Structure, function and interactions of the PufX protein

Biochimica et Biophysica Acta 1777 (2008) 613–630

Contents lists available at ScienceDirect

Biochimica et Biophysica Acta j o u r n a l h o m e p a g e : w w w. e l s ev i e r. c o m / l o c a t e / b b a b i o

Review

Structure, function and interactions of the PufX protein Kate Holden-Dye, Lucy I. Crouch, Michael R. Jones ⁎ Department of Biochemistry, School of Medical Sciences, University of Bristol, University Walk, Bristol, BS8 1TD, United Kingdom

A R T I C L E

I N F O

Article history: Received 28 January 2008 Received in revised form 6 April 2008 Accepted 10 April 2008 Available online 18 April 2008 Keywords: Photosynthesis Purple bacteria PufX Light-harvesting Photosystem

A B S T R A C T The PufX protein is an important component of the reaction centre–light-harvesting 1 (RC–LH1) complex of Rhodobacter species of purple photosynthetic bacteria. Early studies showed that removal of the PufX protein causes changes in the structure of the RC–LH1 complex that result in a loss of the capacity for photosynthetic growth, and that this loss can be overcome though further mutations that change the structure of the LH1 antenna. More recent studies have examined interactions of the PufX protein with other components of the RC–LH1 complex. This review considers our current understanding of the structure and function of the PufX protein, how this protein interacts with other components of the photosynthetic membrane, and its influence on the oligomeric state of the RC–LH1 complex and the larger-scale architecture of the photosynthetic membrane. © 2008 Elsevier B.V. All rights reserved.

1. The purple bacterial photosystem The photosystem in purple photosynthetic bacteria consists of a reaction centre (RC) and one or more types of light-harvesting (LH) antenna complex housed in a lipid bilayer membrane (for recent reviews, see [1–7]). The LH pigment-proteins are formed from oligomers of at least two types of polypeptide chain that encase multiple molecules of bacteriochlorophyll (BChl) and carotenoid. These pigments are responsible for capturing light energy in the form of a pigment excited electronic state, and funneling that energy to the RC to initiate photochemistry [8–11]. The simplest RCs are also formed from BChl, carotenoid and at least three polypeptide chains, together with two quinone molecules. The arrival of excited state energy in the RC triggers a membrane-spanning electron transfer reaction, the ultimate product of which is the double reduction and double protonation of a quinone to form quinol, and the oxidation of a water soluble redox protein — usually a c-type cytochrome (see for reviews [12–17]). The RC can therefore be thought of as a light-powered cytochrome c:quinone oxidoreductase. Lightdriven electron transfer in the bacterium is completed by a partner quinol:cytochrome c oxidoreductase, the membrane-embedded cytochrome bc1 complex, which oxidises quinol, reduces cytochrome c and operates a Q-cycle [18–20] (and see [21–24] for recent reviews). By arranging sites of quinone reduction/protonation and quinol oxidaAbbreviations: AFM, Atomic Force Microscopy; BChl, bacteriochlorophyll; Bl., Blastochloris; EM, electron microscopy; LH, light-harvesting; LH1, core antenna complex; LH2, peripheral antenna complex; ORF, open reading frame; PDB, Protein Data Bank; PS+, photosynthetic phenotype; PS−, non-photosynthetic phenotype; Psp., Phaeospirillum; Rb., Rhodobacter; RC, reaction centre; Rps., Rhodopseudomonas; Rsp., Rhodospirillum ⁎ Corresponding author. Tel.: +44 117 3312135; fax: +44 117 3312168. E-mail address: [email protected] (M.R. Jones). 0005-2728/$ – see front matter © 2008 Elsevier B.V. All rights reserved. doi:10.1016/j.bbabio.2008.04.015

tion/deprotonation on opposite sides of the membrane, light-powered cyclic electron transfer is linked to proton translocation across the membrane, forming the protonmotive force. The RC from purple bacteria was the first membrane protein to be structurally characterised to a high-resolution by X-ray crystallography. The first structure to be described was that for the RC from Blastochloris (Bl.) viridis (formerly Rhodopseudomonas (Rps.) viridis) [25–27], followed shortly after by that for the RC from Rhodobacter (Rb.) sphaeroides [28–33]. An overview of the latter is shown in Fig. 1A and B. More recently a structure has been published for the RC from the moderate thermophile Thermochromatium tepidum [34,35]. These structures have aided the development of a detailed picture of the mechanism of light-powered charge separation, quinone reduction and cytochrome oxidation [14]. X-ray crystal structures have also been published for the so-called LH2 peripheral antenna complex from Rps. acidophila [36] and Phaeospirillum (Psp.) molischianum (formerly Rhodospirillum (Rsp.) molischianum) [37]. These show two concentric membrane-spanning cylinders of protein, each formed from multiple (nine or eight respectively) copies of α- and β-polypeptides, that encase rings of BChl and carotenoid cofactors. The structure of the LH2 from Rps. acidophila is shown in Fig. 1C. In all characterised species of purple photosynthetic bacteria the RC is associated with an LH1 antenna pigment-protein, forming the socalled RC–LH1 core-complex. To date a high-resolution X-ray structure has not been forthcoming for such a core-complex, but a variety of data have established that the LH1 component forms a hollow cylindrical structure around the RC. In early low-resolution work, electron microscopy (EM) on membranes from Bl. viridis and Ectothiorhodospira halochloris showed a central RC surrounded by the LH1 antenna, the latter displaying a six or twelve fold symmetry [38–40]. Subsequent studies involving EM of aggregated LH1 complexes from Rsp.

614

K. Holden-Dye et al. / Biochimica et Biophysica Acta 1777 (2008) 613–630

Fig. 1. Structures of photosynthetic proteins from the purple photosynthetic bacteria. (A) In the Rb. sphaeroides RC the ten cofactors (grey spheres) are held in place by a scaffold consisting of PufL (grey ribbons) and PufM (white ribbons). The H-polypeptide (white tube) has a single membrane-spanning helix and a cytoplasmic domain. The protein is embedded in the photosynthetic membrane (grey box). (B) The RC BChl, BPhe and ubiquinone cofactors (sticks) form two membrane-spanning branches. For clarity the hydrocarbon side chains have been removed and the BA and BB BChls are highlighted in grey. Mg atoms of BChl and non-heme Fe are shown as spheres. The dotted line shows the axis of two-fold symmetry and the black arrows show the route of electron transfer. The QB ubiquinone is dissociable (grey arrows). Panels A and B were prepared using structure 2BOZ [153]. (C) The LH2 complex from Rps. acidophila (structure 1NKZ [154]). Concentric cylinders of membrane-spanning helices of LH2 α and β polypeptides are shown as ribbons, and BChls are shown as sticks. (D) The RC–LH1 complex of Rps. palustris (structure 1PYH [52]) viewed from the periplasmic side of membrane. Central RC shown as in (A). Membrane-spanning helices of LH1 α (grey) and β polypeptides (white), and W-polypeptide (white) are shown as ribbons, and LH1 BChls are shown as spheres, alternating white and grey.

marina [41] and single particle analysis on individual RC–LH1 complexes from Psp. molischianum [42] reinforced this picture of LH1 forming a hollow cylinder around a central RC. In 1995, Karrasch et al. reported structural information from EM at a resolution of 8.5 Å on 2-D crystals of the LH1 complex from Rsp. rubrum [43]. The density maps were interpreted as showing 16 α and β LH1 polypeptides arranged in concentric cylinders with intervening density attributed to a ring of 32 BChl molecules, this interpretation being guided by the X-ray crystal structures of the LH2 antenna [36,37]. The central cavity in this structure of LH1 was sufficiently large to accommodate a central RC. This basic architecture was further demonstrated in a number of EM studies using LH1 complexes from a variety of purple bacteria [44–47]. More recently, Atomic Force Microscopy (AFM) has been used to examine the topography of photosynthetic membranes from several species of purple photosynthetic bacteria. In those studies that achieved a sufficiently high-resolution it has been possible to count the number of pairs of α- and βpolypeptides that constitute each LH1 or LH2 complex, which present as ring-shaped structures in topographs (see [5,7] for recent reviews). Topographs of membranes from Bl. viridis, Rsp. rubrum, Rsp. photometricum and Psp. molischianum reveal LH1 as consisting of a continuous 16-member ring surrounding a central feature assigned as the RC [48–51]. In some cases it has been reported that LH1 forms an elliptical ring around the RC, matching the roughly elliptical crosssection of the RC in the membrane (e.g. see [46,48,49]). To date the most detailed picture of an RC–LH1 core is a 4.8 Å resolution X-ray structure for the complex from Rsp. palustris [52]. In this structure, the LH1 cylinder has an elliptical cross-section and is formed by 15 α and β polypeptides that encase 30 BChls. A repre-

sentation of this structure is shown in Fig. 1D. The LH1 cylinder is incomplete, with an additional transmembrane α-helix, termed W, being located near the gap between LH1 polypeptides. The structure of the Rps. palustris core-complex is therefore somewhat different from that described above for complexes from species such as Bl. viridis and Rsp. rubrum, where LH1 forms a closed 16-member ring around the RC. Scheuring et al. have recently examined photosynthetic membranes from Rps. palustris by AFM and also described the core-complex as consisting of a RC surrounded by an incomplete elliptical ring of 15 α/β pairs [53]. This is consistent with the X-ray crystal structure, although Scheuring et al. reported that the position of the gap in the LH1 ring was variable in relation to the long axis of the elliptical core-complex, in contrast to data from X-ray crystallography. As discussed below, it has been speculated that the W-polypeptide may play a similar role to the PufX protein that is found in Rhodobacter species of purple bacteria, and is the subject of the remainder of this review. 2. The PufX polypeptide The L- and M-polypeptides of the RC and α- and β-polypeptides of the LH1 antenna are encoded by the pufL, pufM, pufA and pufB genes respectively. During sequencing of the puf operon in Rb. capsulatus it was noted that the gene sequence pufBALM was followed by an open reading frame (ORF) of 237 base pairs designated C2397 [54]. The location of this ORF 13 nucleotides downstream of the stop codon for pufM, and the existence of a hairpin transcription terminator downstream of C2397, led to the suggestion that C2397 was part of the puf operon. Analysis of mRNA transcripts subsequently showed this to be the case, and the ORF was renamed pufX [55,56]. Cloning and analysis

K. Holden-Dye et al. / Biochimica et Biophysica Acta 1777 (2008) 613–630

of the puf operon from Rb. sphaeroides also identified a pufX gene transcriptionally-linked to pufBALM [57,58]. By complementing puf operon deletion mutants with a version of the puf operon lacking the pufX gene it was shown that the PufX protein is essential for photosynthetic competence in both Rb. capsulatus and Rb. sphaeroides [59–61]. Strains lacking PufX failed to grow under photosynthetic conditions despite the presence of normal levels of the RC (PS− phenotype). Experiments with antibodies subsequently confirmed that PufX is an integral membrane protein, and a component of the RC–LH1 corecomplex [62]. To signify this, in the following the core-complex in Rhodobacter species is referred to as the RC–LH1–X complex, to distinguish it from the PufX-deficient version that is assembled in species or strains that lack a pufX gene either naturally or through genetic modification (referred to as RC–LH1). An additional point that should be noted is that study of PufX-deficient strains of Rb. sphaeroides or capsulatus is possible because these bacteria assemble their photosynthetic apparatus when grown in the dark under conditions of low aeration (referred to as semiaerobic/dark). Under conditions of high aeration assembly of the photosynthetic apparatus is repressed (see [63] and references therein). The pufX gene encodes a polypeptide of 78 amino acids in Rb. capsulatus [54] and 82 amino acids in Rb. sphaeroides [58] with hydropathy plots indicating a single membrane-spanning region. Alignments of the protein sequences from the two species show a low degree of identity, reported to be between 23 and 29% depending on the exact details of the alignment [61,64,65]. The pufX gene has been identified with certainty only in bacteria of the genus Rhodobacter, which consists of the species sphaeroides, capsulatus, blasticus, azotoformans and veldkampii [66], but it is conceivable that bacteria from other genera also contain a PufX-like protein that has not been identified due to a very low sequence identity (see below). An example could be the W-polypeptide modelled into the low-resolution X-ray crystal structure of the RC–LH1 complex from Rps. palustris [52] (see above). The low degree of identity between PufX polypeptides is in marked contrast to that shown by other components of the Rhodobacter RC– LH1–X complex. For example, the degree of identity across the five Rhodobacter species is approximately 59% for the PufL and PufM polypeptides of the RC. One interpretation of the low sequence identity displayed by PufX polypeptides is that some general property of the polypeptide is important for function rather than the specific sequence of amino acids. Fulcher et al. have tested this idea by replacing the PufX from Rb. sphaeroides with the protein from Rb. capsulatus, and found that the resulting strain was capable of photosynthetic growth (PS+ phenotype) [65]. This supports the idea that PufX function is dependent

615

on some feature of the polypeptide that is relatively insensitive to precise sequence, or could indicate that a small region (or regions) of PufX where local sequence identity is high is key to its function. Fig. 2 shows the central region of an alignment of PufX polypeptides presented by Tsukatani et al. [66], from residue X20 to X70 (Rb. sphaeroides numbering, assigning the N-terminal Met as residue X0 in accord with the convention for the RC polypeptides). This region aligns with no gaps, and includes all eight residues that are absolutely conserved across all five species (grey highlight); in contrast the N- and C-terminal regions align poorly (not shown). Acidic and basic residues are shown in bold in this alignment, and it is noticeable that the section between the conserved Gly residues X29 and X51 consists very largely of apolar amino acids, suggesting the presence of a membrane-spanning α-helix. The likelihood of PufX containing such a helix was pointed out when the sequence of the protein was published for Rb. sphaeroides [58] and Rb. capsulatus [54], and the results of CD spectroscopy are consistent with PufX having sufficient α-helical content to include a membrane-spanning region [64,67]. It has been demonstrated that the N-terminus of the Rb. sphaeroides PufX polypeptide is exposed on the cytoplasmic side of the membrane, giving an Nin–Cout orientation equivalent to that of the α- and β-polypeptides of the LH1 and LH2 antenna, which also span the membrane once [68]. Analysis of the PufX polypeptides extracted from cells of Rb. sphaeroides and Rb. capsulatus has shown that they are processed at their N- and C-termini. N-terminal sequencing has established that the mature protein from both species lacks the N-terminal Met [64], and this is also the case for the protein from Rb. veldkampii [69]. Mass spectrometry and protein sequencing identified the C-terminal amino acid of the mature Rb. sphaeroides polypeptide as Ala X69, demonstrating that 12 amino acids were removed from the C-terminus [64]. Mass spectrometry also indicated removal of 9 amino acids from the C-terminus of the mature Rb. capsulatus protein, making the C-terminal residue Ala X69 (Rb. sphaeroides numbering). In both species, therefore, it is likely that the mature protein consists of 69 amino acids. The PufX protein from Rb. veldkampii has also been analysed by mass spectrometry and shown to lack the N-terminal Met and three amino acids at the C-terminus [69], leaving a protein of 79 amino acids. 3. Interactions of the PufX polypeptide As outlined above, experiments with antibodies demonstrated that PufX is a component of the RC–LH1–X complex in Rhodobacter species [62]. To look at the number of PufX proteins per core-complex, Francia et al. employed quantitative Western blotting using antibodies to a poly-His tag attached to the C-terminus of the Rb. sphaeroides PufX

Fig. 2. Alignment of PufX genes between residues X20 and X70. Conserved residues discussed in the text are highlighted in grey, acidic and basic amino acids are shown in bold. Sequences of PufX polypeptides are from Tsukatani et al. [66], for the PufX-like polypeptides from Venter et al. [123], Béjà et al. [124] and Waidner and Kirchman [125] and for the PufC polypeptides from Hucke et al. [126].

616

K. Holden-Dye et al. / Biochimica et Biophysica Acta 1777 (2008) 613–630

polypeptide to estimate the amount of PufX, and a spectroscopic assay to determine the concentration of RCs [70]. A stoichiometry of one PufX per RC was determined from these assays, and this ratio is generally accepted to be correct and has been supported by other studies (e.g. see [71]). One slightly surprising outcome from the study of Francia et al. was the conclusion that the poly-His tag was not removed by the C-terminal post-translational modification that has been reported for the Rb. sphaeroides PufX [64]. The explanation for this is not clear, but possibly the presence of the poly-His tag prevented this processing in the particular strains used. If so, then it can be concluded that failure to carry out this processing does not affect the basic function of PufX, as strains with the His-tagged PufX protein were not impaired in photosynthetic growth [70]. Insights into interactions of the PufX polypeptide with other components of the RC–LH1–X complex have come from studies involving fractionation of photosynthetic membranes with mild detergents, and other studies in which LH1 was assembled in vitro from its component polypeptides and BChl. Experiments involving deletion strains of Rb. sphaeroides in which LH1, RC and RC–LH1–X complexes were extracted from membranes with mild detergents and subjected to sucrose gradient centrifugation showed that PufX is most tightly associated with intact RC–LH1–X complexes, less tightly bound to preparations of RC-free LH1 complexes, and least tightly associated with preparations of LH1-free RCs [68]. In qualitative agreement with this, it has been reported that the accumulation of immuno-detectable levels of PufX in Rb. capsulatus strains lacking LH2 but containing the RC requires the additional presence of both the LH1 α- and β-polypeptide [72]. The presence of either polypeptide without the other did not lead to detectable levels of PufX, and levels were also lowered in Rb. capsulatus strains that contained LH1 but lacked the RC. These data point towards a situation where native levels of PufX are only obtained when all three components of the RC–LH1–X complex are present. When either the RC or LH1 antenna is absent PufX levels are decreased, particularly so in Rb. capsulatus where the expression levels of the RC are much more dependent on the presence of LH1, and vice versa, than is the case in Rb. sphaeroides [72]. Further information on the interactions of the PufX polypeptide has come from studies of the reconstitution of LH1 antenna complexes from purified α- and β-polypeptides and BChl. Such studies involve the initial formation of a building block “subunit complex”, absorbing maximally around 820 nm (denoted B820) [73], followed by formation of a native LH1 aggregate absorbing around 870 nm. In the case of Rb. capsulatus, addition of the PufX polypeptide decreased the amount of both the B820 subunit complex and native LH1 that could be reconstituted [67], with 50% inhibition at a ratio of 0.5 PufX per LH1 α-polypeptide. A similar result was seen for inhibition of assembly of the Rb. sphaeroides LH1 complex by the Rb. sphaeroides PufX [67]. In contrast, neither PufX was capable of inhibiting formation of a non-native B820 complex between β-polypeptides. These findings, together with an observed tendency for the PufX protein to co-purify with the LH1 α-polypeptide, led to the proposal that PufX preferentially interacts with the LH1 α-polypeptide, interfering with in vitro assembly of the LH1 antenna [67]. By analogy with the LH2 antenna [36,37], a ring of α-polypeptides is expected to form the inner wall of the hollow cylinder of LH1 protein surrounding the RC, with a ring of β-polypeptides forming the outer wall. To look at the roles of different regions of PufX, Parkes-Loach et al. chemically synthesised polypeptides corresponding to the N- and C-terminal regions of the Rb. sphaeroides PufX and the central region of the proteins from both Rb. sphaeroides and Rb. capsulatus [64]. The N- and C-terminal segments, corresponding to amino acids X1–X26 and X48–X81, had no effect on reconstitution of the Rb. sphaeroides LH1 complex, but a central region corresponding to X25–X49 inhibited this reconstitution by 50% at a ratio of 3 PufX per LH1 α-polypeptide. Inhibition of in vitro formation of the Rb. capsulatus LH1 was achieved using a polypeptide corresponding to the central and C-terminal regions

of the Rb. capsulatus PufX (X24–X66, Rb. capsulatus numbering). In addition, and in contrast with findings using native PufX proteins [67], cross-species inhibition was seen using the Rb. sphaeroides X25–X49 polypeptide and the Rb. capsulatus X24–X66 polypeptide [64]. These data indicate a particular role for the central membrane-spanning region of PufX in the interaction with LH1. The influence of PufX on the in vivo assembly of LH1 has also been examined, using a Rb. capsulatus strain that lacked both the LH2 antenna and the RC. Overexpression of PufX was found to have no effect on the rate and level of assembly of LH1 following a switch from aerobic/dark to semiaerobic/dark growth [74], which is at odds with the inhibitory effect on in vitro assembly described above [64,67]. Regarding interactions of PufX with pigment cofactors, evidence has been presented in support of the proposal that the X25–X49 fragment of the Rb. sphaeroides PufX and the X24–X66 fragment of the Rb. capsulatus PufX bind BChl in vitro with an affinity similar to that shown by the corresponding LH1 α-polypeptide [75]. Residues His X58/Gln X59 (Rb. sphaeroides) and the equivalent Asn X59/Gln X60 (Rb. capsulatus) have been suggested as possible BChl ligands [75]. However, it remains to be established whether PufX binds BChl in vivo. In addition Aklujkar and Beatty [74] have reported that overexpression of PufX causes changes in the absorbance spectrum of a strain with an RC-only phenotype when either the LH1 α- or β-polypeptide is present, and attributed the changes to alteration of BChl binding by the RC or binding of BChl by PufX and/or the LH1 polypeptide. However these spectra were of rather poor quality due to the fact that expression levels of the Rb. capsulatus RC are low in the absence of LH1. Finally, Aklujkar and Beatty have also examined the propensity of the Rb. capsulatus PufX to bind to itself, using the TOXCAT assay [76]. The segment between Gly residues X29 and X51 was found to have a moderate propensity for self association, raising the possibility that PufX could form a homodimer [74]. Possible implications of this are returned to below. 4. Effects of PufX on membrane morphology and organisation Section 5 looks at the influence of PufX on the structure and organisation of the RC–LH1–X complex. However, before dealing with this it is relevant to discuss how altering the pigment-protein composition of Rb. sphaeroides affects the morphology and organisation of the photosynthetic membrane. In wild-type strains of Rb. sphaeroides grown photosynthetically, or in the dark with relatively low aeration, the photosynthetic membrane consists predominantly of near spherical invaginations of the cytoplasmic membrane [77]. Isolated membranes consist of sealed vesicular structures, the classic “chromatophore” [78], with the cytoplasmic face of the membrane exposed to the external phase and material from the periplasmic space, including cytochrome c2, trapped in the vesicle interior. However, in mutants where the LH2 antenna is absent a major change in membrane morphology is observed [79–82], with images taken from thin sections of such cells being dominated by tubular membrane structures of ~100 nm in diameter, and up to 2 μm in length, that can cause elongation of the bacterial cell. EM of isolated tubular membranes has shown that they contain highly-ordered arrays of protein [83–85], and spectroscopic analysis has shown that they contain high concentrations of RC–LH1–X complexes. An important feature of the organisation of the photosynthetic apparatus in such LH2-deficient strains, highlighted by an analysis carried out by Siebert et al. [86], is that only a proportion of the RC– LH1–X complexes is located in these highly-ordered tubular membranes. The remainder occupy other regions of the cytoplasmic membrane that surrounds the cell and can be isolated as vesicles or membrane fragments. A striking difference between the two types of membrane is the absence of the cytochrome bc1 complex in highly pure preparations of tubular membranes [86]. Although only a limited

K. Holden-Dye et al. / Biochimica et Biophysica Acta 1777 (2008) 613–630

number of proteins have been analysed, this result suggests that the tubular membranes are highly enriched in the RC–LH1–X complex, these forming well ordered para-crystalline arrays in the absence of large amounts of other proteins. Presumably these tightly packed RC– LH1–X complexes induce a curvature that results in the tubular architecture for the membrane (see [87]). In contrast, the remaining areas of membrane contain RC–LH1–X complexes at lower concentrations, together with the cytochrome bc1 complex and the variety of other membrane proteins that are required for prosperity of the cell. Tubular membranes are not observed in strains that lack LH2 and also lack the PufX protein. Membranes isolated from such cells consist of a mixture of vesicles and fragments similar to those obtained with PufX-containing strains, together with larger vesicles of up to 1 μm in diameter that also fold out into large open sheets [86]. Thin sections of PufX-deficient cells show large vesicular structures of up to 400 nm in diameter. These larger vesicles and sheets contain PufX-deficient RC– LH1 complexes organised in hexagonally-packed arrays, and lack the cytochrome bc1 complex [86]. The high degree of order displayed by RC–LH1–X complexes in tubular membranes also extends to the RC component. Interrogation of oriented tubular membranes with horizontally and vertically polarized light has yielded spectra consistent with a uniform orientation of the RCs relative to the orientation axis, indicating that each RC has a fixed orientation relative to the long axis of the tubular membrane [88]. Interestingly, when LH2-deficient membranes that also lacked the PufX polypeptide were exposed to the same orientation procedure (squeezing in two dimensions) these did not show evidence of long-range ordering of the RC component of the RC– LH1 complexes. The reason for this is discussed below. Additional data on how loss of PufX affects membrane morphology in strains containing LH2 have been reported [87,89,90]. Tubular membranes are not observed in the presence of LH2, but LD spectroscopy and AFM have provided evidence for ordered arrays of RC–LH1–X complexes, the degree of order extending to the RC components within these complexes [87,89,90]. These ordered regions of RC–LH1–X complexes are interspersed with domains of LH2 complexes. In the absence of PufX the membrane appears to also segregate into RC–LH1 domains and LH2 domains, with again a loss of order of the RC component of the RC–LH1 complex. The reason for this finding is also considered below. Frese et al. have recently used Monte Carlo simulations to examine how membrane curvature and the segregation of RC–LH1–X and LH2 complexes into domains with different degrees of order/fluidity is driven by factors such as protein geometry and crowding [87]. This work has shown that domain formation does not necessarily have to involve specific molecular interactions [87]. Additional evidence that the organisation of the photosynthetic membrane changes in the absence of PufX has come from fluorescence-based measurements that show decreased levels of connectivity between RC–LH1 complexes in the absence of PufX [91], and spectroscopic data indicating a near doubling in the size of the quinone pool per RC in the absence of PufX [91].

617

understanding of the structure and organisation of this complex in Rhodobacter species. As outlined above, early studies on a range of purple photosynthetic bacteria established a general pattern where a central RC is surrounded by a hollow cylindrical LH1 antenna. For Rhodobacter species a significant breakthrough was made in 1999 when Jungas et al. published the results of a cryoEM study of negatively-stained photosynthetic membranes from a strain of Rb. sphaeroides devoid of the LH2 antenna [84]. These extensive tubular membranes contain arrays of highly-ordered RC–LH1–X complexes (see above), and projection maps at a resolution of 20 Å revealed an Sshaped structure interpreted as two RCs each of which was associated with an incomplete C-shaped arc of LH1. A crude cartoon of the proposed structure based on Fig. 5 in Ref. [84] is shown in Fig. 3. Based on estimates of the amount of LH1 BChl per RC, each C-shaped LH1 antenna was proposed to consist of approximately 12 pairs of α- and β-polypeptides with 24 BChls. Each RC was proposed to occupy a welldefined orientation within the LH1 arc, rather than being arranged randomly. The density at the interface between the two RC–LH1–X units was tentatively assigned to a cytochrome bc1 complex, consistent with proposals that in Rb. sphaeroides the photosynthetic electron transfer apparatus is organised in supercomplexes consisting of two RCs and one cytochrome bc1 complex (see [93] and references therein). However this assignment seems questionable given the finding described above that highly pure tubular membranes do not contain the bc1 complex [86], and the interpretation of subsequent, higher-resolution data (see below). It was further suggested that the gap in the LH1 arc could be important for quinone diffusion between the RC and the cytochrome bc1 complex [84]. An obvious interpretation of the arrangement of RC–LH1–X complexes presented by Jungas et al. is that this complex is dimeric. This interpretation agrees with findings from experiments involving extraction of complexes from Rb. sphaeroides membranes and fractionation using sucrose density gradient centrifugation [70]. Such an experiment conducted on a strain containing a normal complement of RCs, LH1, LH2 and PufX produced four pigmented bands, corresponding to (in order of increasing mass) the LH2 antenna, LH1 complexes that were (almost) free of RCs, monomeric RC–LH1–X complexes and dimeric RC–LH1–X complexes. The last of these assignments was confirmed by EM of single particles. Intriguingly, an equivalent fractionation experiment conducted on a strain lacking PufX produced the three pigmented bands corresponding to LH2, free LH1 and monomeric RC–LH1 complexes, but gave no indication of dimeric RC–LH1 complexes. The conclusions drawn from this work, therefore, were that the Rb. sphaeroides RC–LH1–X complex can indeed be isolated in a dimeric form, but that the presence of these dimers is dependent on the presence of the PufX protein [70]. Fig. 4 shows fractionation of RC–LH1 and LH2 complexes extracted from strains of Rb. sphaeroides either possessing or lacking the PufX protein grown under semiaerobic conditions in the dark (Jones, M.R and Crouch, L.I., unpublished data). Complexes were extracted from intracytoplasmic membranes using 4% β-dodecylmaltoside [94] and

5. Influence of PufX on the structure and organisation of RC–LH1–X core-complexes It has been clear for many years that the PufX protein has a significant influence on the structure and organisation of individual RC–LH1–X complexes. Many of the early studies of the effects of deletion of the pufX gene reported changes in the expression levels of the antenna complexes with, in particular, increases in the LH1 complex relative to LH2 [59–62,92]. Subsequently a number of studies included a quantitative estimate of the increase in the amount of LH1 BChl per RC consequent on removing PufX, and these are discussed in the next section. Understanding of the influence of PufX on the properties of RC– LH1–X complexes has developed alongside improvements in our

Fig. 3. Cartoon of dimeric core-complex structure based on data from Jungas et al. [84].

618

K. Holden-Dye et al. / Biochimica et Biophysica Acta 1777 (2008) 613–630

Fig. 4. Fractionation of photosynthetic complexes from PufX-containing (X+) and PufXdeficient (X−) strains of Rb. sphaeroides. Complexes were extracted from membranes using 4% β-dodecylmaltoside and separated by sucrose gradient centrifugation (see text).

separated by sucrose gradient centrifugation. In the case of the PufXcontaining strain, the RC–LH1–X complexes migrate as two bands corresponding to a monomeric and dimeric form, whereas in the absence of PufX only the monomeric form is resolved. As an aside, the two closely spaced bands higher in each gradient shown in Fig. 4 correspond to the LH2 antenna, the complexes in the two bands having identical absorbance spectra (not shown). Gubellini et al. [69] have reported that the LH2 antenna from Rb. veldkampii also migrates as two bands on sucrose gradients, and on the basis of size exclusion chromatography and electron microscopy have concluded that the two bands correspond to a 7 nm diameter LH2 similar to the octameric or nonomeric complexes characterised by X-ray crystallography [36,37] and a larger (12 nm diameter) complex. Thus it would appear that, at least in these two species, there is some heterogeneity in the size of the LH2 complexes, but this variation does not affect the details of their absorbance spectra.

Returning to the properties of the RC–LH1 complex, an indication of the effect of removal of PufX on the structure and monomer/dimer organisation of the RC–LH1 complex came from cryoEM of 2D crystals of a PufX-deficient Rb. sphaeroides complex [44]. Projection maps and single particle analysis showed a central feature attributed to the RC completely surrounded by a circular feature attributed to LH1. Together with the data described above this suggested that in the presence of PufX a dimeric RC–LH1–X complex is formed consisting of two RCs encased in an S-shaped antenna, whereas in the absence of PufX a monomeric RC–LH1 complex is formed consisting of a single RC encased in a circular antenna. This was shown to indeed be the case in a direct comparison of PufX-containing tubular membranes and PufXdeficient open-sheet membranes isolated from LH2-deficient strains, and analysed by EM [86]. The PufX-deficient membranes were shown to consist of hexagonally-packed monomeric RC–LH1 complexes in which the antenna component completely surrounded the central RC. As outlined above, and shown in Fig. 4, sucrose gradient fractionation of detergent-extracted RC–LH1 complexes yields a single (monomer) band, whereas fractionation of RC–LH1–X complexes yields two bands attributable to monomers and dimers. Given that dimeric RC–LH1–X complexes have been visualised in intact tubular membranes, the latter pattern either indicates a mixed population of monomeric and dimeric RC–LH1–X complexes in the cell or conversion (breakage) of dimers into monomers during the extraction process. Scheuring et al. have examined the structure of monomeric and dimeric RC–LH1–X complexes isolated in this way from a wild-type strain of Rb. sphaeroides [95]. Experiments involving cryoEM on 2D crystals formed from either monomers or dimers showed arrays of S-shaped dimeric RC–LH1–X complexes similar to those reported by Jungas et al. [84] and Siebert et al. [86] for complexes in native tubular membranes. In the case of the 2-D crystals an alternating up-down arrangement of adjacent rows of dimeric complexes resulted in an approximately flat 2-D sheet, rather than the highly curved morphology seen in native membranes where all core-complexes have the same orientation. In assigning the density features in projection maps, Scheuring et al. commented on a peak of density at the point of contact between the two monomers in the dimer, in the region of antenna density located in-between the two RCs, and assigned this peak to two molecules of PufX [95]. The main basis for this assignment was the observation that a dimer of PufX protein is expected to be approximately 60% larger than an α/β heterodimer of LH1 polypeptides. A cartoon of the proposed structure, based on Fig. 5 from [95], is shown in Fig. 5A. Each

Fig. 5. Cartoons of core-complex structure. (A) Dimeric RC–LH1–X complex from Rb. sphaeroides based on data from Ref. [95]. (B) Monomeric core-complex from Rps. palustris based on data from Ref. [52], including the W-polypeptide. (C) Dimeric RC–LH1–X complex from Rb. sphaeroides based on data from Ref. [94]. (D) Dimeric RC–LH1–X complex from Rb. blasticus based on data from Ref. [98]. In B, C and D the RC is shown as an ellipse with the long axis indicated by the line and the approximate position of the QB binding site indicated. In all panels the assigned position of the PufX or W-polypeptide is shown.

K. Holden-Dye et al. / Biochimica et Biophysica Acta 1777 (2008) 613–630

C-shaped LH1 domain was proposed to consist of 12 α/β heterodimers, based on size and measured ratios of LH1 BChls per RC. No orientation was assigned for the RC components, but these were proposed to adopt fixed, symmerical orientations within the dimer. For comparison, Fig. 5B shows a cartoon of the monomeric corecomplex from Rps. palustris [52]. Qian et al. [94] have reported a projection map at 8.5 Å resolution from EM of 2D crystals of the Rb. sphaeroides RC–LH1–X complex. The map was interpreted as an S-shaped dimeric complex, with each RC encased by a C-shaped LH1 antenna consisting of 14 α/β heterodimers. A cartoon of this structure based on the model presented in Fig. 4 of [94] is shown in Fig. 5C. The orientation of the RC was determined by calculating a projection map of the RC at 8.5 Å resolution using the X-ray crystal structure and matching this to the experimental projection map for the RC–LH1–X complex. This produced an orientation for the RC primary donor BChls relative to the short axis of the dimer that was within 5° of that predicted by LD spectroscopy [88] (see above), and on the basis of calculations Geyer [96] has recently concluded that this orientation is optimal for coupling of the LH1 BChls to the RC primary donor BChls. The PufX protein was assigned to a strong density feature sandwiched between the RC and the terminal α-polypeptide of the LH1 antenna. In addition to EM, AFM has brought increasingly detailed information on the structure of the RC–LH1–X complex from a variety of bacterial species (for recent reviews see [5,7]), and fuelled debate on the location and function(s) of PufX. So, for example, Scheuring et al. have presented AFM images of fused fragments of membrane from Rb. blasticus, a species closely related to Rb. sphaeroides that also contains a PufX protein [98]. Images of RC–LH1–X complexes from this species were interpreted in terms of an S-shaped dimeric complex in which each C-shaped LH1 antenna consisted of 13 α/β heterodimers. A cartoon of this structure based on the model presented in Fig. 6A of [98] is shown in Fig. 5D. As with their previous model of the Rb. sphaeroides complex derived from data from EM (see above — Fig. 5A), topographical features at the monomer:monomer interface were interpreted as two molecules of PufX. In this analysis of fused Rb. blasticus membranes, some 25% of the observed RC–LH1–X complexes were monomeric in form, consisting of a RC within an incomplete arc of LH1 [98]. On the basis of their open morphology these were interpreted as being halves of broken dimers rather than a discrete monomeric form, leading to the conclusion that all of the RC–LH1–X complexes in membranes from this species are dimeric. Together with the observation that PufX-deficient Rb. sphaeroides RC–LH1 complexes are monomeric, as are complexes from species that lack a PufX protein such as Rsp.rubrum and Bl. viridis, this could suggest a correlation between the presence of PufX and a dimeric organisation for the RC–LH1–X complex. However this does not appear to be the case. Levy and co-workers have reported a monomeric organisation for the RC–LH1–X complex from Rb. veldkampii, based on data from sucrose gradient fractionation of complexes isolated from membranes using mild detergents, together with EM of individual complexes or AFM of complexes reconstituted into planar lipid bilayers [69,97,99]. Electron microscopy data collected on single Rb. veldkampii RC–LH1–X complexes at a resolution of 12 Å has been interpreted in terms of a monomeric complex very similar in architecture to that of Rps. palustris, with a central RC surrounded by an incomplete ring of 15 α/β pairs of LH1 polypeptides, and the PufX protein located near a gap in the LH1 cylinder, close to the QB site of the RC [99]. The principal difference between the two structures was a much less elliptical arrangement of LH1 proteins around the RC in Rb. veldkampii than is seen in the Rps. palustris complex (long/short axis ratios of 1.03 and 1.19, respectively [99]). This finding with Rb. veldkampii would suggest that there is no simple correlation between the presence of PufX and dimeric organisation, and this may also have relevance to findings described above for Rps. palustris where the RC–LH1 complex contains a W-polypeptide

619

that may be a PufX analog [52] but where the in vivo organisation is monomeric [53]. One caveat to this is that the organisation of RC–LH1–X complexes in intact membranes of Rb. veldkampii has not been reported, and it is conceivable that the monomer:monomer interaction is unusually weak in this species. This said, EM images of single Rb. veldkampii RC–LH1–X complexes seem to show a more-or-less complete ring of protein surrounding the RC [69], similar to the arrangement in Rps. palustris [52,53], rather than the more open C-shaped antenna seen in structures interpreted as monomers arising from broken dimers in fused membranes of Rb. blasticus [98]. Given this, it seems probable that the Rb. veldkampii RC–LH1–X complex is indeed monomeric in native membranes. AFM has also been used to investigate the organisation of RC and antenna complexes in membranes from a wild-type strain of Rb. sphaeroides [89]. Chromatophore vesicles were flattened onto a substrate through the addition of a low concentration of detergent (0.03% β-dodecyl maltoside), sufficient to disrupt the vesicular structure of the chromatophore without removing complexes from the membrane. This study also revealed a dimeric structure for the majority of RC–LH1–X complexes, with dimers tending to be organised in rows several dimers long. These rows (termed “linear arrays”) were interspersed with LH2 complexes, such that adjacent RC–LH1–X arrays were separated by an average of two LH2 units. An account of variations in this arrangement in response to growth conditions has been published [109], and the AFM and EM studies of the structure of the RC–LH1–X membrane and the larger photosynthetic membrane outlined in this section have inspired attempts to model an entire chromatophore membrane system in silico in both structural and functional terms [110–112]. A feature of all of the proposed models for the RC–LH1–X complex from Rb. sphaeroides and Rb. blasticus is a fixed geometry of the RC components relative to the long axis of the dimer. This arrangement is in qualitative agreement with the data obtained from LD spectroscopy of oriented photosynthetic membranes from Rb. sphaeroides, described above, that show that all of the RCs have a common orientation relative to the orientation axis of the sample (but see below). To summarise the data described in this section, in Rb. sphaeroides and Rb. blasticus the RC–LH1–X complex is dimeric, with two RCs that are surrounded by an LH1 antenna that adopts a characteristic S-shape in images obtained through EM and AFM. In Rb. sphaeroides these dimers are organised in linear arrays of several complexes in the roughly spherical membranes found in wild-type strains, and in the absence of LH2 form more extensive highly-ordered aggregates that impose a tubular structure on the membrane. In LH2-deficient strains of Rb. sphaeroides that also lack PufX there is a switch to a monomeric RC–LH1 complex with a larger LH1 antenna that seems to completely surround the RC. This also changes the gross morphology of the membrane to very large vesicles, the monomeric RC–LH1 complexes forming a hexagonally-packed array with the central RCs adopting random orientations. The proposed models of the dimeric Rb. sphaeroides and Rb. blasticus RC–LH1–X complexes shown in cartoon form in Fig. 5C and D are similar in general terms, but differ in a number of key features that include the location of the PufX polypeptide. Both models envisage a dimeric complex of roughly 20 nm× 10 nm with two RCs and an S-shaped encircling LH1 antenna. However a major difference concerns the orientation of the RC with respect to the long axis of the dimer, and to illustrate this in Fig. 5C and D the long axis of the RC is represented by a line. In the model of the Rb. blasticus RC–LH1–X complex (Fig. 5D) the long axis of each (roughly oval) RC is arranged approximately perpendicular to the long axis of the dimer [98], whereas in the model of the Rb. sphaeroides RC–LH1–X complex (Fig. 5C) the long axis of each RC is much closer to being parallel to the long axis of the dimer [94]. Despite this difference, in both models the PufX polypeptide is positioned adjacent to the long axis of the RC, on the opposite side to

620

K. Holden-Dye et al. / Biochimica et Biophysica Acta 1777 (2008) 613–630

the membrane-spanning helix of the H-polypeptide and close to the QB binding pocket. This positioning is similar to that modelled for the Wpolypeptide in the 4.5 Å resolution structure of the monomeric Rps. palustris RC–LH1 complex depicted in Fig. 5B. In the case of the model of the Rb. blasticus RC this places PufX at the dimer interface (Fig. 5D), whereas in the model of the Rb. sphaeroides RC it places PufX along the long axis of the dimer, close to the break in the continuity of the LH1 antenna (Fig. 5C). It remains to be seen whether the major differences between these models reflect the challenge of modelling a structure based on relatively low-resolution data, or whether they represent genuine differences between the architecture of the dimer in the two species. Although the second of these possibilities might seem unlikely, a cautionary note is sounded by the proposals described above, that the RC–LH1–X complex from Rb. veldkampii is monomeric [69], and so there may be considerable variation in the morphology of the RC–LH1–X complex between different species of Rhodobacter. Regarding the difference in organisation of the RC–LH1–X complex in Rb. sphaeroides and Rb. blasticus (dimeric) versus Rb. veldkampii (monomeric), Busselez et al. [99] have recently pointed out that a GXXXG dimerisation motif found in Rb. sphaeroides PufX (involving Gly X31 and X35 — see Fig. 2) is not present in the Rb. veldkampii PufX, where the equivalent residues are Gly and Val, respectively. It has been proposed that this GXXXG motif is important for dimerisation of PufX at the interface of the Rb. sphaeroides RC–LH1–X dimer [99], as such a motif has been found to drive dimerisation of glycophorin A and a number of other membrane proteins [100]. The monomeric structure of the Rb. veldkampii RC–LH1–X complex has therefore been ascribed to replacement of this GXXXG sequence with GXXXV, the bulkier Val X35 preventing formation of a PufX dimer in line with experimental observations with glycophorin A [101]. However, this argument is undermined by the observation that in Rb. blasticus, where the RC– LH1–X complex is known to be dimeric, the equivalent sequence is FXXXV, and both Gly residues of the GXXXG motif are replaced by bulkier amino acids. Thus it seems rather unlikely that this “motif” is relevant to dimerisation of PufX and/or RC–LH1–X complexes. As can be seen in the alignment in Fig. 2, the GX31XXXGX35 sequence is unique to Rb. sphaeroides, with either one or both Gly residues being absent in the remaining Rhodobacter PufX sequences. Finally, appreciation of the effect of removal of PufX on the structure of the RC–LH1 complex assists in the interpretation of data on membrane morphology and organisation discussed in the last section. In the absence of LH2, dimeric RC–LH1–X complexes form extensive ordered linear arrays, and the shape of the RC–LH1 complex gives rise to an overall tubular membrane morphology. As the RC components have fixed and symmetrical orientations relative to the long axis of the dimer then the RC shows a strong degree of order relative to the long axis of such tubular membranes [88]. In the absence of both LH2 and PufX the organisation of the RC–LH1 complexes switches to a monomeric complex with a closed ring of LH1 protein. This packs into an ordered hexagonal array but the change in protein shape between S-shaped dimers and O-shaped smaller monomers and the change in protein packing no longer leads to a tubular membrane architecture. The RC adopts a random orientation within each O-shaped LH1 ring, and the RC poulation therefore no longer shows long-range order [88]. In the presence of LH2 the RC–LH1 and LH2 components segregate into separate domains, PufX-deficient RC–LH1 complexes forming hexagonal arrays of monomers, and RC–LH1–X complexes forming linear arrays of dimers [87,90]. Individual domains of RC–LH1 or RC–LH1–X complexes are interconnected by domains of LH2 antenna [89]. One additional factor is the observation that some domains of RC–LH1 complexes do show evidence of an ordered arrangement of central RCs — this is proposed to arise in domains where RC–LH1 complexes are very tightly packed, the normally O-shaped monomers being forced into a more elliptical conformation within the array resulting a more uniform orientation of the RC component [87].

6. Effects of PufX on the composition of RC–LH1–X core-complexes As outlined at the start of the last section, it is well established that removal of the PufX polypeptide causes an increase in the amount of LH1 BChl present per RC [59–61,92]. In qualitative terms this is in agreement with the structural studies described above that have shown a switch from an open arc of LH1 protein surrounding the RC in the presence of PufX to a larger closed circle of LH1 protein in its absence. A number of reports have attempted to quantify the amount of LH1 BChl per RC in PufX-containing membranes or isolated complexes, and the effect on this ratio of removal of PufX. Abresch et al. have reported a ratio of 26.6 ± 0.9 LH1 BChls per RC [71] based on measurements on purified RC–LH1–X complexes from Rb. sphaeroides, whereas Ueda et al. have estimated this number to be between 28 and 32 [102], and Franke and Amesz reported a ratio of 27.5 ± 3.2 [103]. The first two studies used strains with native red/ brown carotenoids, but using semiaerobic/dark- and photosynthetically-grown material, respectively, whereas the last study used photosynthetically-grown cells of the carotenoid-deficient R-26 strain of Rb. sphaeroides. A study using strains of Rb. sphaeroides with a single genomic copy of an intact or pufX-deficient puf operon has specified a 17% increase in the level of LH1 BChl per RC in the absence of PufX, from 41 LH1 BChls per RC to 48, based on spectra of membranes that also contained LH2 [104]. These strains also contained non-native green carotenoids rather than the native red-brown carotenoids (see Section 11 for a discussion on carotenoids and their interactions with PufX). Using similar green strains, Francia et al. analysed RC–LH1–X and RC–LH1 complexes isolated on sucrose gradients, and reported ratios of LH1 BChl per RC of 27 for dimeric RC–LH1–X complexes and 36 for RC–LH1 complexes [70]. In a subsequent report, Francia et al. reported values for RC–LH1–X and RC–LH1 complexes isolated in the same way of 21 and 33, respectively [105]. McGlynn et al. also estimated the amount of LH1 BChl per RC in membranes from strains of Rb. sphaeroides lacking the LH2 antenna protein [106]. In experiments carried out on strains with native carotenoids the amount of LH1 BChl per RC increased from approximately 31 to 35, whereas in experiments with strains containing green carotenoids the ratio increased from 24 to 35. To summarise these data, it seems clear that removal of PufX causes an increase in the amount of LH1 BChl relative to the RC. However estimates of the extent of this increase vary somewhat, as do estimates of the actual amount of LH1 BChl per RC in native RC–LH1–X complexes. Part of this variation is likely to come from the difficulty in accurately estimating amounts of LH1 and RCs in membranes that also contain LH2, which could account for the very high values reported in Ref. [104]. The data reported for green strains by Francia et al. (on solubilised complexes) and McGlynn et al. (on LH2-deficient membranes) agree reasonably well, with values of LH1 BChl per RC of 27, 24 and 21 for RC–LH1–X complexes and 36, 35 and 33 for RC–LH1 complexes. The three sets of data suggest an increase in antenna size corresponding to 9–12 LH1 BChls in the absence of PufX. In contrast the data reported by McGlynn et al. for strains with native carotenoids reported only a small increase of 4 LH1 BChls in the absence of PufX. This much smaller change was due to a higher value of LH1 BChl per RC in the PufX-containing strain with native carotenoids (31, as opposed to 24 in the equivalent green strain), and indicates an influence of carotenoid type on the composition of the RC–LH1–X complexes. This point is returned to in Section 11. In the absence of PufX the estimated amounts of LH1 BChl per RC show good agreement in the data of Francia et al. [70,105] and McGlynn et al. [106], with values of 36 and 33 (green carotenoids) and 35 (both native and green carotenoids). This suggests that the composition of the enlarged RC– LH1 complex assembled in the absence of PufX does not vary according to the type of carotenoid present. One point to note is that it is conceivable that ratios of LH1 per RC could be variable in strains where the native composition of the

K. Holden-Dye et al. / Biochimica et Biophysica Acta 1777 (2008) 613–630

photosystem is disrupted. For example, Bahatyrova et al. [107] have reported variability in the size of LH1 rings visualised by AFM in 2-D crystals formed from LH1 complexes isolated from an LH1-only strain of Rb. sphaeroides. 7. Effects of PufX deletion on photosynthetic growth, and suppression mutations As outlined above, a number of studies have shown that deletion of the pufX gene leads to a loss of photosynthetic growth in both Rb. capsulatus and Rb. sphaeroides. However, on detailed analysis in liquid culture it was noted that a capacity for photosynthetic growth (PS+ phenotype) was spontaneously restored to strains lacking the pufX gene following incubation of cultures for several tens of hours [60– 62,108]. It was shown that this restoration was attributable to spontaneous mutations that change the properties of the PufXdeficient RC–LH1 complex, and a number of these mutations were mapped to the α- and β-polypeptides of LH1 [113,114]. It should be mentioned that these experiments were carried out using strains in which part or all of the genes of the puf operon had been removed from the bacterial genome and a pufX-deficient version of the puf operon introduced on a low-copy-number plasmid. These strains therefore contained RCs, LH1 and LH2 but no PufX, and contained several copies of the pufBALM genes. Following DNA sequencing the puf operon with the mutant pufB/A gene was reintroduced into a strain lacking additional copies of pufBA. The precise nature of these suppressor mutations, and their effect on the photosynthetic apparatus, has been discussed in detail [113,114], and in the light of more recent analyses they can be classified into four categories: The first comprises Pro and Phe mutations of residue Ser 2 of the Rb. capsulatus LH1 α-polypeptide (denoted Ser α2), and a Trp β47 Y Gly mutation in Rb. sphaeroides. These mutations restore the PS+ phenotype whilst retaining expression of an LH1 complex at the high levels characteristic of the parental PufX-deficient mutant [72,113]. A possible interpretation of this phenotype is that the mutations alter the structure of the enlarged LH1 antenna arising from the absence of PufX, such that it no longer disrupts the process(es) required to support photosynthetic growth. Further insights into the Rb. capsulatus mutations are considered in Section 12. The second category comprises Ser α47 Y Phe and His β20 Y Arg mutations in Rb. sphaeroides. These restore a weak PS+ phenotype and expression of an LH1 complex occurs but at a somewhat reduced level [113]. This phenotype is similar to one reported for strains of Rb. sphaeroides with truncations of the C-terminus of pufA that lower the level of LH1 expression relative to that of the RC, and which also exhibit a PS+ phenotype in the absence of PufX [115]. In these strains the PS− phenotype seems to be overcome by decreasing the extent of the LH1 antenna surrounding the RC. The third category comprised a Trp β47 Y Arg mutation in Rb. sphaeroides that restored the PS+ phenotype but produced a loss of expression of the LH1 antenna. This represents a more extreme version of the previous category, and the properties of this strain are consistent with the finding that the impairment of photosynthetic growth consequent on removal of PufX is not seen in strains that lack an LH1 antenna due to complete removal of the pufBA genes. McGlynn et al. reported that photosynthetic growth of a so-called RC-only strain, lacking both LH1 and LH2, was not affected by the presence or absence of the pufX gene [106]. The fourth category showed photosynthetic growth only when a second wild-type copy of the pufBA genes was present, in contrast to the preceeding three types. The mutations include Trp α43 Y STOP and Trp β44 Y STOP in Rb. sphaeroides [113]. In the presence of a second copy of pufBA these mutants had normal levels of LH1 and were PS+, but in the absence of the additional copy they did not assemble LH1 and were PS−. The interpretation of this pheotype is that

621

the wild-type copy of the α- or β-polypeptide is required for assembly of LH1, with the mutant α- or β-polypeptide disrupting the structure of LH1 in a way that overcomes the absence of PufX. These mutations remove 15 and 4 amino acids, respectively, from the C-terminus of the α- and β-polypeptide. The last two categories of mutation provide an interesting contrast. It is well established that Rb. sphaeroides will grow under photosynthetic conditions in the absence of PufX if the genes for pufBA are also absent [106]. Such strains grow more rapidly and to a higher optical density than strains that lack both LH1 and LH2, which indicates that LH2 provides excitation energy to the RC in the absence of LH1 [106,116]. However, the results obtained with the Trp α43 Y STOP and Trp β44Y STOP mutations expressed in the absence of extra wild-type copies of pufBA show that the mutant α- or β-polypeptide prevents photosynthetic growth of the bacterium despite the presence of RCs and LH2 complexes. Although the reasons for this are not clear, the finding suggests that the mutant α- or β-polypeptide interferes with the function of the RC when normal LH1 α- and β-polypeptides are not present. 8. Importance of PufX for cyclic electron transfer A number of experiments have provided evidence that the absence of PufX produces an impairment in cyclic electron transfer [61,62,91,104,117]. A study by Barz et al. [104,117] looking at different aspects of cyclic electron transfer over a range of redox potentials concluded that the absence of PufX impaired release of ubiquinol from the RC–LH1–X complex (or its diffusion to the cyt bc1 complex) at high redox potentials. This study also identified a particular importance of PufX for multiple turnover of the cyclic electron transfer chain at low redox potential, conditions relevant to growth in the absence of oxygen or alternative electron sinks. This was interpreted as a limitation in the rate of ubiquinone binding by the RC–LH1–X complex under conditions where the quinone pool is predominantly reduced [104,117]. Thus PufX seemed to be important for ubiquinol release and/or diffusion to the bc1 complex under oxidising conditions, and ubiquinone diffusion and/or binding to the RC under reducing conditions. Under neutral conditions, when the quinone pool consisted of a mixture of ubiquinone and ubiquinol, a lack of PufX had minimal effects. More recently it has been reported that the absence of PufX causes an approximately 2–3 fold slower rate of turnover of quinone at the QB site of the RC (i.e. double reduction/protonation of QB and replacement of the quinol by a new quinone) [91]. This effect was seen both in membranes and with isolated complexes (dimeric RC–LH1–X complexes compared with monomeric RC–LH1 complexes) showing that it is due to a defect in the core-complex itself, rather than to a larger range effect on the organisation of the photosynthetic membrane. The rate of diffusion of quinol from the RC to the cytochrome bc1 complex was also approximately 2-fold slower in the absence of PufX [91]. A prediction of these experiments is that the PS− phenotype displayed by PufX-deficient strains could be overcome by poising the intramembrane quinone pool at redox potential where over-reduction or over-oxidation is avoided, and indeed evidence has been presented that this is the case. Two studies have shown that a PufX-deficient strain will grow at near-to-normal rates under photosynthetic conditions if the growth medium is supplemented with an auxiliary oxidant such as dimethylsulphoxide (DMSO) or trimethylamine oxide (TMAO) [117,118]. This finding parallels an observation made with wild-type strains of Rb. capsulatus, that photosynthetic growth under anaerobic conditions in the presence of highly reducing carbon sources such as propionate or butyrate is only possible when the growth medium is supplemented with an auxiliary oxidant such as TMAO or DMSO [119,120]. It has been proposed that this occurs because highly reducing conditions cause reduction of QB and QA in the RC, shutting down light-driven cyclic electron transfer. On

622

K. Holden-Dye et al. / Biochimica et Biophysica Acta 1777 (2008) 613–630

addition of an auxiliary oxidant electron flow from the quinone pool via TMAO- or DMSO-reductase repoises the intramembrane quinone pool, reoxidising QA and restoring charge separation in the RC [119,120]. A second finding of the study of Comayras et al. [91] was that deletion of PufX brought about an approximately 6-fold decrease in the affinity of the QB site of the RC for the inhibitor stigmatellin (with similar effects seen for other QB inhibitors), a 4-fold decrease in the rate of P+Q−B recombination and a 6-fold decrease in the rate of reduction of Q−B by Q−A. These observations indicate a direct effect of PufX on the properties of the QB site, including a stabilisation of Q−B on removal of PufX. This would suggest that PufX removal does not simply impose an increased barrier to diffusion of quinone (quinol) from (to) the RC, but rather also changes the structure of the RC–LH1 complex in the immediate vicinity of the QB site in such a way that alters the properties of the bound quinone. This is interesting in light of the observation described above that mutant α- or β-polypeptides can interfere with functioning of the RC [113], and suggests that a role of PufX may be to stop one or more components of the LH1 antenna from interacting too intimately with the RC in the region of the QB pocket. In the light of their experimental findings, Comyras et al. have pointed out that it is unlikely that the observed two-fold reductions in the rate of turnover of quinone at the QB site of the RC and the rate of diffusion of ubiquinol from the RC to the cytochrome bc1 complex are responsible for the PS− phenotype displayed by PufX-deficient strains [91]. A possible source of this phenotype is proposed to be a greater tendency for QA to be reduced in PufX-deficient strains under the anaerobic, reducing conditions pertaining during assays of photosynthetic growth, this tendency arising from a lowered equilibrium constant for the reaction Q−AQ−B Y QAQBH2 resulting from LH1-dependent structural changes affecting the QB site following removal of PufX [91]. The caveat “LH1-dependent” refers to the observation that photosynthetic growth is not impaired in PufX-deficient strains that also lack an LH1 antenna [106]. 9. Mutagenesis of the PufX polypeptide To date there has been only one report on the effects of site-directed mutagenesis of PufX, involving progressive truncations of the N- and C-terminus of the protein [121]. At the N-terminus the N-terminal Met was retained but residues 1–3, 1–6, 1–18 or 1–25 were removed, whilst the C-terminus 2, 7, 11, 15 or 29 residues were removed. For the Cterminal deletions, removal of two amino acids had no effect, but truncations of 7, 11 and 15 residues produced a PS− phenotype, and the dimeric form of the core-complex could not be resolved on sucrose gradients. Analysis of the detergent-extracted monomeric corecomplexes from the latter strains with antibodies showed the PufX protein to be absent. This suggests a role for the C-terminus in assembly of PufX into the membrane, consistent with its Nin–Cout topology. Curiously, the PufX protein was found to be present in monomeric complexes from the strain with a 29 amino acid C-terminal deletion, and this strain was capable of photosynthetic growth. Dimeric core-complexes were not resolved on sucrose gradients, but a complex with sedimentation properties intermediate between monomers and dimers was obtained. For the N-terminal truncation mutants, removal of residues 1–3 caused a slowing of photosynthetic growth, but dimeric corecomplexes were resolved on sucrose gradients [121]. Removal of residues 1–6 slowed the rate of photosynthetic growth, and lowered the yield of dimers on sucrose gradients, whereas removal of residues 1–18 also slowed the rate of photosynthetic growth but dimeric corecomplexes were not resolved. In all three of these mutants PufX was detected in antibody analysis of the monomeric core-complex. Finally, removal of residues 1–25 produced a PS− phenotype, no dimers were resolved on sucrose gradients and monomeric core-complexes did not contain the PufX protein.

The kinetics of flash-induced reduction of the cytochrome bc1 complex were also determined for the two sets of truncation mutants. Most of the mutants that showed slowed or no photosynthetic growth also exhibited a marked slowing in the delay before the onset of cytochrome b reduction, indicating an impairment in cyclic electron transfer. Although the range of effects seen in this study were complex, it did show that there is no simple correlation between the presence of the PufX protein and a dimeric organisation for the corecomplex, or between this dimeric organisation and the PS+ phenotype [121]. 10. Distribution and evolutionary origins of the PufX polypeptide The composition and organisation of the puf operon displays considerable diversity across photosynthetic bacteria [122]. As discussed above, amongst species that have been cultured in the laboratory the pufX gene is found only in the genus Rhodobacter. In most other species the pufM gene is followed by a gene (pufC) that encodes a tetra-heme cytochrome subunit of the RC, a structural feature found in a wide range of photosynthetic bacteria but not in Rhodobacter species. A small number of exceptions to this pattern have been identified amoungst cultured species [122], including Rsp. rubrum and Rps. palustris which both have pufM as the last gene of the puf operon. Although the pufX gene is restricted to five cultured species of Rhodobacter, possible pufX-like genes have been identified in DNA sequences from environmental samples collected from the Sargasso sea [123] and off the coast of California [124]. Each of these genes was located immediately downstream of pufM and the corresponding protein sequence differed from the five known sequences of PufX (see above). Yutin and Béjà [122] have presented an alignment of five such PufX-like polypeptides with the sequences of the known PufX polypeptides, and the central part of these sequences is shown in Fig. 2. The sequences labelled proteins 1–4 are from the study of Venter et al. [123] and that labelled Ebac60d04 is from the study of Béjà et al. [124]. Also shown is a sequence of a putative PufX polypeptide (GenBank Accession number AY912081) from a sample collected from the Delaware river, New Jersey, USA [125]. With the caveat that the origins and metabolic significance of these putative PufX polypeptides remain to be established, such sequences from environmental clones have the potential to increase our appreciation of the diversity displayed by PufX sequences, and help in the task of homing in on functionally-important features that are likely to be conserved. In the case of the data summarised in Fig. 2, the putative PufX polypeptides all include a sequence of around 20 amino acids that could consititute a single membrane-spanning α-helix. In addition, when these sequences are taken into account the number of absolutely conserved amino acids across the whole alignment drops to 5, namely Met X26, Gly X29, Gly X51, Leu X54, and Pro X64 (Rb. sphaeroides numbering). Given the (almost complete) lack of gaps in this region of the alignment, one can postulate a “PufX motif” of MXXG(X)21GXXL(X)9P. Although there are no other absolutely conserved residues between Gly X29 and Gly X51, it is striking that this 23 amino acid sequence contains a large number of low-volume residues. Analysis of the PufX sequences shown in Fig. 2 shows that the five Rhodobacter proteins have between 9 and 14 Gly, Ala or Val residues in this region, including Gly X29 and Gly X51. One possibility is that this region of α-helical structure is required to be generally “skinny”, perhaps to facilitate quinone diffusion by a number of “routes” along its length, but that the exact position of the low-volume residues is not critical and so is not conserved. Alternatively it is possible that the exact position of the low-volume residues is influenced by the precise surface topology of the adjacent RC and/or LH1 α-polypeptides. As explained above, in the majority of cultured species of purple photosynthetic bacteria a pufC gene is located immediately downstream

K. Holden-Dye et al. / Biochimica et Biophysica Acta 1777 (2008) 613–630

of pufM, and no pufX gene is present. The PucC polypeptides in Bl. viridis and a number of other species include a Cys residue some 20–25 positions from the N-terminus (see [126] for a discussion), and in Bl. viridis it has been shown that this Cys becomes the N-terminal residue following processing by a signal peptidase [127]. The PucC polypeptide is then anchored to the membrane via a diacylglycerol group that is attached to this N-terminal Cys. In other species, including Roseobacter (Rsb.) denitrificans, this Cys residue is absent and the N-terminus of PufC is extended by between 15 and 20 amino acids. In Rsb. denitrificans it has been shown that N-terminal processing of PufC does not take place, and the N-terminal region contains a stretch of around 20 predominantly hydrophobic amino acids indicative of a membrane-spanning α-helix [126,128]. Given this, it has been proposed that in Rsb. denitrificans (and several other species where the N-terminal Cys is absent) the cytochrome subunit is attached to the membrane by a membrane-spanning α-helix rather than a lipid group [126,128]. The arrangement proposed for Rsb. denitrificans places the DNA encoding the putative membrane-spanning segment of PucC immediately downstream of the pufM gene, in a position occupied by pufX in Rhodobacter species. Hucke et al. have constructed an alignment of the PufX polypeptides of Rb. sphaeroides and Rb. capsulatus with the N-terminal regions of the PufC polypeptides from Rsb. denitrificans and Rhodovulum (Rv.) sulfidophilum and, despite only weak overall similarity, have proposed a phylogenetic link between the two based on sequence similarities localised mainly to the region of proposed αhelical structure [126]. Using an analysis of relationships between 14 species of purple photosynthetic bacteria reported by Matsuura and Shimada [129], Hucke et al. proposed that the pufX gene arose from incomplete deletion of the pufC gene in a common ancestor of Rb. sphaeroides and Rb. capsulatus [126]. If the phylogenetic relationship between PufX and the N-terminal region of membrane-embedded PufC polypeptides is valid, then one wonders whether the latter may not only act as membrane anchors but also display functional properties similar to those described above for the PufX protein. As yet a high-resolution X-ray crystal structure is not available for a RC with a cytochrome subunit that includes an N-terminal membranespanning helix, but in the X-ray structures of the Bl. viridis and Tch. tepidum RCs the N-terminal Cys residue of PucC is located close to the membrane interface at a position approximately symmetrical to the periplasmic end of the membrane-spanning helix of the H-polypeptide. This would be consistent with a location for the N-terminal membrane-spanning helix of PucC symmetrical to the membranespanning helix of the H-polypeptide, which in turn is consistent with the expected position of PufX close to the QB site of the RC (see below) and the position of the polypeptide labelled as W in the X-ray crystal structure of the Rps. palustris RC–LH1 complex (Fig. 1D). Finally, there is the intriguing question of whether this Rps. palustris W-polypeptide is also a PufX-like protein. As indicated above there is no pufX gene immediately downstream of pufM in Rps. palustris, but it is conceivable that an equivalent gene could be located elsewhere in the bacterial genome, as photosynthetic bacteria show considerable variation in the organisation of the genes of the puf operon [122]. The Rps. palustris genome has been sequenced and analysed but the presence of a pufX gene has not been commented on [130]. Given the evidence from the X-ray crystal structure it seems likely that an equivalent of the pufX gene is indeed present, but has eluded identification due to the low sequence identity displayed by this group of genes. In this regard Yutin and Béjà [122] have pointed out the presence of an open reading frame corresponding to a 74 amino acid polypeptide located immediately upstream of pufB. The deduced amino acid sequence does not align with the sequences for pufX and putative pufX genes shown in Fig. 2, and does not contain the MXXG(X)21GXXL(X)9P motif described above, but its presence is interesting and worthy of further investigation.

623

11. PufX and the assembly of the RC–LH1–X complex In marked contrast to the very large literature on the functional properties of the RC–LH1–X complex and its component parts, relatively little is known about how this complex is assembled. However, with regard to the involvement of PufX, studies of the timedependence of assembly of the RC–LH1–X complex have shown that formation of a small complex consisting of the RC, PufX and a small number of LH1 α- and β-polypeptides is followed by an enlargement of the LH1 component to encircle the RC [68]. This suggests that PufX is present at an early stage of core-complex formation, and so may influence the final structure attained. Strains lacking PufX have been used in studies of gene products thought to be involved in assembly of the RC–LH1–X complex. In Rb. capsulatus the puhA gene encoding the RC H-polypeptide is arranged with three other genes in a puhABCE operon, the puhE gene product being a negative modulator of BChl synthesis [131]. Aklujkar et al. [132] have reported that disruption of puhB produces a ~6-fold reduction in levels of RCs and LH1, and the core-complexes present are associated with very low levels of PufX. Strains deficient in PuhB were capable of photosynthetic growth only after a long lag phase, and were found to possess increased levels of PufX once photosynthetic growth had taken place. This suggested that in the absence of PuhB the reduced complement of core-complexes were largely PufX-deficient in character, and during the lag phase these were replaced with RC– LH1–X complexes that were capable of supporting photosynthetic growth. The absolute level of RCs and LH1 complexes did not increase however. The puhB gene product exerts its effect subsequent to puf operon translation, suggesting that it acts as an assembly factor for the core-complex. When assessed separately, the absence of PuhB produced lowered levels of the RC but not the LH1 antenna, indicating that PuhB is an assembly factor for the RC. The data also reinforced earlier findings that the expression level of LH1 is linked to that of the RC in Rb. capsulatus, such that when expression of the RC is reduced then that of LH1 follows suit. The puhB gene encodes a protein of 214 amino acids that is predicted to have three membrane-spanning αhelices, and which may act as a dimer [132]. Aklujkar et al. have also examined the influence of the puhC gene product on the properties of the RC–LH1–X complex [133]. This gene encodes a polypeptide of 162 amino acids with a single putative membrane-spanning α-helix. Strains lacking puhC gene were found to have a reduced complement of RC–LH1–X complexes, but experiments with strains expressing only LH1 or only RCs showed that PuhC did not affect the accumulation of either component. PuhC did not have any influence on the accumulation of RCs in strains lacking the LH1 and LH2 antenna during a 20 hour period after switching from aerobic/dark growth to semiaerobic/dark growth, and the same result was obtained with strains that also contained the LH1 α-polypeptide or LH1 β-polypeptide. However, despite the presence of RCs the rate of photosynthetic growth was markedly slowed in the absence of PuhC in the strains that also contained either the LH1 α- or β-polypeptide. This suggested that PuhC prevents impairment of RC function in strains that contain only one of the two polypeptides required to assemble an LH1 antenna. In strains that had a deletion of PuhC the rate of photosynthetic growth was strongly impaired (but not abolished) despite the presence of the genes for the RC, LH1 and PufX, but this impairment was partially overcome in strains with multiple copies of the pufX gene. Given that PuhC did not have any influence on the assembly of the LH1 and RC components, this suggested that PuhC is involved in some way in the efficient assembly of the RC– LH1–X complex from its component parts, or in reorganisation of the complex to include PufX. 12. PufX and carotenoid Recent studies of PufX have drawn attention to a possible influence of the carotenoid content of RC–LH1–X complexes [72]. In fact, a

624

K. Holden-Dye et al. / Biochimica et Biophysica Acta 1777 (2008) 613–630

connection between PufX and carotenoid was apparent from early studies of mutations that suppress the PS− phenotype of PufXdeficient strains of Rb. capsulatus. Lilburn and Beatty [108] described a suppressor mutant named ΔRC6⁎(pTL2) that carried one or more mutations in genomic DNA, rather than the plasmid-borne copy of the puf operon. This mutant lacked the LH2 antenna, and has been reported to be unable to synthesise carotenoid [72]. It is known that LH2-deficient strains are still dependent on PufX for photosynthetic growth, leading to the conclusion that the suppression of the PS− phenotype must have arisen from the lack of carotenoid in the assembled LH1 antenna. As described above, it has been reported that the PS− phenotype observed in a PufX-deficient strain of Rb. capsulatus can be overcome by substitution of Ser 2 of the LH1 α-polypeptide (Sα2) with Phe or Pro. In subsequent work Aklujkar and Beatty [72] showed that this restoration of photosynthetic growth occurs without a decrease in the expression level of the LH1 antenna, which remains significantly higher than that observed in PufX-containing strains. Thus, in the case of these mutants, the restoration of a PS+ phenotype did not appear to be due to decreases in the level of the expression of the LH1 antenna per RC back to, or below, that seen in PufX-containing strains. In seeking an alternative explanation for the effects of the Sα2 mutations, Aklujkar and Beatty reported that the intensity of pigmentation by red carotenoid was reduced in cultures of the two suppressor mutants grown under semiaerobic conditions, indicating a change in the carotenoid complement of the photosynthetic membrane. Decreases in the carotenoid complement of photosyntheticallygrown cultures of the two Sα2 mutants were also observed, by 41% in the Phe mutant and 67% in the Pro mutant relative to the PufXdeficient strain from which they arose. These data suggest that the Ser residue at the α2 position is involved determining the carotenoid content of the LH1 antenna. In addition, it was observed that complementation of the suppressor mutant strains in trans with a copy of the pufX gene restored a deep red colouration, suggesting that PufX is also involved in some way in the binding of a normal complement of carotenoids to the LH1 antenna. The conclusions drawn from this work were that the PS− phenotype of the original PufX-deficient strain was overcome not by a restoration of, or further reduction in, the extent of the LH1 antenna surrounding the RC, but rather through the loss of a part of the carotenoid complement of the enlarged LH1, this fraction being bound to LH1 in a manner dependent on the Ser residue at the α2 position. In addition, Aklujkar and Beatty discussed the intriguing possibility that the presence of PufX in Rhodobacter species is connected with the particular carotenoid composition of LH1 in these species [72]. As reviewed by Hunter [134], the study of photosynthetic energy transduction in purple bacteria has been greatly aided by an ability to modulate the composition of the photosynthetic apparatus to remove unwanted components, or reveal hidden spectroscopic signals [82,134,135]. So, for example, studies of the Rb. sphaeroides RC–LH1–X complex are greatly aided by use of strains that lack the LH2 antenna through deletion of the relevant structural genes. Similarly, it has proven possible to study charge separation in membrane-embedded RCs through the use of RC-only strains that lack both LH1 and LH2, and structure/function studies of the antenna complexes have been aided by the use of LH2-only or LH1-only strains. Such strains can be constructed by the deletion of the genomic copies of the structural genes of the RC, LH1 and LH2 complexes (usually involving replacement with a DNA cassette conferring antibiotic resistance), and complementation in trans with the desired genes on a mobilisable low-copy-number broad-hostrange vector. As their cellular membranes are packed with light-absorbing pigments, photosynthetic bacteria are brightly coloured. An early indication that the composition of the photosynthetic apparatus could be varied came from studies of Griffiths and co-workers [136,137], who isolated a range of coloured mutants of Rb. sphaeroides produced

either spontaneously or in response to UV illumination. The native colour of Rb. sphaeroides is brown when grown anaerobically and red– purple when grown in the presence of oxygen, and the isolated mutants exhibited variations in the colour or intensity of this pigmentation indicative of changes in the carotenoid content of the photosynthetic apparatus, or the profile and/or quantity of pigmentprotein complexes in the cell. In the period before genetic manipulation of the photosynthetic apparatus became possible, strains of Rb. sphaeroides or Rb. capsulatus with such alterations in carotenoid synthesis were widely used for a number of purposes. Perhaps the best known of these are the blue–green carotenoid-deficient R-26 and R-26.1 strains of Rb. sphaeroides [134,138–141]. These strains have a lesion in a gene encoding an early stage of carotenoid synthesis that results in an absence of carotenoid and strongly reduced levels of the LH2 complex. This greatly assists the purification of the RC, and a great many experimental studies of RC structure and mechanism have used carotenoid-deficient R-26 or R-26.1 RCs. Another widely used group are strains of Rb. sphaeroides or Rb. capsulatus that have a lesion in the crtD or crtC genes that encode the the last steps of carotenoid synthesis (see [142] for a review). Such strains have a green colouration arising from the carotenoid neurosporene (plus methoxy and hydroxy derivatives in crtD mutants), as opposed to the brown or red colouration of native strains attributable to spheroidene (in anaerobically-grown cells) or spheroidenone (in semiaerobicallygrown cells). Green strains have been used for a number of purposes, and in particular for studies of cytochrome oxidation by the RC as green strains have a much lower background carotenoid absorbance in the region of c-type cytochrome absorbance than native red-brown strains (e.g. see [104] for a study relevant to PufX). The findings outlined above that details of the carotenoid composition of the RC–LH1 complex may influence whether a lack of PufX produces a PS− phenotype suggest that carotenoid type is a factor that needs to be taken into account in studies of the RC–LH1–X complex. As discussed by Aklujkar and Beatty [72] a significant proportion of studies of the PufX protein has involved strains with green carotenoids, rather than native brown/red carotenoids, but there is increasing evidence that the two backgrounds may not necessarily be equivalent. In a rare case where the effects of PufX deletion in native and green carotenoid backgrounds were compared, McGlynn et al. [106] observed that this deletion caused a much greater increase in the amount of LH1 BChl per RC in the green carotenoid background than in the native background (see above for values). In fact, this was mainly because the amount of LH1 per RC was lower in the green PufX-containing strain than in its red/brown counterpart, levels increasing to approximately equal values in the two backgrounds on removal of PufX. Another potential complicating factor is variations in carotenoid content arising from growth conditions used, with spheroidene being synthesised under anaerobic conditions and spheroidenone under semiaerobic conditions. Again, the impact of this difference on the composition and organisation of RC–LH1–X complexes is not known. As strains that lack PufX are non-photosynthetic, studies of the role of PufX tend to use material from cells grown under dark/semiaerobic conditions. However, much of the material used in recent studies of the structure of the Rhodobacter RC–LH1–X complex has come from cells grown under illuminated/anaerobic conditions. In this regard, Francia et al. have reported that the RC–LH1–X complexes from semiaerobically-grown cells (with green carotenoids) contain 21 LH1 BChls per RC, whereas the same complexes from photosynthetically-grown cells with green carotenoids contain 31.5 LH1 BChls per RC [105]. Thus, it would appear that growth conditions are an additional factor that needs to be taken into account when trying to rationalise findings from different studies, the presence or absence of oxygen possibly exerting an influence on the composition of the RC–LH1–X complex throughout the types of carotenoid present in the membrane.

K. Holden-Dye et al. / Biochimica et Biophysica Acta 1777 (2008) 613–630

13. Overexpression and structural characterisation of PufX The PufX protein has been overexpressed in Escherichia coli by a number of laboratories [70,143,144]. Most recently, Onodera et al. expressed a 69 amino acid version of the Rb. sphaeroides protein (residues X1–X69) with a His6 tag attached to the C-terminal end via a connecting Leu and Glu [144], and Tunnicliffe et al. expressed the full-length Rb. sphaeroides PufX protein [143]. In both cases the protein dissolved in organic solvent was used for structural studies by NMR spectroscopy. Fig. 6 shows representations of the average structures for the Rb. sphaeroides PufX deposited in the Protein Data Bank (PDB) by Tunnicliffe et al. ([143]; PDB code 2NRG) and Wang et al. ([145]; PDB code 2DW3). The approximate position of the membrane is shown by the horizontal lines, and the position of PufX in the membrane is based on data from Fig. 2C of Ref. [143], and the assignment from Ref. [145] that residues Lys X28 to Gly X35 are located at the centre of the membrane. The two structures are in agreement in showing a polypeptide with a predominantly α-helical structure between residues Asn X13 and Met X53, but with disordered N- and C-termini. However, the structures show a clear difference in that the PufX of Wang et al. shows an approximately straight helix whereas that of Tunnicliffe et al. shows significant distortion between Gln X25 and Ala X33, such that the helix is bent with an average inter-helical angle of 120°. In the structure of Tunnicliffe et al. the α-helical region of PufX is envisaged as being positioned within the ~40 Å span of the membrane, the bend in the helix permitting this despite its ~41 amino acid

625

length. In contrast, in the structure of Wang et al. the residues at the centre of the helix are postulated to be at the centre of the membrane, with the result that the α-helix would be expected to protrude from the membrane by a couple of helical turns at both the C- and N-terminal ends. Fig. 6 also shows the positions of the eight residues that are absolutely conserved in the sequences of PufX proteins from the five Rhodobacter species. In both PufX structures five of these amino acids (three Pro, one Leu and one Ala) are located in the C-terminal unstructured region that is presumed to be located outside the membrane. In the structure of Wang et al. the remaining three amino acids are located in the central α-helix. Met X26 and Gly X29 are located near the centre of the helix, whilst Gly X51 is located close to the C-terminal end of the helix. In the structure of Tunnicliffe et al. Gly X51 is also located close to the C-terminal end of the helical region, at the periplasmic surface of the membrane, whilst Met X26 and Gly X29 are located close to the bend in the helix. The position of Gly X29 is particularly interesting as it is located on the inside of the bend, and modelling shows that a larger residue at this position could not be accommodated without steric clashes with neighbouring side chains. The fact that this Gly is absolutely conserved in all known sequences for PufX, and for that matter in all of the sequences of putative PufXlike proteins in the alignment in Fig. 2, could indicate that this bent helical structure is relevant to the structure of PufX in vivo. In analysing their structure for PufX, Wang et al. have drawn attention to a section of the α-helical region between positions X29

Fig. 6. Average NMR structures for the PufX protein from Rb. sphaeroides. Structure 2NRG [143] is shown on the left and structure 2DW3 [145] on the right. In both structures the region of the protein assigned as a helix is shown as a ribbon, and residues conserved amongst the five identified Rhodobacter PufX proteins are shown as spheres with Pro and Met coloured white, Leu, Gly and Ala shown in dark grey. For structure 2NRG the position of the protein in the membrane is as proposed by Tunnicliffe et al. in Fig. 2C of Ref. [143]. Structure 2DW3 is shown on the same scale, with residues X28–X35 located at the centre of the membrane, as proposed by Wang et al. [145]. In both studies the protein was dissolved in a 1:1 (v/v) mixture of CDCl3 and CD3OH, and data were collected at 25 °C.

626

K. Holden-Dye et al. / Biochimica et Biophysica Acta 1777 (2008) 613–630

and X35 that is rich in low-volume Gly and Ala residues, with the sequence GAGWAGG. Wang et al. have proposed that these lowvolume side chains constitute a site for the exchange of quinone and quinol between the interior of the core-complex and the external membrane phase and/or a region that interacts with LH1 α- or β-polypeptides [145]. Although this sequence of low-volume residues is striking, it should be noted that with the exception of R. azotoformans it is not conserved in other PufX proteins, including the protein from Rb. capsulatus which can substitute for the Rb. sphaeroides PufX in facilitating photosynthetic growth [65]. In Rb. capsulatus the equivalent residues have the sequence GAFLGSI. The conserved Gly X29 aside, of the remaining five low-volume residues found in this part of the Rb. sphaeroides PufX, none is conserved as a low-volume residue (Gly, Ala or Val) in all five PufX proteins. As discussed above, residues X31 to X35 are consistent with the helix-dimerisation GXXXG motif [100], but this is not conserved in any other Rhodobacter species. 14. Analogies to photosystem II The counterpart of the purple bacterial RC in oxygenic photosynthetic organisms is the photosystem II (PSII) RC, which catalyses membrane-spanning charge separation using pheophytin and quinone as electron acceptors. Plastoquinone is doubly reduced and doubly protonated at the QB binding site, and then migrates to the cytochrome bf complex where it is oxidised and deprotonated on the opposite side of the membrane. The PSII RC is surrounded in the membrane by antenna pigment-protein complexes, and in recent years a number of X-ray crystal structures for the cyanobacterial PSII have been published [146–150]. These show a central RC that is completely surrounded by the polypeptides, chlorophylls and carotenoids of the associated antenna proteins. Just as ubiquinol produced at the QB site of the purple bacterial RC has to diffuse through the surrounding LH1 antenna to reach the cytochrome bc1 complex, so plastoquinol produced at the QB site of the PSII RC has to diffuse through the surrounding antenna to reach the cytochrome bf complex. A recent X-ray crystal structure of the PSII complex from Thermosynechococcus elongatus has revealed structural features that may be relevant to this process [150]. This structure shows four lipids, two digalactosyldiacylglycerol (DGDG), one sulphoquinovosyldiacylglycerol (SQDG) and one monogalactosyldiacylglycerol (MGDG) lining a cavity that forms an antechamber to the QB site of the central RC domain, and the tail of the QB plastoquinone emerges from the interior of the RC domain into this cavity. In their analysis of this structure Loll et al. have noted an intramembrane portal for release of plastoquinol, or entry of plastoquinone, between the membrane-spanning helices of two minor antenna polypeptides assigned as PsbE and PsbJ [150]. This portal is created by small-volume amino acid residues on one side of the membrane-spanning α-helix of the PsbJ polypeptide, which make it thin on one side for approximately three helical turns. The resulting gap between the helices of PsbE and PsbJ is sufficiently large to allow passage of a quinone molecule. This arrangement is resonant with ideas concerning the structure and role of the PufX protein in purple bacteria, given the large number of small-volume residues in the membrane-spanning helix and the impairment in quinone-mediated cyclic electron transfer seen in mutants that lack PufX (see [151] for a recent discussion). 15. So, what is the role of PufX? PufX is clearly an important component of the RC–LH1–X corecomplex in Rb. sphaeroides and Rb. capsulatus. The observations that removal of PufX produces a PS− phenotype, increases the amount of LH1 BChl per RC and causes the LH1 antenna to assemble as a closed ring around the central RC has led to proposals that PufX is responsible for keeping the LH1 ring in an open state, facilitating rapid quinone/

quinol exchange between the RC and the cytochrome bc1 complex. Observations that the PS+ phenotype can be restored to PufX-deficient strains through mutations that further disrupt the structure of the RC–LH1 complex, including lowering the expression level of the LH1 antenna, seem to support this idea. However the situation is clearly more complex than this, as PufX-deficient strains are capable of photosynthetic growth in the presence of auxiliary oxidants such as TMAO, and measurements of cyclic electron transfer show that the absence of PufX has only modest effects on the kinetics of quinone/ quinol exchange. A number of key questions concerning the PufX protein remain to be answered. What is the principal function of PufX? Given the accumulated data on the PufX protein outlined above, it seems possible that its principal role is to prevent the LH1 antenna from engaging in too intimate a contact with the QB binding pocket of the RC. In order to achieve this PufX interacts with the RC, possibly at a position roughly symmetrical to the membrane-spanning helix of the H-polypeptide, and also binds to one or more α-polypeptides of the LH1 antenna. The result is that the LH1 aggregate does not follow the contour of the RC in the region adjacent to the QB site but rather is held away somewhat. In order to achieve this without blocking quinone/quinol diffusion the membrane-spanning region of PufX has a high number of low-volume amino acids, the exact identity and position of which is not conserved but varies according to the precise structure of the adjacent RC and LH1 proteins. The contacts made between PufX and the adjacent LH1 α-polypeptide(s) prevent the latter from making the usual proteinprotein contacts with the next repeating unit in the LH1 complex, thus stopping further aggregation and creating a break in the continuity of the LH1 ring surrounding the RC. In the absence of PufX the LH1 antenna makes a more intimate contact with the RC surface, and some component of LH1 interferes with normal operation of the QB site in such a way that the RC is sensitised to over-reduction of the quinone pool [91]. Under the anaerobic, reducing conditions pertaining to photosynthetic growth experiments in the laboratory this interference leads to a modest slowing in the rate of quinone/quinol exchange at the QB site but, more importantly, sensitisation to the reduced state of the quinone pool leads to a reduction of QA and a shutting down of photosynthetic charge separation. Comayras et al. have proposed that this sensitisation of QA comes from a lowered equilibrium constant for the reaction Q−AQ−− B Y QAQBH2, derived from an increase in the redox potential for the QB/Q−B redox couple [91]. The blockage in charge separation can be overcome either through the addition of an auxiliary oxidant to raise the redox potential of the quinone pool, drawing electrons away from QA, or through changes in the structure of LH1 that remove the interference with normal QB site function. The latter include a partial or complete reduction in the expression level of the LH1 antenna, the incorporation of structurally altered LH1 α- or β-polypeptides, or a change in the carotenoid content of the antenna. The last of these points to the possibility that the structural component of LH1 responsible for disturbing normal operation of the QB site may be a carotenoid. A second question concerns the connection between PufX and the dimeric nature of the Rb. sphaeroides RC–LH1–X complex. A possible clue lies in the finding that truncation of PufX at the N-terminus leads to a loss of dimeric RC–LH1–X complexes but photosynthetic growth, albeit with a reduced rate, is retained [121]. This suggests that some part of the PufX protein between residues X4 and X18 is important for the assembly of dimers. One possibility is that a dimeric structure for the Rb. sphaeroides RC–LH1–X complex is dictated at an early stage in assembly when the developing complex consists only of the RC, a PufX polypeptide and one or two B820-type LH1 subunits. If, at this stage, the N-termini of two PufX proteins engage in an interaction, bringing two RCs into close proximity (and perhaps a fixed mutual orientation), then this could dictate the formation of the S-shaped antenna seen in EM and AFM images. An alternative (or additional) interaction could involve the membrane-spanning helix of PufX, which exhibits a weak propensity for self association. What advantage this dimeric core-

K. Holden-Dye et al. / Biochimica et Biophysica Acta 1777 (2008) 613–630

complex structure confers, relative to the type of monomeric complex seen in most other purple photosynthetic bacteria, remains to be determined. Gubellini et al. have reported that the RC–LH1–X complex from Rb. veldkampii is monomeric [69], which would seem to suggest that the PufX-mediated dimeric organisation seen in Rb. sphaeroides and Rb. blasticus may be a peculiarity of these two species, and this organisation of RC–LH1–X complexes into dimers is not a core function of the PufX protein. Examination of the alignment of PufX sequences presented by Tsukatani and co-workers [TY04] shows that the PufX protein from Rb. veldkampii is shorter at the N-terminus by four amino acids relative to the protein from Rb. sphaeroides, which may be relevant to the different organisation of the RC–LH1–X complex in the two species. Then there is the question of why the primary sequence of PufX is so poorly conserved between the five Rhodobacter species identified to date. One answer could be that PufX interacts with surface-exposed regions of the RC and LH1 complex that are themselves poorly conserved across species. Certainly in the case of the RC PufL and PufM polypeptides the majority of absolutely conserved residues are important for the protein fold or protein-cofactor interactions internal to the protein, and there is weaker conservation of peripheral, surfaceexposed residues. As outlined above, variations in the precise structure of the RC and/or LH1 polypeptides could explain why the low-volume residues that seem to be a feature of the membranespanning region of the PufX protein are not positionally conserved. However, in addition to this there is the possibility that the structural and, perhaps, functional role played by PufX varies even between species of Rhodobacter. For example, it is intriguing that the positioning of PufX relative to the RC is similar in the models of the dimeric RC–LH1–X complex developed for Rb. sphaeroides and Rb. blasticus, but due to different assigned orientations for the RC the resulting position of the PufX proteins in the structure of the dimer is very different. This difference could simply be due to the challenges of correctly modelling a structure based on relatively low-resolution experimental data, but a second possibility is that both structures are correct, and that the structures of the RC–LH1–X complex in the two species are genuinely different (with the monomeric RC–LH1–X complex from Rb. veldkampii providing a third variation on the theme). Thus it may be that the core function of the PufX protein is to keep the region of the RC around the QB pocket free from interference from the LH1 antenna protein and, this achieved, a variety of monomeric and dimeric structures are then possible for the RC–LH1–X complex. Finally one might ask why, if the PufX protein is so important for RC function in Rhodobacter species, is it not found in other purple photosynthetic bacteria. Perhaps the first point to make is that the role of PufX in Rb. blasticus, Rb. azotoformans and Rb. veldkampii has not been tested, and so although it is generally assumed that PufX is essential for photosyntheic growth of these species this remains to be proven. This aside, it seems possible that some species of purple bacteria, such as Rps. palustris, might contain a protein that fulfils the function of PufX but has no significant sequence identity with the Rhodobacter proteins. Similarly it may be that in species that contain a RC with a cytochrome subunit with a membrane-spanning helix, this helix fulfils the function of PufX in preventing close approach of the LH1 antenna to the QB site of the RC. Finally, although a great deal of sequence data on the puf operon has been collected for purple photosynthetic bacteria, and this has shown the pufX gene to be confined to the Rhodobacter genus, there is relatively little biochemical data on the protein composition of the RC–LH1 complex in different species. In the absence of such data it cannot be ruled out that the core-complexes from some of these species contain a minor polypeptide that fulfils a PufX-like function. Such a polypeptide might be difficult to detect in the presence of a much larger number of similarly-sized LH1 polypeptides, and difficult to visualise through EM or AFM in species where the LH1 antenna forms a closed ring around the RC.

627

Acknowledgements Research in the laboratory of the authors is supported by the Biotechnology and Biological Sciences Research Council of the United Kingdom. Figs. 1 and 6 were produced using PyMOL [152].

References [1] X. Hu, A. Damjanović, T. Ritz, K. Schulten, Architecture and mechanism of the light-harvesting apparatus of purple bacteria, Proc. Natl. Acad. Sci. U. S. A. 95 (1998) 5935–5941. [2] X.C. Hu, T. Ritz, A. Damjanović, F. Autenrieth, K. Schulten, Photosynthetic apparatus of purple bacteria, Quart. Rev. Biophys. 35 (2002) 1–62. [3] R.J. Cogdell, A.T. Gardiner, A.W. Roszak, C.J. Law, J. Southall, N.W. Isaacs, Rings, ellipses and horseshoes: how purple bacteria harvest solar energy, Photosynth. Res. 81 (2004) 207–214. [4] C.J. Law, A.W. Roszak, J. Southall, A.T. Gardiner, N.W. Isaacs, R.J. Cogdell, The structure and function of bacterial light-harvesting complexes (Review), Mol. Memb. Biol. 21 (2004) 183–191. [5] S. Scheuring, D. Lévy, J.-L. Rigaud, Watching the components of photosynthetic bacterial membranes and their in situ organisation by atomic force microscopy, Biochim. Biophys. Acta 1712 (2005) 109–127. [6] R.J. Cogdell, A. Gall, J. Kohler, The architecture and function of the light-harvesting apparatus of purple bacteria: from single molecules to in vivo membranes, Quart. Rev. Biophys. 39 (2006) 227–324. [7] S. Scheuring, AFM studies of the supramolecular assembly of bacterial photosynthetic core-complexes, Curr. Op. Chem. Biol. 10 (2006) 387–393. [8] G.R. Fleming, R. van Grondelle, Femtosecond spectroscopy of photosynthetic light harvesting systems, Curr. Opin. Str. Biol. 7 (1997) 738–748. [9] V. Sundström, T. Pullerits, R. van Grondelle, Photosynthetic light-harvesting: reconciling dynamics and structure of purple bacterial LH2 reveals function of photosynthetic unit, J. Phys. Chem. B 103 (1999) 2327–2346. [10] R. van Grondelle, J.P. Dekker, T. Gillbro, V. Sundström, Energy-transfer and trapping in photosynthesis, Biochim. Biophys. Acta-Bioenerg. 1187 (1994) 1–65. [11] M. Yang, R. Agarwal, G.R. Fleming, The mechanism of energy transfer in the antenna of photosynthetic purple bacteria, J. Photochem. Photobiol. A-Chem. 142 (2001) 107–119. [12] N.W. Woodbury, J.P. Allen, The pathway, kinetics and thermodynamics of electron transfer in wild type and mutant bacterial reaction centers of purple nonsulfur bacteria, in: R.E. Blankenship, M.T. Madigan, C.E. Bauer (Eds.), Anoxygenic Photosynthetic Bacteria, Kluwer Academic Publishers, The Netherlands, 1995, pp. 527–557. [13] W.W. Parson, Photosynthetic bacterial reaction centers, in: D.S. Bendall (Ed.), Protein Electron Transfer, BIOS Scientific Publishers, Oxford, UK, 1996, pp. 125–160. [14] A.J. Hoff, J. Deisenhofer, Photophysics of photosynthesis: Structure and spectroscopy of reaction centres of purple bacteria, Phys. Rep. 287 (1997) 2–247. [15] M.E. Van Brederode, M.R. Jones, Reaction centres of purple bacteria, in: N.S. Scrutton, A. Holzenburg (Eds.), Enzyme-Catalysed Electron and Radical Transfer, Kluwer Academic/Plenum Publishers, New York, USA, 2000, pp. 621–676. [16] M.Y. Okamura, M.L. Paddock, M.S. Graige, G. Feher, Proton and electron transfer in bacterial reaction centers, Biochim. Biophys. Acta 1458 (2000) 148–163. [17] C.A. Wraight, Proton and electron transfer in the acceptor quinone complex of photosynthetic reaction centers from Rhodobacter sphaeroides, Front. Biosci. 9 (2004) 309–337. [18] P. Mitchell, Proton motive redox mechanism of the cytochrome bc1 complex in the respiratory chain: proton motive ubiquinone cycle, FEBS Lett. 56 (1975) 1–6. [19] P. Mitchell, Possible molecular mechanisms of the protonmotive function of cytochrome systems, J. Theor. Biol. 62 (1976) 327–367. [20] A.R. Crofts, S.W. Meinhardt, K.R. Jones, M. Snozzi, The role of the quinone pool in the cyclic electron transfer chain of Rps. sphaeroides: a modified Q-cycle mechanism, Biochim. Biophys. Acta 723 (1983) 202–218. [21] C. Hunte, H. Palsdottir, B.L. Trumpower, Protonmotive pathways and mechanisms in the cytochrome bc1 complex, FEBS Lett. 545 (2003) 39–46. [22] A.R. Crofts, The Q-cycle — a personal perspective, Photosynth. Res. 80 (2004) 223–243. [23] A.R. Crofts, The cytochrome bc1 complex: Function in the context of structure, Ann. Rev. Physiol. 66 (2004) 689–733. [24] A.R. Crofts, S. Lhee, S.B. Crofts, J. Cheng, S. Rose, Proton pumping in the bc1 complex: a new gating mechanism that prevents short circuits, Biochim. Biophys. Acta-Bioenerg. 1757 (2006) 1019–1034. [25] J. Deisenhofer, O. Epp, K. Miki, R. Huber, H. Michel, X-ray structure analysis of a membrane protein complex — electron-density map at 3 Å resolution and a model of the chromophores of the photosynthetic reaction center from Rhodopseudomonas viridis, J. Mol. Biol. 180 (1984) 385–398. [26] J. Deisenhofer, O. Epp, K. Miki, R. Huber, H. Michel, Structure of the protein subunits in the photosynthetic reaction center of Rhodopseudomonas viridis at 3 Angstrom resolution, Nature 318 (1985) 618–624. [27] J. Deisenhofer, O. Epp, I. Sinning, H. Michel, Crystallographic refinement at 2.3angstrom resolution and refined model of the photosynthetic reaction center from Rhodopseudomonas viridis, J. Mol. Biol. 246 (1995) 429–457. [28] J.P. Allen, G. Feher, T.O. Yeates, D.C. Rees, J. Deisenhofer, H. Michel, R. Huber, Structural homology of reaction centers from Rhodopseudomonas sphaeroides and Rhodopseudomonas viridis as determined by X-ray diffraction, Proc. Natl. Acad. Sci. U. S. A. 83 (1986) 8589–8593.

628

K. Holden-Dye et al. / Biochimica et Biophysica Acta 1777 (2008) 613–630

[29] J.P. Allen, G. Feher, T.O. Yeates, H. Komiya, D.C. Rees, Structure of the reaction center from Rhodobacter sphaeroides R-26 — the cofactors, Proc. Natl. Acad. Sci. U. S. A. 84 (1987) 5730–5734. [30] J.P. Allen, G. Feher, T.O. Yeates, H. Komiya, D.C. Rees, Structure of the reaction center from Rhodobacter-sphaeroides R-26: the protein subunits, Proc. Natl. Acad. Sci. U. S. A. 84 (1987) 6162–6166. [31] C.-H. Chang, O. El-Kabbani, D. Tiede, J. Norris, M. Schiffer, Structure of the membrane-bound protein photosynthetic reaction center from Rhodobacter sphaeroides, Biochemistry 30 (1991) 5352–5360. [32] U. Ermler, G. Fritzsch, S.K. Buchanan, H. Michel, Structure of the photosynthetic reaction center from Rhodobacter sphaeroides at 2.65-angstrom resolution — cofactors and protein-cofactor interactions, Structure 2 (1994) 925–936. [33] U. Ermler, H. Michel, M. Schiffer, Structure and function of the photosynthetic reaction center from Rhodobacter sphaeroides, J. Bioenerg. Biomemb. 26 (1994) 5–15. [34] T. Nogi, I. Fathir, M. Kobayashi, T. Nozawa, K. Miki, Crystal structures of photosynthetic reaction centre and high-potential iron–sulfur protein from Thermochromatium tepidum: thermostability and electron transfer, Proc. Natl. Acad. Sci. USA 97 (2000) 13561–13566. [35] I. Fathir, T. Mori, T. Nogi, M. Kobayashi, K. Miki, T. Nozawa, Structure of the H subunit of the photosynthetic reaction center from the thermophilic purple sulfur bacterium, Thermochromatium tepidum — implications for the specific binding of the lipid molecule to the membrane protein complex, Eur. J. Biochem. 268 (2001) 2652–2657. [36] G. McDermott, S.M. Prince, A.A. Freer, A. Hawthornthwaite-Lawless, M.Z. Papiz, R.J. Cogdell, N.W. Isaacs, Crystal-structure of an integral membrane light-harvesting complex from photosynthetic bacteria, Nature 374 (1995) 517–521. [37] J. Koepke, X.C. Hu, C. Muenke, K. Schulten, H. Michel, The crystal structure of the light-harvesting complex II (B800–850) from Rhodospirillum molischianum, Structure 4 (1996) 581–597. [38] K.R. Miller, Three-dimensional structure of a photosynthetic membrane, Nature 300 (1982) 53–55. [39] W. Stark, W. Kuhlbrandt, I. Wildhaber, E. Wehrli, K. Muhlethaler, The structure of the photoreceptor unit of Rhodopseudomonas viridis, EMBO J. 3 (1984) 777–783. [40] H. Engelhardt, A. Engel, W. Baumeister, Stoichiometric model of the photosynthetic unit of Ectothiorhodospira halochloris, Proc. Natl. Acad. Sci. U. S. A. 83 (1986) 8972–8976. [41] R.U. Meckenstock, K. Krusche, R.A. Brunisholz, H. Zuber, The light-harvesting core-complex and the B820-subunit from Rhodopseudomonas marina. Part II. Electron microscopic characterisation, FEBS Letts. 311 (1992) 135–138. [42] A.F. Boonstra, L. Germeroth, E.J. Boekema, Structure of the light harvesting antenna from Rhodospirillum molischianum studied by electron microscopy, Biochim. Biophys. Acta 1184 (1994) 227–234. [43] S. Karrasch, P.A. Bullough, R. Ghosh, The 8.5 Å projection map of the lightharvesting complex I from Rhodospirillum rubrum reveals a ring composed of 16 subunits, EMBO J. 14 (1995) 631–638. [44] T. Walz, S.J. Jamieson, C.M. Bowers, P.A. Bullough, C.N. Hunter, Projection structures of three photosynthetic complexes from Rhodobacter sphaeroides: LH2 at 6 Angstrom LH1 and RC–LH1 at 25 Angstrom, J. Mol. Biol. 282 (1998) 833–845. [45] H. Stahlberg, J. Dubochet, H. Vogel, R. Ghosh, Are the light-harvesting I complexes from Rhodospirillum rubrum arranged around the reaction centre in a square geometry? J. Mol. Biol. 282 (1998) 819–831. [46] S.J. Jamieson, P.Y. Wang, P. Qian, J.Y. Kirkland, M.J. Conroy, C.N. Hunter, P.A. Bullough, Projection structure of the photosynthetic reaction centre-antenna complex of Rhodospirillum rubrum at 8.5 Angstrom resolution, EMBO J. 21 (2002) 3927–3935. [47] I. Ikeda-Yamasaki, T. Odahara, K. Mitsuoka, Y. Fujiyoshi, K. Murata, Projection map of the reaction center-light harvesting 1 complex from Rhodopseudomonas viridis at 10 Angstrom resolution, FEBS Letts. 425 (1998) 505–508. [48] S. Scheuring, J. Seguin, S. Marco, D. Levy, B. Robert, J.L. Rigaud, Nanodissection and high-resolution imaging of the Rhodopseudomonas viridis photosynthetic core complex in native membranes by AFM, Proc. Natl. Acad Sci U. S. A. 100 (2003) 1690–1693. [49] D. Fotiadis, P. Qian, A. Philippsen, P.A. Bullough, A. Engel, C.N. Hunter, Structural analysis of the reaction center light-harvesting complex I photosynthetic core complex of Rhodospirillum rubrum using atomic force microscopy, J. Biol. Chem. 279 (2004) 2063–2068. [50] S. Scheuring, J.N. Sturgis, V. Prima, A. Bernadac, D. Lévy, J.L. Rigaud, Watching the photosynthetic apparatus in native membranes, Proc. Natl. Acad. Sci. U. S. A. 101 (2004) 11293–11297. [51] R.P. Gonçalves, A. Bernadac, J.N. Sturgis, S. Scheuring, Architecture of the native photosynthetic apparatus of Phaeospirillum molischianum, J. Struct. Biol. 152 (2005) 221–228. [52] A.W. Roszak, T.D. Howard, J. Southall, A.T. Gardiner, C.J. Law, N.W. Isaacs, R.J. Cogdell, Crystal structure of the RC–LH1 core complex from Rhodopseudomonas palustris, Science 302 (2003) 1969–1972. [53] S. Scheuring, R.P. Gonçalves, V. Prima, J.N. Sturgis, The photosynthetic apparatus of Rhodopseudomonas palustris: structures and organization, J. Mol. Biol. 358 (2006) 83–96. [54] D.C. Youvan, E.J. Bylina, M. Alberti, H. Begusch, J.E. Hearst, Nucleotide and deduced polypeptide sequences of the photosynthetic reaction center, B870 antenna, and flanking polypeptides from R. capsulata, Cell 37 (1984) 949–957. [55] J.G. Belasco, J.T. Beatty, C.W. Adams, A. Vongabain, S.N. Cohen, Differential expression of photosynthesis genes in R. capsulata results from segmental differences in stability within the polycistronic rxcA transcript, Cell 40 (1985) 171–181.

[56] Y.S. Zhu, J.E. Hearst, Regulation of expression of genes for light-harvesting antenna proteins LH-I and LH-II – reaction center polypeptides RC-L, RC-M, and RC-H – and enzymes of bacteriochlorophyll and carotenoid biosynthesis in Rhodobacter capsulatus by light and oxygen, Proc. Natl. Acad. Sci. U. S. A. 83 (1986) 7613–7617. [57] T.J. Donohue, P.J. Kiley, S. Kaplan, The puf operon region of Rhodobacter sphaeroides, Photosynth. Res. 19 (1988) 39–61. [58] J.K. Lee, B.S. DeHoff, T.J. Donohue, R.J. Gumport, S. Kaplan, Transcriptional analysis of puf operon expression in Rhodobacter sphaeroides 2.4.1. and an intercistronic transcription terminator mutant, J. Biol. Chem. 264 (1989) 19354–19365. [59] G. Klug, S.N. Cohen, Pleiotropic effects of localised Rhodobacter capsulatus puf operon deletions on production of light-absorbing pigment-protein complexes, J. Bact. 170 (1988) 5814–5821. [60] J.W. Farchaus, H. Gruenberg, D. Oesterhelt, Complementation of a reaction center deficient Rhodobacter sphaeroides pufLMX deletion strain in trans with pufBALM does not restore the photosynthesis-positive phenotype, J. Bact.172 (1990) 977–985. [61] T.G. Lilburn, C.E. Haith, R.C. Prince, J.T. Beatty, Pleiotropic effects of pufX gene deletion on the structure and function of the photosynthetic apparatus of Rhodobacter capsulatus, Biochim. Biophys. Acta 1100 (1992) 160–170. [62] J.W. Farchaus, W.P. Barz, H. Gruenberg, D. Oesterhelt, Studies on the expression of PufX polypeptide and its requirement for photoheterotrophic growth in Rhodobacter sphaeroides, EMBO J. 11 (1992) 2779–2788. [63] C.L. Tavano, T.J. Donohue, Development of the bacterial photosynthetic apparatus, Curr. Op. Microbiol. 9 (2006) 625–631. [64] P.S. Parkes-Loach, C.J. Law, P.A. Recchia, J. Kehoe, S. Nehrlich, J. Chen, P.A. Loach, Role of the core region of the PufX protein in inhibition of reconstitution of the core light-harvesting complexes of Rhodobacter sphaeroides and Rhodobacter capsulatus, Biochemistry 40 (2001) 5593–5601. [65] T.K. Fulcher, J.T. Beatty, M.R. Jones, Demonstration of the key role played by the PufX protein in the functional and structural organization of native and hybrid bacterial photosynthetic core complexes, J. Bact. 180 (1998) 642–646. [66] Y. Tsukatani, K. Matsuura, S. Masuda, K. Shimada, A. Hiraishi, K.V.P. Nagashima, Phylogenetic distribution of unusual triheme to tetraheme cytochrome subunit in the reaction center complex of purple photosynthetic bacteria, Photosynth. Res. 79 (2004) 83–91. [67] P.A. Recchia, C.M. Davis, T.G. Lilburn, J.T. Beatty, P.S. Parkes-Loach, C.N. Hunter, P.A. Loach, Isolation of the PufX protein from Rhodobacter capsulatus and Rhodobacter sphaeroides: evidence for its interaction with the alpha-polypeptide of the core light-harvesting complex, Biochemistry 37 (1998) 11055–11063. [68] R.J. Pugh, P. McGlynn, M.R. Jones, C.N. Hunter, The LH1-RC core complex of Rhodobacter sphaeroides: interaction between components, time-dependent assembly, and topology of the PufX protein, Biochim. Biophys. Acta-Bioenerg. 1366 (1998) 301–316. [69] F. Gubellini, F. Francia, J. Busselez, G. Venturoli, D. Lévy, Functional and structural analysis of the photosynthetic apparatus of Rhodobacter veldkampii, Biochemistry 45 (2006) 10512–10520. [70] F. Francia, J. Wang, G. Venturoli, B.A. Melandri, W.P. Barz, D. Oesterhelt, The reaction center-LH1 antenna complex of Rhodobacter sphaeroides contains one PufX molecule which is involved in dimerization of this complex, Biochemistry 38 (1999) 6834–6845. [71] E.C. Abresch, H.L.A. Axelrod, J.T. Beatty, J.A. Johnson, R. Nechushtai, M.L. Paddock, Characterization of a highly purified, fully active, crystallizable RC–LH1–PufX core complex from Rhodobacter sphaeroides, Photosynth. Res. 86 (2005) 61–70. [72] M. Aklujkar, J.T. Beatty, The PufX protein of Rhodobacter capsulatus affects the properties of bacteriochlorophyll a and carotenoid pigments of light-harvesting complex 1, Arch. Biochem. Biophys. 443 (2005) 21–32. [73] J.F. Miller, S.B. Hinchigeri, P.S. Parkes-Loach, P.M. Callahan, J.R. Sprinkle, J.R. Riccobono, P.A. Loach, Isolation and characterization of a subunit form of the light-harvesting complex of Rhodospirillum rubrum, Biochemistry 26 (1987) 5055–5062. [74] M. Aklujkar, J.T. Beatty, Investigation of Rhodobacter capsulatus PufX interactions in the core complex of the photosynthetic apparatus, Photosynth. Res. 88 (2006) 159–171. [75] C.J. Law, J. Chen, P.S. Parkes-Loach, P.A. Loach, Interaction of bacteriochlorophyll with the LH1 and PufX polypeptides of photosynthetic bacteria: use of chemically synthesized analogs and covalently attached fluorescent probes, Photosynth. Res. 75 (2003) 193–210. [76] W.P. Russ, D.M. Engelman, TOXCAT: a measure of transmembrane helix association in a biological membrane, Proc. Natl. Acad. Sci. U. S. A. 96 (1999) 863–868. [77] G. Drews, J.R. Golecki, Structure, molecular organization, and biosynthesis of membrane of purple bacteria, in: R.E. Blankenship, M.T. Madigan, C.E. Bauer (Eds.), Anoxygenic Photosynthetic Bacteria, Kluwer Academic Publishers, The Netherlands, 1995, pp. 231–257. [78] A.B. Pardee, H.K. Schachman, R.Y. Stanier, Chromatophores of Rhodospirillum rubrum, Nature 169 (1952) 282–283. [79] C.N. Hunter, J.D. Pennoyer, J.N. Sturgis, D. Farrelly, R.A. Niederman, Oligomerization states and associations of light-harvesting pigment protein complexes of Rhodobacter sphaeroides as analyzed by lithium dodecyl-sulfate polyacrylamide-gel electrophoresis, Biochemistry 27 (1988) 3459–3467. [80] J.R. Golecki, U.R. Heinrich, Ultrastructural and electron spectroscopic analyses of cyanobacteria and bacteria, J. Microsc. 162 (1991) 147–154. [81] J.R. Golecki, S. Ventura, J. Oelze, The architecture of unusual membrane tubes in the B800–850 light-harvesting bacteriochlorophyll-deficient mutant 19 of Rhodobacter sphaeroides, FEMS Micro. Letts. 77 (1991) 335–340. [82] M.R. Jones, R.W. Visschers, R. Van Grondelle, C.N. Hunter, Construction and characterisation of a mutant of Rhodobacter sphaeroides with the reaction centre as the sole pigment-protein complex, Biochemistry 31 (1992) 4458–4465.

K. Holden-Dye et al. / Biochimica et Biophysica Acta 1777 (2008) 613–630 [83] M. Sabaty, J. Jappé, J. Olive, A. Verméglio, Organization of electron-transfer components in Rhodobacter sphaeroides forma sp. denitrificans whole cells, Biochim. Biophys. Acta 1187 (1994) 313–323. [84] C. Jungas, J.L. Ranck, J.L. Rigaud, P. Joliot, A. Verméglio, Supramolecular organization of the photosynthetic apparatus of Rhodobacter sphaeroides, EMBO J. 18 (1999) 534–542. [85] W.H. Westerhuis, J.N. Sturgis, E.C. Ratcliffe, C.N. Hunter, R.A. Niederman, Isolation, size estimates, and spectral heterogeneity of an oligomeric series of lightharvesting 1 complexes from Rhodobacter sphaeroides, Biochemistry 41 (2002) 8698–8707. [86] C.A. Siebert, P. Qian, D. Fotiadis, A. Engel, C.N. Hunter, P.A. Bullough, Molecular architecture of photosynthetic membranes in Rhodobacter sphaeroides: the role of PufX, EMBO J. 23 (2004) 690–700. [87] R.N. Frese, J.C. Pàmies, J.D. Olsen, S. Bahatyrova, C.D. van der Weij-de Wit, T.J. Aartsma, C. Otto, C.N. Hunter, D. Frenkel, R. van Grondelle, Protein shape and crowding drive domain formation and curvature in biological membranes, Biophys. J. 94 (2008) 640–647. [88] R.N. Frese, J.D. Olsen, R. Branvall, W.H.J. Westerhuis, C.N. Hunter, R. van Grondelle, The long-range supraorganization of the bacterial photosynthetic unit: A key role for PufX, Proc. Natl. Acad. Sci U. S. A. 97 (2000) 5197–5202. [89] S. Bahatyrova, R.N. Frese, C.A. Siebert, J.D. Olsen, K.O. van derWerf, R. van Grondelle, R.A. Niederman, P.A. Bullough, C. Otto, C.N. Hunter, The native architecture of a photosynthetic membrane, Nature 430 (2004) 1058–1062. [90] R.N. Frese, C.A. Siebert, R.A. Niederman, C.N. Hunter, C. Otto, R. van Grondelle, The long-range organization of a native photosynthetic membrane, Proc. Natl. Acad. Sci. U. S. A. 101 (2004) 17994–17999. [91] R. Comayras, C. Jungas, J. Lavergne, Functional consequences of the organization of the photosynthetic apparatus in Rhodobacter sphaeroides – II. A study of PufX− membranes, J. Biol. Chem. 280 (2005) 11214–11223. [92] W.H.J. Westerhuis, J.W. Farchaus, R.A. Niederman, Altered spectral properties of the B875 light-harvesting pigment-protein complex in a Rhodobacter-sphaeroides mutant lacking PufX, Photochem. Photobiol. 58 (1993) 460–463. [93] P. Joliot, A. Joliot, A. Verméglio, Fast oxidation of the primary electron acceptor under anaerobic conditions requires the organization of the photosynthetic chain of Rhodobacter sphaeroides in supercomplexes, Biochim. Biophys. Acta-Bioenerg. 1706 (2005) 204–214. [94] P. Qian, C.N. Hunter, P.A. Bullough, The 8.5 angstrom projection structure of the core RC–LH1-PufX dimer of Rhodobacter sphaeroides, J. Mol. Biol. 349 (2005) 948–960. [95] S. Scheuring, F. Francia, J. Busselez, B.A. Melandri, J.L. Rigaud, D. Lévy, Structural role of PufX in the dimerization of the photosynthetic core complex of Rhodobacter sphaeroides, J. Biol. Chem. 279 (2004) 3620–3626. [96] T. Geyer, On the effects of PufX on the absorption properties of the light-harvesting complexes of Rhodobacter sphaeroides, Biophys. J. 93 (2007) 4374–4381. [97] P.-E. Milhiet, F. Gubellini, A. Berquand, P. Dosset, J.-L. Rigaud, C. Le Grimellec, D. Lévy, High-resolution AFM of membrane proteins directly incorporated at high density in planar lipid bilayer, Biophys. J. 91 (2006) 3268–3275. [98] S. Scheuring, J. Busselez, D. Lévy, Structure of the dimeric PufX-containing core complex of Rhodobacter blasticus by in situ atomic force microscopy, J. Biol. Chem. 280 (2005) 1426–1431. [99] J. Busselez, M. Cottevieille, P. Cuniasse, F. Gubellini, N. Boisset, D. Lévy, Structural basis for the PufX-mediated dimerization of bacterial photosynthetic core complexes, Structure 15 (2007) 1674–1683. [100] W.P. Russ, D.M. Engelman, The GxxxG motif: a framework for transmembrane helix–helix association, J. Mol. Biol. 296 (2000) 911–919. [101] M.A. Lemmon, J.M. Flanagan, H.R. Treutlein, J. Zhang, D.M. Engelman, Sequence specificity in the dimerization of transmembrane alpha-helices, Biochemistry 31 (1992) 12719–12725. [102] T. Ueda, Y. Morimoto, M. Sato, T. Kakuno, J. Yamashita, T. Horio, Isolation, characterization, and comparison of a ubiquitous pigment-protein complex consisting of a reaction center and light-harvesting bacteriochlorophyll proteins present in purple photosynthetic bacteria, J. Biochem. 98 (1985) 1487–1498. [103] C. Francke, J. Amesz, The size of the photosynthetic unit in purple bacteria, Photosynth. Res. 46 (1995) 347–352. [104] W.P. Barz, F. Francia, G. Venturoli, B.A. Melandri, A. Verméglio, D. Oesterhelt, Role of PufX protein in photosynthetic growth of Rhodobacter sphaeroides. 1. PufX is required for efficient light driven electron-transfer and photophosphorylation under anaerobic conditions, Biochemistry 34 (1995) 15235–15247. [105] F. Francia, M. Dezi, A. Rebecchi, A. Mallardi, G. Palazzo, B.A. Melandri, G. Venturoli, Light-harvesting complex 1 stabilizes P+Q−B charge separation in reaction centers of Rhodobacter sphaeroides, Biochemistry 43 (2004) 14199–14210. [106] P. McGlynn, C.N. Hunter, M.R. Jones, The Rhodobacter sphaeroides PufX protein is not required for photosynthetic competence in the absence of a light-harvesting system, FEBS Letts. 349 (1994) 349–353. [107] S. Bahatyrova, R.N. Frese, K.O. van der Werf, C. Otto, C.N. Hunter, J.D. Olsen, Flexibility and size heterogeneity of the LH1 light harvesting complex revealed by atomic force microscopy — functional significance for bacterial photosynthesis, J. Biol. Chem. 279 (2004) 21327–21333. [108] T.G. Lilburn, J.T. Beatty, Suppressor mutants of the photosynthetically incompetent PufX deletion mutant Rhodobacter capsulatus DRC6(pTL2), FEMS Micro. Letts. 100 (1992) 155–159. [109] C.N. Hunter, J.D. Tucker, R.A. Niederman, The assembly and organisation of photosynthetic membranes in Rhodobacter sphaeroides, Photochem. Photobiol. Sci. 4 (2005) 1023–1027. [110] T. Geyer, V. Helms, A spatial model of the chromatophore vesicles of Rhodobacter sphaeroides and the position of the cytochrome bc1 complex, Biophys. J. 91 (2006) 921–926.

629

[111] T. Geyer, V. Helms, Reconstruction of a kinetic model of the chromatophore vesicles from Rhodobacter sphaeroides, Biophys. J. 91 (2006) 927–937. [112] M.K. Sener, J.D. Olsen, C.N. Hunter, K. Schulten, Atomic-level structural and functional model of a bacterial photosynthetic membrane vesicle, Proc. Natl. Acad. Sci. U. S. A. 104 (2007) 15723–15728. [113] W.P. Barz, D. Oesterhelt, Photosynthetic deficiency of a Pufx deletion mutant of Rhodobacter-sphaeroides is suppressed by point mutations in the light-harvesting complex genes pufB or pufA, Biochemistry 33 (1994) 9741–9752. [114] T.G. Lilburn, R.C. Prince, J.T. Beatty, Mutation of the Ser2 codon of the lightharvesting B870 alpha-polypeptide of Rhodobacter capsulatus partially suppresses the PufX phenotype, J. Bact. 177 (1995) 4593–4600. [115] P. McGlynn, W.H.J. Westerhuis, M.R. Jones, C.N. Hunter, Consequences for the organization of reaction center-light harvesting antenna 1 (LH1) core complexes of Rhodobacter sphaeroides arising from deletion of amino acid residues from the C terminus of the LH1 alpha polypeptide, J Biol. Chem. 271 (1996) 3285–3292. [116] S. Hess, K. Visscher, J. Ulander, T. Pullerits, M.R. Jones, C.N. Hunter, V. Sundstrom, Direct energy transfer from the peripheral LH2 antenna to the reaction centre in a mutant of Rhodobacter sphaeroides that lacks the core LH1 antenna, Biochemistry 32 (1993) 10314–10322. [117] W.P. Barz, A. Verméglio, F. Francia, G. Venturoli, B.A. Melandri, D. Oesterhelt, Role of the PufX protein in photosynthetic growth of Rhodobacter sphaeroides 2. PufX is required for efficient ubiquinone ubiquinol exchange between the reaction-center QB site and the cytochrome bc1 complex, Biochemistry 34 (1995) 15248–15258. [118] A. Verméglio, P. Joliot, Supramolecular organisation of the photosynthetic chain in anoxygenic bacteria, Biochim. Biophys. Acta-Bioenerg. 1555 (2002) 60–64. [119] D.J. Richardson, G.F. King, D.J. Kelly, A.G. McEwan, S.J. Ferguson, J.B. Jackson, The role of auxiliary oxidants in maintaining redox balance during phototrophic growth of Rhodobacter capsulatus on propionate or butyrate, Arch. Micro. 150 (1988) 131–137. [120] M.R. Jones, D.J. Richardson, A.G. McEwan, S.J. Ferguson, J.B. Jackson, In vivo redox poising of the cyclic electron transport system of Rhodobacter capsulatus and the effects of the auxiliary oxidants, nitrate, nitrous oxide and trimethylamine N-oxide, as revealed by multiple short flash excitation, Biochim. Biophys. Acta 1017 (1990) 209–216. [121] F. Francia, J. Wang, H. Zischka, G. Venturoli, D. Oesterhelt, Role of the N- and C-terminal regions of the PufX protein in the structural organization of the photosynthetic core complex of Rhodobacter sphaeroides, Eur. J. Biochem. 269 (2002) 1877–1885. [122] N. Yutin, O. Béjà, Putative novel photosynthetic reaction centre organizations in marine aerobic anoxygenic photosynthetic bacteria: insights from metagenomics and environmental genomics, Environ. Microbiol. 7 (2005) 2027–2033. [123] J.C. Venter, K. Remington, J.F. Heidelberg, A.L. Halpern, D. Rusch, J.A. Eisen, D.Y. Wu, I. Paulsen, K.E. Nelson, W. Nelson, D.E. Fouts, S. Levy, A.H. Knap, M.W. Lomas, K. Nealson, O. White, J. Peterson, J. Hoffman, R. Parsons, H. Baden-Tillson, C. Pfannkoch, Y.H. Rogers, H.O. Smith, Environmental genome shotgun sequencing of the Sargasso Sea, Science 304 (2004) 66–74. [124] O. Béjà, M.T. Suzuki, J.F. Heidelberg, W.C. Nelson, C.M. Preston, T. Hamada, J.A. Eisen, C.M. Fraser, E.F. DeLong, Unsuspected diversity among marine aerobic anoxygenic phototrophs, Nature 415 (2002) 630–633. [125] L.A. Waidner, D.L. Kirchman, Aerobic anoxygenic photosynthesis genes and operons in uncultured bacteria in the Delaware River, Environ. Microbiol. 7 (2005) 1896–1908. [126] O. Hucke, E. Schiltz, G. Drews, A. Labahn, Sequence analysis reveals new membrane anchor of reaction centre-bound cytochromes possibly related to PufX, FEBS Letts. 535 (2003) 166–170. [127] K.A. Weyer, W. Schafer, F. Lottspeich, H. Michel, The cytochrome subunit of the photosynthetic reaction center from Rhodopseudomonas viridis is a lipoprotein, Biochemistry 26 (1987) 2909–2914. [128] C. Kortlüke, K. Breese, N. Gad'on, A. Labahn, G. Drews, Structure of the puf operon of the obligately aerobic, bacteriochlorophyll a-containing bacterium Roseobacter denitrificans OCh114 and its expression in a Rhodobacter capsulatus puf puc deletion mutant, J. Bact. 179 (1997) 5247–5258. [129] K. Matsuura, K. Shimada, Evolutionary relationships between reaction center complexes with and without cytochrome c subunits in purple bacteria, in: M. Baltscheffsky (Ed.), Current Research in Photosynthesis, vol. I, Kluwer Academic Publishers, The Netherlands, 1990, pp. 193–196. [130] F.W. Larimer, P. Chain, L. Hauser, J. Lamerdin, S. Malfatti, L. Do, M.L. Land, D.A. Pelletier, J.T. Beatty, A.S. Lang, F.R. Tabita, J.L. Gibson, T.E. Hanson, C. Bobst, J.L.T.Y. Torres, C. Peres, F.H. Harrison, J. Gibson, C.S. Harwood, Complete genome sequence of the metabolically versatile photosynthetic bacterium Rhodopseudomonas palustris, Nature Biotech. 22 (2004) 55–61. [131] M. Aklujkar, R.C. Prince, J.T. Beatty, The puhE gene of Rhodobacter capsulatus is needed for optimal transition from aerobic to photosynthetic growth and encodes a putative negative modulator of bacteriochlorophyll production, Arch. Biochem. Biophys. 437 (2005) 186–198. [132] M. Aklujkar, R.C. Prince, J.T. Beatty, The PuhB protein of Rhodobacter capsulatus functions in photosynthetic reaction center assembly with a secondary effect on light-harvesting complex 1, J. Bact. 187 (2005) 1334–1343. [133] M. Aklujkar, R.C. Prince, J.T. Beatty, The photosynthetic deficiency due to puhC gene deletion in Rhodobacter capsulatus suggests a PuhC protein-dependent process of RC/LH1/PufX complex reorganization, Arch. Biochem. Biophys. 454 (2006) 59–71. [134] C.N. Hunter, Genetic manipulation of the antenna complexes of purple bacteria, in: R.E. Blankenship, M.T. Madigan, C.E. Bauer (Eds.), Anoxygenic Photosynthetic Bacteria, Kluwer Academic Publishers, The Netherlands, 1995, pp. 473–501.

630

K. Holden-Dye et al. / Biochimica et Biophysica Acta 1777 (2008) 613–630

[135] D.C. Youvan, S. Ismail, E.J. Bylina, Chromosomal deletion and plasmid complementation of the photosynthetic reaction center and light-harvesting genes from Rhodopseudomonas capsulata, Gene 38 (1985) 19–30. [136] M. Griffiths, W.R. Sistrom, G. Cohen-Bazire, R.Y. Stanier, Function of carotenoids in photosynthesis, Nature 176 (1955) 1211–1214. [137] M. Griffiths, R.Y. Stanier, Some mutational changes in the photosynthetic pigment system of Rhodopseudomonas spheroides, J. Gen. Micro. 14 (1956) 698–715. [138] R.K. Clayton, C. Smith, Rhodopseudomonas-spheroides — high catalase and bluegreen double mutants, Biochem. Biophys. Res. Commun. 3 (1960) 143–145. [139] D.W. Reed, R.K. Clayton, Isolation of a reaction center fraction from Rhodopseudomonas spheroides, Biochem. Biophys. Res. Commun. 30 (1968) 471–475. [140] D.W. Reed, Isolation and composition of a photosynthetic reaction center complex from Rhodopseudomonas spheroides, J. Biol. Chem. 244 (1969) 4936–4941. [141] E. Davidson, R.J. Cogdell, The polypeptide composition of the B850 lightharvesting pigment-protein complex from Rhodopseudomonas sphaeroides, R26.1, FEBS Letts. 132 (1981) 81–84. [142] G.A. Armstrong, Genetic analysis and regulation of carotenoid biosynthesis: structure and function of the crt genes and gene products, in: R.E. Blankenship, M.T. Madigan, C.E. Bauer (Eds.), Anoxygenic Photosynthetic Bacteria, Kluwer Academic Publishers, The Netherlands, 1995, pp. 1135–1157. [143] R.B. Tunnicliffe, E.C. Ratcliffe, C.N. Hunter, M.P. Williamson, The solution structure of the PufX polypeptide from Rhodobacter sphaeroides, FEBS Letts. 580 (2006) 6967–6971. [144] S. Onodera, H. Suzuki, Y. Shimada, M. Kobayashi, T. Nozawa, Z.Y. Wang, Overexpression and characterization of the Rhodobacter sphaeroides PufX membrane protein in Escherichia coli, Photochem. Photobiol. 83 (2007) 139–144. [145] Z.Y. Wang, H. Suzuki, M. Kobayashi, T. Nozawa, Solution structure of the Rhodobacter sphaeroides PufX membrane protein: Implications for the qui-

[146]

[147]

[148]

[149] [150]

[151] [152] [153]

[154]

none exchange and protein–protein interactions, Biochemistry 46 (2007) 3635–3642. A. Zouni, H.T. Witt, J. Kern, P. Fromme, N. Krauss, W. Saenger, P. Orth, Crystal structure of photosystem-II from Synechococcus elongatus at 3.8 Å resolution, Nature 409 (2001) 739–743. N. Kamiya, J.-R. Shen, Crystal structure of oxygen-evolving photosystem II from Thermosynechococcus vulcanus at 3.7-Angstrom resolution, Proc. Natl. Acad. Sci. U. S. A. 100 (2003) 98–103. J. Biesiadka, B. Loll, J. Kern, K.D. Irrgang, A. Zouni, Crystal structure of cyanobacterial photosystem II at 3.2 Angstrom resolution: a closer look at the Mn-cluster, Phys. Chem. Chem. Phys. 6 (2004) 4733–4736. K.N. Ferreira, T.M. Iverson, K. Maghlaoui, J. Barber, S. Iwata, Architecture of the photosynthetic oxygen-evolving centre, Science 303 (2004) 1831–1838. B. Loll, J. Kern, W. Saenger, A. Zouni, J. Biesiadka, Towards complete cofactor arrangement in the 3.0 angstrom resolution structure of photosystem II, Nature 438 (2005) 1040–1044. M.R. Jones, Lipids in photosynthetic reaction centres: structural roles and functional holes, Prog. Lip. Res. 46 (2007) 56–87. W.L. DeLano, The PyMOL Molecular Graphics System, DeLano Scientific, 2002 San Carlos, CA, USA, http://www.pymol.org. J.A. Potter, P.K. Fyfe, D. Frolov, M.C. Wakeham, R. van Grondelle, B. Robert, M.R. Jones, Strong effects of an individual water molecule on the rate of primary charge separation in the Rhodobacter sphaeroides reaction centre, J. Biol. Chem. 280 (2005) 27155–27164. M.Z. Papiz, S.M. Prince, T. Howard, R.J. Cogdell, N.W. Isaacs, The structure and thermal motion of the B800–850 LH2 complex from Rps. acidophila at 2.0 Å resolution and 100 k: new structural features and functionally relevant motions, J. Mol. Biol. 326 (2003) 1523–1538.