Amino Acids and the Racemization "Problem"

Amino acids come in two different "flavors" depending on the orientation of atoms bound to the central α-carbon. The two possibilities are L- and D- configurations. In the examples shown here, you can see that the two forms of serine, L-serine and D-serine, are mirror images of each other. These forms are called stereoisomers because they contain the same atoms with different mirror image arrangements.

Stereoisomers cannot be interconverted without breaking covalent bonds. They are distinct molecules. Almost all amino acids in living organisms are L-amino acids. Proteins are almost exclusively composed of L-amino acids and not D-amino acids.

The α-carbon atom of amino acids is chiral, or asymmetric. You need at least one chiral atom in a molecule in order to have stereoisomers. One amino acid (glycine) does not have a chiral α-carbon so there is only one configuration of glycine.

When amino acids are synthesized in a chemistry laboratory, you often end up with a mixture of equal amounts of L- and D- stereoisomers. When you examine the amino acids found in meteorites and in the vicinity of stars, you find a mixture of both stereoisomers. These are called racemic mixtures since the process of converting one stereoisomer into another is called racemization.

Now, the fact that amino acids in living organisms are all L- forms is not a problem since the L-amino acids are the only ones that are synthesized in any appreciable amounts. All of the amino acid biosynthesis pathways produce only L- forms and not D- forms. This is not unusual since enzyme catalyzed reactions are usually sterospecific. It's not a surprise that modern proteins are composed of L-amino acids because those are the only ones available inside the cell.

The "problem" arises when we start to think about how life arose in the first place. The general assumption is that life arose in a warm pond containing a racemic mixture of L- and D- amino acids. If that is true then how did life evolve to select exclusively L-amino acids? Most of the proposed solutions to these questions make assumptions about how the primordial soup could have spontaneously come to have a preference for L-amino acids over D- amino acids.

I'd like to propose another way of thinking about this problem.¹

Let's assume there was a primordial soup where amino acids came together spontaneously to form short peptides. In the beginning, the soup contained racemic mixtures of the D- and L-forms of amino acids. These molecules were formed spontaneously by the kinds of chemical reactions that are simulated in the laboratory.

Some of the random peptides acted as catalysts for chemical reactions. This is observed in modern-day experiments. One kind of reaction, amino acid synthesis, would have been especially favorable since it created more amino acids and led to more peptides.

The simplest pathway to more amino acids is the formation of glycine, probably by adding an amino group to acetate or glycerol. (This pathway no longer exists.) The next simplest is the conversion of pyruvate (a common three carbon organic acid) to alanine—a fairly simple transamination reaction.

In modern cells, this reaction is catalyzed by sophisticated transaminases but in the beginning it would have been catalyzed by short peptides that formed spontaneously in the primordial soup. Such reactions are stereospecific, the modern reaction only produces L-alanine and never D-alanine (well, hardly ever!). Let's assume that a similar reaction in the beginning produced, by chance, L-alanine.

Another simple pathway is from oxaloacetate (a common four carbon organic acid) to aspartate. Both of these reactions require a relatively simple addition of ammonia to a keto group and both reactions could have been catalyzed (inefficiently) by the same peptide.


As I mentioned above, enzyme catalyzed reactions tend to be stereospecific so it's likely that the early products were L-alanine and L-aspartate from the same enzyme. They could have been D-alanine and D-aspartate, but they weren't. As the concentrations of glycine, L-alanine and L-aspartate increased there were more and more peptides formed and the new peptides were enriched in these two particular L-amino acids.

Other simple amino acid synthesis reactions were catalyzed in the primordial soup. The most likely one is the synthesis of serine from glycerol or glycerate (common three carbon organic alcohols or organic acids). Again, the enzyme catalyzed reactions will only produce one isomer of the amino acid and there might have been selection for those parts of the soup that made L-serine (instead of D-serine) because the L-serine could more easily combine with L-alanine and L-aspartate to make many more peptides. In this case, the specificity of the reaction derives from selecting D-glycerate over L-glycerate as the substrate.

L-serine is the precursor to L-cysteine so it's likely that L-cysteine was also one of the early amino acids to accumulate in the primordial soup. This was an important addition to the repertoire since L-cysteine has a sulfur group and that leads to many more possibilities for catalytic active sites in the peptides. Note that once L-serine began to accumulate in the soup it led directly to the stereospecific L-cysteine. You can't make D-cysteine from L-serine so there's no racemization problem once L-serine accumulates.

L-glutarate (from alpha-ketoglutartic acid, a common five-carbon organic acid) is another good candidate for the primitive amino acids. (It's quite possible that L-alanine, L-asparate, and L-glutamate were all made by the same primitive enzyme using very similar 3, 4, and 5-carbon substrates.)


At this point there would have been all kinds of peptides containing various combinations of L-alanine, L-aspartate, L-serine, glycine, L-cysteine, and L-glutamate since these six amino acids have become much more abundant that the ones formed spontaneously by uncatalyzed reactions that produce a racemic mixture. This is probably the time when there was a shift to encoding peptides in a sequence of nucleotides.

This is an important point. The shift to more and more complex peptides did not have to take place in a random mixture of both forms of all 20 amino acids. It could have taken place under conditions where there was already a significant enrichment of a small number of L-amino acids due to catalytic biosynthesis from non-amino acid precursors.

There's some suggestive evidence to indicate that the primitive genetic code was much simpler than the one we see today and may have only had codons for the six initial amino acids. The other L-amino acid synthesis pathways arose later on and the genetic code expanded when codons were "stolen" from the precursors of these new L-amino acids.

One of the primitive codons for aspartate, for example, might have been AXX (any codon beginning with A). L-aspartate is the precursor to: L-lysine (AAA, AAG), L-asparagine (AAU, AAC), L-threonine (ACX), L-methionine (AUG), and L-isoleucine (AUU, AUC, AUA). The idea is that the new amino acids were originally synthesized on L-aspartate that was attached to its tRNA and they were incorporated into proteins at some positions in place of L-asparate. (This hypothesis on the origin of the genetic code was developed by my former colleague Jeff Wong. The idea came to him while teaching an undergraduate course in biochemistry ... but that's another story.)

Note that many amino acids are made from pre-existing amino acids. Once you have a supply of L-aspartate, for example, it follows that the derivatives will also be L- forms. There's no need to postulate that the preferential use of L-asparagine, L-threonine, L-methinione, and L-isoleucence, in contrast to the D- forms, arose independently. This greatly reduces the probability problem that most people are hung up on.

I don't have any good ideas about how the transition to encoded peptides happened but that's not the real point of this speculative posting.

The real points are ....

1. The most primitive catalysts were probably not very big. They were probably composed of mixtures of L- and D-amino acid residues.
2. The first important step was synthesis of new stereospecific amino acids which meant that the process was no longer dependent on the original pool of compounds that formed spontaneously.
3. The first peptides and polypeptides (proteins) probably contained only six amino acids. These are the amino acids that can be easily made from readily available precursors.

If you think about the origin of life in this way it will help you to understand why biochemists don't think the "racemization problem" is a real problem. This scheme will also help you to understand why some *particular* amino acids came to be enriched in proteins and not all of the other amino acids that were in the primordial soup in the very beginning. (There are far more than 20 amino acids.)

The original choice of the first L-amino acids over their D-isomers was probably an accident. It could just as easily have been the D- amino acids.

UPDATE: I now believe that Metabolism First and the Origin of Life is a more likely explanation for the origin of life. Please ignore references to "primordial soup" in the essay above. My conversion doesn't change the point. In the beginning very simple amino acids were spontaneously synthesized in restricted environments around thermal vents. By chance, the first chiral amino acid, alanine?, may have been L-alanine. All other may have been synthesized using L- amino acid precursors and this explains the the racemization problem.

---

1. This is a modified version of an article that was originally posted on talk.origins in January, 2004.