Early Archean origin of heterodimeric Photosystem I

Abstract

When and how oxygenic photosynthesis originated remains controversial. Wide uncertainties exist for the earliest detection of biogenic oxygen in the geochemical record or the origin of water oxidation in ancestral lineages of the phylum Cyanobacteria. A unique trait of oxygenic photosynthesis is that the process uses a Type I reaction centre with a heterodimeric core, also known as Photosystem I, made of two distinct but homologous subunits, PsaA and PsaB. In contrast, all other known Type I reaction centres in anoxygenic phototrophs have a homodimeric core. A compelling hypothesis for the evolution of a heterodimeric Type I reaction centre is that the gene duplication that allowed the divergence of PsaA and PsaB was an adaptation to incorporate photoprotective mechanisms against the formation of reactive oxygen species, therefore occurring after the origin of water oxidation to oxygen. Here I show, using sequence comparisons and Bayesian relaxed molecular clocks that this gene duplication event may have occurred in the early Archean more than 3.4 billion years ago, long before the most recent common ancestor of crown group Cyanobacteria and the Great Oxidation Event. If the origin of water oxidation predated this gene duplication event, then that would place primordial forms of oxygenic photosynthesis at a very early stage in the evolutionary history of life.

Keywords: Biochemistry; Evolution; Genetics.

Fig. 1

Maximum Likelihood tree of Type I reaction centre proteins. The tree is characterised by a deep split of reaction centre proteins, which separates those employed in anoxygenic phototrophy from those employed in oxygenic photosynthesis (grey spot). Cyanobacteria and photosynthetic eukaryotes are the only known phototrophs to have a heterodimeric Type I reaction centre made of two subunits known as PsaA and PsaB. All extant Cyanobacteria descended from a common ancestor that already had highly divergent PsaA and PsaB subunits (red spot). The gene duplication that led to PsaA and PsaB occurred at an earlier point in time (orange spot), which predated the most recent common ancestor of Cyanobacteria by an unknown period of time. It is hypothesised that the gene duplication that led to PsaA and PsaB occurred as a specialisation to oxygenic photosynthesis, therefore water oxidation should have originated before this gene duplication event (arrow).

Fig. 2

Percentage of sequence identity of PsaA and PsaB as a function of time. The data points are listed in Table 1. Dots in orange represent the level of sequence identity of PsaA from one species compared to the same subunit in a different species. Dots in grey represent PsaB for similar pairs of species. The blue bar represents the GOE. The oldest pair of dots are those from G. kilaueensis and are placed arbitrarily at about the start of the GOE. The second oldest node was set at 1.45 Ga and taken as the average of the estimated divergence time for the clade containing Cyanidioschizon merolae as calculated in Ref. and , dated in the range of 1.1–1.8 Ga respectively. The red bars represent the level of uncertainty on the estimated divergence times of the clades that contain these species. The blue dashed line marks the average level of seq_id between PsaA and PsaB in all known oxygenic phototrophs, at about 42.5%.

Fig. 3

Bayesian relaxed molecular clock of Type I reaction centres. The tree was calculated assuming that Type I reaction centres had originated by 3.5 Ga (grey dot). The orange dot indicates the duplication event that allowed the divergence of PsaA and PsaB. Red dots highlight the nodes that were calibrated as described in Materials and Methods. The light grey bars along the nodes show the uncertainty, 95% confidence interval, on the estimated divergence time. Sequences marked as hA denote those from strains of Heliobacteria, while those marked as cA from Chlorobi and Acidobacteria. Sequences marked as A and B represent PsaA and PsaB from Cyanobacterial and eukaryotic PSI. The blue bar marks the GOE.

Fig. 4

Effect of calibration choice on the estimated divergence time. (A) Comparison of divergence time estimates of a tree calculated using a root of 3.5 Ga and a minimum age for the MRCA of Cyanobacteria of 2.45 Ga (x axis) against a tree calculated using a minimum age for the MRCA of 3.00 Ga (y axis). There is no deviation from the slope (red dashed line) meaning that the calibration did not have a strong effect on the estimated divergence time. The orange dot represents the divergence time of the duplication that led to PsaA and PsaB. (B) Divergence times of a tree calculated using a root of 3.5 Ga and a minimum age for the MRCA of Cyanobacteria of 2.45 Ga (x axis) against a tree calculated without a calibration on this point (y axis). (C) Divergence times of a tree calculated using a root of 3.5 Ga and a minimum age for the MRCA of Cyanobacteria of 2.45 Ga (x axis) against a tree calculated without a calibration on points 8, 7, 9 (y axis). (D), (E), and (F) were calculated as (A), (B), and (C) respectively, but employing a root of 3.8 Ga instead. In all scenarios the duplication that led to PsaA and PsaB is the oldest node after the root. Removing the oldest calibration constraints led to younger ages on a number of nodes, however the estimated age of the PsaA/PsaB duplication event (orange dot) is set in the early Archean in every scenario.

Fig. 5

Rates of evolution as a function of divergence time. Grey dots represent divergence times of a tree calculated using a root constraint of 3.5 Ga (shown in Fig. 3), while orange dots represent those calculated using a root constraint of 3.8 Ga. The fastest rate of evolution occurs at the PsaA/PsaB duplication event. The plot shows that higher rates of evolution, measured as amino acid changes per site per million years, are necessary to explain the divergence of PsaA and PsaB at any point in time in the history of life.

Fig. 6

Structural comparisons of cofactor positions in Photosystem I and a homodimeric Type I reaction centre. (A) Overlap of PsaA (grey lines and sticks) and PsaB (orange lines and stick). Conserved carotenoids between PsaA and PsaB are highlighted with red arrows. Betacarotene molecules are shown in thick sticks and chlorophylls in thin lines. The phylloquinone cofactor is marked with A₁, the Fe₄S₄ cluster is marked with F_X, and the redox chlorophyll that makes the ‘special pair’ with P. (B) Overlap of PsaB and PshA from the anoxygenic phototrophic anaerobe Heliobacterium modesticaldum. The homodimeric Type I reaction centre lacks a quinone binding-site, has bacteriochlorophyll g as the main pigment, and features a single carotenoid molecule per monomer (4,4′-diaponeurosporene). The PDB ID for the structure of PSI is 1jb0 and for the structure of the heliobacterial reaction centre is 5v8k .