It's a very interesting time for origin of life speculation. You can look around at the work coming from Lane's group as well as the excellent work by Koonin's group and consider what fits together logically from the various scenarios on offer. None are completely satisfying alone but a mixture of the ideas gives a very pleasant brew. Let's give Koonin some time now:
How do you build an ATP synthase complex? On the surface it looks to be utterly incomprehensible that a lipid embedded, ion gradient driven rotor should turn a stalk to a set of molecular forceps each of which forces a phosphate on to an ADP to give ATP as the rotor turns. It's particularly striking that Nick Lane considers this to be one of the core molecular machines present in LUCA before the separation in to the populations which gave rise to the archaeal and bacterial lineages. Running on a proton gradient. It would need to have been a very early development.
If we buy in to the alkaline hydrothermal vent scenario (Koonin doesn't) for the origin of life we have microscopic "pockets" of metabolism in iron (+/- nickel) sulphide walled chambers at the boundary of alkaline vent fluid with acidic ocean fluid. In the progress towards something akin to life there has to be the development of both protein and RNA. In a vent system, where conditions at the microscopic scale vary from time to time, the ability for a given replicator (RNA based) to spread itself over the maximum number of protocells would provide a survival benefit. The development of any sort of cell membrane is clearly a hindrance to this process, potentially terminal. The formation of simple pores of protein derivation are not inconceivable and would allow the continued spread of successful RNA to a maximum number of protocells. Pores do not have to be complex. RNA spread can be concentration based.
RNA will form double helices. Not with the solidity of those derived from DNA, but double helices never the less. There is a large family of proteins which unwind helices to allow replication, be that of DNA or RNA. Many of them use ATP to facilitate the process. They are very basic machines by evolutionary standards and there is no reason why they were not as available early in evolution as simple membrane pores.
Bear in mind that one of the core products of protometabolism is an acetyl thioester capable, energetically speaking, of producing ATP (was that as long ago as February?).
If we have these basics in mind we are in a position to go to this rather nice speculative opinion piece by Koonin's group. Here is figure 2:
Across the top we have the possibility of an ATP driven RNA helicase sticking itself to a membrane pore and using ATP to power RNA export through a proto membrane bound protocell. The helicase, if it sticks itself to the membrane pore, will extrude an RNA strand out of the protocell. As any helicase passes down an RNA strand, it rotates. If the helicase is fixed to the pore, the RNA will rotate as it passes through the pore. The right hand diagram shows the next development with an RNA strand replaced by a protein strand, not an impossible transformation and still potentially of benefit to a proto replicator.
Working along the middle row from right to left we have a protein translocator using ATP to export a protein using the same rotary machinery inherited from the RNA helicase/translocase. The + and - within the circles represent electrostatic charges which normally hold the pore and helicase stationary relative to each other. The middle diagram has a red protein which is physically stuck in the pore. The rotary component, powered by ATP, pushes/twists against the membrane bound section while the electrostatic charges resist rotation. Something has to give and in the process an ion or more is squeezed outwards as the pore derived section moves against the stator section.
The machine extrudes ions by consuming ATP. Ion gradients have many uses to protocells so the development may well have been advantageous at the time. The machine also works perfectly well in reverse so, given an appropriate ion gradient, it will generate ATP as ions enter the protocell. The ion gradient itself is a whole different ball game. Just for now recall that in those archaea and bacteria which lack cytochromes, the gradient is of Na+ ions, not H+ ions. Another post there.
This is a plausible derivation for the bacterial/mitochondrial F type ATP synthase.
The archaeal V type ATP synthase has very similar ion pump and ATP synthase sections to the F type, but the central stalk is quite different. In fact it doesn't look much like a stalk at all, see the bottom left diagram in the above figure 2. It looks more like a protein used to stabilise the pore and helicase sections against each other. It does the same job as the central stalk of the F type synthase but its origins are clearly different.
The ATP synthase complex was developed twice.
It looks very much as if the pore/helicase combination was ubiquitous in LUCA but the system of locking them together is different in the archaeal vs bacterial lineages. That seems very profound to me when thinking about LUCA.
The scenario does not have the ATP synthase as primordial. The system evolved as an ATP consuming machine, not an ATP generating machine. It has no hallmarks of the core power producing system in LUCA. Where did the ATP come from to power the engine which was eventually to become the dynamo?
I commented in my post on the reduction of CO2 to CO that the acetyl thioester produced was quite capable of substrate level phosphorylation. It is able to produce molecules in the energetic range of ATP and undoubtedly did so, because we are here today we are and running on ATP.
ATP is undoubtedly what worked best for the RNA helicase as it appears to have carried this preference forwards to the protein translocase and hence the ion pump which reversed role to become a dynamo. If the helicase had been powered by acetyl phosphate rather than ATP I can see no reason why metabolism might not be based around such a molecule. We use ATP because that's what worked best in the motor which, once it reversed to become a dynamo, unleashed a ubiquitous supply of energetic substrate based on an ion gradient.
Just to recapitulate: A proton gradient is essential to power protometabolism because without it you cannot reduce CO2 to CO to develop acetyl thioester and its assorted phosphorylated derivatives. This substrate level phosphorylation is utterly proton gradient driven but does not require proton translocation per se, although proton translocation may have been used to produce a localised environment which reduced ferredoxin. A source of direct ATP synthesis was required to power an immediately useful translocase which allows development of an ion pumping motor which could subsequently be reversed to provide an ATP generating dynamo as the membrane energetics changed.
The change in membrane energetics was probably related to a biologically generated Na+ gradient in addition to the geochemical H+ gradient. If the sodium ion gradient was supported by anti porting on a geothermal H+ gradient, it was both free and nearly unlimited in availability. So we have to think about antiporters.