Proof-reading and specificity
Sun, 31 Oct 1999 15:39:24 EST

As I have argued before, I think specified complexity is indeed
positive evidence of design. So are there any good examples of
specified complexity amidst life? Identifying complexity is
usually not a problem, as one only needs to determine the
number of interacting components needed to generate a
biological function. And modern biology continues to surprise
scientists by finding that life at the molecular and cellular
level is more complex than previously thought. But specificity
is turning out to be just as characteristic of life. For example,
even in bacteria, we are finding more and more examples
where the *location* of the proteins is essential to the

However, I'd like to consider two recent scientific papers that
outline a very good example of the importance of specificity.
These two papers have to deal with the phenomena of proof-reading.
As I see it, proof-reading is prima facie evidence of specificity,
as the very act of proof-reading exists only *because* of the
importance of specificity. If specificity was not important,
proof-reading would be irrelevant.

Synthetase Structure with tRNA and Murpirocin. Science 285;

I find it quite fascinating to realize that the phenomena of proof-reading
is more common than once thought. Most students of molecular biology
learn that DNA replication involves proof-reading, where misincorporated
bases are removed before synthesis of the newly-forming strand of DNA
proceeds. Yet proof-reading also seems to be involved in the charging
of tRNAs (when the proper amino acid is attached to the proper tRNA).
An extreme example concerns the tRNAs for isoleucine, threonine,
and alanine, where amino acids must be discriminated on the basis of
one methyl group. Pauling estimated that the enzyme which attached
isoleucine to its proper tRNA could only distinguish between isoleucine
and valine by a factor of 5, yet fewer than 1 in 3000 errors occurs. This
has always suggested a proof-reading component to charging, whereby
mismatched amino acids are removed.

Silivan et al. have used crystal data and modeling to propose one
such proof-reading mechanism. It turns out that the enzyme that
attaches amino acids to their appropriate tRNAs have two active sites,
a synthetic domain and an editing domain, separated by about 34
angstroms. The basic story seems to be that the acceptor stem of the
tRNA (the region where amino acids are bound) can adopt two
conformations, such that mismatched amino acids end up in the editing
domain where hydrolysis takes place and properly matched amino acids
remain in the synthetic domain. The exact mechanism of this shifting
between two domains is not currently understood.

What is interesting about this study is not just the proof-reading that
takes place, but that the general mechanism of proof-reading is analogous
to that which is used by the DNA polymerase when it synthesizes DNA.
If the Silivan model holds up, this means that the process of DNA replication
and tRNA charging employ a similar mechanism not due to common descent.
Put simply, we see an abstract engineering-like principle at work here, where
the same basic logical strategy (a dynamic competition between synthetic
and editing functions) is being employed in very different proteins carrying
out very different processes. For me, it is hard to resist the subtle
implications of design, where not only is proof-reading itself an echo of
design, but that the same basic logic of a proof-reading mechanism is
employed in different systems only serves to amplify this echo.

by Ribosomal RNA. Science 285; 1722-1725.

Proof-reading not only occurs when DNA is replicated and when tRNA
is charged with amino acids, but also seems to be involved in the codon-
anticodon interaction that takes place between mRNA and tRNA during
the process of translation (protein synthesis). Translation takes place in
the ribosome and the proof-reader appears to be 16S rRNA (a component
of the ribosome). Various experiments have found that there are two
residues in 16S rRNA that are essential, such that mutations of these
residues are lethal. The residues are two adenines, at positions 1492
and 1493.

Yoshizawa use an elegant approach to uncover the reason why these
residues are so important - they apparently form the core of the proof-
reading. Here's the basic story. The adenines at positions 1492 and
1493 are positioned in the A site of the ribosome (the region where new
tRNAs shuttle in to contribute their amino acids). Charged tRNAs bind to
mRNA in the A site through the interaction between the 3-bp codon of
the mRNA and the 3-bp anticodon of the tRNA. If the correct tRNA
binds to the correct codon, the 2' OH groups on the mRNA can form
hydrogen bonds with the N1 component of the adenines at positions
1492 and 1493. This is thought to stabilize the condon-anticodon
interaction and thus transmit changes to further the elongation process.
On the other hand, if an improper codon-anticodon interaction takes
place, the mispairing of bases alters the conformation of the mRNA
so that it doesn't interact with the adenines on the 16S rRNA and no
stabilization occurs. This is thought to increase the dissociation rates
(and slow elongation) of the improperly-paired tRNA and codon.

The moral to the story seems to be that the very core biotic processes
that make life as it is intimating involve sophisticated, efficient,
and elaborate proof-reading mechanisms. This clearly supports that
the notion that cellular life depends not only on these very basic
processes, but also on ensuring the processes exceed a certain
specificity-threshold. Clearly, an ID scientist might want to
employ various experiments to identify and quantify this specificity
threshold. But why it is that intelligent design would be
flippantly dismissed as the causal origin behind such phenomena
is beyond me. Is there a better explanation? Is there some law
of self-organization that would add editing domains to the
machinery behind these processes? There is not a shred of
evidence to indicate this. Might not mutation and natural selection
have added these editing domains/functions after the basal machinery
arose? Sure, but entailed in this musing is the assumption of life
existing without these basic and core proof-reading mechanisms and
there is not a shred of evidence to indicate that such a state ever did
exist, not to mention could exist. Why do I get the feeling it is
philosophy, and not evidence, that is behind the rejection of ID
as a cause behind the proof-reading that seems to make life