Evolutionary Transition States of the Syntrophy Hypothesis and Hydrogen Hypothesis in Review: A work on Eukaryogenesis.
[Eukaryogenesis is] arguably one of the most important events in the history of life, after the origin of life itself.
— Daniel Mills, Ludwig-Maximilians-Universität München
Foreword
I met Ayden a few months before planning his foray into writing this piece. While serving as his mentor, I found that he is a voracious reader, eager to understand the concepts placed before him in the metabolic puzzles of life. His attention to the detail of papers which he has consumed and discussed below takes a deep critique of the greatest arguments regarding this mystery. Ayden’s work has placed two phenomenal arguments regarding the origin of eukaryotic life and identified real challenges in their comparison yet to be answered by current theoreticians. With the work that I’ve seen him complete, I am confident that with the right guidance and mentorship from true experts, his interests and passions will carry him into an amazing career future scientist.
- Jonathan Felix, PhD Candidate at the Keck Graduate Institute for Applied Life Sciences, https://www.linkedin.com/in/jon-felix/
A Word From The Author
A while ago, I picked up a copy of the 2018 Edition of Nick Lane’s Power, sex, suicide Mitochondria and the meaning of life. And it was a fascinating read that gave me a new perspective on mitochondria in general and eukaryogenesis — the development of the eukaryotic cell — in particular. Specifically, he introduced me to the H2 Hypothesis, which directly contradicted what I had learned in the classroom (the Archezoa Hypothesis–in which a primitive amitochondriate eukaryote acquired an alpha-proteobacterial endosymbiont, which eventually became the mitochondrion, via phagocytosis).
With my interest in eukaryogenesis well and truly piqued, I dove deeper into the scientific literature discussing this topic. One paper I found particularly interesting was Purificación López-Garcìa and David Moreira’s The Syntrophy hypothesis for the origin of eukaryotes revisited, which put forward a different theory in order to explain eukaryogenesis, the HS Syntrophy Hypothesis. These 2 theories are part of a broader debate regarding the origin of the modern eukaryotic cell that forms the basis of all true multicellular life on earth, one that still rages on to this day. Unwilling to be a bystander in this debate, I decided to conduct my own literature review on this topic, which will be the subject of this series.
As a young conscript whose highest educational certification consists of a GCE ‘A’ Level certificate, I am extremely grateful for the mentorship of Mr Jonathan Felix. Through our discussions, he taught me many invaluable lessons in searching for scientific literature, in cross-referencing multiple works against each other, and in composing formal scientific reports. He also instilled in me a confidence to trust my own abilities and take my own stance in the many debates that this project will touch on. Nevertheless, the conclusions I drew and thus any mistakes made in this paper are my own.
All in all, given that this is still my first literature review on this topic, I am sure that a fair number of my criticisms of both hypotheses are off-target. As such, I would like to emphasise the fact that this series is not intended as an attack on any of the great scientists whose work I critique. Rather, it serves to document the start of my journey into scientific research, and to introduce readers to the scientific debate surrounding this seminal event in the history of life. Furthermore, as this work is adapted from a currently unpublished manuscript written with the more rigid structure and specialised language of academic work, I must apologise in advance if one finds this work a difficult read.
Abstract
Among evolutionary problems, eukaryogenesis is one of the most interesting, with many published theories proposing different evolutionary transition states in an effort to explain how it may have occurred in a way that aligns with current empirical evidence 1. Among the most detailed and well known theories is the Hydrogen Hypothesis, first published in 1998, updated in 2015 and in the meantime popularised in the public conscience in no small part by Nick Lane’s 2005 book Power, Sex, Suicide: Mitochondria and the meaning of life. While the Hydrogen Hypothesis played a significant part in popularising symbiogenetic theories for eukaryogenesis, it is not the only such theory out there. Another equally detailed theory is the Syntrophy Hypothesis, also published in 1998. And while the 1998 Syntrophy Hypothesis arguably less convincing than the 2015 H2 Hypothesis 1, its 2020 revision does warrant a reevaluation of its strength vis a vis the H2 Hypothesis. Hence, using 1, the author has specified 9 questions that the hypotheses must answer, with 6 regarding historicals (Metabolism of host, Metabolism of symbiont, Initial relationship, Mechanism of inclusion, Early selective advantage, Vertical transmission), the seventh regarding eukaryotic singularity, the eight regarding endomembrane development, and the ninth regarding the implications of Archaeal and Bacterial contributions to the chimeric eukaryotic genome on timing of inclusion. The author then compared how well the evolutionary transition states proposed by both hypotheses answers these questions, and how well they align with contemporary primary research. In conclusion, the HS Syntrophy Hypothesis does appear stronger in its explanation for 2 of the questions– the implications of the eukaryotic Chimeric genome on timing of inclusion and the origin of the eukaryotic cell surface membrane, and the evolution of the nuclear membrane. However, it fails to offer a plausible explanation for eukaryotic singularity, which the Hydrogen Hypothesis does. With regard to the remaining 6 questions–metabolism of the Asgard archaea and bacterial symbionts, Initial relationship, Early selective advantage, vertical transmission and the mechanism of inclusion, both hypotheses offer equally convincing–or unconvincing–explanations. Thus, it seems that there are more questions that the HS Syntrophy Hypothesis provides convincing answers to compared to the H2 Hypothesis.
Section 1: The H2 Hypothesis
The Hydrogen Hypothesis 2, in its updated and expanded form 3 , posits that eukaryotes arose through the symbiotic association of a hydrogen dependent, strictly anaerobic and probably autotrophic 2–4 host and a metabolically versatile, heterotrophic symbiont able to perform both cellular aerobic respiration and anaerobic H2 and CO2 producing fermentations 2,3. Phylogenetic results strongly support an alpha- proteobacterial ancestor of the symbiont 1, while more recently Asgard archaea was suggested as the closest relative of the host as its composite genome encodes most homologues of eukaryotic signature proteins5.
The initial relationship between host and symbiont at the start of mitochondrial symbiosis was posited to be one of anaerobic syntrophy, in which H2 and CO2 produced by the endosymbiont was taken up by the host for carbon fixation 2,3 . (The host and symbiont must meet in anaerobic environments where CO2 and geological H2 are abundant so as to ensure that the host is viable from the start 2). This enables the pair to survive in and thus colonise environments in which geological hydrogen is unavailable, so long as the host and symbiont remain together. Should separation occur, the host starves immediately 2 . Thus, hosts with cell shapes of large surface area that surround the endosymbiont (in order to ensure that the hosts and symbiont adhere to each other) are more likely to survive and thus are selected for 3. Another benefit of this is to increase the surface area in contact with each other so that more H2, CO2 can be supplied from the symbiont to the host 2 . (Due to being limited to the limited surface area of the bioenergetic cell surface membrane, Martin et al. assume that the host cell is unlikely to have been able to produce the requisite chemical energy for phagocytosis 2,4,6.)
With gradual engulfment of the endosymbiont by the host, the surface area of the symbiont in contact with the surroundings is reduced, greatly limiting the rate at which the reduced organic substrates for H2 generation can diffuse 2,3,6 into the symbiont. (At this stage, the host lacks transport proteins capable of allowing fermentable organic compounds across its cell surface membrane). Thus, 2 changes must now occur in the hosts for symbiosis to remain viable. Firstly, hosts would acquire such proteins via gene transfer from the symbiont to the host 2,3,6. This assumes that the bacterial transport proteins encoded by the symbiont’s transferred genes are functional in the archeal membrane, as is the case for haloarchaea 7. Secondly, given the fact that the autotrophic host’s carbon metabolism is specialised towards anabolic pathways 2,3,6 , the host now needs to acquire the ability to oxidise fermentable organic substances to intermediates that the symbiont can further oxidise. This is most likely achieved via the transfer of genes for the symbiont’s heterotrophic carbon metabolism–i.e Glycolysis–to the host’s chromosomes 3. In other words, glycolysis would be transferred out of the symbiont to the host. The net effects of the above 2 changes would be to allow the reduced organic compounds that are imported (and partially broken down) by the host to be directed to the symbiont and confer heterotrophy upon the host. Most importantly, it has now eliminated the host’ dependence on hydrogen provided by the endosymbiont, since the host can now meet both its carbon and energy demands from imported organic substances 2. Further, the host is also no longer confined to anaerobic environments given the symbiont’s ability to carry out cellular aerobic respiration by virtue if its metabolic diversity 1–3 , hence the additional ATP produced via the Krebs cycle and oxidative phosphorylation by the symbiont in the presence of O2 would now confer a selective advantage upon the host. (ATP pumps were then evolved soon after endosymbiosis 6 to enable the symbiont to supply the host with chemical energy in the form of ATP) 2,6.
At this stage of endosymbiosis, bacterial vesicles (phospholipid bilayer) that bud off from the endosymbiont fuse with the archeal membrane (made up of isoprene lipids) to form a chimeric cell surface membrane 3. This is supported by an experimental study that indicates that mixed liposomes of bacterial and archeal lipids are stable 8. Further extensive gene transfers from the facultatively aerobic bacterial endosymbiont to the host 3 now occur to replace many of the indigenous archeal genes encoding archeal pathways with their bacterial counterparts 3. In particular, genes encoding bacterial lipid synthesis are transferred from the symbiont to the host 3,6, allowing the host to synthesise bacterial lipids such as phospholipids to further speed up the conversion of the host cell surface membrane from consisting of an isoprene phospholipid monolayer (characteristic of archaea) to one consisting of a bacterial phospholipid bilayer 6. (The archaeal lipid synthesis pathway has been retained by eukaryotes, however, for the synthesis of isoprenes in general 3.) Thus the host’s archeal cell surface membrane is now replaced by a bacterial cell surface membrane characteristic of today’s eukaryotes.
To explain the development of the nuclear membrane, Martin et al. endorse 3 the Splicing Hypothesis, which contends that the formation of the nucleus is precipitated by such gene transfers-in particular, the transfer of self-splicing Group II introns present in the prokaryotic endosymbiont’s genome 9,10 to the host’s genome. Such introns are mobile, can insert themselves into many sites in the host’s genome 3,9,11 and encode maturases that help catalyse self-splicing 12. Given the similarity of their splicing mechanism to that of spliceosomal intron removal 10,13, Group II introns have long been viewed as the precursors of both eukaryotic spliceosomal introns and their associated snRNA spliceosomes 3,10. In particular, it has been proposed that mutations in some of the Group II introns now present across the host’s genome could have resulted in no functional maturase being encoded, causing such introns to no longer be self splicing 10. Thus, in order to ensure that only the coding exons are transcribed, the host must evolve an endonuclease (i.e a protospliceosome) to excise such introns. Martin et al. proposes that the SnRNA (of the protospliceosome) was likely encoded by sections of the Group II introns, while other novel accessory proteins were recruited to form a primitive protosplicesome 10. However, this now presents an issue-ribosomes translate the mRNA template far faster than the protospliceosome can excise all introns 3,10. And whereas in group II introns maturase synthesised in the first ribosomal passage block the 5’ end of the mRNA until the introns are removed 3, spliceosomal introns lack an equivalent protein that could block ribosomes from translating before splicing is complete. The net result is that both mRNA introns (that have yet to be spliced) and exons are both transcribed to form non functional protein gene products at hundreds of gene loci simultaneously 3, which would most likely be fatal for the organism. To correct this, the near simultaneous transcription and translation characteristic of prokaryotes must be abandoned. Instead, the initial pressure that conferred upon organisms with a nuclear membrane the selective advantage was the need to physically separate active chromatin undergoing transcription and post transcriptional modification (especially the slow splicing process) and active ribosomes which would carry out the rapid translation process 3, which would ensure that splicing is complete before translation could occur.
A mechanism for the formation of the nuclear membrane is presented by Nick Lane, a proponent of the Hydrogen Hypothesis 6. Lane proposes that while genes encoding proteins for bacterial lipid synthesis were acquired by host’s genome via gene transfer from the symbiont, the mechanism for targeting such proteins such that the phospholipids end up in the cell surface membrane was not transferred from the symbiont to the host (or the symbiont’s version was incompatible with the host) 6. Thus, the formed bacterial phospholipids remain around the nucleus where they form hydrophobic interactions with each other, precipitating as vesicles that coalesce around the nucleus 6. (Pore complexes were also evolved to ensure that mature mRNA could exit the nucleus after processing 10).
The acquisition of mitochondria freed the ancestral eukaryote from the constraints that bioenergetic membranes impose on energy generation 1,3,6. Prokaryotes are limited to the surface area of their cell surface membranes to embed respiratory enzymes 6. And as prokaryote size rises, their surface area to volume ratio and thus chemical energy (ATP) generating capacity relative to their size falls 6. This becomes critical once one considers the fact that gene expression consumes most of a cell’s energy budget 14, meaning that the number of genes a prokaryote can express is fundamentally limited, making them relatively small and phenotypically simple compared to eukaryotes 6,14. Furthermore, this lack of energy for the synthesis and maintenance of large genomes also selects for prokaryotes with as little as possible non-coding or regulatory DNA sequences, to the point that multiple structural genes are grouped under a single promoter and operator in operons that are rarely separated from each other by non-coding regions 14. Moreover, the median length of prokaryotic proteins is also kept short to minimise the number of DNA nucleotides needed to encode proteins 14. This results in prokaryotes having a high gene density of 500–1000 genes per Mb 14.Even so, energetic limitations cap bacterial genomes at around 9000 genes, with prokaryotes that discard unnecessary genes being selected for 6. By contrast, eukaryotes are freed from this constraint as ATP generation could be increased by simply stuffing the cell full of more mitochondria, whose bioenergetic membranes are intracellular 6. Indeed, according to Lane and Martin’s calculations, a prokaryote with a genome, volume and shape similar to an average unicellular eukaryote (protozoan) would have around 230000 times less energy available per gene compared to said average eukaryote 14. In other words, at a similar energy per gene, eukaryotes can energetically support a genome more than 200 000 times larger than that of Prokaryotes 14. Indeed, this massive increase in energy enabled a large intermediate expansion of the eukaryotic genome, allowing the cell to experiment with new gene families 1,15, resulting in Eukaryotes eventually evolving to encode around 3000 novel gene families 16,17 that conferred upon the host many of the hallmarks of eukaryotic complexity-a wide array of novel protein folds, protein interactions and regulatory cascades 14.
In particular, the evolution of new protein folds in eukaryotes was the most intense period of gene invention since the evolution of life itself 18. Eukaryotes invented 5 times as many protein folds as bacteria, and 10 times as many as archaea 17. Moreover, median protein length is 30% greater in eukaryotes than prokaryotes 19. The ability of eukaryotes to energetically support large genomes also enabled them to reduce their gene density to around 12 Genes per Mb 14, freeing up large sections of DNA to harbour large numbers of regulatory (to the point that almost every eukaryotic gene is afforded its own promoter) and non coding regions, along with sequences that encode regulatory microRNAs 14. All in all, the energy from the endosymbiont was necessary in order to allow eukaryotes to achieve a level of complexity unmatched by any other domain of life.
Lane and Martin are aware 14 of the fact that many prokaryotes, such as cyanobacterium, nitrifying bacteria and large sulfur bacteria do have extensively invaginated inner membranes on which ATPsynthases can be embedded 20,21. However, they contend that they lack the necessary genome organisation to exploit the benefits of intracellular bioenergetic membranes. To explain this, Martin et al. cite the Co-location for rRedox Regulation Hypothesis, which contends that there exists a primary subset of genes (especially those encoding electron carriers 22) whose expression is regulated by the redox state of their gene products (or the electron carriers with which their gene products interact) 23,24. This ensures that rapid environmental changes (, such as a change in partial pressure of O2) leading to changes in the position of the redox equilibrium and thus membrane potential of mitochondria can be dealt with in order to restore the optimal membrane potential 23,24. (Allowing the membrane potential to fall too drastically would result in a collapse in energy charge, blocking active transport of protons across the cell membrane, and ultimately a rise in free radical leakage 14 to form reactive oxygen species 6, culminating in apoptosis 25. On the other hand, allowing the membrane potential to rise too far may result in the electron transport chain becoming a significant producer of reactive oxygen species in its own right 26.) And thus it is essential that such critical genes be co-located in close proximity to the electron transport chains 6,23,24. In eukaryotes, the small mitochondrial genome encoding such genes (while other genes are transferred to the nucleus in low copy number 14) serves this function 6,14,23,24. By contrast, as we shall soon see, this is where the organisation of prokaryotic genomes limit them. In particular, Lane and Martin argue that the lack of many giant prokaryotic plasmids (associated with the prokaryote’s bioenergetic membranes) that encode the aforementioned critical redox regulated genes is a major limitation 14. (prokaryotes tend to either have many small plasmids or a few large ones 27 ). Instead,the few large prokaryotes like Thiomargarita namibiensis and Thiomargarita magnifica get around this via extreme polyploidy, distributing many copies of their genomes across the cell 14,21. (Whereas most prokaryotes are limited in size so that their genome can be in close proximity to their bioenergetic membranes). And thus increased energy generation granted by the complex internal membrane structures of such bacteria 21 is wasted replicating thousands of copies of a relatively small genomes rather than enabling the prokaryote to support a large genome encoding many genes 14, thus precluding the possibility of a prokaryote with similar complexity to the average eukaryote 14.
Section 2: The HS Syntrophy Hypothesis
2.0 A Note on Organisation
In describing their own hypothesis, Lopez Garcia and Moreira have opted to divide eukaryogenesis into 5 stages 28. However, given the tightly linked nature of stages 3 and 4, the author of this paper has opted to describe them together in section 2.3. Thus, the subsequent content is organised as follows.
2.1: Initial context
The Syntrophy Hypothesis, in its updated HS form, posits that eukaryotes arose from 2 symbiosis events. The initial symbiotic association is between a sulfate reducing delta-proteobacterial host and a heterotrophic, hydrogen producing Asgard archeal symbiont (the precursor of the nucleus). This is followed by symbiosis between the initial pair of cells and a metabolically versatile, mixotrophic (or possibly even photosynthetic), alpha-proteobacterium (mitochondria precursor) capable of aerobic respiration, sulfide oxidation, among others 28. While Lopez Garcia and Moreia’s hypothesis acknowledges the alpha-proteobacterial origin of mitochondria 1 and the contribution of Asgard archea to the eukaryotic genome 5, it further holds that the presence of a significant number of bacterial genes in the eukaryotic genome (outnumbering archeal genes)– of which a significant proportion are non alpha-proteobacterial 29, with most of the non alpha-proteobacterial genes in eukaryotes being delta-proteobacterial in nature 29 — supports the presence of a third, delta-proteobacterial symbiont.
The fact that the alpha-proteobacterial symbiont has evolved the ability to respire aerobically suggests the presence of aerobic environments near the site of eukaryogenesis 28, while the strictly anaerobic nature 30 of Asgard archaea, conversely, points to an anaerobic habitat 28. Thus, Lopez-Garcia and Moreia propose the phototrophic microbial mats that dominate shallow aquatic and terrestrial habitats-redox transition environments where anaerobic, microoxic and aerobic zones are in close proximity-as the initial environment for endosymbiosis 28. Such mats can broadly be divided into 5 layers, as seen in Figure 2.1. Whereas the upper layers (a) to (c) are at least oxic if not microoxic, the deeper layers (c) and (d) are anoxic (anaerobic) and rely on syntrophy 31 to break down the carbon fixed by the organisms of the upper layers (a) and (b). Indeed, metabolically versatile delta-proteobacteria have been observed to either produce hydrogen in syntrophy with methanogens (that used that hydrogen for carbon fixation to form methane) or consume hydrogen to reduce sulfates in synthropy with methanotrophic archaea 32.
2.2: Initial Integration Stage
Thus, this supports the hypothesis, which starts with the initial facultative symbiosis stage. At the outset of this stage, there exists an initial syntrophy between a sulfur reducing, hydrogen requiring, myxobacterial like delta-proteobacteria and a hydrogen producing, heterotrophic Asgard archaea 28 within the deeper, anoxic layers of the microbial mat. The Asgard archaea takes up fermentable complex organic compounds from the surroundings, which it then oxidises, producing hydrogen which it supplies to the delta-proteobacteria which uses it to reduce sulfate 28.
This initial facultative symbiosis was then formalised by the internalisation of the Asgard symbiont in the first integration stage 28. In the original Syntrophy Hypothesis, this would be the eventual result of extensive membrane to membrane contact between the delta-proteobacteria and the Ashard symbiont developed to facilitate metabolite exchange 33. The bacterial cell would have ended up encircling the Asgard symbiont and fusing its membrane around it. This would not have been considered a true endosymbiosis as the symbiont’s cytoplasm would not have been incorporated into the host until much later, and was proposed as such as it was then believed that phagocytosis did not occur in prokaryotes 33. However, the subsequent discovery endosymbionts within prokaryotes 34–39 has challenged this aspect of the hypothesis, and as such Lopez-Garcia and Moreia no longer specify a mechanism for endosymbiosis, only that it occurred to result in the full incorporation of the Asgard within the host’s cytoplasm as an obligate symbiont 28. Gene transfer between host to the symbiont then occurred 28,33,40, initially in both directions. However, eventually, gene transfer from the host to the Archaea eventually won out 33. The authors propose that this could happen due to a random transfer of a few essential genes from the host (in which genome these genes were subsequently lost) to the Asgard 28,33, paving the way for the progressive centralisation of the genome within the archaea 28 and the eventual formation of the nucleus. The Asgard’s own archaeal bioenergetic membrane is retained as a separate metabolic compartment 28. The evolution of the endomembrane system also begins at this stage. Since complex fermentable organic substances taken up from the surroundings are hydrolysed in the delta-proteobacteria’s periplasm before the simpler products (along with amino acids) are supplied to the Asgard archaea, the invagination of the inner host membrane around the Asgard archaea such that a layer of periplasm now surrounds the Asgard would facilitate transfer of simpler organics from the host to the symbiont and is thus selected for 28. (While there are currently no known delta-proteobacteria with endomembranes, Lopez Garcia and Moreia contend that their diversity is far from fully explored and that they have membrane remodelling potential 28. Indeed, developed cytoskeletons necessary for membrane modelling are present in delta-proteobacteria (myxobacteria) as part of protruding membrane tubes that connect cells 41.)
Concurrently with the first integration stage (that entails the endosymbiosis of Asgard archaea, gene transfer from the host to the symbiont and the opening stages of endomembrane development), the entire consortium (consisting of the delta-proteobacteria and Asgard archeal symbiont) moves up from the deeper anoxic layers to the shallower, transition between the anoxic and oxic layer 33, where it came into contact with a second, metabolically versatile alpha-proteobacterium 28. While the authors argue that they are likely photosynthetic or mixotrophic (since H2S dependent anoxygenic bacteria dominate these sections of the microbial mat 42,43), it is nonetheless assumed that they are also capable of aerobic respiration and sulfide oxidation. And the latter is initially more important. In particular, the alpha-proteobacteria takes up sulfides from the consortium and then oxidises it to sulfate which it then supplies to the consortium. This sequence of events is favoured by the authors as they contend that it is more in line with genomic evidence suggesting a late acquisition of mitochondria 29.
2.3: Second and Advanced Integration Stages
This is followed by the second integration stage, which entails the internalisation of the alpha-proteobacteria, which, being capable of the far more efficient process of aerobic respiration (compared to the anaerobic archeal metabolism and bacterial sulfate reduction), would not only enable the tripartite consortium to colonise aerobic environments relying on aerobic respiration alone 28,33, but would also render anaerobic archeal metabolism and bacterial sulfate reduction (and the bioenergetic membranes that support them) redundant, leading to their gradual loss via natural selection 28,33. (ATP pumps were then evolved soon after endosymbiosis 15 to enable the symbiont to supply the host with chemical energy in the form of ATP). Thus, by the time of the advanced integration stage, bacterial sulfate metabolism and archaeal energy metabolism will have been fully lost 28.
Meanwhile, endomembrane development proceeds apace during the second and advanced integration stages as more genes are transferred from the delta-proteobacteria (alpha-proteobacteria endosymbiosis is still ongoing) to the archaea, including those encoding hydrolytic enzymes to be secreted into the periplasm to hydrolyse complex organics 28. This selected for the development of a transport system to allow such bacterial hydrolytic enzymes through the archeal membrane in order to eventually reach the periplasm without entering the cytosol and degrading cytoplasmic components 28. First, bacterial proteins may have been inserted into the archaeal membrane to allow such bacterial gene products through the archaeal membrane 28. Eventually, as genes encoding complex, essential bacterial gene products were transferred to the Asgard, pore-complexes on the bacterial membrane which bounds the periplasm to allow the export of more complex gene products would have to be developed. Such pore-complexes would likely have been able to communicate with the transport (channel) proteins in the archaeal membrane. (Given the fact that Archaea can establish intracellular cytoplasmic bridges and fuse 28,44, while Myxobacterial delta-proteobacteria can undergo membrane fusion 45,46 or develop contact dependent coordinated gliding via junctional pore complexes 47,48, it is logical to assume that such communicating pores, as described previously, would be entirely possible to develop.) In other words, the endomembrane system, derived from the host’s inner membrane, now serves to transport gene products (proteins) synthesised in the Asgard throughout the cell. Ribosomes then began to congregate around these proto-nuclear pores in order to enable the synthesised gene products to be quickly exported 28. Thus, the near simultaneous transcription and translation characteristic of prokaryotes begins to decouple, with transcription taking place towards the interior of the Asgard in its nucleoid region, whereas translation takes place around the nuclear pores 28. Moreover, the development of the cytoskeleton accompanies the emergence of the endomembrane system 33, and in turn enables the development of mitosis and meiosis by the end of eukaryogenesis 33.
2.4: LECA Stage
And thus the stage is set for the final series of events that would give rise to the Last Eukaryotic common ancestor. 2 events now happen at around the same time. Firstly, fusion of the delta-proteobacteria derived endomembrane system leads to the full internalisation of the periplasm (which eventually becomes the cisternae lumen) 28, while some of the digestive functions once performed by the periphery are transferred to independent vesicles 28. As mitochondrial endosymbiosis completes along with the loss of the delta-proteobacteria and Asgard’s bioenergetic membranes, the archaeal membrane now provides no advantage while coming with a large resource cost. Thus, cells that retain it are selected against, leading to the loss of the Archeal membrane 28. (The bacterial membrane around the former Asgard archaea is now the proto-nuclear membrane). This coincided with a period of rapid genome evolution 28,33 involving gene transfer from the (alpha-proteobacterial) endosymbiont (and what remains of the delta-proteobacteria’s genome) to the former archaea’s (now a proto nucleus) genome. Duplication of genes and their reshuffling led to further genetic changes 28. Moreover, Lopez Garcia and Moreia now cite the Splicing Hypothesis described earlier. Mobile group II self splicing introns from the alpha-proteobacteria endosymbiont 10,28 insert themselves into the proto-nuclear genome and then fragment, giving rise to spliceosomal introns and spliceosome encoding genes 10. The far faster rate at which ribosomes translate mRNA compared to the rate of splicing 10 makes the strict decoupling of transcription and translation paramount 10,28 to ensure that only introns are translated to form functional protein gene products. Thus, this creates the selection pressure for further decoupling of transcription and translation beyond the spatial separation within the former Asgard achieved previously. Thus, the authors propose that the former Asgard ribosomes eventually migrate out of the proto-nucleus into the cytoplasm via the proto-nuclear pore complexes, eventually associating along the endomembrane system (future endoplasmic reticulum), where translation now takes place 28. The proto-nuclear membrane’s primary function becomes one of enforcing this decoupling of transcription and translation 28. The genes encoding ribosomal RNA, however, remain in the proto-nucleus as the nucleolus, where it is transcribed to rRNA. Ribosomal proteins synthesised in the cytoplasm re-enter the nucleus in which they assemble with rRNA to form ribosomal subunits that are then exported from the nucleus. All proteins involved in ribosomal synthesis are of archaeal origin 49. And thus we get the last eukaryotic common ancestor, with an Asgard derived nucleus, genome and macromolecular synthesising machinery, a delta-proteobacteria derived membrane, cytoplasm and proteins involved in cell signalling and a mitochondria capable of aerobic respiration.
Section 3: Criteria for Comparison of Theories.
As outlined by Zachar and Szathmáry 1, a compelling hypothesis for the the incorporation of mitochondria, accompanied by eukaryogenesis, must be able to explain the following:
- Eukaryotic singularity: The fact that eukaryogenesis only occurred once, and no similarly complex sister group to eukaryotes has been observed.
- Absent intermediates: The fact that the signature eukaryotic features (endomembrane system, organelles, sex and syngamy) seemingly appeared simultaneously without any apparent intermediate transitional forms.
- Archaeal and Bacterial contributions to the chimeric eukaryotic genome
- Loss of membrane bioenergetics: The loss of prokaryotic ATP synthesis machinery embedded in the cell surface membrane must have occurred during eukaryogenesis.
- Inability of mitochondria to photosynthesize: Despite the fact that Alpha-proteobacteria precursors of mitochondria may have been capable of photosynthesis 1,28
- Origin of mitochondria related organelles: Either via divergent evolution from a metabolically versatile proto-mitochondria that retained various functions in specific lineages or via subsequent evolution of the proto-mitochondrion acquiring functions selected for by selection pressures found in a particular ecological niche1
- Metabolism of host
- Metabolism of symbiont
- Initial relationship: between the host and symbiont
- Mechanism of inclusion: Of the symbiont into the host
- Early selective advantage: of the initial symbiosis despite the initial lack of host ATP transporters and metabolic compartmentalisation necessary for the symbiont to immediately supply ATP to the host.
- Vertical transmission: At any point during eukaryogenesis the newly forming chimeric cell had to divide so that its selective advantage was heritable and cooperation could fixate 1 .
Criteria 1–6 deal with observations of modern eukaryotes that a theory must be able to explain, whereas criteria 7–12 deal with historical evolutionary transition stages that must be explained by any theory for eukaryogenesis 28. Given the focus of this paper on evolutionary transition states, the former will not be investigated in detail in this paper, and is instead the subject for future work. Nevertheless, the importance of eukaryotic singularity is such that the author has decided to consider it as well. Moreover, the Archaeal and Bacterial contributions to the chimeric eukaryotic genome will be considered when weighing the proposed timings of inclusion and nature of LECA’s cell surface membrane. Lastly, given that differences between each theory’s proposed metabolic pathways for the host and symbiont(s) are in no small part influenced by the fact that the Asgard archaea features prominently as the Hydrogen Hypothesis’s proposed host 3,4 while, on the other hand, serving as a symbiont in the HS Syntrophy Hypothesis, criteria 7 and 8 will be modified to consider the metabolism of the Asgard archaea and the other bacterium in the consortium respectively.
Furthermore, an additional evolutionary transition stage will be considered by the author.
13) Evolution of the nuclear membrane.
Of criteria 7–12, 9), 11) and 12) are considered dealt with equally well by both theories, whose answers to these questions are summarised in the table below.
Thus, section 4 of this paper would only deal with the following criteria to which the HS Syntrophy and Hydrogen Hypothesis provide contrasting answers:
Section 4: Points of Contention Between The Theories
4.0 Eukaryotic Singularity
The Hydrogen Hypothesis proposes that eukaryotes only evolved once due to the singular incorporation of mitochondria 1, which was solely responsible for providing the cell with the energetic surplus necessary for it to develop into a eukaryote 2,3,50. However, the energetic implications 51 of the Hydrogen Hypothesis being a mitochondria early 1 hypothesis weaken this argument.
Apart from the fact that, as argued in section 5, modern eukaryotes have a fairly modest energetic advantage over prokaryotes 51 compared to Lane and Martin’s claims 14, incorporating mitochondria early reduces that advantage even further. To summarise, Schavemaker and Muñoz-Gómez convincingly argue that an aerobic eukaryote’s energetic advantage over an aerobic prokaryote is solely due to the better genomic organisation a eukaryote enjoys 51, something that Lane and Martin indeed propose 14. However, the early incorporation of mitochondria necessarily entails its incorporation into a prokaryote, with its inefficient genomic organisation still intact 51. Thus, the energetic advantage of a hypothetical cell which acquired mitochondria early compared to an equivalent prokaryote (with ATP synthase and thus aerobic respiration) is eliminated 52 [see extended figure 5].
While such a eukaryote would likely still have an energetic advantage over a the prokaryote proposed in the Hydrogen Hypothesis (an Asgard hydrogen dependent autotroph), the above fact does not preclude a prokaryotic cell having enough energy to develop eukaryotic complexity 51. Thus, the energetic barrier to multiple eukaryogenesis events that Lane and Martin propose is undermined by the nature of their theory.
Moreover, their explanation does not preclude multiple acquisitions of mitochondria. If the fact that eukaryogenesis only happened once was due to the fact that mitochondria, needed to overcome the energetic barrier to eukaryogenesis, was only acquired once, then what prevented another cell line from subsequently acquiring mitochondria? 1 Lane and Martin have yet to provide an adequate answer, and thus the Hydrogen Hypothesis’s explanation of eukaryotic singularity is incomplete.
To be fair, it is currently unclear to the author of this paper if the HS Syntrophy Hypothesis, like its HM counterpart 1, provides an adequate explanation for eukaryotic singularity.
Thus, however flawed the H2 Hypothesis’s argument for eukaryotic singularity is, its mere existence vis a viz the lack of any compelling equivalent argument from the HS Syntrophy Hypothesis gives it an advantage.
4.1 Metabolism of Host and Symbionts.
4.1.1: The Asgard Archaea
Whereas both theories acknowledge the Asgard Archaea as a core member of the endosymbiotic relationship 2,3,28,33, they diverge regarding the role of the Asgard archaea. The Hydrogen Hypothesis insists on the Asgard archaea being a strictly anaerobic hydrogen consuming, carbon reducing (autotrophic) host 1–4 that enters into syntrophy with a hydrogen and carbon dioxide source. Indeed, Martin et al provide evidence indicating that Lokiacheron–at the time thought to be the closest ancestors of eukaryotes–possesses the hallmark pathway of hydrogen-dependent anaerobic autotrophs: the archeal tetrahydromethanopterin (H4MPT)-dependent acetyl-Coenzyme A pathway (Wood-Ljungdahl pathway) 4,30.
By contrast, the HS Syntrophy Hypothesis proposed that the Asgard archaea is a hydrogen producing, carbon oxidising (organoheterotrophic) symbiont 2,3. Hydrogen production provides the impetus for syntrophy with the hydrogen consuming delta-proteobacteria host, whereas the Asgard’s heterotrophy provides an initial selection pressure for endomembrane development (to supply the Asgard with simple organics) 28.
Indeed, there is more recent research in favour of this interpretation. A recent molecular phylogenetic analysis has indicated that Heimdallacheron (particularly the order Hodarchaeales) , not Lokiacheron, are the closest archaeal relatives 30,53. Furthermore, Eme, L., Tamarit, D., Caceres, E.F. et al’s ancestral modelling suggests that the Wood-Ljungdahl pathway was lost before the last common ancestor of heimdallarchea , indicating that Heimdallarchea and thus Hodarchaeales were heterotrophic fermenters 53. In particular, they are predicted to have a complete Embden-Meyerhof-Parnas pathway for glycolysis. Furthermore, Hodarchaeales are predicted to have a complete tricarboxylic acid cycle and an electron transport chain using nitrate as a terminal electron acceptor 53. In other words, the last archeal and eukaryotic common ancestor generated ATP via the oxidation of monosaccharides 53 using Nitrate as final electron acceptor , making it predominantly a heterotroph 53. Group 4 [NiFe]-Hydrogenases have also been identified in Hodarchaeales’s cell surface membranes, allowing it to convert protons to diatomic hydrogen (and vice versa) 53. Thus, while Eme, L., Tamarit, D., Caceres, E.F. et al concede that further analysis is necessary to determine if the [NiFe] Hydrogenase acted primarily to produce or use up hydrogen, it is thus at least conceivable that the last archaeal and eukaryotic common ancestor produced hydrogen while oxidising carbon.
It must be noted, however, that the loss of the Wood-Ljungdahl pathway in the last Archaea and Eukaryote common ancestor is not a universally agreed upon fact. Indeed, other molecular phylogenetic analyses have predicted (in addition to the heterotrophic pathways necessary to obtain carbon sources by degrading carbohydrates and peptides) a complete Wood-Ljungdahl pathway (both THF and THMPT dependent versions) in Heimdallarchaea 30, suggesting that the Asgard archaeal superphylum lives a mixotrophic lifestyle–potentially allowing the Heimdallarchael host to use hydrogen to litiautotrophically form reduced compounds (suiting the Hydrogen Hypothesis) or to produce hydrogen as a by-product of fermenting molecules with higher oxidation states 53 (suiting the Syntrophy Hypothesis).
Thus, while recent research seems to favour a heterotrophic 30,53 Heimdallarchaea as the last archaeal and eukaryotic common ancestor, favouring the Syntrophy Hypothesis , the lack of conclusive agreement regarding the loss of the Wood-Ljungdahl pathway means that the Hydrogen Hypothesis cannot be precluded yet.
4.1.2 Metabolism of Bacteria in Consortium
Both theories are able to provide equally plausible explanations of the metabolism of their proposed bacterial symbionts. The Hydrogen Hypothesis proposes a metabolically versatile alpha-proteobacterial endosymbiont capable of performing both cellular aerobic respiration and anaerobic H2 and CO2 producing fermentations 2,3. And there is evidence supporting this argument that the ancestral mitochondria-precursor symbiont was at least a facultative aerobe, capable of oxidative phosphorylation under low oxygen conditions 1.
Meanwhile the Syntrophy hypothesis proposes an even more metabolically versatile alpha-proteobacterial symbiont (capable of not only aerobic respiration, but also of sulfur oxidation and anaerobic photosynthesis), while the delta-proteobacterial host consumes hydrogen and reduces sulfate 28). While more radical than the Hydrogen Hypothesis here, the fact that metabolically versatile delta-proteobacteria have been observed to either produce hydrogen in syntrophy with methanogens (that used that hydrogen for carbon fixation to form methane) or consume hydrogen to reduce sulfates in synthropy with methanotrophic archaea 32 supports the HS Syntrophy Hypothesis. 28 also cites the position of the alpha-proteobacteria in the upper layers of the microbial mats where they interact with sulfur reducing bacteria in sulfur cycling as evidence for their interpretation for the alpha-proteobacteria’s metabolism 28.
4.2 Mechanism, Timing and Nature of inclusion
4.2.1 Mechanism of Inclusion
Another point of contention between the 2 theories is the mechanism by which the symbiont(s) are taken up by the host cell. In the Hydrogen Hypothesis, while Martin and Muller concede in 1998 that their theory remains viable even if the host has acquired a cytoskeleton and by extension phagocytosis before association with the symbiont 2, the Hydrogen Hypothesis does strongly favour a non-phagocytic mechanism of inclusion more in line with the energetic arguments made by Lane and Martin 2,4,6,14. In particular, Lane and Martin argued that the development of a cytoskeleton and thus phagocytosis required an enormous investment of energy that could only be provided by mitochondria 2,4,6. Thus, the alpha-proteobacterial symbiont could not have been incorporated via phagocytosis, and was instead incorporated due to selection for host surface membranes cells with greater surface area in contact with the symbiont (both to ensure they remain together and to facilitate exchange of substances), leading to slow syntrophic engulfment. Further, Martin and Muller argue for the early incorporation of mitochondria 1–3.
However, as argued in section 5 of this paper, there exists a compelling challenge to the energetic argument that phagocytosis was impossible without mitochondria, since a LECA sized (in terms of physical and genomic size) prokaryote capable of aerobic respiration was shown 51 to be energetically viable. Furthermore, the mitochondrial early nature of the Hydrogen Hypothesis further damages it, since as argued in section 4.0, the early incorporation of mitochondria necessarily entails its incorporation into a prokaryote, with its energetically inefficient genomic organisation still intact 51. Thus, the energetic advantage of a hypothetical cell which acquired mitochondria early compared to an equivalent prokaryote (with ATP synthase and thus aerobic respiration) is eliminated 51 [see extended figure 5].
Nevertheless, this is not fatal for the Hydrogen Hypothesis given the non-aerobic hosts proposed by both hypotheses, since a LECA sized Asgard archaea (H2 Hypothesis) or delta-proteobacterial host (Syntrophy Hypothesis) capable of phagocytosis may very well be energetically unviable. Further investigation into the respective host is necessary to prove this to be the case. While, adopting such an argument would weaken the H2 Hypothesis’s case for eukaryotic singularity (since a different, LECA sized prokaryote capable of phagocytosis would be well capable of taking up an alpha-proteobacterial symbiont), it does mean that the Hydrogen Hypothesis’s argument that phagocytosis is energetically unfeasible for its proposed host is, at least somewhat, viable.
By contrast, the HS Syntrophy Hypothesis does not insist on a specific mechanism of inclusion of the Asgard and alpha-proteobacteria symbionts into the delta-proteobacteria host 28. Thus, this leaves the possibility of inclusion of symbionts via phagocytosis. Indeed, Lopez Garcia and Moreia do cite a recent report of prey engulfment by Planctomycetes, suggesting that bacterial phagocytosis exists, albeit based on different molecular grounds than eukaryotic phagocytosis 54. Moreover ,Lopez Garcia and Moreia argue that delta-protobacteria have membrane remodelling potential 28. In particular, developed cytoskeletons necessary for membrane modelling are present in delta-proteobacteria (myxobacteria) as part of protruding membrane tubes that connect cells 41. And indeed, no syntrophic case is known where strong metabolic coupling actually led to obligate endosymbiosis among prokaryotes 1.
Nevertheless, there is still sufficient doubt about the possibility of delta-proteobacteria phagocytosis that Lopez-Garcia and Moreia refrain from outright proposing it as a mechanism of inclusion 28. Indeed, any hypothesis that proposes that the symbiont was taken up via phagocytosis by the host must explain why the symbiont was not degraded 1. While it has been proposed that the pre-existing syntrophy between the symbiont and the host selected against symbiont degradation, an exact mechanism by which symbiont digestion was prevented has yet to be proposed1.
Furthermore, while the Hydrogen Hypothesis in general and the energetic argument it adopts to explain eukaryotic singularity and the lack of intermediates in particular would be weakened by should phagocytosis be the mechanism of inclusion, it would not be fatally compromised 2. Hence, neither theory could be significantly faulted for its proposed mechanism of inclusion(s).
4.2.2 The Implications of the Eukaryotic Chimeric Genome on Timing of Inclusion
The timing of alpha-proteobacterial inclusion is another point of contention, one that can only be answered by looking at the makeup of eukaryotic genomes and the nature of its membranes. The Hydrogen Hypothesis proposes an early mitochondrial inclusion 1–3. Indeed, the inclusion of mitochondria and the resultant energetic boost to the cell precipitates eukaryogenesis. By contrast, the Syntrophy Hypothesis proposes a relatively late inclusion of mitochondria, at which point the opening stages of endomembrane development are well underway 28.
And when one considers the makeup of the eukaryotic genome, the available evidence is firmly in favour of the latter. Indeed, molecular phylogeny has revealed that LECA protein families of alpha-proteobacteria origin and of mitochondrial localisation are closest in phylogenetic distance to their counterparts in the closest prokaryotic relatives, followed by LECA protein families associated with the endomembrane system. LECA protein families derived from archaea associated with the nucleus and ribosomes, by contrast, were the most distantly related (longer stem length) from their homologues in their closest prokaryotic relatives, indicating an earlier acquisition of such protein families by LECA (allowing more time for differences to accumulate between LECA and the prokaryote from which LECA and its precursors acquired the genes encoding said protein family) 29. Indeed, Pittis and Gabaldon’s study revealed that LECA protein families whose closest homologues were non alpha-proteobacteria bacteria, such as delta-proteobacteria, were more distantly related to their prokaryotic counterparts than LECA protein families derived from alpha-proteobacteria. Furthermore, these non alpha-proteobacteria protein families were predominantly associated with endomembrane related compartments 29.
All in all, this indicates that the Archaeal contribution to the eukaryotic genome was the most ancient, followed by the acquisition of non alpha-proteobacteria genes encoding protein families functionally related to the endomembrane system, which therefore must have begun development before mitochondrial incorporation. Only after the above 2 steps was the alpha-proteobacterial symbiont acquired, leading to alphaproteobacteria protein families (mainly involved in metabolism) diverging from their counterparts in alphaproteobacteria to the lowest extent.
And this matches up with the HS Syntrophy Hypothesis far better than it does with the Hydrogen Hypothesis. While the HS Syntrophy Hypothesis has a delta-proteobacterial host, the hypothesis does hold that gene transfer from the host to the Asgard endosymbiont won out, resulting in the centralisation of the consortium’s genome within the Asgard, which develops into the nucleus 28. Thus, it is the Asgard’s genome to which the delta-proteobacterial genes from the host are transferred, meaning that the original Asgard’s archaeal genome not only is the most ancient part of the consortium’s genome, but is also disproportionately associated with the nucleus 28, as indicated by Pittis and Gabaldon’s study. After all, the Asgard archaea becomes the nucleus 28. Furthermore, the hypothesis’s proposed mechanism for endomembrane development– Invagination of the inner delta-proteobacterial membrane, combined with the aforementioned transfer of delta-proteobacterial genes to the Asgard’s genome, lines up with Pittis and Gabaldon’s discovery that non alpha-proteobacteria protein families–mostly associated with the endomembrane system– are the second most ancient component of the eukaryotic genome 29. Only after the internalisation of the Asgard and the opening stages of endomembrane development does the incorporation of the alpha-proteobacterial symbiont occur 28, resulting in the alpha-proteobacterial component of LECA’s genome being the most recently incorporated, and thus the corresponding encoded protein families in LECA are the most closely related to their homologues in other alpha-proteobacteria 29.
4.2.3 Chimeric Eukaryotes, Membrane Replacement and the Origin of the Eukaryotic Cell Surface Membrane
Furthermore, the chimeric nature of eukaryotes also raises questions regarding what happens to the host’s cell surface membrane after inclusion of the endosymbiont(s) 1. And here the 2 theories diverge again. The Hydrogen Hypothesis proposes that the Asgard host’s archaeal membrane is fused with by vesicles of bacterial phospholipid bilayers (initially from the symbiont and then from the host’s genome after bacterial phospholipid genes are transferred from the host to the symbiont) to form a chimeric cell surface membrane 1–3,6. Over time, as more bacterial vesicles fuse with the archaeal membrane, the archaeal membrane with archaeal type isoprene lipids is slowly converted to a bacterial type cell surface membrane characteristic of bacteria and eukaryotes 1–3,6. By contrast, the HS Syntrophy Hypothesis proposes a delta-proteobacterial host, whose cell surface membrane is retained throughout eukaryogenesis 28.
And here, the available evidence seems to slightly favour the HS Syntrophy Hypothesis. While there exists an experimental study in which E. coli were genetically engineered such that they could synthesise archaeal phospholipids to form a chimeric cell surface membrane consisting of 30% of archaeal lipids without side effects of a such a severity as to render such a cell unviable, the study also revealed that over-induction of archaeal lipid gene expression such that the archaeal lipid content of the chimeric cell surface membrane exceeded 30% –which is necessary for membrane replacement– would result in severe growth retardation and drastic changes to cell morphology 55. This, combined with the fact that there is no selective advantage conferred by replacing the archaeal membrane with a bacterial one 1, argues against the Hydrogen Hypothesis’s proposal of membrane replacement, and for the Syntrophy Hypothesis’s proposed delta-proteobacteria host.
4.3 Evolution of the Nuclear Membrane and Endomembrane.
The 2 theories clearly diverge regarding the development of LECA’s nucleus and endomembrane system. The Hydrogen Hypothesis argues for the origin of endomembranes via de novo vesicle formation 1,3,56 after the incorporation of the alpha-proteobacterial symbiont as mitochondria. Alpha-proteobacterial genes for fatty acid ester lipid biosynthesis were transferred into the host’s genome 1,3, enabling the host to synthesise bacterial lipids that spontaneously formed lipid micelles that gradually differentiated into the eukaryotic nuclear envelope and endomembrane system 6,56. The selection pressure for the evolution of the nuclear membrane was intron invasion 3. Mobile Group II introns from the alpha-proteobacterial endosymbiont insert themselves through the host’s genome 3,10, before fragmenting ,giving rise to spliceosomal introns and spliceosome encoding genes 3,10. The far faster rate at which ribosomes translate mRNA compared to the rate of splicing 10 makes the strict decoupling of transcription and translation paramount 3,10 to ensure that only introns are translated to form functional protein gene products. Thus, this creates the selection pressure for further decoupling of transcription and translation–which is achieved through the development of a nuclear membrane, complete with nuclear pores, to physically separate active chromatin undergoing transcription and post transcriptional modification (especially the slow splicing process) and active ribosomes which would carry out the rapid translation process 3, which would ensure that splicing is complete before translation could occur.
By contrast, the Syntrophy Hypothesis proposes that the nucleus and endomembrane system originated from the endosymbiosis of the Asgard archaea (and predates alpha-proteobacterial endosymbiosis), which initially existed in metabolic syntrophy with the delta-proteobacterial host 28. Since complex fermentable organic substances are hydrolysed in the delta-proteobacterial periplasm before being transferred to the Asgard for further breakdown, increasing the surface area of the periplasm in contact with the Asgard would facilitate the aforementioned movement of substances and thereby confer upon the consortium advantages in metabolic efficiency 28. Thus, the inner host membrane invaginates around the Asgard such that the periplasm now surrounds the Asgard 28. (See Fig 2.3).
In the Second and Advanced integration steps according to 28 (Fig 2.4) of the HS Syntrophy Hypothesis, a second selection pressure now begins to take precedence. The transfer of delta-proteobacteria genes from the host into the Asgard endosymbiont now requires the development of a transport system enabling the delta-proteobacterial gene products, especially the bacterial hydrolytic enzymes responsible for the aforementioned breakdown of complex carbohydrates in the periplasm, to be transported out of the Asgard through the Asgard’s own membrane to their target areas 28. Moreover, since the transport system would be handling hydrolytic enzymes, it must prevent their entry into the cytosol in order to prevent degradation of cytoplasmic components 28. And this now selects for the development of a proto-endomembrane system. As to how this occurs, the HS Syntrophy Hypothesis proposes that initially, bacterial proteins may have been inserted into the archaeal membrane in order to allow simpler bacterial gene products to pass through 28. This is then followed by the development of pore complexes (on the bacterial membrane which bounds the layer of periplasm around the Asgard) able to communicate with the Archaeal membrane’s transport proteins in order to allow more complex substances to be transported out of the Asgard proto-nucleus. A side effect of this is the congregation of ribosomes around the proto-nuclear pores in order to enable the translated gene products to be quickly exported 28. By contrast, transcription occurs primarily in the Asgard’s interior 28.
This spatial separation of transcription and translation soon proves invaluable in the LECA stage of eukaryogenesis (section 2.5). The loss of the Archaeal membrane is followed by the same process of intron invasion cited by the Hydrogen Hypothesis 10,28, requiring a clearer separation of translation from transcription and the attendant post-transcriptional modifications (splicing) 28 than the aforementioned spatial separation. This third selection pressure then selects for consortia in which the former Asgard Ribosomes migrate out of the proto-nucleus (across the proto-nuclear pores) and eventually associate along the endomembrane system to form a proto rough endoplasmic reticulum on which translation now takes place 28. Meanwhile, the periplasm that onced lined the delta-proteobacteria now becomes fully internalised (with some of its digestive functions now taken over by independent vesicles 28), eventually becoming the cisternae lumen.
The fact that the HS Syntrophy Hypothesis’s argument for nuclear and endomembrane evolution is far more elaborate than that proposed by the Hydrogen Hypothesis does not escape the author’s attention. Indeed, whereas the Hydrogen Hypothesis proposes a single selection pressure (intron invasion) for nuclear membrane evolution (and no compelling selection pressure for the evolution of an endomembrane system with secretory functions) 3, the 3 selection pressures proposed by the Syntrophy Hypothesis (efficiency of diffusion, export of proto-nuclear gene products and intron invasion) more clearly explain the impetus behind both the development of the nuclear membrane (export of proto nuclear products and intron invasion) and the other parts of the endomembrane (Efficiency of diffusion, export of proto nuclear products) 28. The evolution of the endomembrane’s digestive (specialised vesicles/organelles) and secretory (Endoplasmic reticulum, Golgi body, exocytotic and endocytotic vesicles) functions is also more adequately explained by the HS Syntrophy Hypothesis, while the Hydrogen Hypothesis mostly ignores the above.
Furthermore, it is unlikely that intron invasion (from the alpha-proteobacterial genome to the host’s genome) alone can explain the development of a nuclear membrane 28. Since, as the Hydrogen Hypothesis points out, the net result of intron invasion, ceteris paribus, would be the translation of unspliced mRNA to form non functional gene products across hundreds of gene loci, which would most likely be fatal 3. Thus, either the population of organisms rapidly develops a nuclear membrane by which the simultaneous transcription and translation characteristic of prokaryotes can be decoupled or individual organisms in which alpha-proteobacterial endosymbiosis and thus introin invasion occur are selected against. And the long list of intermediate steps that eventually results in the formation of the nuclear membrane (and the contiguous endomembrane system) makes it unlikely that the former possibility could occur quickly enough before at least one intron was inserted into an essential gene, with disastrous consequences for the organism as a whole 28. Considering the above, it makes sense for Lopez-Garcia and Moreia, as well as the author of this paper, to regard the latter possibility as much more likely. Hence, the argument that the existence of at least some level of transcription-translation uncoupling (as occurs in phase 2 of the HS Syntrophy Hypothesis) must predate intron invasion (which then selects for a more tightly enforced decoupling of transcription and translation) 28 is more convincing.
Further, the Hydrogen Hypothesis’s proposed mechanism for nuclear membrane formation has also been questioned 56. Firstly, there is a contradiction in the Hydrogen Hypothesis’s argument. Not only does the Hypothesis argue that the Alpha-proteobacterial genes for fatty acid ester lipid biosynthesis transferred into the host’s genome 1,3 enabled it to synthesise phospholipids that made it to the cell surface membrane, which was replaced archaeal membrane lipids 2,3,6, but the Hypothesis also adopts Nick Lane’s argument that while alpha-proteobacterial lipids synthesis genes were transferred into the host cell’s genome, the mechanism for targeting such proteins such that the phospholipids end up in the cell surface membrane was not transferred from the symbiont to the host were not, preventing the bacterial phospholipids synthesised by the host from being transferred to the host’s cell surface membrane 6, leaving them to instead spontaneously form lipid micelles that gradually differentiated into the eukaryotic endomembrane system 56. These 2 arguments are clearly logically inconsistent with each other.
The Hydrogen Hypothesis also fails to explain how the aforementioned lipid micelles eventually developed the secretory functions of the modern endomembrane system 56. The prerequisite for secretion and transmembrane protein synthesis is a system of transmembrane protein insertion and translocation, which in most organisms (bacteria, archaea and eukaryotes alike) is mediated by the universal protein conducting channel, the trimeric Sec61/SecY complex 56. But to insert this complex into a membrane, other protein conducting channels must pre exist in the membrane 56, thus making the insertion of the universal protein conducting channel into the purely lipid micelles postulated by the Hydrogen Hypothesis unlikely. While it is entirely possible that such an insertion could happen anyway, as it almost certainly did at the origin of the first cellular lipid membranes 56, the chance of such a rare event occurring again is, according to 56, unlikely.
However, this is not to say that the HS Syntrophy Hypothesis’s argument for nuclear membrane evolution is without problem. Indeed, 56 argues that there are issues with theories that propose an endosymbiotic origin of the nucleus, which includes the Syntrophy Hypothesis in both its older HM and newer HS forms. Firstly, the study argues that the loss of the Archaeal symbiont’s membrane would likely be fatal 56 for the consortium, since membrane loss could not have occurred gradually and the sudden loss of the the membrane would mean the loss of all membrane associated functions–from hydrogen production and DNA segregation to transport 56. While the HS Syntrophy Hypothesis argues that the alpha-proteobacterial endosymbiont’s aerobic respiration would have made the bioenergetic capacity of the archaeal membrane irrelevant 28, the HS Syntrophy Hypothesis has a harder time explaining how the consortium can survive the sudden loss of the other functions of the archaeal membrane. In particular, given the fact that the proto-nuclear pores in phase 2 (second and advanced integration stage in 28) coordinate with membrane transport proteins in the archaeal membrane 28 in order to export gene products synthesised in the archaeal endosymbiont, it is questionable if the proto-nuclear pores would still be able to effectively function given the immediate loss of the archaeal membrane proteins co-ordinating the the proto-nuclear pore. Thus, while membrane loss can occur occasionally during evolution, such as at the origin of Gram positive bacteria 56, the difficulty and rarity of it is such that even Lopez-Garcia and Moreia admit that it is a major difficulty faced by any theory proposing an endosymbiotic origin to the nucleus, including theirs 57.
Secondly, the existence of the Archaeal membrane presents difficulties for the gene transfer from the delta-proteobacterial host to the Asgard endosymbiont proposed by the HS Syntrophy Hypothesis 56. Lopez-Garcia and Moreia propose no specific mechanism for gene transfer between the Asgard symbiont and delta-proteobacterial host, only that it did occur from the host to the symbiont 28. By contrast, Jékely adopts the argument 58 that gene transfer best occurs when the symbiont lyses and releases its genome to integrate with the host genome. This method is clearly inapplicable in the opposite direction, as lysis of the host will kill the whole consortium 56. And even if some other means of gene transfer was possible 56, such as the host’s mRNA entering the Asgard and being reverse transcribed, it would require the embedding of transport proteins within the Asgard’s membrane and the development of a proto-nuclear pore complex to allow host mRNA into the Asgard 56. And it is here where the Syntrophy Hypothesis runs into issues. In particular, the hypothesis’s argument that gene transfer from the host to the symbiont is what eventually gives rise to the selection pressures that provide impetus for the evolution of the proto-nuclear pore complex 28 is now unviable. In other words, the proto-nuclear pore equivalent must predate gene transfer, which in the HS Syntrophy Hypothesis occurs concurrently with endomembrane development.
Thirdly, the protocoatmer model 59,60 that Jékely adopts for the origin of the nuclear pore 56 raises further challenges for the HS Syntrophy Hypothesis. The model postulates that the Nuclear Pore complex evolved from vesicle coatmer complexes 56,59. The architecture of Coatomer proteins, consisting of beta propellor domains at the N-terminus and alpha-solenoid domains the C-terminus, along with components consisting of alpha-solenoids and beta-propellers alone, is highly flexible and allows for a wide range of structures to be formed 59. Not only are they involved in facilitating the budding off of vesicles (COPI, COPII vesicle coats) transporting substances in between the Golgi body and the Endoplasmic reticulum, but coatomer proteins have also been found to make up well over half the mass of the Nuclear Pore complex 59. Thus, it has been suggested that the vesicle coat and Nuclear pore were the evolutionary descendents of a protocoatomer complex originally restricted to forming vesicle coat complexes 56,59,60. In particular, the protocoatomer evolved into 2 major classes, type I (typified by clathrin-coated vesicles and COPI transport vesicles) and type II (typified by COPII transport vesicles) 59. The proto Nuclear pore then evolved via an amalgamation of both type I and II coatomer proteins 60.This, in turns, implies that the evolution of an already differentiated endomembrane system, complete with the precursors of modern type I and type II coatomers , must predate the evolution of the nucleus and the nuclear pore 60. This sequence of events, the evolution of the endomembrane system, followed by the evolution of the nuclear pore, which allows gene transfer between the proto-nucleus and the host, is in contrast to that proposed by the HS Syntrophy Hypothesis, which holds that the endomembrane evolved roughly concurrently with gene transfer from the host to the Asgard, which is then followed by the evolution of the nuclear pore 28.
Nevertheless, the 3 arguments presented in 56 do not fatally compromise the HS Syntrophy Hypothesis. For HS Syntrophy Hypothesis to remain viable,the protocoatomer complex must have evolved by the time of the first endosymbiosis of the Asgard (phase 1), when the inner delta-proteobacterial membrane invaiginates around the Asgard to facilitate transfer of substances 28. However, the argument that these bacterial proteins could be inserted into the Asgard’s membrane as a proto-nuclear pore to facilitate gene transfer 28 is invalidated by considerations of membrane topology 56. Since membrane proteins are inserted into the lipid bilayer from the cytoplasmic side by the action of a cytoplasm-oriented protein machinery, the protein conducting channel 56,any protein inserted into the Asgard membrane ought to come from the Asgard’s cytoplasm and not the delta-proteobacteria’s. Hence, in order for gene transfer between the host and symbiont to be viable in time to stabilise endosymbiosis in phase 1 (first endosymbiosis stage) , the Asgard must evolve an analogous structure to the nuclear pore before or soon after its endosymbiosis (which is stabilised by gene transfer). The selection pressure for the evolution of this Asgard nuclear pore analogue will not be to facilitate the movement of substances between the Asgard and host, but to enable endosymbiotic gene transfer to stabilise endosymbiosis. This is entirely possible since Asgard archaea are suggested to possess genes encoding proteins similar to the beta-propellor protein Sec13 and alpha-solenoid containing proteins as well as Rab-like GTP-ases, the building blocks required of (primitive) vesicle coat complexes 60 that may have evolved into a nuclear pore analogue on the Asgard membrane. Thus, by the time of the second and advanced integration steps (phase 2), these Asgard membrane proto-nuclear pores would be the ones communicating with the host’s own proto-nuclear pores integrated in the bacterial membrane bounding the Asgard to facilitate the export of complex gene products 28. Given the availability of a well developed endomembrane system, complete with protocoatomers, by this stage, the cell could now likely survive the loss of the archaeal membrane without compromising its transport and DNA segregation functions too badly. While this sequence of events in general, in particular the convergent evolution of nuclear pore analogues by both the Asgard endosymbiont and delta-proteobacteria host, is unlikely, it is, as far as the author of this paper knows, viable. Thus, although Jékely’s arguments certainly weaken the HS Syntrophy Hypothesis, they do not entirely invalidate its arguments for endomembrane development.
Thus, while there are issues with both hypotheses’ arguments for nuclear and endomembrane origins, the issues faced by the Hydrogen Hypothesis’s argument are judged to be greater due to the inherent contradiction within the argument outlined earlier. By contrast, it is possible to construct a sequence of events (see previous paragraph), however unlikely, in which the Syntrophy Hypothesis’s more extensive argument for endomembrane and nuclear envelope evolution is viable.
Section 5: The Energetic Argument of Eukaryotic Origin-an Analysis
Central to the Hydrogen Hypothesis’s explanation of the mechanism of inclusion, eukaryotic singularity and eventual eukaryotic diversification is the energetic argument put forward by Lane and adopted by Martin 6,14,15 and laid out in the Section 1 of this paper. However, certain key aspects of the argument have been challenged.
In particular, Lane and Martin’s use of the concept of “energy per gene” has come under criticism 15,51,52,61. By using idealised averages for the ATP production and genome sizes of modern eukaryotes and prokaryotes, the average spherical eukaryotic cell was projected to have around 200 000 times more ATP production per gene than an equivalent sized and shaped prokaryote 14. Thus, they argue that the larger size and more complex genomes of eukaryotes compared to prokaryotes is enabled by mitochondrial energy surplus eukaryotes enjoy over prokaryotes 14.
On the surface, there is some evidence to back this up, namely the fact that eukaryotes have larger genomes (mostly between 8–10000 Mbp) than prokaryotes (mostly <1–16 Mbp). Eukaryotes (mostly 10 to 10^8 µm^3) are also physically larger than prokaryotes (mostly between 10^–2 to 10^2 µm^3) 51. Thus, overlaps (8–16 Mbp genome, 10 to 10^2 µm^3 cell volume) notwithstanding, the data suggests that there is an upper limit to the size, both physical and genomic, that amitochondriate cells like prokaryotes can attain 51.
However, Lane and Martin’s methodology has increasingly come under criticism. In particular, Lynch and Marinov 61 argue that the concept of ‘energy per gene’ ignores the energy used by eukaryotic cells for biomass incorporation, the energetic expenditure for the synthesis of mitochondrial membrane lipids ,and different generation lengths 61. Furthermore, Schavemaker and Muñoz-Gómez also argue that the concept also unfairly penalises large prokaryotes whose gene numbers increase with polyploidy 51. Since a cell’s energetic demands are solely dependent on their volume, a prokaryote and a eukaryote with the same volume and number of genes would see their genes expressed at a similar rate. Thus, both cells should have a similar energetic cost for gene expression 51. Furthermore, dividing the average eukaryote’s energy production by the number of gene copies in its nuclear genome ignores the many copies of mitochondrial genomes that the eukaryote must energetically support 51. Hence, it is unsurprising that quite a few recent analyses have concluded that the 200 000 fold eukaryotic energetic advantage predicted by Lane and Martin is scientifically untenable 51,61.
Lane and Martin then offer 2 explanations for the energetic advantage of eukaryotes: 1) The internalisation and of bioenergetic membranes within mitochondria released eukaryotes from the constraints of the limited cell surface membrane area for chemical energy production 2) The evolution of highly reduced and specialised mitochondrial genomes which conferred upon eukaryotes am advantageous genomic asymmetry (Due to the CoRR Hypothesis, whereas the ploidy Prokaryotes (and by extension copy number of all prokaryotic genes minus those in plasmids) must scale up with cell volume, in eukaryotes the nuclear genome has a constant copy number and only the mitochondrial genome copy numbers rise with cell volume) 6,51 . And while the degree to which these 2 factors provide eukaryotes with an energetic advantage is disputed, the fact that they do (especially the second factor) is less disputed 51.
In particular, the results of Schavemaker and Muñoz-Gómez’s modelling are indicative. Depending on the proportion of cell surface membrane devoted to ATP production, volume, shape and frequency of cell division, a prokaryote relying exclusively on its cell surface membrane would not be constrained by membrane surface area until it reaches a size between 10 (spherical cell that devotes 8% of its membrane to ATP production that divides every 1h and 10^5 µm^3 (oblong cell that devotes 18% (largest possible fraction of respiratory membrane in E. coli) of membrane to respiration and divided every 10h). Such a size is well within range for eukaryotic cells (10 to 10^8 µm^3) 51 , and indeed the estimated volume of the last eukaryotic common ancestor (23 to 57 µm^3) falls within this range 51 .
Furthermore, while Schavemaker and Muñoz-Gómez 51 do concur with Lane and Martin 14 that eukaryotic genome organisation does confer upon eukaryotes an energetic advantage 14,51, Schavemaker and Muñoz-Gómez’s modelling does reveal that the advantage is far less pronounced than Lane and Martin’s predictions of a 200 000 fold increase in ‘energy per gene’. By comparing the now the proportion of ATP a cell can devote to anything other than the synthesis of its genome varies with haploid genome length in prokaryotes and eukaryotes, Schavemaker and Muñoz-Gómez estimate an energetic advantage of less than 10% for a eukaryote with a 10^6 to 10^7 bp long nuclear genome compared to a eukaryotic (10 to 10^8 µm^3) sized prokaryote with the same genome length 51. For larger (haploid) genome sizes (10^7 to 10^8 base pairs), the energetic advantage of eukaryotes rises to around 200% 51. Prokaryotic genomes are limited to 3*108 base pairs since the energy to synthesise a larger genome would exceed the total ATP available to a cell during a single cell cycle 51. However, an amitochondriate prokaryote could, by devoting 10% of its ATP production per cell cycle to DNA synthesis, support a haploid genome 3*10^7 base pairs long ,equivalent to many small single celled eukaryotes. (though such a prokaryote would have about 20% less energy to devote to processes other than DNA replication) 51. Indeed, comparative genomic analyses have estimated that the Last Eukaryotic common ancestor had 4431 gene domains, corresponding to a genome size of 2*10^7 to 5*10^7 base pairs (and a volume of 23 to 57 µm^3), well within the limits for prokaryotic genome and physical size described in this and the previous paragraph 51.
Martin and Lane’s counter-argument to the above rests on the notion that it takes more energy to evolve complex eukaryotic features as it does to maintain them 50. Since the evolution of new genes entails the exploration of a broad protein space, many protein variants, a significant number of which are non-functional, would be produced, incurring an energetic cost that only the presence of mitochondrial ATP could fulfil 14. Indeed, Lane argues that a large intermediate (around 10 fold) expansion of the genome to allow experimentation with novel gene families was necessary for eukaryogenesis 1,15.
However, while this argument certainly deserves further scrutiny, there is reason to believe that it is incorrect 51,52.A sudden multiplication of genome size would result in an unacceptable number of replicative errors 1. Furthermore, the evolution of photosynthesis, which would likely have been as costly as the evolution of certain eukaryotic hallmarks like phagocytosis, did occur in prokaryotic cyanobacteria 1, indicating that the evolution of at least some complex eukaryotic traits would not have necessitated mitochondrial energy. Finally, there is no theoretical or comparative evidence to support the requirement of a transient appearance of a huge number of genes exceeding the final count by up to an order of magnitude in order to evolve complex processes 62. And while Booth and Doolittle do concede that an energy surplus could facilitate a higher rate of evolution, they nevertheless remain unconvinced by this argument 63. Since evolution is ultimately driven by selection pressures, they argue that should selection pressures favour eukaryotic complexification, it would have been evolved regardless of whether there was a large energy surplus 63. Thus, while further research will be needed to definitively prove whether the counter-argument presented above is true or false, many in the scientific community currently lean towards the latter interpretation.
Moreover, even if Lane and Martin are correct regarding the cost of evolving novel gene families, the energetic advantage provided by mitochondria is far less than they assume it to be (as argued above). Indeed, the estimated genome size of LECA (2*10^7 to 5*10^7 base pairs) falls well below the theoretical limit of 3*10^8 base pairs for an amitochondriate prokaryote of any volume, allowing the possibility of an amitochondriate cell with a genome size of 2–5*10^7 base pairs (estimated LECA genome length range) but the energetic capacity to support 3*10^8 base pairs 51, giving it the theoretical capacity to expand its genome about 6 to 15 fold for the experimentation Lane and Martin deem important. However, the prokaryote in this case can have a volume and cell division timing radically different compared to that of LECA. However, even when we restrict a prokaryote to the dimensions of LECA, there is little to suggest that such a prokaryote would have been under a crippling energetic disadvantage compared to LECA. Indeed, the calculated energetic advantage, based on formulas in 51 of even the largest, most complex LECA with 5*10^7 base pairs and a volume of 57 µm^3 compared to a similar sized prokaryote ranges from 26%(mitochondrial incorporation late) to 7.78% (mitochondrial incorporation early). While more research is needed to determine if such an energetic disadvantage would be enough to cripple the ability of a prokaryote to evolve to the level of complexity achieved by LECA ( a fully complex, fully eukaryotic cell possessed of most eukaryotic hallmarks (actin-based cytoskeleton, complex endomembrane system, nuclear envelope, vesicle trafficking, Golgi body, lysosomes, autophagosomes, etc 1), such numbers seem much too low to justify the above argument. Moreover, the claim that a prokaryote could not evolve any of the complex eukaryotic structures (phagocytosis, nucleus, cytoskeletons, vesicle trafficking, mitosis) is rendered even more tenuous.
Thus, while Lane and Martin’s counter argument that the evolution of genes requires an intermediate genome expansion that in turn necessitates an energy surplus that only mitochondrial energy can provide 1,14,15,50 is yet to be entirely disproven (or proven), the fact that the mitochondrial energy surplus is much less significant than they anticipate for cells of LECA’s size 51, combined with the fact that the idea that it takes much more energy to evolve than to maintain a new gene is disputed, makes it unlikely that this would have been a significant barrier to the evolution of at least some complex eukaryotic traits. Moreover, Schavemaker and Muñoz-Gómez’s calculations do demonstrate that a eukaryote of LECA’s size would not have been constrained by limited bioenergetic membrane surface area, meaning that the first eukaryotes world have been energetically viable with or without mitochondria 51. Any energetic advantage eukaryotes have compared to prokaryotes is, indeed, due to the superior eukaryotic genomic organisation 51. To be fair, Schavemaker and Muñoz-Gómez concede that mitochondria became much more important in energetically enabling eukaryotes to evolve from the small, heterotrophic flagellate that the Last Eukaryotic common ancestor (LECA) probably was 51 to become progressively larger, faster dividing and more complex. Nevertheless, it is clear that a prokaryote of LECA size that was capable of aerobic respiration would be energetically able to support many complex processes thar Lane and Martin claim are out of reach for prokaryotes, including phagocytosis as a mechanism for inclusion, the development of a nuclear envelope and the formation of the cytoskeleton.
Section 6: Conclusions
The Hydrogen Hypothesis proposes that eukaryotes only evolved once due to the singular incorporation of mitochondria 1, which was energetically necessary for the development of eukaryotic complexity 2,3,50. By contrast, the HS Syntrophy Hypothesis does not propose any convincing theory to explain eukaryotic singularity. Thus, the H2 hypothesis does a better job than the HS Syntrophy Hypothesis here simply by having a plausible explanation for eukaryotic singularity where the HS Syntrophy Hypothesis seems to have none. Further work should correct the flaws in the H2 Hypothesis’s arguments in section 4.0, or give the HS Syntrophy Hypothesis an argument to explain eukaryotic singularity.
With regard to the metabolism of host and symbiont, there are clear contrasts between the 2 theories. The H2 Hypothesis proposes a strictly anaerobic, hydrogen consuming, autotrophic host 1–4 and a metabolically versatile alpha-proteobacterial endosymbiont capable of performing both cellular aerobic respiration and anaerobic H2 and CO2 producing fermentations. By contrast, the HS Syntrophy Hypothesis proposes a tripartite consortium consisting of a hydrogen producing, carbon oxidising (organoheterotrophic) Asgard archaea symbiont, an even more metabolically versatile alpha-proteobacterial symbiont (capable of not only aerobic respiration, but also of sulfur oxidation and anaerobic photosynthesis), and a delta-proteobacterial host that consumes hydrogen and reduces sulfate 28. Unfortunately, conflicting primary research precludes a judgement being made on which hypothesis’s proposed metabolic syntrophy is superior, and future phylogenetic studies that could provide a conclusive picture of the metabolic pathways present in Asgard archaea, alpha and delta-proteobacteria would contribute to our understanding of eukaryogenesis. Nevertheless, it is perhaps a testament to the strength of endosymbiotic theories for eukaryotic origin that both hypotheses propose metabolic syntrophy as the initial transition state in eukaryogenesis.
With regard to the mechanism, timing and nature of inclusion, the HS Syntrophy clearly shines. While the differences in proposed mechanisms of inclusion are not fatal to either theory, the HS Syntrophy Hypothesis provides a superior explanation of the implications of the eukaryotic Chimeric genome on timing of inclusion. Molecular phylogeny has revealed that LECA protein families of alpha-proteobacteria origin and of mitochondrial localisation are closest in phylogenetic distance to their counterparts in the closest prokaryotic relatives 29. This favours a late mitochondrial Incorporation as proposed in the Syntrophy Hypothesis over the early mitochondrial incorporation proposed by the Hydrogen Hypothesis. Finally, the Syntrophy Hypothesis’s delta-proteobacterial host sideskirts the need for membrane replacement that bedevils the Hydrogen Hypothesis.
Finally, the Syntrophy Hypothesis proposes a more elaborate and clearly superior mechanism by which the nuclear membrane and endomembranes are evolved compared to the H2 Hypothesis. Although not without flaw, the Syntrophy Hypothesis proposes a more extensive and convincing set of selection pressures to explain endomembrane development in stages across transition states, whereas the Hydrogen Hypothesis only proposes a single selection pressure for endomembrane development. Worse, the Hydrogen Hypothesis, in its attempts to explain both cell surface membrane replacement and nuclear membrane development, ends up contradicting itself. This, in the author’s eyes, fatally undermines its argument for endomembrane development. Thus, while both hypotheses have been questioned in this criteria 56, the Syntrophy Hypothesis still comes out on top in this comparison.
With regard to the remaining criteria–initial relationship, early selective advantage, vertical transmission, both hypotheses offer equally convincing (or unconvincing, in the case of vertical transmission) explanations. The results of the comparison are summarised in the table below.
In Summary, of the 9 criteria by which the 2 theories are compared against each other,in 6 of them both theories are considered equal (metabolism of the Asgard archaea and bacterial symbionts, Initial relationship, Early selective advantage, vertical transmission (which both fail to explain1) and the mechanism of inclusion). The Hydrogen Hypothesis wins in one criteria (Eukaryotic singularity), whereas the Syntrophy Hypothesis wins in 2 criteria (Evolution of the nuclear membrane and the implications of Chimeric eukaryotes on the eukaryotic membrane and timing of inclusion). Thus, While neither theory can provide adequate explanations for all the historical transition states in eukaryogenesis, along with their implications, the proposed transition states of the HS Syntrophy Hypothesis does appear , overall, to be more in line with current evidence compared to the Hydrogen Hypothesis. Hence, the HS Syntrophy Hypothesis is scored more highly (at 7) than the Hydrogen Hypothesis (with a score of 2) in this comparison.
While these conclusions seem to favour the updated HS Syntrophy Hypothesis over the H2 Hypothesis, it must be borne in mind that the HS Syntrophy Hypothesis constitutes a more recent work–being proposed in 2020–than the H2 Hypothesis–updated in 2015. Indeed, this constitutes a reversal of the verdict in 1 where the Hydrogen Hypothesis came out ahead of the older HM Syntrophy Hypothesis. How both hypotheses, especially the Hydrogen Hypothesis, could be updated to better reflect more recent evidence would be an interesting undertaking, but one that falls outside the scope of this paper.
Section 7: Future work
In this work, the author has elected to focus on the evolutionary transition states proposed by these 2 hypotheses. Thus, the theories were not compared based on their ability to explain current observations of eukaryotes. Such observations include the Loss of membrane bioenergetics, the Inability of mitochondria to photosynthesize, the Origin of mitochondria related organelles like hydrogenosomes, and the absence of Intermediates between prokaryotes and LECA. Thus, the author intends the above to be the subject of a subsequent study. Furthermore, this paper only compares 2 hypotheses, the H2 and HS Syntrophy Hypotheses, both of which are endosymbiotic theories. In the future, this work can be expanded to compare more hypotheses of varying types, such as those proposing a parasite as the origin of mitochondria. Lastly, the lack of conclusive phylogenetic data to indicate the precise metabolic pathways present in Asgard archaea also limits the author’s ability to conclusively determine which hypothesis’s proposed host is more accurate. Thus, a comprehensive phylogenetic study of Asgard archaeal metabolism would be on the cards as future work.
Section 8: References
(11) Lambowitz AM, Zimmerly S. 2004 Mobile Group II Introns. Annu. Rev. Genet. 38, 1–35.
(14) Lane N, Martin W. The Energetics of Genome Complexity. Nature. 2010;467:929–34.