What was the nature of the first photosynthetic reaction centers? How similar were they to modern RCs? Were they integral membrane proteins similar to the ones we study in organisms that are alive today? Unfortunately, there is almost nothing that can be concluded for certain about these earliest RCs, which almost certainly were found in organisms that are long extinct. However, some informed speculations about these first RCs - 18, 106] can be made, but the uncertainties far outweigh any firm conclusions that can be drawn.
While information on the very earliest reaction centers may be lost forever, some of the story can be picked up and some reasonably reliable conclusions can be drawn by doing a comparative analysis of the structures and gene sequences of all known reaction centers, and by looking for conserved elements. Assuming that these complexes are products of divergent, and not convergent, evolution it has been concluded that these conserved elements are derived from an ancestor of all known reaction centers. This ancestral complex may have been very different from the earliest RCs, presumably much more sophisticated and more structurally complex. This ancestral complex, having a clear relationship to the modern RCs, will be called the Ur-reaction center. The properties of this Ur-RC are discussed in this section.
Figure 13.1 summarizes the general structural features ofknown modern RCs, especially with respect to the protein subunits. The Ur-RC was a dimeric integral membrane protein. It almost certainly was a homodimer, such as is found in the modern heliobacteria and GSB. Three independent gene duplication and divergence events have given rise to the heterdimeric RCs found in most modern RCs; the simplest conclusion is that the form of the Ur-RC was a homodimer. Whether an even earlier monomeric RC ever existed is uncertain, as the dimeric nature of all known modern RCs is essential to their function. Perhaps the RC only became a functional electron transfer complex when it dimerized, having some other function when it was a monomer. This would represent a case of gene recruitment, in that the more ancient protein, whose gene was recruited to become the reaction center, had a different function. Currently, there are no good candidates for what might have been this precursor. Meyer  and Xiong and Bauer  have suggested that it might have been the cytochrome b subunit of the cytochrome bc1 complex (see Chapter 7) , but detailed structural comparisons do not reveal any significant similarity in structure, so this now seems unlikely . The core structure of the RC has been remarkably well conserved during the course of evolution from the Ur) RC to the present day. This is shown in Figure 13.5) in which the known structures of ten protein chains from five RC complexes are superimposed. However, it is important to keep in mind that there are no structures for the RCs of three of the groups of photosynthetic organisms, the GSB, the FAP, and the heliobacteria, which are all among the simplest RCs in terms of protein composition and, therefore, possibly closer to the Ur-RC. There is still a long way to go to accumulate a robust dataset from which to draw firm conclusions.
The Ur-RC probably had a bidirectional electron transfer chain, using chlorophyll-type pigments, although likely not the exact pigments that we know today. These pigments are clearly the product of a long evolutionary process that has optimized them for light absorption and efficient electron transfer properties. There is every reason to expect that the impressive quantum efficiencies of close to one for modern RCs developed over a long time, and that earlier complexes were significantly less efficient. As Erasmus said, "In the land of the blind, the one-eyed man is king," so even an inefficient light-storage system is better than none at all. Evolution subsequently relentlessly improved the quantum efficiency so that there remains little more room for improvement. The same cannot be said for the energy efficiency of photochemistry, which is significantly less -in most cases, it is only a few tenths of percent. While the dramatic difference in the evolutionary optimization of quantum efficiency, but not energy efficiency, is not well understood, it probably at least in part represents some fundamental limits imposed by the second law of thermodynamics.
Could the similar structures of the reaction centers shown in Figure 13.5 have arisen by convergent evolution? Convergent evolution occurs when organisms independently evolve similar traits or structures, rather than descending from a common ancestor, called divergent evolution. Many clear cases of convergent evolution are known, for example, wing development in insects and birds. However, when they can be seen at an increasing resolution, especially down to the molecular level, their independent origin usually becomes readily apparent. Relatively few cases of convergent evolution at the molecular level are known and these are readily recognizable. The classic example is the apparently independent evolution of the active site catalytic triad of the two broad classes of serine protease enzymes found in bacteria and eukaryotes [109, 110]. Here, the active site geometries of the critical triad of amino acids that form the active site are similar, but the protein folds that position the active site residues are entirely distinct. The proteins have no apparent structural or sequence similarity, with the active site residues coming from different parts of the primary sequence in the two different classes of proteases. In this case, it appears that the chemistry of the reaction that the enzymes
Figure 13.5 (Left) Unrooted phylogenetic tree constructed using RMSDs derived from structural alignments directly as proxies for evolutionary distances. (B) Unrooted Neighbor-Joining phylogenetic tree of photosynthetic reaction centers based on a sequence alignment derived from the structural alignments shown at the bottom right. The red stars represent inferred gene duplication events. The colored boundaries enclose proteins sharing a particular percentage of similarity. (Bottom right) Structural alignments of all photosynthetic reaction center proteins: a-proteobacteria: Rhodobacter sphaeroides (1AIJ), L, M chains; Rhodopseudomonas viridis (1DXR), L, M chains; Thermochromatium tepidum (1EYS), L, M chains; Cyanobacteria: Thermosynechococcus elongatus (1S5L), D1, D2 chains of photosystem II; and Synechococcus elongatus (1JB0), A1, A2 chains of photosystem I. The 6 N-terminal helices that constitute the antenna domain of the photosystem I complex are not shown but were included in the data set used for alignment. The unaligned thread-like portions on the top and the bottom are the loops outside the membranes, joining the transmembrane helices. The left figure shows a front view of the 10 overlaid structures, whereas the right figure shows the side view of the same complexes rotated by 90 (Top right) Unrooted Neighbor-Joining phylogenetic tree based on an extensive set of sequences of photosynthetic reaction centers. Only the C-terminal electron transfer domains of the Type I reaction centers were used in the analysis. The red stars represent inferred gene duplication events. The blue colored region represents sequence space of reaction centers either known or inferred to have a homodimeric core protein structure, whereas all others have a heterodimeric structure. The red colored region represents sequence space of those reaction centers that evolve oxygen. The dashed line indicates a development of oxygen evolution capability that is well after the gene duplication event that led to a heterodimeric reaction center, whereas the solid line indicates an earlier development of oxygen evolution capability. Figures from Sadekar et al., Mol. Biol. Evol, 2006 , with permission.
carry out has dictated a narrow range of potential structures that will do the job; two different evolutionary paths have led to the same cofactor arrangement, but most other aspects of the proteins are entirely different.
In sharp contrast, the RC structures have all the hallmarks of purely divergent evolution - they have not only the same general structure at the level of the protein fold and cofactor placement, but they also have conservation of the order of cofactor binding residues in the primary sequence. One feature that does, however, seem to be describable as convergent is the gene duplication events that produced heterodimeric RCs from homodimeric ones. All evidence suggests that this took place on at least two, and, more likely, three, separate occasions. The gene duplication that gave rise to the heterodimer in PSI is clearly the most recent of these events, as both the structure of the heterodimer PSI subunits and the primary sequences of the core PSI proteins are very similar. Also, as described above, current evidence suggests that electron transfer does go down both sides of PSI, so even its function is closer to the homodimeric case than the type-II RCs, which appear to have a very strong preference for one electron transfer pathway. Presently, no sound understanding exists of the evolutionary pressures that gave rise to heterodimeric reaction centers, especially in PSI where electrons from the two pathways converge at the FX center.
An essential aspect of the Ur- RC that has been much discussed is whether it had 5 or 11 TMH in each half of the protein core. Modern RCs and photosystems with both arrangements are known, and it is not clear which model is the ancestral and which is the derived trait. The last 5 TMH of the 11 helix type-I photosystems are clearly functionally and structurally homologous with the 5 helix type-II RCs. The six N -terminal TMH of the type - I photosystems are also similar to the core antenna proteins in PSII, but nothing similar is found in any of the anaerobic type-II RCs. There are basically two different schools of thought on this question, which can be summarized into the representations of 11 - 6 = 5 or 5 + 6= 11. The 11 -6 = 5 scenario imagines that the Ur-RC had 11 TMH with the first 6 functioning as an antenna and the last 5 as the electron transfer domain, much as in modern PSI. A putative gene fission event separated the two domains into separate proteins. The antenna domain was retained as the CP43 and CP47 core antenna proteins in PSII, but was lost in the anoxygenic RCs, leaving a different type of core antenna to take its place. The 5 + 6 = 11 scenario imagines that the Ur-RC had 5 TMH and a gene fusion of the RC domain with the separately evolved 6 helix antenna domain gave rise to the 11 TMH complexes found in type-I photosystems, while the type-II RCs are closer to the ancestral state. Both scenarios have been extensively discussed by numerous authors [111-119].
There is one significant argument in favor of the 11-6 = 5 scenario. In the PSI and PSII, the antenna domains are remarkably similar to each other, both in terms of overall structure and positions of some of the pigments. Evolutionary analysis of the antenna domains of PSI, PSII, heliobacteria, and GSB show trees with similar topology to those derived from the core electron transfer domains of the RCs - 20, 118] . This suggests that the antenna domains have followed a similar evolutionary trajectory as the electron transfer domain. This, in turn, seems most consistent with the scenario that has these 2 domains developing together in an 11 TMH complex. This further suggests that the 11 TMH complex is the more ancestral version. However, it is important not to take this argument too seriously, as the sequence identity in these comparisons is extremely low and could be confounded by tree construction artifacts such as long- branch attraction.
If the Ur-RC was already a photosystem and had 11 TMH, as many authors have suggested and the above argument supports, then a difficult question arises regarding whether or not that complex was, itself, derived from a fusion of an even more ancient antenna complex with an electron transfer complex. This is well beyond the scope of current knowledge, and may well be lost forever in the long evolutionary development of photosynthesis from the earliest systems to the Ur-RC.
Do all the antenna domains of the type - I photosystems have similar antenna pigment arrangements? How do these relate to the antenna pigments in PSII? PSI has, by far, the most pigments of any of the RCs, ~100 vs 14 for the GSB RC, ~35 in heliobacteria, and ~36 in PSII. In addition, PSI and PSII have 26 Chl molecules in very similar positions in the antenna domain - 120]. although PSI has a significantly larger number of pigments, especially in the region near to the electron transfer domain. Fyfe et al. - 121] have carried out modeling of all the type-I photosystems based on the structure of PSI. Sequence comparisons and measurements of the pigment contents of these photosystems led these authors to propose that 14 BChl in the GSB RCs, and somewhat more in the heliobacterial complex, bind to conserved ligands, principally histidines, in the N-terminal domains. So, there seems to be a core structure of the antenna domains with a modest number of pigments, possibly in similar positions that are common to all type-I photosystems and PSII. However, until such time as detailed structures of the 11 helix photosystems from GSB and heliobacteria are available to compare to those from PSI and PSII, all of these comparisons will remain uncertain and should not be taken too seriously.
Was this article helpful?