I'm about to break my own rule and address the comments of Joseph Mastropaolo, though it might be OK since I am directing my comments at the group rather than him. But I could just kick myself for not thinking of these sooner. And I call myself a protein chemist. Bad boy! (This post is long, but well worth the read.)
Anyway, Joseph claims that the probability of evolving a protein randomly amino acid by amino acid is 2.3 E-75. Well, what do you think would be the chance of performing a "frame-shift" mutation on a functional gene (one that already codes for a functional protein) and getting a completely different functional protein as a result? I wouldn't know, but it has happened.
Back when I first joined this list I submitted a post that gave examples of genetic changes that resulted in increases of information. One example was an enzyme that could digest nylon. This enzyme, called linear oligomer hydrolase, we now know was created by just such a "frame-shift" mutation (Ohne, S. 1984. "Birth of a unique enzyme from an alternative reading frame of the preexisted, internally repetitious coding sequence." _Proceedings, National Academy of Sciences_ 81:2421-2425). Frame-shift mutations cause a change in the "reading frame" of the gene by the addition or subtraction of a single nucleotide. The result is a totally new and essentially random sequence of amino acids that has no structural, antigenic or enzymatic relationship to the original protein. Usually these proteins are useless, but occasionally they are not. Linear oligomer hydrolase was one such frame-shift success.
Three tests were done to determine if the new enzyme was indeed unique (Kinoshita, S, T Terada, T Taniguchi, Y Takene, S Masuda, N Matsunaga, H Okada. 1981. "Purification and characterization of 6-aminohexanoic-acid-oligomer hydrolase of _Flavobacterium sp._ K172." _European Journal of Biochemistry_ 116:547-551). First of all, antibodies made against this new enzyme found no similar antigens among the other enzymes in the bacterium. Secondly, the new enzyme had no activity for biologically derived molecules with similar chemical structures. Thirdly, the new enzyme had an enzymatic efficiency for the 6-aminohexanoic acid oligomer bond of only 2%, compared to enzymes that catalyzed similar reactions with biologically derived substrates. Had the new enzyme simply evolved as the result of a minor modification of a preexisting enzyme there would have been much stronger antigenic and enzymatic cross-reactivity, and its catalytic efficiency would have been much higher. This last espe
cially demonstrates that the new enzyme was derived from randomly ordered amino acids, because low efficiency is exactly what you would expect of such a newly formed enzyme. As time goes along, we would expect natural selection to improve this efficiency by selecting for enzymes with point mutations that improve its catalytic function.
This alone should put Joseph's argument to final rest, but I can give another example. If you took a number of (supposedly) phylogenetically closely related homologous genes (members of a gene family), from different species no less, cut them up into pieces, shuffled them together and let them randomly recombine into what are called chimeric genes, what would be the chance that any of these chimeric genes would have any function whatsoever, much less a useful function? Again, I wouldn't know, but it has been done. The scheme I just described, called family shuffling (Crameri, A, S-A Raillard, WPC Stemmer. 1998. _Nature_ 391:288-291), is being used by pharmaceutical companies, in combination with more traditional drug design methods, to create proteins with novel therapeutic functions. Though there have been few successes, the principle is sound. Proof of concept is provided by the work of Kumamaru _et al._ (Kumamaru, T, H Suenaga, M Mitsuoka, T Watanabe, K Furukawa. 1998. "En
hanced degradation of polychlorinated biphenyls by directed evolution of biphenyl dioxygenase." _Nature ssed using evolution-derived techniques. These take advantage of the blind nature of the "evolutionary algorithm" of mutation and natural selection to obtain improved and/or novel catalytic functions. Four requirements have to be satisfied for success: the desired function must be physically possible; the function must be biologically or evolutionarily possible; it must be possible to make chimeric libraries complex enough to contain rare beneficial combinations; and there must be a way to rapidly screen or select for the desired function. The third requirement is routine work in any molecular biology lab throughout the world. Requirements one and two can be satisfied by the family shuffling concept or by taking advantage of proteins that already have a binding activity (like antibodies) to evolve catalytic activity as well (Fuji, I, _et al._ 1998. _Nature Biotechnology_ 16:
463-472). Requirement four is the most tricky, bu
water-in-oil emulsion to trap a transcription/translation reaction mixture in bacterium-size vesicles may be a step in the right direction (Tawfik, DS, AD Griffiths. 1998. "Man-made cell-like compartments for molecular evolution." _Nature Biotechnology_ 16:652-656).
After all this, it's difficult to see how anyone could still harbor any certainty for the claim that it is too improbable for novel proteins to evolve from random amino acid sequences. But this discussion also puts the lie to another creationist claim, that evolution has contributed nothing to modern society. Very shortly now it may in fact provide the very basis for creating new therapeutics to fight communicable, metabolic and genetic diseases. I thank God that in all His wisdom, He chose to create evolution.
Kevin L. O'Brien
"Good God, consider yourselves fortunate that you have John Adams to abuse, for no sane man would tolerate it!" William Daniels, _1776_