Well, something accumulates, but not a canonical scenario for how all this happened. Beyond the simple textbook story of endosymbiosis lies an unresolved mess:
Chlamydia species are prokaryotes, and they live inside of human cells. They are obligate intracellular parasites which means that Chlamydia species can’t reproduce on their own outside of the host cell. How can it be a mystery when we can watch in happening in real time in the modern era?
The whole point is that when you build new adaptations on previous adaptations then the earlier adaptations then the earlier adaptations are harder to change without lowering fitness. Therefore, it is wrong to assume that a whole system had to come about in one fell swoop.
Did you read the articles I posted?
http://mmbr.asm.org/content/81/3/e00008-17.abstract
These authors (and many others I did not cite) argue that the origin of the mitochondrion – which is what we’re discussing, after all, not obligate parasitic prokaryotes and the like – was literally a singular, or one-off event, of vanishingly small probability, and therefore unrelated to (for instance) Bdellovibrio infection, or your favorite example, Chlamydia infection, which happen all the time.
From Michael Lynch and Georgi Marinov (2017, p. 10; emphasis added):
“The origin of the mitochondrion was a singular event, and we may never know with certainty the early mechanisms involved in its establishment, nor the order of prior or subsequent events in the establishment of other eukaryotic cellular features.”
From Martin et al. (2017, pp. 25-6; emphasis added):
“Rarity is furthermore a desirable property of endosymbiosis, because mitochondria arose only once in 4 billion years, roughly the same rate at which life and the solar system arose. Bacterial endosymbionts in the cytosol of phagocytosing eukaryotic cells are nothing unusual; on the contrary, they are extremely common. Examples include the many known cases of proteobacterial endosymbionts of insects (293–295), the methanogenic endosymbionts of anaerobic ciliates (296), the purple endosymbionts of the ciliate Strombidium (297), the sulfur-metabolizing symbionts of clam gills (298), the chemosynthetic endosymbiont consortia of gutless tubeworms (299), endosymbionts that live within the endoplasmic reticulum of diatoms (300), or the cyanobacterial endosymbi- onts of sponges (301), to name just a few. The commonplace occurrence of bacterial endosymbionts in phagocytic cells stands in diametric contrast to the very rare origin of mitochondria (a singular event among the ancestors of all microbes that have left known descendants in 4 billion years of evolution). Thus, we can safely say that phagocytosis promotes the frequency with which endosymbionts can come to reside within the eukaryotic cytosol, but it has no bearing whatsoever on the rate at which mitochondria arise from endosymbionts. This is one more (strong) reason why phagocytosis is unlikely to have anything to do with mitochondrial origin.”
“Of course, microbial cells snuggling up to one another should not evoke the impression that they are only one step away from one getting inside and becoming a mitochondrion. The endosymbiosis that gave rise to eukaryotes was rare. How rare? Whitman et al. (319) estimated that roughly 10exp30 prokaryotic cells exist on Earth today. If we are granted a simplifying assumption, namely, that the environment has harbored roughly the same number of cells over the last 2 billion years, and furthermore granted a pure guess that an average cell has a doubling time of about 2 months in nature (some are slower, and some are faster [27]), we obtain a rough but round estimate of about 10exp40 prokaryotic cells that have lived in the last 2 billion years. Most or all of them had a partner from the other domain nearby. This represents a very large number of opportunities to create eukaryotes, opportunities where nothing other than metabolic interactions and occasional interdomain gene transfer (127, 320) ever happened, except once during a fateful encounter at eukaryote origin. Eukaryote origin was a very rare event.”
Now why would these evolutionary biologists argue that this event – the origin of the mitochondrion – was so rare, as to be “singular” and “once during a fateful encounter” in the history of Earth?
Exercise for the reader. The answer leads into a very deep and dark hole, within which resides an unhappy paradox for the cogency of evolutionary theory.
Plenty of mystery here. Chlamydia infection is a distraction.
Right. Riedl’s theory of “burden,” and Wimsatt’s theory of “generative entrenchment,” among others, describe this general pattern and its significance.
So how did Aphidius ervi, the parasitic wasp whose embryogenesis I mentioned above, evolve holoblastic cleavage, when the ancestral mode of development in both the Diptera and Hymenoptera was a syncytium, where diffusion or active transport of a morphogen such as bicoid is essential for normal development, namely, establishing the A-P axis of the embryo? In A. ervi, a marker dye smaller in molecular dimensions than bicoid (a DNA-binding protein), cannot diffuse between cells. See the 1998 PNAS paper I cited above.
Evolutionary principles such as “build on what came before” work fine, until one looks at the actual diversity of living things. Many supposedly deeply-entrenched features, such as cleavage patterns in the Metazoa, have been replaced entirely by functionally equivalent, but non-homologous, features.
Assuming your disputation holds water (and @T_aquaticus seems happy to dispute it; but I have no axe to grind here, and no expertise to get between you two in any case) --but assuming you found an intractable mystery: don’t you mean instead … “an unhappy paradox for the cogency of abiogenesis”?
Thinking that evolutionary theory is contingent on solving abiogenesis is just as bizarre as thinking heliocentric astronomy must be held in question until we’ve explained the first nanoseconds of the big bang. How is that bizarre logic sustained?
No – the endosymbiotic origin of the mitochondrion occurred long after the first cell came to be (i.e., after abiogenesis occurred).
The puzzle I’m alluding to concerns the use of singularities (unique events) in evolutionary explanation. That is, saying “X is so improbable that it happened only once, but it DID happen once.”
Well – that makes it a little more interesting then … carry on!
I suppose the alleged endosymbiosis is thought to have occurred very early in the development of life (generally speaking)? I.e. if it is still “a” … “biogenesis” kind of issue (not to be confused with the abiogenesis event itself) then my distinction would still stand – as in we’re still trying to resolve issues that are within a “few moments” of the analogous “big bang”. Sorry for my layman’s ignorances here, but your patient explanations go a long way.
He’s talking about the transition from single-celled to multi-cellular organisms. There are singularities at many (every?) transitional stages in evolutionary history, and there are probably just as many prior to the appearance of life on Earth. The National Geographic series “One Strange Rock,” narrated by Will Smith, focuses on a lot of those events to make the case that Earth actually might be unique – a singularity, as @paulnelson58 would say – in the universe.
I think you will agree that evolutionary theory is not complete. Which doesn’t mean its wrong. What we do know is that any theory, simple enough to be expressed with some equations and a paragraph of text is not likely to describe the total history of life. The truest statement I can think of in this regard is “Biology is not physics, and not even Chemistry.” The singularity of mitochondrial endosymbiosis is not the only example of non continuity in evolutionary history, and indeed refutes the (strawmannish) argument that all of that history can be fit into well described patterns of slow progressive accumulation of incremental increases in fitness. Punctuated equilibrium accomplishes the same refutation, as does neutral theory (in a different manner). My point is that using singularities (or even less rare exceptions to Darwin’s original proposal) do not destroy the concept of evolution by natural selection, but only serve to force us to improve it.
The problem turns on how one knows that a singularity (a unique, or what the British call “a one-off” event) has actually occurred in evolutionary history – and how to maintain that very low probability at a stable value, in the face of increasing biological knowledge. All phylogenetic inference requires these low probability events, to “pull together” otherwise disparate groups into monophyletic clades. However, as the probabilities of the evolutionary transition or transformation move away from zero (0.0) towards one (1.0), the signal of history – of unique occurrences – is lost.
In two of the papers I linked (above), the authors say that we can know with certainty that all eukaryotes share a common ancestor (the Last Eukaryotic Common Ancestor, abbrev. LECA). Why? Because the original incorporation of an alphaproteobacterium into an unknown prokaryotic ancestor, to become the progenitor of mitochondria, happened once, and only once, in LECA. Thus any cell with mitochondria shares common ancestry with any other such cell, at the LECA singularity.
But this small probability, at LECA, stands on a razor’s edge. Far from being a stable probability, the event is just about as unstable (uncertain) as can be.
That is, the event needs to be so improbable as to be literally unique in Earth history – yet NOT so improbable as not to happen at all. One can discover this for oneself by simply asking, “Well, if the incorporation of a bacterium eventually to become a mitochondrion happened once, why couldn’t it have happened more than once? After all, cells exist in populations, and surely there existed many more candidates for the transition, and many more such populations, than simply one cell in one population. Right?”
Now the probability wants to start moving towards 1.0, and that creates havoc with phylogenetic inference. As the German molecular systematist Johann-Wolfgang Wägele points out, “The fact that the probability of a multiple evolution of identical complex structures [e.g., the origin of mitochondria] is low does not mean that complex organisms should not evolve at all.” (Foundations of Phylogenetic Systematics [Pfeil Verlag, 2005, p. 148]) But that is a major problem, because now we must very carefully return the probability to the razor’s edge of absolute uniqueness – making it incredibly vulnerable. As Elliott Sober explains in his monograph Evidence and Evolution (Cambridge U. Press, 2008, p. 314), the less we know about how an evolutionary transition occurred, at the origin of character X – in terms of mechanistic detail – the more confident we can be that any organism possessing X shared a common ancestor with any other X-bearing organism. “[I]t is a point in favor of the common-ancestry hypothesis,” writes Sober, “that it says the evidence is very improbable.”
This is genuinely a deep and dark hole, and it exists because the vast majority of major evolutionary transitions in Earth history are not understood at all, with respect to how they actually happened. One has therefore the appearance of phylogenetic knowledge – e.g., “all chordates share a common ancestor, because the first origin of the chordate body plan was highly improbable” – only because of how little one really knows about the relevant events.
I agree with most of what you wrote here, especially, of course, the quotes from Wägele and Sober.
The part I (and most other evolutionary biologists) don’t agree with is the idea that very rare or even unique transitional events suggest a “deep and dark hole” for evolution. It’s true that we do not know all the details of mechanisms for many of the major transitions in biological history. But when we do discover such mechanisms we find that they are often one offs, with very low probability of occurrence. The probability of the MER20 retrotransposon inserting itself in a position to rewire gene regulatory networks in a way to contribute to the evolution of mammalian pregnancy is quite low, but it happened. In fact different transposable elements ended up serving as promoters for some pregnancy related genes (like prolactin) in different clades, leading to evolutionary convergence. In each case, probabilities are very low for the particular coin toss result. But as A. Wagner has shown for protein sequences and gene regulatory networks, biological structures show great robustness, meaning that the space of functional variants is vast.
This biological fact renders assumptions based on probability tenuous at best. Certainly, calculations that show a vanishingly low probability of life coming up with the modern sequence for any critical protein are utterly meaningless.
The point is that while some events might have only happened once, this does not rule out the possibility that a large number of similar events could happened instead. The one that did happen is the one we see; if it works well, there is no evolutionary good reason for it to happen again in exactly the same way.
I believe you made a good point in your answer to Dennis’ analogy of electrification of cities, and to the origin of the trucking industry. I agree that the missing ingredient in these examples was an intelligent agency. So lets consider a different analogy. Suppose you were to design a robot that could (using some form of AI and complex system engineering) improve its functioning. (such machines exist, and interestingly enough often use a form of natural selection called genetic algorithms to work). No intelligence is needed for their evolution, only for their original construction.
I don’t know how you could calculate any probabilities when they also make this statement:
“The origin of the mitochondrion was a singular event , and we may never know with certainty the early mechanisms involved in its establishment , nor the order of prior or subsequent events in the establishment of other eukaryotic cellular features.”
If you don’t know the pathways or mechanisms, how can you determine if this was a low probability or high probability event?
What paradox would that be?
You don’t seem to have a grasp on how probability works. If you shuffle a deck of cards really well and lay out the cards one by one, the order of those cards is unique in the whole history of the universe, and that exact order will probably never occur again . . . and yet it happened. In fact, every time you shuffle a deck of cards you produce a highly, highly improbable order of cards, and yet it is easy to do. Every moment of every day is marked by nearly infinite number of events that are one-off events, and yet we are able to get through our day without any problems.
What you seem to be stumbling on is called the Texas Sharpshooter fallacy. This is where you shoot a gun at a distant barn and then draw a bulls eye around the bullet hole. The probability of something happening after it happens is 1 in 1, because it happened. The evolution of the early eukaryotes could have gone in a nearly infinite number of directions, but it would have gone in one of those directions.
It is completely certain because it happened. Again, you are using the Texas Sharpshooter fallacy. There are nearly an infinite number of evolutionary pathways that early eukaryotes could have taken, and the odds of them taking one of those is nearly unavoidable even though each individual pathway is highly improbable.
It seems to me a good point that needs answer. It also seems to me that the debate here could be characterized as Dr. Nelson claiming that “the bull’s eye” seems to be “pre-specifiable” and that there are very few bull’s eyes (if indeed more than one). While you, @T_aquaticus, respond that there are probably (in fact must be?) infinite numbers of bull’s eyes enabling the arrow to inevitably hit one of them.
To use another analogy, each of our own existences is so improbable as to be impossible. Our genomes required two people out of billions, and even then it required a specific sperm out of billions and a specific egg out of hundreds of thousands to meet each other at just the right time. And that is just for each of us. The same improbability had to occur for each of the parents, and each of their parents, and so on.
Of course, given the human proclivity towards having children there would have been children no matter what. If we rewound the clock of history and restarted it, we would probably see different people as a result, and each of their existences would be just as improbable as ours.
I fully understand the concept of probability (and the sharpshooter fallacy). I’m not disputing any of that. But what I think is being suggested (I trust you will correct me @paulnelson58 if I misrepresent you here) is that there were not infinite possibilities for target locations so that a random arrow must inevitably land on one of them. He seems to be insisting that these “one-offs” can be demonstrated to be pre-specified targets (before any arrow has landed). So, to continue with your people analogy; he seems to insist it is like making up a photograph and all the details of a fictitious person ahead of time, and then gambling that the exact person made up will be born next year and live out the exact predicted life.
So it seems to me the dispute isn’t so much about probability, which I’m sure Dr. Nelson understands as it is you convincing him that there were multitudes of possible pathways that these “singular events” could have successfully taken. Your insistence that there must have been … because we are all here after all … doesn’t address his challenge. Because the question before us (as I see it) is: is life’s existence highly contingent on a knife’s edge event? Or is it a near inevitability with existing processes carrying on as they do?
[edited]
A separate question, but this seems like an interesting paper that seems like a review paper of sorts:
Also, I don’t see how anyone calculates the probability of any of this without absolute knowledge of conditions 3.5+ billion years ago and absolute knowledge of all the possible mechanisms that occur at the microscopic level and even with all of that… it seems quite odd to put a limit on what probability is ‘too small’ other than exactly zero.
The questions put to me about probability estimates for the origin of the mitochondrion via endosymbiosis are misdirected. I am observing how evolutionary inferences to unique (singular) events are made, and the consequences of those inferences for phylogenetic reasoning.
Lynch, Martin, and many other investigators claim to know that an event (i.e., the incorporation of an alphaproteobacterium within an unknown prokaryotic host) happened only once in Earth history. Ask them, not me, how they know this.