Yet another reason why the Wolfe-Simon conclusion is so improbable

Experimental results need to be evaluated from two perspectives, the quality of the data and the probability of the explanation.  I and many others have critiqued the Wolfe-Simon results based on the poor quality of the data, and chemists have critiqued them based on the predicted instability of arsenic bonds in a DNA backbone.  But I don't think anyone has spelled out the improbability from an evolutionary perspective.  So here goes...

We're told that the authors originally thought they might have cultured a member of the hypothetical 'shadow biosphere'.  If such organisms existed they would be descendants of an evolutionary lineage that has been evolving independently of known organisms for billions of years.  Known organisms all belong to a single lineage, and the 'shadow biosphere organisms might either have originated independently of this lineage (A in the sketch below) or diverged from it shortly after its origin (B in the sketch below).

The organisms contained DNA, and PCR amplification using 'universal' primers for the 16S rRNA gene was successful.  This meant that the organisms could not be members of a type A shadow biosphere, as an independent origin wouldn't have produced either DNA or 16S rRNA genes.  But placing the 16S rRNA sequence in a phylogenetic tree showed that the organisms weren't members of a type B shadow biosphere either.  Instead they were members of a conventional group of bacteria, indicated by the dark wedge in the bacterial triangle.

Finding that a shadow biosphere organism had DNA with a backbone containing only arsenic (no phosphorus) would have been very surprising, both because we have no evidence that the shadow biosphere exists and because of the predicted instability of the arsenic ester bonds.  However, the claim that a conventional bacterium can function when its DNA backbone contains a mixture of arsenic and phosphorus is even more improbable.  Here's why:

Many different enzymes interact with DNA, and almost all of them need to do so with high precision.  They can do this because the DNA backbone has a very consistent structure.  Although the four bases that attach to the backbone have different shapes, they can occur with any sequence only because the nature of the base pair does not alter the backbone geometry.  The enzyme DNA polymerase is especially dependent on precise interactions with DNA, because it is responsible for accurate replication of the genetic material.  Typical measures of the fidelity of DNA replication by bacterial DNA polymerases show an error rate of about 10^-9.

The covalent radius of arsenic is about 11% larger than that of phosphorus (see this periodic table tool), and the As-O bond lengths reported in Table 3 are about 13% longer than the P-O bond lengths in DNA (Table S3).  No DNA polymerase could function on a template with such unpredictable geometry.  Even if it could bind to the template strand and proceed along it, the differing bond lengths it encountered would prevent the incoming nucleotides from forming the necessary bond with the growing end.   And even if it could create the backbone of the daughter strand, the error rate due to base mispairing would be so high that reproduction would fail.

Similar problems would affect every enzyme that contributes to production of the DNA and RNA precursors, to DNA and RNA synthesis, to DNA repair, and to energy metabolism.  No matter how stable the arsenic-substituted metabolites might be (and the chemists tell us they are mostly very unstable), the proteins that need to interact with them would not be able to cope with the unpredictability of usinging both arsenic and phosphorus versions.

The fundamental problem isn't that millions of generations would be needed to evolve enzymes that could tolerate a mixture of arsenic and phosphorus, but that the required adaptations are not compatible with the precision needed for function.