Friday, July 27, 2012

NASA Confusion About the Origin of Life: Part II

I'm going to keep hammering on this until it sinks in.

We don't know how life began but one of the more fanciful hypotheses is that it began in a primordial soup of organic molecules supplied by meteors, comets, and violent lightening storms. The idea is that the ocean was full of glucose, amino acids, and nucleotides. Glucose and similar carbohydrates supplied the energy for life. Amino acids spontaneously came together to form proteins. Nucleic acids arose by stringing together pre-existing nucleotides or nucleosides.

In the most extreme version, the ocean itself was the primordial soup and concentrations of organic molecules were sufficient to drive the formation of life. A simple back-of-the envelope calculation indicates that the concentration of typical amino acids would have been about 0.1 nM (10-10 M) [Can watery asteroids explain why life is 'left-handed'?]. This is an unlikely scenario [More Prebiotic Soup Nonsense]. You won't get spontaneous formation of polymers in water at that concentration.

I prefer another hypothesis called "Metabolism First" [Was the Origin of Life a Lucky Accident?] [Metabolism First and the Origin of Life]. (Bill Martin is one of the leading proponents of "Metabolism First." He can be a very persuasive guy!)

In this scenario life arose in restricted environments around thermal vents. The triggering force was the presence of natural proton gradients that could drive spontaneous chemical reactions. This energy source resulted in the formation of simple compounds like glycine, acetate, glcerol. These simple compounds serve as substrates for the synthesis of more complex molecules.

"Metabolism First" postulates that "life" bootstrapped its way up from very simple molecules rather than arising in a primordial soup where those molecules already existed. One of the great advantages of "Metabolism First" is that is offers a reasonable explanation for the preferred chirality of amino acids and sugars. The preference for L-amino acids, for example, may have arisen naturally from primitive catalysis reactions where the original product just happened to be L-alanine or some other simple amino acid [Amino Acids and the Racemization "Problem"] [Nicholas Wade on the Origin of Life ].

NASA has been funding investigations into the origin of life. Some of those projects have been spectacular failures while other failures have been less spectacular. Those that purport to explain the chirality "problem" suffer from a serious flaw because they make the assumption that a preference for L-amino acids has something to do with the concentration of amino acids in asteroids, and therefore the concentration in the primitive ocean. I've pointed out the flaw in those assumptions before [NASA Confusion About the Origin of Life].

NASA has done it again: NASA and university researchers find a clue to how life turned left.
Researchers analyzing meteorite fragments that fell on a frozen lake in Canada have developed an explanation for the origin of life's handedness – why living things only use molecules with specific orientations. The work also gave the strongest evidence to date that liquid water inside an asteroid leads to a strong preference of left-handed over right-handed forms of some common protein amino acids in meteorites. The result makes the search for extraterrestrial life more challenging.

"Our analysis of the amino acids in meteorite fragments from Tagish Lake gave us one possible explanation for why all known life uses only left-handed versions of amino acids to build proteins," said Dr. Daniel Glavin of NASA's Goddard Space Flight Center in Greenbelt, Md. Glavin is lead author of a paper on this research to be published in the journal Meteoritics and Planetary Science.
The problem with this work is not that it's bad science. It's that it automatically assumes a particular view of the origin of life and doesn't take into account the idea that the results may be irrelevant. Even worse, the results consistently fail to provide a reasonable explanation for the amino acid chirality "problem" but it never seems to occur to the researchers that their hypothesis has been falsified. Maybe it's time to look for another explanation?


  1. One of the great advantages of "Metabolism First" is that is offers a reasonable explanation for the preferred chirality of amino acids and sugars.

    Whatever other merits of the idea, I think a better explanation of chirality lies within the stereochemistry of the macromolecules which these chiral subunits make up. Regardless of the form pre-nucleic acid chemistry took, the regularity of DNA and RNA helixes places a constraint on the ribose isomer that works in that context. Replication of the chains would probably be by means of some kind of macromolecule, and to the two catalytic kinds we know, nucleic acid and protein, isomers generally look like chalk and cheese. So even if the original source were environmental and racemic, cells would settle upon one kind as their macromolecular arsenal grew. Likewise with the catalysts of synthesis. To the extent that other sugar pathways were built upon ribose-metabolism precursors, they would inherit their isomeric preference.

    A similar argument would apply to amino acids. The ribosome itself, the enzymes/ribozymes of acid synthesis and tRNA charging etc would all pass on a distinct handedness inherited from that of ribo/enzymes synthesising or using the earliest chiral acid of biological origin or biologically 'filtered' from a racemic mix.

  2. Larry- in this scenario for the origins of biological homochirality, what's the nature of this first catalyst of L-alanine. Specifically, AFAIK, enantioselective catalysts must themselves be chiral, so where did this chirality originate?

    1. David, ANY molecule complex enough to be catalytic is probably asymmetric. That's all it would take.

    2. So, clarification: by asymmetric you mean chiral? A planar molecule can be asymmetric along 2/3 spatial axes but is still achiral.

      I wouldn't say a catalyst has to be complex, heck single metal ions can be catalysts; but for enzyme-like stereoselective synthesis of L-alanine, sure, I'll grant that would probably be complex. But that doesn't seem to me to solve the problem-complex or simple, if your catalyst is enantioselective it has to be chiral. That seems like an "assume we had a can opener" solution: postulating a chiral catalyst to explain homochirality seems problematic, right? How did that catalyst get chirally enriched?

    3. David,

      You could imagine that equal mixtures of L- and D-alanine were synthesized but the first catalytic peptides just happened to include L-alanine and the first major product to be synthesized was L-alanine.

      From there it's easy to see why all other amino acids might be L- forms.

      I don't know if this is what happened but it seems much more plausible to me than surmising that a preference for L- amino acids arose in a dilute soup of dozens of different amino acids each of which had both L- and D- forms.

      Do you have a better "primordial soup" scenario?

    4. ...if your catalyst is enantioselective it has to be chiral [...] How did that catalyst get chirally enriched?

      If one had a hypothetical macromolecular catalyst of mixed L and D subunits, in any proportion, it would typically be enantioselective by default. Any biogenic subunits arising from such catalysts would be of one form only, and conversion pathways would likewise favour that form. There's nothing to stop the opposite isomer also having pathways, but nothing to guarantee it either, while stochastic factors would act against a tidy 50/50 balance in usage of these forms in the catalysts themselves. Catalyst composition would, of course, have to evolve in tandem with the pathways of subunit synthesis, but one would expect them to evolve away from hard-to-glean environmental sources.

      My own preference is for nucleic-acid-first, though I know there are biochemical issues with that. (And it is near-impossible to visualise how a non-chiral xNA molecule could replicate - that constraint could be the ultimate source of sugar chirality). The isomeric mix in which proto-biology 'actually' arose compounds the already-tough Origins problem. No good waiting for something to crawl out of your jar of D-ribose monomers if this can't happen.

      Amino acid chirality may also come from ribozyme asymmetry, if early catalysts were nucleic acid. There is also the asymmetry of the ribosome itself - but that probably reflects pre-existing synthesis of L acids, whatever they were used for prior to the invention of the ribosome.

    5. My apologies, but I'm at a conference with sporadic email access, so detailed responses may take a bit. But briefly:
      1. I think you guys are vastly overestimating the potential for short peptides to accomplish highly stereoselective catalysis. Modest levels? Sure, but a peptide that catalyzes transamination with >90%ee? And what level is needed to drive things toward homochirality?

      2. I would say even primordial soup fans agree that the current crop of biological amino acids is derived from a much smaller subset. There may be a fraction who try to demonstrate chemical methods to make L-arginine or tryptophan, but most understand that this is silly as an argument for the prebiotic origin of these amino acids.

      3. The mechanism of chiral amplification you are proposing isn't impossible, but is not experimentally demonstrated or very plausible. Asteroid delivery of chiral amino acids isn't either, but those studies do show chiral bias can occur in complex abiotic environments, and work like Donna Blackmond's has started to show general methods by which this can happen.

    6. We don't have many facts and there's a lot of bias in what we have. Any answers are highly uncertain.

      I like nucleic-acid-first because transcription is built around RNA with proteins propping them up. If we had lots of proteins when we were learning how to transcribe proteins, why not build ribosomes mostly from proteins? Instead we built ribosomes from RNA and later improved it with proteins.

      The genetic code is a gray code. Single base mutations tend to have minimal effect because similar amino acids have similar codes. Maybe that's a spandrel and it just had to be that way. But if we once had multiple genetic codes and this one won out -- that seems like a whole lot of selection.

      Maybe we used to have a two-base 15-amino-acid code with ornithine, and later we switched. It takes a whole lot of adaptation to build two successive genetic codes. I don't know how long that would take but my intuition says a long time.

      But probably we got life soon after we had liquid water.

      So my prejudice says it probably didn't happen here.

      When our sun first formed, how much gas and debris was around it? There would be a band where liquid water was possible, assuming the gas pressure was high enough. Maybe for awhile the gas pressure was that high. Then we could have water droplets of various size stretched all around the earth's orbit. A volume perhaps larger than 300 billion cubic miles. Depending on how quickly the droplets separate and coalesce, life that started one place might spread around the whole ring. So evolution could somewhat trade space for time. Later some life might survive in comets etc while the earth cooled enough to start a tiny puny renaissance in our own dinky ocean.

      The really nice thing about this idea is that we know much less about the solar system back in those days than we know about the earth after the oceans cooled and before there was life. So it's much harder to falsify.

  3. You wrote: "A simple back-of-the envelope calculation...."
    This would yield an average concentration, right? If so, then isn't it possible that the local variations in concentrations resulting from various physical processes, currents, tidal basins, etc and so on, could have created the concentrations necessary?

  4. @David,

    You're thinking about bulk solutions, where we expect an equal mixture of both enantiomers. Maybe if you thought of single molecules...

  5. I think you guys are vastly overestimating the potential for short peptides to accomplish highly stereoselective catalysis.

    Well, if it started with peptides, then yeah ... but if it started with nucleic acid, which also has a noncatalytic requirement for sugar homochirality ... the problem does not disappear, of course - just substitute 'short ribozyme' for 'short peptide'.

    But I still don't think we necessarily have to look to look outside of biology. We are seeing everything through the 'filter' of LUCA. Its lineage had long since fixed a nucleic acid replicative system, and biosynthetic amino acids/sugars. Biosynthesis is an inevitable source of chirality in these monomers. As these pathways evolved and diversified, they retained their ancestral chirality as different side chains were glued on/modified.

    Having said which ... I don't propose to solve here the problem of how a replicative nucleic acid with D-ribose arose in a racemic world, soupy or otherwise!

  6. @Larry–

    OK, this is a bit late, and I’m not sure if anyone’s still reading comments here but to expand on my earlier statements–

    Just to be clear, I’m not in favor of prebiotic soup scenarios over metabolism first - both make some things easier to explain, but both have big glaring problems too. For balance, I’ll mention Leslie Orgel’s last published article “The implausibility of metabolic cycles on the prebiotic Earth” in PLoS Biology, which summarizes some of the problems with metabolism first theories, but is quite clear on how they would help if proven to work.

    In regards to homochirality, my issue is that this “catalyst that enantioselectively makes L-amino acids” model is not particularly plausible, and certainly not much better than many others under consideration. As I first alluded to, the question of how you get a chiral catalyst to start things off isn’t trivial, and nor will be the question of how initial symmetry breaking gets amplified into homochirality.

    In your model as I now understand it, initial stochastic variation from sampling bias yields a non-racemic amount (as low as 1 molecule) of a catalytic peptide. The peptide selectively favors formation of L-alanine, and is itself partially composed of L-alanine, thereby enhancing its own formation by biasing the L-alanine concentration. Further amplification may happen if it or other peptides can catalyze stereoselective synthesis of other amino acids; once a few amino acids reach homochirality, all products synthesized downstream, including many other amino acids, will retain the same chirality due to stereospecific catalysis.

    My points are 1) What kind of peptide would we need to get to start this off? How many peptides must be synthesized to get 1 stereoselective transaminase, and how active and how specific will it need to be to trigger downstream chiral amplification? A catalyst that’s only changing alanine ee by a few percent can’t bias its own synthesis very well. 2) Even if the peptide is skewing alanine ee significantly, how specific are the reactions producting the peptide – both in the sense of sequence and chirality? Unless both are high, only a small fraction of the alanine ee makes its way back to catalytic peptide synthesis. 3) Further downstream amplification also doesn’t seem obvious – how many different amino acids do we need chirally enriched? Even if one happens first, why should the production of other chirally selective catalysts favor the same stereochemistry in new amino acids?

    We don’t have much experimental evidence to show that any of these issues can be resolved, and what we know of chemistry suggests there are a lot of potential pitfalls. Models for delivery of non-racemic amino acids from space aren’t terribly plausible either (though I wouldn’t be so terribly swift to dismiss local concentration of space-delivered organic compounds-that may be implausible, but on the face of it not much worse of an issue than some I brought up above). That leaves us with not much evidence to favor either scenario (at least from the perspective of better explaining homochirality).

    The areas that I see as being most productive for understand the origins of homochirality are actually pretty neutral with respect to soup vs. metabolism questions. We now know there are plenty of mechanisms that take even small stochastic deviations from 0% ee and amplify it up to 100% ee (Donna Blackmond has had a few relevant reviews in the last few years). These mechanisms might explain the results being found in asteroids, and could just as easily have happened on early earth, in the ocean or in hydrothermal vents; starting with strongly enantioenriched organic substrates helps both prebiotic soup and metabolism first theories. That certainly won’t solve all questions, but its finally moving to experimentally verified scenarios for chiral amplification.

    1. Even if one happens first, why should the production of other chirally selective catalysts favor the same stereochemistry in new amino acids?

      If you had a pathway for making glycine, and a derivative attached a side chain in place of one of the hydrogens, further derivatives would either modify that side chain, or attach a different one at the same place. Gradually, a 'library' would build up by descent. The usage of these amino acids does not have to be, in the first instance, the manufacture of catalysts.

      It may, or it may not, be the case that there was, pre-LUCA, a mechanism for specifying peptide catalysts containing both L and D forms. Equally, early catalysts could have been ribozymal, or nucleic acid-peptide complexes. But in all cases, they would be asymmetric, and initial biosynthetic chiral acids would therefore be one enantiomer or the other, if there was any side-chain specificity deriving from catalyst shape.

  7. @Rosie
    Mostly see my post to Larry above. Stochastic variations can break symmetry, of course, but you still need a plausible method to amplify the initial variation. It’s not at all clear a short peptide will be active enough or selective enough to do so.

  8. @Allan
    Again, if a pathway, even a prebiotic one, is making nonracemic chiral amino acids, it is catalyzed by a nonracemic chiral catalyst.* Setting aside the issue of how active and selective such prebiotic catalysts could be, a nonracemic chiral catalyst can arise either by chance due to low total number (if you only have 1 molecule, it can either be L- or D-, not racemic) or by specific synthesis by some other pathway. In the former case, that randomly synthesized catalyst will be destroyed or diluted away rather quickly - it must be amplified to remain effective, which means that its activity must feed back into its production, or in this case, amino acid synthesis must increase specific production of the catalyst, which is not all that plausible for reasons I've discussed above. In the latter case, the specific synthesis of a chiral catalyst must again use another chiral catalyst - the origin of homochirality has just been displaced to an earlier event.

    But in all cases, they would be asymmetric, and initial biosynthetic chiral acids would therefore be one enantiomer or the other.
    Again, restating myself here, but 1) these catalysts must be specifically chiral, not just asymmetric, so we must explain how they were synthesized stereoselectively, or how very small amounts could be amplified subsequently; and 2) one enantiomer or the other are the extremes, real world catalysts, unless they are extremely complex (i.e. enzymes), achieve something in between, some fraction L and some fraction D; now the question is, how much bias is needed? The answer does not seem obvious, and is unlikely to be anything above 0 bias.

    * Another possibility, as I mentioned in a previous comment, are some chemical and physical processes which do not have a built in stereoselectivity, but rather take any small deviation from racemic mixtures and amplify it to high enantiopurity. These are cool, they're a promising lead, but they need to be fit into and tested in the context of specific OoL models, and they don't really relate to the chiral catalysis scenarios we've been discussing here.

  9. David, I'll repeat as much of your material as fits my needs, to see if I understand it.

    You point out that biological materials are usually either right-handed or left-handed, because that's how biological enzymes make them. Usually these enzymes are the only way we know to make one or the other instead of a mixture. So before there was life, there would have been racemic mixtures, and the first life would have to deal with that. It might have to be made of racemic stuff because that's what there was for it to use.It couldn't make right-handed enzymes to make only right-handed components, unless it already had right-handed components to make them from. Every attempt to explain it away, just takes the problem a step farther back and leaves it unsolved.

    My response is to take it as far back as possible, and leave it there. Assume a molecule that can replicate itself, slowly and with lots of errors. Assume an environment which gives it everything it needs to do that. I'd call it RNA because that's the only candidate I know of, but it could be something unknown that later created RNA. This molecule sometimes replicates, and sometimes it folds itself into shapes that can somehow aid its replication. Let's say that most of the time when it replicates it fails and creates something else, but 1% of the time it makes a good copy. Rather, it makes a good copy often enough that in some time period the population increases 1% despite the random losses. Eventually its numbers could rise pretty high.

    I want to hypothesize that it was all one form and not the other, because I can't see mixed D and L RNA backbones doing easy reproduction. It follows that if it could produce good copies 1% faster than they're distroyed, it would be selected to find ways to incorporate the correct form and avoid the lethal form.

    So the useful form of monomer would drop in concentration, leaving more of the lethal stuff. Forms that better avoid the bad monomers will get selected, and each time they do the concentration of good monomers goes down until nothing else can survive. Forms that convert bad monomer to good monomer would be selected, because the concentration of good monomer would be higher around them. Between selection for producing good monomer, versus solution for being close to a molecule that produces good monomer, which does better?

    Once you get a single molecule that can preferentially reproduce itself using a single enantiomer, then the rest follows. I don't know how unlikely it was to get that first one. I don't know how an environment could be created that encouraged random nucleic acids to collect and join together. But the self-reproducing molecule seems like it would be the first step to life. All the random catalytic structures that happen before that would be just random catalytic structures that might happen repeatedly if they are simple enough to be repeated, but which could never get beyond that step. If reproduction is the central thing it has to do, then reproduction is the behavior it has to perform in a racemic environment. That plausibly does not work with racemic components, so that plausibly is the initial behavior that has to distinguish between different monomers.