuged for 10 sec at 9000 × g. The supernatant was carefully removed and transferred into a fresh tube. One hundred microlitres chloroform (water saturated) were added to the supernatant, vortexed and centrifuged for 10 sec at 9000 × g. Two hundred microlitres of deionized water were added, vortexed and centrifuged for 1 min at 9000 × g. The protein precipitate formed the interphase. The upper phase was carefully removed and discarded, then 300 μl methanol were added to the sample, vortexed and centrifuged for 2 min at 9000 × g. The pellet was washed twice with 500 μl 95% (v/v) methanol and lyophylized. For 2-DE the pellet was solubilized with solubilization buffer (see below).
Two-dimensional gel electrophoresis
The procedure for two-dimensional gel electrophoresis was modified from (Pasquali et al., 1997). Thylakoid membrane proteins, that had been precipitated as described above, were solubilized in 380 μl of solubilization buffer (2 m thiourea, 8 m urea, 4% (w/v) 3-([3-Cholamidopropyl]dimethylammonio)-1propane-sulphonate (CHAPS), 20 mm Tris-Base, 30 mm dithiothreithole (DTE), 0.5% (v/v) immobilized pH gradient (IPG) buffer, pH 3–10 (Amersham), 0.05% β-dodecyl-maltoside und 0.5% (v/v) bromphenol blue) and incubated for 1 h at room temperature, then centrifuged at 9000 × g for 5 min. Three hundred and fifty microlitres of the supernatant were carefully removed, filled into a fresh tube and 0.5% (v/v) IPG buffer (Amersham Pharmacia Biotech, Freiburg, Germany) was added again. This solution was loaded into the groove of a ceramic strip tray (Amersham). The IPG strip, with a linear pH gradient from 3.0 to 10.0, was overlayed upside down, covered with paraffin and incubated for 12 h at 20°C. After rehydration of the IPG strips in the presence of the sample, electrophoresis was performed using an IPGphor apparatus (Amersham) at 15°C until 60 kV h were reached, using the following programme: 300 mV, 15 min; 500 mV, 30 min; 1000 V, 60 min; 3000 V, 60 min; 8000 V, 7 h.
After electrophoresis in the first dimension was finished, the IPG strips were removed from the tray and equilibrated for SDS-PAGE using two different equilibration solutions. The IPG strips were first incubated with solution 1 (50 mm Tris, pH 6.8; 6 m urea; 30% (v/v) glycerol; 2% (w/v) SDS; 2% (w/v) DTE) for 12 min and afterwards with solution 2 (Tris, pH 6.8; 6 m urea; 30% (v/v) glycerol; 2% (w/v) SDS; 2.5% (w/v) iodoacetamide; 0.5% (w/v) bromphenol blue) for 5 min. The strips were then briefly washed with water, loaded on top of a prepared SDS-PAG [13% acrylamide; 0.42% piperazine diacrylamide (PDA)] and covered with 0.5% agarose. The SDS-PAGE was run at 8°C and 30 mA per gel using a Bio-Rad multicell apparatus.
Staining procedures with Coomassie brillant blue and silver were performed according to the manufacturer's instructions (Amersham). Stained 2-DE maps were further analysed using Phoretix 2-D software.
Tryptic digestion of 2-DE separated spots
Spots were cut out of Coomassie blue-stained 2-DE gels and digested in-gel according to (Mortz et al., 1994). Lyophylized tryptic peptides were dissolved in 5% formic acid and put into an ultrasonic water bath for 2 min, concentrated and desalted using approximately 500 nl of Poros R2 material (PerSeptive Biosystems, Cambridge, MA) that was placed in the tip of a pulled glass capillary as described (Wilm and Mann, 1996). The Poros R2 material was equilibrated with 5% formic acid and the bound peptides where washed twice with 5 μl 5% formic acid. The peptides where directly eluted into a nanospray capillary with 1–2 μl 50% methanol, 5% formic acid.
Nanoelectrospray mass spectrometry
All mass spectra were obtained on a Finnigan MAT TSQ700 tandem quadrupole mass spectrometer (Thermo-Finnigan, San Jose, CA, USA) on which the electrospray source head was replaced by a self-made nanoelectrospray source.
Nanospray needles and the Poros R2 desalting columns were self made by pulling borosilicate glass capillaries (GC120F-10, Harvard Apparatus, Kent, UK) with a micropipette puller (Model P-87 Puller, Sutter Instrument Co., Novato, CA, USA). Nanospray needles were gold-coated in batches of approximately 20 needles. For each sample a new desalting column and a new nanospray needle was used. Nano ES measurements were performed with about 600–1500 V applied to the spraying needle. For MS/MS mode approximately 2–3 mtorr Argon was used in the collision cell and the collision energy was set to 10–15 eV for daughter and to 35–50 eV for parent ion scans. Q1 resolution was set to 2 units and Q3 was set to 1 unit resolution. Plugged needles could be reopened by tipping the needle to the front side of the heated capillary with the needle at 200 V and the capillary at 40 V according to (Wilm and Mann, 1996).
Western blot analysis
Peptide synthesis and coupling of peptides to bovine serum albumine for antibody production
The synthesis of peptides was performed by solid-phase peptide synthesis in a Milligen Model 9050 peptide synthesizer (Milligen/Millipore, Eschborn, Germany). Cleavage and purification of peptides by preparative HPLC were performed as described (Rau et al., 1998). The following peptides were synthesized: CEEKSIAKVDRSKDQL; peptide p14.1 (Bassi et al., 1992), CAADRLWAPGVVAPEYK; peptide 15.1 (Bassi et al., 1992); MAFALASRKALQVTLKATGKKTAAKC, peptide preLhcb1 (Imbault et al., 1988). In contrast to the amino acid sequence of Lhcb1 (Imbault et al., 1988) Cys15 is replaced by Leu in peptide preLhcb1. Peptides 14.1, 15.1, and preLhcb1 were coupled to bovine serum albumin using the crosslinker MBS via N-or C-terminal Cys (see above, the additional Cys is indicated in italics) as described (Sambrook et al., 1989) and sent to EUROGENTEC, where rabbits were used for the production of antibodies against the coupled peptides.
We would like to thank R. Loyal for excellent technical assistance, Drs A. Boschetti, J.D. Rochaix and S. Merchant for generous donation of antibodies and/or strains; Dr W. Haehnel for continuing support of the work and Dr R. Bock for critical reading the manuscript. M.H. acknowledges support from the Deutsche Forschungsgemeinschaft (Hi739/1–1 and SBF388 A1 to W.H.)
- 1989) Properties of a chloroplast enzyme that cleaves the chlorophyll a/b binding proteins precursor. Plant Physiol. 96, 1220–1227. , , (
- 2000) Chloroplast proteases: possible regulators of gene expression? Biochimie, 82, 647–654. (
- 2000) A prediction of the size and evolutionary origin of the proteome of chloroplasts of Arabidopsis. Trends Plant Sci. 5, 141–142. , , (
- 1992) Charactarization of chlorophyll a/b proteins of Photosystem I from Chlamydomonas reinhardtii. J. Biol. Chem. 267, 25714–25721. , , , , (
- 1997) The chloroplast ycf3 and ycf4 open reading frames of Chlamydomonas reinhardtii are required for the accumulation of the photosystem I complex. EMBO J. 16, 6095–6104. , , , , (
- 1988) Chloroplast transformation in Chlamydomonas with high velocity microprojectiles. Science, 240, 1534–1538. , , , et al. (
- 1989) Mutations at the transit peptide-mature protein junction separate two cleavage events during chloroplast import of the chlorophyll a/b-binding protein. J. Biol. Chem. 264, 17544–17550. , , (
- 1990) Loss of efficient import and thylakoid insertion due to N- and C-terminal deletions in the light-harvesting chlorophyll a/b binding protein. Plant Cell, 2, 173–184. , , (
- 1991) Determinants for cleavage of the chlorophyll a/b binding protein precursor: a requirement for a basic residue that is not universal for chloroplast imported proteins. J. Cell Biol. 114, 681–688. , , , (
- 1988) Membrane insertion, proteolytic processing assembly into LHCII and localisation to appessed membranes occurs in chloroplasts lysates. Plant Physiol. 86, 1120–1126. & (
- 2000) Synthesis, assembly and degradation of thylakoid membrane proteins. Biochimie, 82, 615–634. , , , (
- 1975) Thylakoid membrane polypeptides of Chlamydomonas reinhardtii: wild-type and mutant strains deficient in photosystem II reaction center. Proc. Natl Sci. USA, 72, 2175–2179. , , , (
- 2000) Resolution and identification of the protein components of the photosystem II antenna system of higher plants by reversed-phase liquid chromatography with electrospray-mass spectrometric detection. J. Chromatogr. A, 21, 111–121. , , , , , (
- 1989) The argininosuccinate lyase gene of Chlamydomonas reinhardtii: an important tool for nuclear transformation and for correlating the genetic and molecular maps of the ARG7 locus. EMBO J. 8, 2803–2809. , , , , (
- 1987) Characterisation of a chloroplast in the psA2 gene of Chlamydomonas reinhardtii. Curr. Genet. 12, 489–495. , , , , , , (
- 2000) Sequence submitted to dbEST. , , , , , , , , , (
- 1989) The Chlamydomonas Sourcebook. San Diego, CA: Academic Press. & (
- 1989) Identification of the plastocyanin binding subunit of Photosystem I. FEBS Lett. 250, 280–284. , , , (
- 1997) Fast Electron Transfer from Cytochrome c6 and Plastocyanin to Photosystem I of Chlamydomonas reinhardtii Requires PsaF. Biochemistry, 36, 6343–6349. , , , , , (
- 1998) Chlamydomonas genetics. A tool for the study of bioenergetic pathways. Biochim. Biophys. Acta, 1367, 1–62. , , , , (
- 2000) Limitation in electron transfer in photosystem I donor side mutants of Chlamydomonas reinhardtii. Lethal photo-oxidative damage in high light is overcome in a suppressor strain deficient in the assembly of the light harvesting complex. J. Biol. Chem. 275, 5852–5859. , , , , , , (
- 1988) Structure of the Chlamydomonas reinhardtii cabII-1 gene encoding a chlorophyll-a/b-binding protein. Gene, 73, 397–407. , , , , , , (
- 1999) A guide to the Lhc genes and their relatives in Arabidopsis. Trends Plant Sci. 4, 236–240. & (
- 1989) Stable nuclear transformation of Chlamydomonas using the Chlamydomonas gene for nitrate reductase. J. Cell Biol. 109, 2589–2601. , , , , , (
- 1992) Identification of the photosystem I antenna polypeptides in barley. Isolation of three pigment-binding antenna complexes. Eur. J. Biochem. 206, 209–215. , , , , (
- 1990) Movement of newly imported light-harvesting chlorophyll-binding protein from unstacked to stacked thylakoid membranes is not affected by light treatment or absence of amino-terminal threonines. J. Biol. Chem. 265, 2118–2123. , , , (
- 1994) Atomic model of plant light-harvesting complex by electron crystallography. Nature, 367, 614–621. , , , , (
- 1982) A simple method for displaying the hydrophadic character of a protein. J. Mol. Biol. 157, 105–132. & (
- 1990) Stable nuclear transformation of Chlamydomonas reinhardtii by using a C. reinhardtii gene as the selectable marker. Proc. Natl Acad. Sci. USA, 87, 2087–2091. , , , (
- 2000) Two-dimensional electrophoresis of membrane proteins using immobilized pH gradients. Anal Biochem. 280, 1–10. & (
- 1985) Structure and developmental regulation of a wheat germ encoding the major chlorophyll a/b binding protein. Mol. Cell. Biol. 5, 1370–1378. , , , , (
- 1987) Processing of a wheat light-harvesting chlorophyll a/b protein precursor by a soluble enzyme from higher plant chloroplasts. J. Cell Biol. 105, 2641–2648. , , , (
- 1998) Posttranslational assembly of photosynthetic metalloproteins. Annu. Rev. Plant Physiol. Plant Mol. Biol. 49, 25–51. , , , (
- 1999) Title . In Advances in Photosynthesis, the Molecular Biology of Chloroplasts and Mitochondria in Chlamydomonas (Rochaix, J.D.Goldschmidt-Clermont, M. and Merchant, S. eds) Dordrecht, Netherlands: Kluwer Academic Publisher Group . , , , , (eds.) (
- 1983) A chlorophyll b-less mutant of Chlamydomonas reinhardtii lacking in the light-harvesting chlorophyll a/b-protein complex but not in its apoproteins. Biochim. Biophys. Acta, 725, 417–424. , , , , (
- 1991) Tandem mass spectrometry identifies sites of three post-translational modifications of spinach light-harvesting chlorophyll protein II. J. Biol. Chem. 266, 17584–17591. , , , , , , (
- 1994) Identification of proteins in polyacrylamide gels by mass spectrometric peptide mapping combined with database search. Biol. Mass Spectrom. 23, 249–261. , , , , , (
- 2000) The Crd1 gene encodes a putative di-iron enzyme required for photosystem I accumulation in copper deficiency and hypoxia in Chlamydomonas reinhardtii. EMBO J. 15, 2139–2151. , , , , , (
- 1983) The amino acid sequence of the polypeptide segment which regulates membrane adhesion (grana stacking) in chloroplasts. J. Biol. Chem. 258, 9941–9948. & (
- 2001 ) Functional studies of Ycf3: role in assembly of photosystem I and interactions with some of its subunits. Plant Cell, in press. , , , , (
- 2000) Proteomics of the chloroplast: systematic identification and targeting analysis of lumenal and peripheral thylakoid proteins. Plant Cell, 12, 319–341. , , , , , , , , (
- 1997) Preparative two-dimensional gel electrophoresis of membrane proteins. Electrophoresis, 18, 2573–2581. , , , , (
- 1991) Biochemical composition and organization of higher plant photosystem II light-harvesting pigment proteins. J. Biol. Chem. 266, 6745–16754. , , , (
- 1987) The tomato Cab-4 and Cab-5 genes encode a second type of CAB polypeptides localized in photosystem II. Plant Mol. Biol. 9, 109–120. , , , , , , , (
- 1989) Determination of accurate extinction coefficients and simultaneous equations for assaying chlorophylls a and b extracted with four different solvents: verification of the concentration of chlorophyll standards by atomic absorption spectroscopy. Biochim. Biophys. Acta, 975, 384–394. , , , , (
- 1998) Modular synthesis of de novo-designed metalloproteins for light-induced electron transfer. Proc. Natl Acad. Sci. USA, 95, 11526–11531. , , , , (
- 1998) A chloroplast processing enzyme functions as the general stromal processing peptidase. Proc. Natl Acad. Sci. USA, 95, 7463–7468. , , , (
- 2000) Chloroplasts of the green alga Chlamydomonas reinhardtii possess at least four distinct stromal processing proteases. Photosyn. Res. 63, 249–258. , , , (
- 1989). Molecular Cloning: a Laboratory Manual. Cold Spring Harbor. NY: Cold Spring Harbor Laboratory Press. , , , , (
- 2000) Membrane proteins and proteomics: un amour impossible? Electrophoresis, 21, 1054–1070. , , , , (
- 1997) In vitro reconstitution of the photosystem I light-harvesting complex LHCI-730: Heterodimerization is required for antenna pigment organization. Proc. Natl Acad. Sci. USA, 94, 7667–7672. , , , , , (
- 1999) TAKs, thylakoid membrane kinases associated with energy transduction. J. Biol. Chem. 274, 9137–9140. , , , (
- 1991) Directed chloroplast transformation in Chlamydomonas reinhardtii: insertional inactivation of the psaC gene encoding the iron sulfur protein destabilizes photosystem I. EMBO J. 10, 2033–2040. , , , , , , (
- 1986) Lateral distribution of the main protein complexes of the photosynthetic appartus in Chlamydomonas reinhardttii and in spinach: an immunocytochemical study using intact thylakoid membranes and PS II enriched membrane preparation. Photobiochem. Photobiophys. 12, 203–220. , , , , (
- 1998) Protein phosphorylation and redox sensing in chloroplast thylakoids. Curr. Opin. Plant Biol. 1, 217–223. , , , , (
- 2001) Mass spectrometric resolution of reversible protein phosphorylation in photosynthetic membranes of Arabidopsis thaliana. J. Biol. Chem. 276, 6959–6966. , , , , , (
- 1998) Toward the bilayer proteome, electrospray ionization-mass spectrometry of large, intact transmembrane proteins. Proc. Natl Acad. Sci. USA, 96, 10695–10698. , , , , , , , , (
- 1984) A method for the quantitative recovery of protein in dilute solution in the presence of detergents and lipids. Anal Biochem. 138, 141–143. , , , (
- 1996) Analytical properties of the nanoelectrospray ion source. Anal. Chem. 66, 1–8. , , , (
- 1996a) Parent ion scans of unseparated peptide mixtures. Anal. Chem. 68, 527–533. , , , , (
- 1996b) Femtomole sequencing from polyacrylamide gels by nano-electrospray mass spectroscopy. Nature, 379, 466–469. , , , , et al. (
- 1984) Correlation between changes in light energy distribution and changes in thylakoid membrane polypeptide phosphorylation in Chlamydomonas reinhardtii. J. Cell Biol. 98, 1–7. & (
- 1999) The biogenesis and assembly of photosynthetic proteins in thylakoid membranes. Biochim. Biophys. Acta, 1411, 21–85. , , (