Relations of phototrophic bacteria and the evolution of photosynthesis

I am very interested in the evolution of photosynthesis. From my research, I have come to conclude that photosynthesis evolved near the root or at the root of the tree of life of the domain Bacteria. In other words, the core components of photosynthesis such as reaction center proteins or the enzymatic components of the chlorophyll and bacteriochlorophyll synthesis pathway, originated before most of the known phyla of bacteria appeared for the first time.

A way to understand how and when photosynthesis originated we need to understand the phylogenetic relations of bacteria. Known phototrophic bacteria are found in seven phyla of bacteria: Cyanobacteria, Firmicutes, Chloroflexi, Chlorobi, Proteobacteria, Acidobacteria, and Gemmatimonadetes. In addition to this, it has been shown that the phylum Actinobacteria might have been ancestrally capable of phototrophy, because some of the strains in this phylum seem to have a vestigial chlorophyll synthesis pathway.

An interesting aspect about the evolution of photosynthesis is that all the phyla containing phototrophic bacteria are clustered within non-phototrophic groups.

Within the phylum Chlorobi there are at least two classes, Chlorobea and Ignavibacteria. All described members of the Ignavibacteria class lack phototrophy. The phylum Chlorobi is very closely related to the phylum Bacteroidetes and no phototrophic Bacteroidetes have been described yet.

Within the phylum Firmicutes, phototrophy is only found in the family Heliobacteraceae (Heliobacteria). Heliobacteria is closely related to the family Peptococcaceae, with the closest strains to Heliobacterium modesticaldum (the only fully sequenced strain of Heliobacteria) being Desulfitobacterium and Synthrophobotulus, which are not phototrophic. What is more, the phylum Firmicutes is subdivided in several classes including, Clostridia, Bacilli, and others. The Heliobacteraceae family belongs to the class Clostridia.

Within the phylum Proteobacteria, only the Alpha-, Beta-, and Gammaproteobacteria are phototrophic, but this is not a universal trait among them. There are many representative of these classes that are not phototrophic. Currently there are no described strains in the class Delta- and Epsilonproteobacteria  with phototrophy.

Only a single strain in the phylum Acidobacteria has been described with phototrophy. Chloracidobacterium thermophilum, but it is probably not the only one. All other strains described in this phylum are non-phototrophic. It should be said though, that only very few strains in this phylum have been characterized. Acidobacteria is closely related to the Proteobacteria as demonstrated by many phylogenomic analysis, sometimes the phylum actually clusters within the Deltaproteobacteria. It is very likely that Acidobacteria and Proteobacteria shared a common ancestor.

Cyanobacteria were thought to be made of only phototrophic strains, however the discovery of Melainabacteria has shown that also Cyanobacteria have very close non-phototrophic relatives. It has been suggested that Melainabacteria should be classified within the phylum Cyanobacteria, and the known photosynthetic Cyanobacteria should be downgraded to class level.

The phylum Chloroflexi is made of about eight different classes, only two of them, the Chloroflexia and the Anaerolineae, have been shown to contain phototrophic strains.

Only one strain in the phylum Gemmatimonadetes has been described to be capable of phototrophy. It appears that this strain obtained photosynthesis via horizontal gene transfer from proteobacteria. However, metagenomic analysis has demonstrated that there are more strains in this phylum with phototrophy.

It is usually considered that the scattered distribution of phototrophy in the tree of life of bacteria is the outcome of horizontal gene transfer, but this is probably not completely correct. So it is suggested for example, that phototrophy could have evolved in the phylum Chlorobi and then via horizontal gene transfer it seeded all other branches. In the case of Cyanobacteria, it is common to suggest that the phylum obtained both reaction centers from two distinct phyla of anoxygenic phototrophic bacteria. This is just one example but almost all possible origins have been considered, including Heliobcteria, Chloroflexi, Cyanobacteria, and Proteobacteria as the original innovators of photosynthesis.

There is absolutely no data or piece of evidence to suggest that photosynthesis originated in any of the described phototrophic groups. All reaction center proteins and all core proteins of the chlorophyll synthesis pathway (e.g. Mg-chelatase, Mg-protoporphyrin IX methyl transferase, protochlorophyllide a oxidoredutase, chlorophyllide a oxidoreductase) share a common origin. They should have descended from ancestral forms that existed more than 3.5 billion years ago. Therefore these ancestral forms should have existed before they diversified into the forms commonly found in the extant groups that we have right now. None of the groups of phototrophs we know today carry these ancestral forms of reaction center proteins or chlorophyll synthesis genes. So for example, suggesting that photosynthesis evolved in the Chlorobi implies that the phylum Chlorobi already existed 3.5 billion years ago; not only that, but it implies that the divergence of Ignavibacteria and Chlorbia had already occurred then, and the divergence of Bacteroidetes and Chlorbi too, and in consequence it has the ultimate implication that the vast majority of groups of bacteria had already differentiated at that time: which is virtually impossible. The same applies if we select any of the phototrophic groups as possible birth places of photosynthesis. In fact, it is very likely that the divergence events that caused the diversification of the group of bacteria mentioned above: Chlorobi/Bacteroidetes, Cyanobacteria/Melainabacteria, Acidobacteria/Proteobacteria, Chlostridia/Bacilli occurred around the Great Oxygenation Event or after. Most likely, photosynthesis is an ancestral trait of the domain Bacteria that has been lost as the different groups of microbes adapted to an oxygenic world, to heterotrophic lifestyles (due to global primary production being relegated to oxygenic photosynthesis), and to live in symbiotic or parasitic relationships with all eukaryotes on Earth. This is valid even if horizontal gene transfer has occurred among ancestor of today’s phyla at some point in time. The only convincing and unambiguous case of horizontal gene transfer of phototrophy is the case of Gemmatimonas and this probably occurred only recently.

Can rubisco be used in a technology for carbon capture?

How much rubisco is needed to capture 37 Gigatonnes of CO₂ in one year? 37 Gigatonnes is the current amount of human emissions per year.

Assuming that 1 single rubisco enzyme fixes 3 CO₂ molecules per second, then in a year 1 mole of rubisco, weighing about 0.49 tonnes, could fix 4163.69 tonnes of CO₂. To fix 37 Gigatonnes of CO₂ in one year, it will be necessary 4.35 Megatonnes of rubisco. To produce this amount of rubisco in one year, the rate of production needed is 0.138 tonnes of rubisco per second. I found somewhere that in nature about 1000 tonnes of rubisco are produced per second, so the rate of production would be equivalent to 0.01% of nature’s.

In comparison 300 Megatonnes of plastic are produced each year as of 2013, equivalent to a rate of production of 9.51 tonnes per second.

If you spread the production of 0.138 tonnes of rubisco per second across countries all over the world, then it doesn't really seem like much.

The above example assumed that 1 rubisco could be active for an entire year, but in reality the half-life of rubisco is of several days, which is actually pretty stable. This means the production rate has to be higher than those values to compensate for the inactivation of the enzyme. On the other hand, there are also better rubiscos that can fix more than 3 CO₂ molecules per second, so it all balances itself out.

The big question is: can a hybrid system that combines genetic engineer, materials, surfaces, and other industrial technologies be implemented to use the reactions of the Calvin-Benson-Bassham cycle as a CO₂ sequestration mechanism? My colleagues at Imperial would probably say NO… but I’m hopeful that we’re only scratching the surface of the technologies that we’ll be able to develop in the near and far future. This system would have to be fully powered by renewables of course (e.g. solar) and it would have to be better than biomass accumulation.

What do you think?

Nature's carbon capture enzyme, also known as rubisco

A cyanobacterium with an anoxygenic Type II reaction center from purple bacteria? (Contamination)

With the surge of genome sequences it is crazy what you can find. As I was doing a few BLASTs searching for new photosynthetic reaction center proteins in the database I noticed that the genome of Lyngbya confervoides BDU141951, which was reported early this year (link), contained genes that encode an anoxygenic Type II reaction center from Alphaproteobacteria (Purple Bacteria).

The three genes for the L, M and H subunit.

Name	Accession	Start	Stop	Strand	GeneID	Locus	Locus tag	Protein product	Length	Protein name
Un	NZ_JTHE01000129.1	1447	2301	+	-	-	QQ91_RS05655	WP_039723522.1	284	photosynthetic reaction center subunit L
Un	NZ_JTHE01000129.1	2315	3298	+	-	-	QQ91_RS05660	WP_039723523.1	327	photosynthetic reaction center subunit M
Un	NZ_JTHE01000254.1	3008	3799	-	-	-	QQ91_RS10590	WP_039724487.1	263	photosynthetic reaction center subunit H

1. The best hit for the L subunit was to:

Photosynthetic reaction center subunit L [Oceanibaculum indicum]
Sequence ID: ref|WP_008944428.1|

Score	Expect	Method	Identities	Positives	Gaps
442 bits(1137)	2e-153	Compositional matrix adjust.	222/280(79%)	240/280(85%)	1/280(0%)

2. The best hit for the M subunit was to:

Photosynthetic reaction center subunit M [Ahrensia sp. R2A130]
Sequence ID: ref|WP_009758337.1|

Score	Expect	Method	Identities	Positives	Gaps
457 bits(1175)	5e-158	Compositional matrix adjust.	221/308(72%)	260/308(84%)	2/308(0%)

3. The best hit for the H subunit was to:

Photosynthetic reaction centre, H-chain [Hoeflea phototrophica]
Sequence ID: ref|WP_007196606.1|

Score	Expect	Method	Identities	Positives	Gaps
282 bits(721)	5e-91	Compositional matrix adjust.	139/258(54%)	175/258(67%)	2/258(0%)

What is going on?

When I BLASTed the BchZ subunit of chlorophyllide a reductase from Rhodobacter sphaeroides, which makes bacteriochlorophyllide a from chlorophyllide a... the best hit was to this same strain of cyanobacteria: but it seems it is a fragment only (see image below, never mind the annotation). A close homologous to BchY did not retrieve anything out of the ordinary, the same for BchX. So this cyanobacterium cannot really make bacteriochlorophyll... I wonder if the anoxygenic Type II reaction center in this strain could be active using chlorophyll a along side PSII and PSI.

The fact that the best hits to the reaction center proteins are below 80% means that the alphaproteobacterium that donated this genes represents at least a new genus or something.

UPDATE: It turned out after all to be contamination. It might be that the cultures used were not completely axenic. It's a pity.

Horizontal gene transfer from gammaproteobacteria to cyanobacteria and others

It seems to me that Gammaproteobacteria can transfer a lot of their photosynthesis related genes to other bacteria.

Recent phylogenetic studies (Bryant and Liu, 2013; Sousa et al., 2013) showed that an ancestor of marine Synechococcus and Prochlorococcus strains obtained a set of bchLNB genes from gammaproteobacteria. Such an event of lateral gene transfer from a gammaproteobacterium into an ancestor of the Synechococcus/Prochlorococcus group also included other genes, encoding proteins such as CmpA involved in circadian output (Dvornyk, 2006), carboxisome proteins, rubisco (Marin et al., 2007), threonyl tRNA synthetase, and quite possibly many more (Zhaxybayeva et al., 2006; Zhaxybayeva et al., 2009).

Gemmatimonas photoautotrophica acquired photosynthesis via horizontal gene transfer from a gammaproteobacterium (Zeng et al., 2014) and the same seems true for a Firmicutes of the genus Alkalibacterium. I wonder why...

Microbial mat

References
Bryant, D.A., and Liu, Z.F. (2013). Green bacteria: Insights into green bacterial evolution through genomic analyses. Advances in Botanical Research 66, 99-150.

Dvornyk, V. (2006). Subfamilies of cpmA, a gene involved in circadian output, have different evolutionary histories in cyanobacteria. Microbioliology 152, 75-84.

Marin, B., Nowack, E.C.M., Glockner, G., and Melkonian, M. (2007). The ancestor of the Paulinella chromatophore obtained a carboxysomal operon by horizontal gene transfer from a Nitrococcus-like gamma-proteobacterium. BMC evolutionary biology 7.

Perreault, N.N., Greer, C.W., Andersen, D.T., Tille, S., Lacrampe-Couloume, G., Lollar, B.S., and Whyte, L.G. (2008). Heterotrophic and autotrophic microbial populations in cold perennial springs of the high arctic. Appl. Environ. Microb. 74, 6898-6907.

Sousa, F.L., Shavit-Grievink, L., Allen, J.F., and Martin, W.F. (2013). Chlorophyll biosynthesis gene evolution indicates photosystem gene duplication, not photosystem merger, at the origin of oxygenic photosynthesis. Genome biology and evolution 5, 200-216.

Zeng, Y.H., Feng, F.Y., Medova, H., Dean, J., and Koblizek, M. (2014). Functional Type 2 photosynthetic reaction centers found in the rare bacterial phylum Gemmatimonadetes. P. Natl. Acad. Sci. U. S. A. 111, 7795-7800.

Zhaxybayeva, O., Doolittle, W.F., Papke, R.T., and Gogarten, J.P. (2009). Intertwined evolutionary histories of marine Synechococcus and Prochlorococcus marinus. Genome biology and evolution 1, 325-339.

Zhaxybayeva, O., Gogarten, J.P., Charlebois, R.L., Doolittle, W.F., and Papke, R.T. (2006). Phylogenetic analyses of cyanobacterial genomes: Quantification of horizontal gene transfer events. Genome Research 16, 1099-1108.

Evolution of oxygenic photosynthesis - New D1 sequences

Early this year I published a study on the evolution of all D1 proteins (Cardona et al., 2015). Interestingly, there are a number of sequences that were very atypical and appeared to be early evolving. Because of their phylogenetic position and sequence characteristics, I suggested that it is possible that these D1 sequences evolved before the water oxidizing complex had reached its standard configuration in PSII.

Back then there were about 40-45 atypical sequences. Since them more have appeared, there are a total of 62 sequences... so here I show a Maximum Likelihood tree for all the atypical D1 forms. See Figure 1. Of particular interests is the fact that they seem to follow an evolutionary pattern consistent with vertical descent and loss, although some likely events of later gene transfer can also be identified.

Figure 1. Updated tree of atypical D1 sequences.

Of particular interest is the appearance of numerous D1 fragments. Some of them are from incompletely sequenced genes, but some of them seem to be legitimate proteins, probably originating from partial gene duplication followed by divergence (Figure 2). A couple of these fragmented D1 seems to have phylogenetic affinity for the early evolving forms.

Figure 2. Sequence alignments of different D1. That marked with Fr. is the D1 fragment, found in the genome of some of the earliest evolving cyanobacteria strains, Synechococcus sp. PCC 7336. It seems to have some affinity for the G0 and early branching forms.

Is a 'tree of life' also a 'tree of death'?

Imagine that you could build an evolutionary tree that included every single organism that has ever existed since the origin of life: including every single bacterium, every single archaeum: every individual organism.

The divisions between groups of organisms would then become blurred. For example, it would be impossible to tell where Homo neanderthalensis ends and where Homo sapiens begins. It would be impossible to detect the exact moment when non-avian dinosaurs turned into birds. This also applies at the unicellular level.

The reason we can distinguish groups in the trees of life that represent evolutionary events, is because we do not see most of the branches: the branches that have gone extinct. It is the absence of these individuals, which did not successfully passed on their genomes to the next generation, that allows for the classification and distinction of different types of organisms.

Therefore, a phylogenetic tree of organisms not only gives insight into the groups of organisms in question, but it also shed lights into the organisms that must have existed but did not make it. You could say that a tree of life is also a tree of death.

It also implies that using a phylogenetic tree we could calculate the amount of diversity that has been lost between the two most closely related branches. Assuming that we know the rate of divergence, and that we could somehow put a number to “amount of diversity”.

Darwin's Tree of Life, 1837