Sulfoquinovosyl diacylglycerol is required for dimerisation of the Rhodobacter sphaeroides reaction centre-light harvesting 1 core complex

Abstract

The reaction centre-light harvesting 1 (RC-LH1) core complex is indispensable for anoxygenic photosynthesis. In the purple bacterium Rhodobacter (Rba.) sphaeroides RC-LH1 is produced both as a monomer, in which 14 LH1 subunits form a C-shaped antenna around 1 RC, and as a dimer, where 28 LH1 subunits form an S-shaped antenna surrounding 2 RCs. Alongside the five RC and LH1 subunits, an additional polypeptide known as PufX provides an interface for dimerisation and also prevents LH1 ring closure, introducing a channel for quinone exchange that is essential for photoheterotrophic growth. Structures of Rba. sphaeroides RC-LH1 complexes revealed several new components; protein-Y, which helps to form the quinone channel; protein-Z, of unknown function and seemingly unique to dimers; and a tightly bound sulfoquinovosyl diacylglycerol (SQDG) lipid that interacts with two PufX arginine residues. This lipid lies at the dimer interface alongside weak density for a second molecule, previously proposed to be an ornithine lipid. In this work we have generated strains of Rba. sphaeroides lacking protein-Y, protein-Z, SQDG or ornithine lipids to assess the roles of these previously unknown components in the assembly and activity of RC-LH1. We show that whilst the removal of either protein-Y, protein-Z or ornithine lipids has only subtle effects, SQDG is essential for the formation of RC-LH1 dimers but its absence has no functional effect on the monomeric complex.

Keywords: Rhodobacter sphaeroides; RC-LH1; light harvesting; lipids; photosynthesis; reaction centre.

Figure 1.. Structure of the dimeric Rba. sphaeroides RC-LH1 core complex.

(A) Surface view of the complex from the lumenal (periplasmic) face. The RC subunits are shown in orange (RC-L), magenta (RC-M) and cyan (RC-H). LH1 subunits are in yellow (α) and blue (β). Additional subunits are in red (PufX), green (protein-Y) and purple (protein-Z). Lipids and cofactors are shown in stick representation in green (BChl), magenta (carotenoids) and grey (SQDG). Unassigned density for lipid 2 is shown as a grey surface. Dashed boxes illustrate areas enlarged in panels (B), (C) and (D). (B) Enlarged view highlighting one of the SQDGs and lipid 2 bound at the dimer interface. The subunits from the left monomer (PufX, RC-L, α1 and β1), and those from the right monomer (α1′, β1′ and Z1′) are labelled. (C) A further enlarged view of the SQDG lipid with the protein in ribbon representation. Hydrogen bonds between the SQDG head group and PufX Arg49 and Arg53, and the backbone of RC-L residues Leu75 and Gly140 are shown. (D) Enlarged view of the two protein-Z subunits bound to the left monomer (Z1 and Z2, purple) and protein-Y of the right monomer (Y′, green) in ribbon representation. The rest of the protein is in surface representation.

Figure 2.. Generation of strains deficient in SQDG or ornithine lipids.

(A) Sulfoquinovosyl diacylglycerol (SQDG) is synthesised in two steps. The first step is the addition of sulphite to UDP-glucose (UDP-Glc) to produce UDP-sulfoquinovose (UDP-SQ) by SqdB. Next, diacylglycerol (DG) is added and UDP is removed by SqdA, C and D to produce SQDG [47]. (B) Ornithine lipids are synthesised by the acyltransferases OlsB and OlsA, which sequentially add a 3-hydroxyacyl group then an acyl group to L-ornithine using 3-hydroxyacyl-ACP and acyl-ACP as substrates, respectively [48,49]. (C) Structure of the Rba. sphaeroides sqdBDC operon. The region labelled ΔsqdB (Rsp_2569) was removed to abolish SQDG biosynthesis. (D) Structure of the Rba. sphaeroides olsBA operon. The labelled region spanning olsB (Rsp_3826) and olsA (Rsp_3827) was removed to prevent OL biosynthesis. (E) Agarose gel of ethidium bromide-stained PCR products showing size differences for the amplified regions spanning the sqdB and olsBA genes showing a clear reduction in size in the knockout strains relative to the wild type. (F) TLC plate showing loss of SQDG in the ΔsqdB strain. By comparison to a SQDG standard, SQDG is present in membranes from ΔcrtA but not ΔcrtA ΔsqdB, confirming the loss of SQDG biosynthesis in the absence of SqdB. An uncropped image showing the standards for the other lipids is shown in Supplementary Figure S1.

Figure 3.. The effect of removing protein-Y, protein-Z, SQDG or ornithine lipids on the RC-LH1 monomer: dimer ratio and activity.

(A,B) Separation of LH2, RC-LH1 monomer and RC-LH1 dimer complexes from detergent-solubilised membranes on sucrose step gradients. (A) The wild-type strain (WT), a control strain that does not produce dimeric RC-LH1 (PufX R49L R53L), a strain lacking protein-Y (ΔpuyA), a strain lacking protein-Z (ΔpuzA), a strain that cannot produce ornithine lipids (ΔolsBA), and a strain that cannot produce SQDG (ΔsqdB). (B) Sucrose gradients for the strains in (A) in a background that is also deficient in the crtA gene encoding spheroidene monooxygenase (ΔcrtA). Two further repeats of the gradients in panels (A) and (B) are shown in Supplementary Figure S2. (C) UV/Vis/NIR spectra for monomer and dimer bands harvested from gradients for the ΔcrtA strain, the RC-LH1 dimer-deficient ΔcrtA PufX R49L R53L strain, and the ΔcrtA ΔolsBA and ΔcrtA ΔsqdB strains. The LH1 maximum is at 873 nm. (D) Turnover assays for monomeric complexes from WT, PufX R49L R53L, ΔolsBA and ΔsqdB mutants. Rates show moles of cyt c₂ oxidised (reported as electrons) per second per mole of RC during illumination using an 810 nm LED of a solution containing 0.05 μM RC-LH1 and 5 μM cyt c₂. Unpaired two-tailed t-tests (for three technical replicates) were performed relative to rates for monomeric or dimeric WT complexes where ns denotes a p-value >0.05. (E) UV/Vis/NIR spectra for monomer and dimer bands harvested from gradients for the ΔcrtA strain, and those also harbouring the ΔpuyA and ΔpuzA mutations. The LH1 maximum is at 873 nm, with low levels of LH2 contamination contributing to the slight shoulder at 850 nm. (F) Turnover assays for monomeric and dimeric RC-LH1 complexes from the ΔcrtA, ΔcrtA ΔpuyA and ΔcrtA ΔpuzA strains under the same conditions as in (D) except the RC concentration, which was 0.01 μM. A p-value <0.05 is denoted by * and ns denotes a p-value >0.05.

Figure 4.. Phylogenetic analysis of the proteins involved in RC-LH1 dimer formation.

(A) Sequence alignment of PufX polypeptides. Red outlined boxes indicate the two arginine residues that bind SQDG in Rba. sphaeroides. (B) 16S rRNA phylogeny of selected PufX-producing species of the Rhodobacteriales order for which the oligomeric state of their RC-LH1 complex is known. The root of the Cereibacter subgroup is labelled. Caulobacter vibrioides, a non-photosynthetic alphaproteobacterium, is included as an outgroup to root the tree. The table shows the presence (green + symbols) or absence (red — symbols) of genes encoding the LH2 complex, PufX, protein-Y and protein-Z, and whether the species produces dimeric RC-LH1 complexes and SQDG (the latter inferred by the presence of the sqdB gene).