I wish to higlight here and comment on the recent paper by Jagoda Jabłońska & Dan S. Tawfik titled The evolution of oxygen-utilizing enzymes suggests early biosphere oxygenation published in Nature Ecology and Evolution (25 February 2021). A really exciting paper and a lot of hard work! Something that I wished I could have done at some point myself, for sure… Congratulations to the authors!
The authors summarize in the abstract that: “About 60% of the O2-enzyme families whose birth can be dated appear to have emerged at the separation of terrestrial and marine bacteria (22 families, compared to two families assigned to the last universal common ancestor).”
This ancestor (at the separation of “marine” and “terrestrial” bacteria) represents a point in time that predates Cyanobacteria. The authors don’t say it explicitly, but it is clear now that—at the very least—a substantial fraction of the domain Bacteria likely had an ancestor that was capable of oxygenic photosynthesis. They call this ancestor “The Last Universal Oxygen Ancestor”.
I think we all need to take a moment to think about this and have a conversation about the origin of Bacteria now.
Two things puzzled me about how the authors interacted with my research. The authors say:
“Other niche families could serve as proxies for the emergence of biogenic oxygen, foremost the L and M proteins of Photosystem II; however, members of this family are often found in phototrophic organisms that do not perform oxygenic photosynthesis18. Thus, despite their potential insight, we excluded these families to ensure rigorous assignment (Supplementary Table 2).”
It puzzled me because L and M in anoxygenic photosynthesis are easily distinguishable from those used in oxygenic photosynthesis, D1 and D2. They are not only extremely well understood at a structural and functional level, but their phylogeny is simple. The authors should have been aware of this, since they read and cited my Geobiology paper on the evolution of Photosystem II. A missed opportunity, but I suppose that hesitation is attributable to the bad influence of that cited ref. 18 ;) haha
The second thing is the following, the authors say: “Specifically, a relaxed molecular clock of the Type II reaction centre proteins dates the last common ancestor of the oxygenic system at 3.1 Ga (ref. 30)”
Thank you for the citation! However, we didn’t provide a single absolute age for the origin of type II reaction centres. We studied many models and scenarios, and concluded that oxygenic photosynthesis could predate Cyanobacteria by well over a billion years and that this origin likely occurred in the early Archean. We didn’t specifically say 3.1 Ga. In fact, in our main figure, where we showed one of many scenarios we tested, the age of the respective node was 3.22 Ga.
I should warn the readers however that the “3.1 Ga” date should be interpreted with extreme caution, and should not be taken literally, but only to mean that this ancestor predates the diversification of most known groups of bacteria. Thus, most telling is the place of the ancestral node itself, regardless of the exact date; what the authors called the Last Universal Oxygen Ancestor (LUOA).
In any case, the new paper adds a bit of extra weight to what I’ve been saying for a few years now since before the publication of the Geobiology paper. For those of you who have not seen it yet, the follow-up to it, Oliver et al. 2021, is out now in BBA-Bioenergetics, and it is Open Access. Go have a look: Time-resolved comparative molecular evolution of oxygenic photosynthesis. We showed that Photosystem II and water oxidation [oxygenic photosynthesis] can be as old as life itself, by a very comfortable margin.
This is Figure 9 from our recent paper in BBA-Bioenergetics. It shows the evolutionary relationships of the photosystem core subunits in comparison with the evolution of bacteria to the right. I’m quite pleased with the final result, if I may say so myself. Check the paper for all the details!
On that note, I wish to highlight a text that came published in the supplementary materials of the Geobiology paper, some of it overlaps with the observations reported in the Nature Ecology and Evolution work. In this additional text we discussed some of the biological evidence consistent with oxygenic photosynthesis having started from very early on.
I copy-pasted the text below for your own pleasure and entertainment (it was also published Open Access, references are in the original file in the publisher’s website). Notice the first sentence in the last paragraph… that was a bit of a disclaimer, but perhaps we should think twice about this.
Peculiar oxygen anomalies
The following compilation is not meant to be exhaustive.
A long-standing debate on the early evolution of life is whether aerobic respiration predates oxygenic photosynthesis. This debate started with the realization that the core proteins of the cytochrome c oxidases, which catalyzes the reduction of oxygen to water, are almost universally conserved and appear to be much older than Cyanobacteria, see A. L. Ducluzeau et al. (2014) for a critical review. One example of this can be found in Brochier-Armanet et al. (2009) who wrote: “A-O2Red are the most ancient O2Red and were already present prior to the divergence of major present-day bacterial and archaeal phyla, thus before the emergence of Cyanobacteria and oxygenic photosynthesis.” Many ways to rationalize this anomaly have been put forward, including the possibility that O2 reductases originated to use the traces of oxygen (few nanomolar or less) that are expected in a world without oxygenic photosynthesis (Stolper et al., 2010) or that they originated from NO reductases (A. L. Ducluzeau et al., 2014).
Weiss et al. (2016) reported an automated minimal reconstruction of the proteome of the last universal common ancestor (LUCA). They suggested that this was an anaerobe, methanogenic and capable of nitrogen fixation. However, out of 355 predicted proteins 8 turned out to use oxygen or to deal with reactive oxygen species (ROS): more than those categorized as “nitrogen metabolism” (seven proteins) and “energy metabolism” (two proteins). Nevertheless, Weiss et al. (2016) dismissed these as artefacts.
These 8 proteins are the following:
1. Heme/copper-type cytochrome/quinol oxidase
2. Cu/Zn superoxide dismutase
3. Peroxiredoxin
4. Rubrerythrin
5. Homogentisate 1,2-dioxygenase
6. Aromatic ring-opening dioxygenase
7. Aromatic ring hydroxylase
8. Betacarotene, NADH:oxygen 3-oxidoreductase
The first one is the core subunit of O2 reductases, as seen above. The next three enzymes handle ROS. The great antiquity of ROS-handling enzymes had been noted before (Zamocky et al., 2000; Ouzounis et al., 2006; Knoops et al., 2007; Slesak et al., 2012). The next three enzymes have a role in the catabolism and oxidation of aromatic rings of tyrosine, phenylalanine and other aromatic and phenolic compounds. The last enzyme is a monooxygenase involved in the hydroxylation of the aromatic rings of carotenoids.
The evolution of cytochrome bc and b6f complexes have also hinted to the possibility of oxygenic photosynthesis appearing at a very early stage. Nitschke et al. (2010) wrote the following: “Our presently most favoured rationalisation for the presence of heme c-i postulates that this cofactor allows a substantial reduction in lifetime of the semiquinone species during Q-i site turnover in the b6f complex, thus adapting the enzyme to an oxygenic environment.
Reasoning along these lines, we cannot help noticing that this hypothesis would impose a mind-boggling evolutionary scenario for the emergence of oxygenic photosynthesis: If heme ci indeed evolved as a response to the presence of oxygen, we have to conclude that photosynthetic O2 production had already evolved prior to the node leading to Firmicutes and Cyanobacteria.”
An examination of the phylogeny of oxygen-tolerant hydrogenases led to the following statement by Pandelia et al. (2012): “The tree topologies shown in Figs. 3 and 4 suggest a provocative conclusion: 6C-hydrogenases [oxygen-tolerant] appear to have emerged prior to the divergence of Chlorobia/Heliobacteria on one and α-, β- and γ-proteobacterial radiations on the other side. In the traditional thinking that Heliobacteria and Firmicutes have diverged prior to the origin of oxygenic photosynthesis, the presence of 6C-enzymes in these species would indicate that their appearance pre-dates the emergence of light-driven O2 production. Therefore, either the 6C cluster appeared without an apparent link to O2-tolerance or we need to rethink our evolutionary history of oxygenic photosynthesis.”
A recent paper by Kacar et al. (2017) indicated that the divergence of methanogenic archaeal Group 3 rubisco from photosynthetic bacterial Group 1 rubisco occurred as the latter acquired structural traits to deal with oxygen and to optimize oxygen-mediated regulation, but see also Nisbet et al. (2007). The phylogeny of these enzymes showed a deep divergence for Group 3 and Group 1 rubisco, which is expected for an archaeal/bacterial split. Rubisco is undoubtedly a very ancient enzyme probably as old as 3.5 Ga (Nisbet et al., 2007; Schopf et al., 2018). In agreement with this, Weiss et al. (2016) reported that rubisco would trace to LUCA if a single ancient event of HGT is allowed. Thus, it is not surprising that this enzyme displays very slow rates of evolution. In fact, between all Group 1 rubisco, which include all cyanobacterial and proteobacterial rubisco, the level of sequence identity is not lower than 78% (Kacar et al., 2017). In contrast, the level of sequence identity between archaeal Group 3 rubisco and bacterial Group 1 rubisco is 39%. At these slow rates, even slower than the rates of evolution of D1 and D2, the divergence of Group 1 and Group 3 rubisco would be placed in the early Archean.
Gold et al. (2017) applied a molecular clock approach to time the evolution of sterol biosynthesis proteins, the product of ancient duplications. They studied squalene monooxygenase and oxidosqualene cyclase, both of which use oxygen. They noted that the ancestral node to these oxygen-using enzymes was timed around the GOE: however, the duplication event leading to the acquisition of oxygen-using capabilities was deep in the early Archean.
A.-L. Ducluzeau and Nitschke (2016) have also pointed out the deep dichotomies that exists in several metabolic processes, which arguably may have occurred for the integration of oxygen-using enzymes. The first example is the dichotomy of the heme synthesis pathway: one goes via protoporphyrin IX and can incorporate several oxygen-using enzymes, and a second via sirohydrochlorin found in many anaerobes. A similar dichotomy exists in the biosynthesis pathway of menaquinone, the men pathway and the futalosine pathway; in the pathway for the assembly of iron-sulfur clusters, the Isc and Suf systems; and in the aerobic and anaerobic pathways for the synthesis of cobalamin. A.-L. Ducluzeau and Nitschke (2016) highlighted that oxygen-using enzymes are asymmetrically distributed in these pathways. It should be noted that the phylogenetic relationship of many enzymes in these metabolic processes remain to be elucidated, yet the distribution of these dichotomic pathways across prokaryotes argue against very recent divergences (e.g. post-GOE).
For example, the aerobic pathway for the synthesis of cobalamin uses an enzyme homologous to Mg-chelatase of the chlorophyll synthesis pathway to insert Co into the precursor ring. In the aerobic pathway this is done towards the end of the pathway after the oxygen-requiring step, while in the anaerobic pathway Co insertion occurs at the beginning of the pathway by a different chelatase unrelated to Mg-chelatase, see Sousa et al. (2013) and reference therein. If Mg-chelatase, at the origin of photosynthesis, emerged from Co-chelatase: would not that imply that the aerobic pathway for the synthesis of cobalamin predated the origin of photosynthesis? On the other hand, if Co-chelatase originated from Mg-chelatase then it could be postulated that the aerobic cobalamin synthesis pathway emerged after the evolution of oxygenic photosynthesis. However, all Mg-chelatase subunits used in photosynthesis are monophyletic to the exclusion of Co-chelatase’s paralogs (Sousa et al., 2013). The oxygen-dependent pathway for the synthesis of cobalamin is catalyzed by two enzymes, CobG or CobZ, both of which use oxygen (Heldt et al., 2005). One could postulate therefore that the aerobic pathway was at first an entirely anaerobic path and that the oxygen-dependent step was catalyzed by an alternative enzyme that did not require oxygen, yet an oxygen-independent enzyme catalyzing this step, to the best of our knowledge, has not been reported. This conundrum is easily resolved if primordial forms of oxygenic photosynthesis appeared early in the evolutionary history of life.
David and Alm (2011) using a combined phylogenomic/molecular clock approach showed that a major diversification event took place in prokaryotes starting about 3.4-3.5 Ga and peaking about 3.2 Ga. The authors claimed that in this period of evolutionary innovation 27% of major modern gene families appeared. The authors stated: “Our chronologies of oxygen and redox-sensitive metal and compound utilization suggest ancient increases in oxygen bioavailability, as well as an Archaean biosphere with some of the basic genetic components required for oxygenic photosynthesis and respiration”, and yet their molecular clock (shown in their Supplementary Figure 13) timed the MRCA of Cyanobacteria at about 2.0 Ga, and the divergence of the Cyanobacteria and Chloroflexi phyla at about 3.0 Ga.
More recently, Granold et al. (2018) reported that the last evolving amino acids were only incorporated into the universal genetic code because of oxygen. The authors stated “Our data indicate that in demanding building blocks with more versatile redox chemistry, biospheric molecular oxygen triggered the selective fixation of the last amino acids in the genetic code.” But how could the universal genetic code have been only “fixed” in the late Archean or around the GOE, when the divergence of Bacteria and Archaea may have occurred more than 3.5 Ga ago (Marin et al., 2017; Schopf et al., 2018)?
We do not claim that oxygenic photosynthesis originated in the LUCA. Nonetheless, all of the above apparent anomalies could be conveniently explained if primordial forms of photosynthetic water oxidation appeared at an early stage in the evolutionary history of Bacteria and photosynthesis. In consequence, limited amounts of bioavailable oxygen would affect metabolic processes and allow the emergence of aerobic respiration and other oxygen-using and ROS-handling enzymes from an early time and long before the most recent common ancestor of described Cyanobacteria, even if other complex geological processes delayed the oxygenation of the atmosphere until the GOE (Smit & Mezger, 2017; Bindeman et al., 2018).