Duplication and divergence of the Psb27 subunit of Photosystem II in the green algal lineage

Abstract

Photosystem II (PSII) is subject to light-induced damage and continuously undergoes repair to restore the damaged photosystems. During repair of PSII, the Psb27 protein interacts with the CP43 subunit of PSII. Two Psb27-like proteins (Psb27-H1 and Psb27-H2) are involved in the repair of PSII in the higher plant Arabidopsis thaliana. Here, we present evidence that duplication and divergence of Psb27 occurred in the green algal lineage and that the Psb27-H2 protein shares c. 30% sequence identity with Psb27-H1. Structural modelling of Arabidopsis thaliana Psb27-H1 and Psb27-H2 indicate that the two proteins have different distributions of surface charge. We suggest that the green algal lineage Psb27-H2 protein occupies a different binding niche than does Psb27-H1.

Keywords:

• Arabidopsis
• biogenesis
• Chlamydomonas
• CP43
• extrinsic proteins
• Photosystem II
• PsbO
• Psb27
• Synechocystis

Introduction

Photosystem II (PSII) is the enzyme responsible for light-driven water oxidation and plastoquinone reduction. The PSII complexes within green algae and higher plants (hereafter viridiplantae, Leliaert et al. 2012) and red algae are of cyanobacterial origin (Larkum et al. 2007; Keeling 2013). As such, the core components of PSII including the reaction centre proteins D1 and D2 and the chlorophyll a-binding proteins CP43 and CP47 are highly conserved between cyanobacteria and photosynthetic eukaryotes (Eaton-Rye & Putnam-Evans 2005; Nixon et al. 2005).

Photosynthetic water oxidation is catalysed by the manganese-calcium (Mn₄CaO₅) cluster of the oxygen-evolving complex (OEC). Both the D1 and the CP43 proteins provide ligands to the Mn₄CaO₅ cluster and are essential for water oxidation (Umena et al. 2011). The OEC is located on the lumenal side of PSII and several proteins that bind to the lumenal side of PSII act to stabilize the Mn₄CaO₅ cluster. The most highly conserved of these is the PsbO protein; PsbO is found in cyanobacteria and in all algal lineages (Roose et al. 2007). The stoichiometry of PsbO to PSII in cyanobacteria and red algae is 1:1 (Umena et al. 2011; Krupnik et al. 2013), whereas in higher plants there is evidence to suggest that two PsbO proteins bind per reaction centre (Xu & Bricker 1992; Popelkova et al. 2002a,b). A key difference between viridiplantae PsbO proteins and cyanobacterial PsbO proteins is the presence of an N-terminal extension on viridiplantae PsbO proteins (Eaton-Rye & Murata 1989; Popelkova et al. 2002a,b). In cyanobacteria and red algae, a ‘PsbQ-like’ protein, PsbU and the PsbV protein join PsbO on the lumenal side of PSII (Bricker et al. 2012) but in viridiplantae, PsbU and PsbV are absent and PsbQ and PsbP, together with the putative second copy of PsbO, appear to fill the role of these proteins. A fourth polypeptide, PsbR, also joins PsbO, PsbQ and PsbP in viridiplantae PSII. At present the localization and stoichiometry of PsbR is not well understood (Bricker et al. 2012).

PSII is subject to light-induced damage and damaged photosystems are repaired (Vass 2011; Tyystjärvi 2013). Both de novo biogenesis and the repair cycle of PSII following photodamage involve a large number of proteins (Nixon et al. 2010; Komenda et al. 2012b; Shi et al. 2012; Nickelsen & Rengstl 2013). One of the proteins that interacts specifically with the lumenal side of inactive PSII complexes is Psb27 (Roose & Pakrasi 2004; Nowaczyk et al. 2006). The complexes with which Psb27 interacts are depleted of the Mn₄CaO₅ cluster and lack the PsbO, PsbQ-like, PsbU and PsbV proteins (Nowaczyk et al. 2006; Mamedov et al. 2007; Liu et al. 2011a). In cyanobacteria, the binding site of Psb27 has been located on the CP43 protein using both chemical cross-linking and two-dimensional blue native SDS–PAGE (Liu et al. 2011b; Komenda et al. 2012a). The Psb27 protein is a four-helical bundle with the helices arranged in a right-handed up–down up–down topology (Cormann et al. 2009; Mabbitt et al. 2009; Michoux et al. 2012). Cross-linking and solvent-protection data suggest that the loop between helices 2 and 3, and the fourth helix of Psb27 interact with CP43 (Liu et al. 2011b, 2013).

Two paralogues of Psb27 have been identified in Arabidopsis thaliana (Chen et al. 2006; Wei et al. 2010). The first protein identified as a Psb27 orthologue (encoded by At1g03600) is more similar in sequence to Synechocystis sp. PCC 6803 Psb27 (33% identity) than is the second protein LPA19 (encoded by At1g05385; 26% identity) (Wei et al. 2010; Fagerlund & Eaton-Rye 2011). The A. thaliana Psb27 and LPA19 proteins are hereafter referred to as Psb27-H1 (inactivated in psb27 mutant plants) and Psb27-H2 (inactivated in lpa19 mutants), respectively. A. thaliana psb27 and lpa19 mutants were phenotypically different. Although psb27 mutants were susceptible to photodamage, their rate of growth was unimpaired (Chen et al. 2006). By comparison, lpa19 mutants have a greatly reduced rate of growth even under low-light conditions. In the lpa19 mutant, the rate of D1 processing from its precursor (pD1) was reduced compared with the rate in wild type (Wei et al. 2010). Furthermore, there is some evidence that the binding site of A. thaliana Psb27-H2 is different to that of cyanobacterial Psb27-H1: in yeast two-hybrid assays, the A. thaliana Psb27-H2 protein interacted with both mature D1 and pD1 (Wei et al. 2010).

Whereas the Psb27-H1 protein appears to have originated in cyanobacteria and has been retained in viridiplantae (Chen et al. 2006), the origin and phylogenetic distribution of the Psb27-H2 protein have not been thoroughly examined. To address this question, we have constructed multiple sequence alignments and phylogenetic trees of Psb27-H1 and Psb27-H2 protein sequences from cyanobacteria, brown algae, red algae, green algae, diatoms and higher plants.

Materials and methods

Multiple sequence alignment

Amino acid sequences for Psb27-H1 and Psb27-H2 were found by a BLAST (Altschul et al. 1990) search of the National Center for Biotechnological Information (NCBI) database using the Synechocystis sp. PCC 6803 Psb27 sequence as the query, followed by a second search using A. thaliana Psb27-H2 (Wei et al. 2010). Sequences were aligned using the multiple sequence comparison by the log-expectation (MUSCLE) algorithm (Edgar 2004). Chloroplast import pathways and signal sequences differ greatly from the thylakoid lumen import pathway in cyanobacteria (Gutensohn et al. 2006; Shi & Theg 2013). All Psb27 sequences were trimmed so that Tyr11 was the first residue, to ensure that only informative regions of sequence were included in the alignment. Multiple sequence alignment figures were prepared using TeXshade (Beitz 2000).

Species and sequences examined

The Psb27 sequences available in the NCBI database from cyanobacteria and higher plants greatly outnumber those from other organisms. For cyanobacteria, representative Psb27 sequences were chosen from phylogenetically diverse species (Gupta & Mathews 2010). For viridiplantae land plants, representative Psb27-H1 and Psb27-H2 proteins from an angiosperm, bryophyte and gymnosperm were chosen. For organisms other than cyanobacteria and land plants, all sequences with an expect value (E) of < 1 × 10⁻⁵ were included.

Amino acid sequences for Psb27 were obtained for the following species (accession codes in brackets): Aureococcus anophagefferens (EGB06728.1), Acaryochloris marina MBIC 11017 (ABW29403.1), Arabidopsis thaliana (Q9LR64 and NM_100418.2), Chlamydomonas reinhardtii (XP_001700736.1 and XP_001701619.1), Chlorella variabilis (EFN58701.1 and EFN58701.1), Coccomyxa subellipsoidea C−169 (EIE25666.1), Cyanidioschyzon merolae strain 10D (BAM80544.1), Ectocarpus siliculosus (CBN76956.1), Emiliania huxleyi CCMP 1516 (EOD15678.1), Galdieria sulphuraria (EME228982.1), Guillardia theta CCMP 2712 (EKX52014.1), Micromonas pusilla CCMP 1545 (XP_003058671.1 and XP_003062962.1), Micromonas RCC 299 (XP_002508067.1 and XP_002508382.1), Nostoc sp. PCC 7120 (NP_485301.1), Ostreococcus lucimarinus CCE 9901 (XP_001419197.1), Ostreococcus tauri (XP_003078942.1), Paulinella chromatophora (YP_002049501.1), Phaeodactylum tricornutum CCAP 1055/1 (XP_002177207.1), Physcomitrella patens (XP_001770919.1 and XP_001776869.1), Picea sitchensis (ABK23991.1 and ABK23129.1), Prochlorococcus marinus CCMP 1986 (NP_892625.1), Prochlorococcus marinus CCMP 1375 (NP_874900.1), Synechocystis sp. PCC 6803 (NP_441782.1), Thermosynechococcus elongatus BP−1 (NP_683253.1), Thalassiosira oceanica CCMP 1005 (EJK47273.1), Thalassiosira pseudonana CCMP 1335 (XP_002288234.1), Trichodesmium erythraeum IMS101(YP_721473.1) and Volvox carteri (EFJ42752.1 and XP_002948422.1).

Phylogenetic tree

Unrooted maximum-likelihood phylograms were estimated using the Phylogeny Inference Package (PHYLIP version 3.695; Felsenstein 2013). Analyses were run allowing global rearrangements using the Jones-Taylor-Thornton probability model of amino acid change (Jones et al. 1992).

Modelling of Arabidopsis thaliana Psb27 proteins

Models of A. thaliana Psb27-H1 and Psb27-H2 were generated using the SWISS-MODEL server (Schwede et al. 2003; Arnold et al. 2006). The solution structure of Synechocystis sp. PCC 6803 Psb27 (Mabbitt et al. 2009) was used as the template structure. The A. thaliana Psb27-H1 and Psb27-H2 protein sequences submitted to the SWISS-MODEL server were trimmed so that Tyr11 was the first residue.

Results and discussion

An unrooted maximum-likelihood phylogram was constructed from a multiple sequence alignment of Psb27-H1 and Psb27-H2 sequences (Fig. 1). The Psb27 sequences present in red algae are all more similar to Psb27-H1 than to Psb27-H2. Viridiplantae genomes encode both Psb27-H1 and Psb27-H2 proteins. The absence of Psb27-H2-like sequences in cyanobacteria and red algae suggests that Psb27-H2 is specific to the green algal lineage (Fig. 1).

Figure 1 Unrooted phylogram of Psb27 amino acid sequences. Green algae are coloured blue, higher plants (together with a bryophyte) are coloured green, red algae are coloured red, brown algae are coloured brown, cyanobacteria are coloured cyan, diatoms are coloured orange. The coccolithophore Emiliania huxleyi is coloured black. Paulinella chromatophora and Guillardia theta CCMP 2712 (coloured pink) have acquired a cyanobacterial or red algal endosymbiont, respectively (Nowack et al. 2008; Curtis et al. 2012). The Psb27-H1-like sequences are above the dashed line and the Psb27-H2-like sequences are below the dashed line. Species are Aureococcus anophagefferens, Acaryochloris marina MBIC 11017, Arabidopsis thaliana, Chlamydomonas reinhardtii, Chlorella variabilis, Coccomyxa subellipsoidea C−169, Cyanidioschyzon merolae strain 10D, Ectocarpus siliculosus, Emiliania huxleyi CCMP 1516, Galdieria sulphuraria, Guillardia theta CCMP 2712, Micromonas pusilla CCMP 1545, Micromonas sp. RCC 299, Nostoc sp. PCC 7120, Ostreococcus lucimarinus CCE 9901, Ostreococcus tauri, Paulinella chromatophora, Phaeodactylum tricornutum CCAP 1055/1, Physcomitrella patens, Picea sitchensis, Prochlorococcus marinus CCMP 1986, Prochlorococcus marinus CCMP 1375, Synechocystis sp. PCC 6803, Thermosynechococcus elongatus BP−1, Thalassiosira oceanica CCMP 1005, Thalassiosira pseudonana CCMP 1335, Trichodesmium erythraeum IMS101 and Volvox carteri.

The structures of Synechocystis sp. PCC 6803 and Thermosynechococcus elongatus Psb27-H1 proteins have been solved and the structures are highly similar (Mabbitt et al. 2009; Michoux et al. 2012). A number of aromatic residues are present in the hydrophobic core of Psb27-H1, these residues appear to be important for the overall fold of Psb27-H1 (Mabbitt et al. 2009; Michoux et al. 2012). Several of these aromatic residues are highly conserved in both Psb27-H1 and Psb27-H2. These conserved residues are located at positions 11, 53, 65, 78 and 79 (Figs 2–3). An aspartic acid residue located at position 14 of Synechocystis sp. PCC 6803 Psb27-H1 acts to stabilize the structure of Psb27-H1 by hydrogen bonding with the N-terminal tail of Psb27 (Mabbitt et al. 2013). This aspartic acid residue is highly conserved amongst Psb27-H1 proteins, but absent from Psb27-H2 proteins (Figs 2–3). A Pro–Φ–Pro motif (where Φ is Ile, Leu or Val) has been identified in Synechocystis sp. PCC 6803 Psb27-H1 (Mabbitt et al. 2009). There is evidence to suggest that the Pro–Φ–Pro motif stabilizes the tertiary structure of Psb27-H1 (Mabbitt et al. 2013). The Pro–Φ–Pro motif is present in Psb27-H1 polypeptides from phylogenetically diverse organisms (positions 86–88, Fig. 2), whereas only the central Φ residue of this motif remains highly conserved in the Psb27-H2 proteins (position 88, Fig. 3). Interestingly, the conservation of the aromatic residues in the hydrophobic core of Psb27-H1 and Psb27-H2 suggests that the proteins retain similar folds; however, the loss of Asp14 and the Pro–Φ–Pro motif suggests that the amino terminus and third loop of Psb27-H2 may be structurally different from those of Psb27-H1.

Figure 2 Comparison of Psb27-H1 protein sequences. Sequences are numbered according to the sequence of mature Thermosynechococcus elongatus BP−1 Psb27. All of the sequences have been trimmed so that Phe or Tyr11 is the first residue in each sequence. The positions of α helices in the solution structure of cyanobacterial Psb27 are indicated below the alignment (Mabbitt et al. 2009). Highly conserved residues are coloured purple, residues with at least 50% conservation or conservative substitutions are coloured yellow. Species are T. elongatus BP−1, Synechocystis sp. PCC 6803, Paulinella chromatophora, Prochlorococcus marinus CCMP 1986, Prochlorococcus marinus CCMP 1375, Acaryochloris marina MBIC 11017, Nostoc sp. PCC 7120, Trichodesmium erythraeum IMS101, Ectocarpus siliculosus, Aureococcus anophagefferens, Phaeodactylum tricornutum CCAP 1055/1, Thalassiosira pseudonana CCMP 1335, Thalassiosira oceanica CCMP 1005, Guillardia theta CCMP 2712, Emiliania huxleyi CCMP 1516, Cyanidioschyzon merolae strain 10D, Galdieria sulphuraria, Ostreococcus lucimarinus CCE 9901, Micromonas sp. RCC 299, Micromonas pusilla CCMP 1545, Physcomitrella patens, Arabidopsis thaliana, Picea sitchensis, Chlorella variabilis, Chlamydomonas reinhardtii and Volvox carteri.

Figure 3 Comparison of Psb27-H2 protein sequences. Sequences are numbered according to the sequence of mature Arabidopsis thaliana Psb27. All of the sequences have been trimmed so that Tyr11 is the first residue in each sequence. Highly conserved residues are coloured purple, residues with at least 50% conservation or conservative substitutions are coloured yellow. Species are A. thaliana, Physcomitrella patens, Picea sitchensis, Ostreococcus tauri, Micromonas sp. RCC 299, Micromonas pusilla CCMP 1545, Volvox carteri, Chlamydomonas reinhardtii, Chlorella variabilis and Coccomyxa subellipsoidea C−169.

Because structurally important residues are conserved between Synechocystis sp. PCC 6803 Psb27 and both A. thaliana Psb27-H1 and Psb27-H2, we modelled the tertiary structures of the A. thaliana proteins using Synechocystis sp. PCC 6803 Psb27 as the template. The quality of the modelled structures was assessed via the QMEAN Z-score. This score gives an indication of how native-like the modelled protein was compared to a database of high-resolution protein structures (Benkert et al. 2011). The QMEAN Z-score for both the Psb27-H1 and Psb27-H2 models was within 1 standard deviation of that observed for the reference dataset (Table S1). Similarly, the QMEAN Z-score for the 2KMF solution NMR structure of Synechocystis sp. PCC 6803 Psb27 was within 1 standard deviation of that observed for the reference dataset (Table S1). The most well modelled regions of both Psb27-H1 and Psb27-H2 were the four helices. The loop between helices 2 and 3 was the least reliable region of both models (Fig. S1).

The A. thaliana Psb27-H1 and Psb27-H2 models retain the right-handed up-down up-down fold of Synechocystis sp. PCC 6803 Psb27. The conserved aromatic residues at positions 11, 53, 65, 78 and 79 occupy equivalent positions in all three structures (Fig. 4). The distribution of surface charge on helices 3 and 4 is markedly different between the Psb27-H1 proteins and the Psb27-H2 protein (Fig. 4). Much of this difference is due to the presence of acidic residues on helix 4 of Psb27-H2 (Glu90, Glu94, Asp97, Asp98, Glu103, Glu104) (Figs 3–4).

Figure 4 Molecular models of A. thaliana Psb27 proteins. The A. thaliana Psb27-H1 and Psb27-H2 proteins were modelled using the SWISS-MODEL server (Schwede et al. 2003; Arnold et al. 2006). The template for modelling was the solution structure of Synechocystis sp. PCC 6803 Psb27 (Mabbitt et al. 2009). A ribbon diagram of residues 11–110 of the Synechocystis sp. PCC 6803 Psb27 protein is shown in cyan (upper). A ribbon diagram of residues 11–110 of the A. thaliana Psb27-H1 protein is shown in green (lower left) and a ribbon diagram of residues 11–106 of the A. thaliana Psb27-H2 protein is shown in orange (lower right). In each of the ribbon diagrams, conserved aromatic residues at positions 11, 53, 65, 78 and 79 are shown as red sticks. Below each ribbon diagram is a representation of the potential surface charge of the protein (DeLano 2002), positively charged positions are coloured blue, negatively charged positions are coloured red.

It has been noted that helix 3 of Synechocystis sp. PCC 6803 Psb27-H1 is devoid of acidic or basic residues (Mabbitt et al. 2009). This is also the case for Psb27-H1 proteins from photosynthetic eukaryotes (Fig. 2). The highly conserved residues on helix 3 of Psb27-H1 are: Tyr69, Thr70, Ala71, Asn73, Ala76 and Gly77. The equivalent positions on Psb27-H2 do not share the same pattern as Psb27-H1; in addition, both acidic and basic residues are found in this region of the Psb27-H2 sequence (Fig. 3). Several highly conserved acidic and basic residues are found on helix 4 of Psb27-H1: Lys90, Lys92, Arg94, Glu98, Glu103 and Arg108. The equivalent positions on Psb27-H2 are either poorly conserved, retained their charge (position 92) or have reversed their charge (position 90) (Figs 2–3). Given the poor sequence conservation of helix 4 between Psb27-H1 and Psb27-H2 it is unlikely that Psb27-H1 and Psb27-H2 occupy the same binding niche.

Conclusions

The complement of extrinsic proteins in viridiplantae is different from that in cyanobacteria and red algae (Roose et al. 2007; Bricker et al. 2012). The appearance of the Psb27-H2 protein in the viridiplantae lineage coincides with the loss of the PsbU and PsbV proteins, and changes in the PsbO protein. At present, the function of Psb27-H1 and Psb27-H2 in green algae is unknown. In the higher plant A. thaliana the Psb27-H2 protein had an effect on the rate of D1 processing and PSII accumulation and it was suggested that Psb27-H2 binds to the D1 protein (Wei et al. 2010). In cyanobacteria, Psb27-H1 binds to the large lumenal loop of CP43 (Liu et al. 2011b, 2013; Komenda et al. 2012a). We suggest that in viridiplantae, Psb27-H1 retains its association with CP43 and that the Psb27-H2 binds at the interface between the D1 protein and CP43.

Supplementary file

Supplementary file 1: Duplication and divergence of the Psb27 subunit of Photosystem II in the green algal lineage.

Acknowledgements

PDM was supported by an Otago University postgraduate scholarship. This research was partially supported by a Marsden grant 08-UOO-043 to JJE-R.