When Enzymes Don't Lie
New research published in Bio-Complexity calls into question some fundamental assumptions of neo-Darwinian theory and enzyme evolution.
Enzymes are proteins that catalyze reactions that are necessary for life. Enzymes play such a fundamental role in life that many researchers are interested in how they originated and how they have evolved. They are composed of strings of amino acids, and the particular sequence of amino acids determines what three-dimensional shape each protein has, and what enzymatic function it carries out. Biologists categorize enzymes into families based on similarity of structure. The more similar the structure, the closer the evolutionary relationship is presumed to be.
It is generally believed that these enzyme families arose by a process of gene duplication followed by divergence of the extra copies over time. If accumulating mutations in an extra gene led to a beneficial change in enzyme function, the gene encoding that enzyme would tend to be preserved. Over time, then, repeated rounds of duplication and divergence would produce the large multi-functional families we see today. Yet for this explanation to be true, converting enzymes to new functions must require only a few mutations in order for the process to be within reach of neo-Darwinian evolution.
Doug Axe and Ann Gauger from Biologic Institute recently published a paper that addresses this pervasive assumption about the ease of enzyme conversion :
Here, we explore this question by asking how many mutations are needed to achieve a genuine functional conversion in a case where the necessary structural change is known to be small relative to the change commonly attributed to paralogous divergence.
As the authors report, they focused "not on minor functional adjustments, like shifts in substrate profiles, but rather on true innovations -- the jumps to new chemistry that must have happened but which seem to defy gradualistic explanation." Their aim was not to establish ancestry between two particular enzymes, but to identify a functional innovation that should be relatively straightforward within a superfamily and then evaluate how evolutionarily feasible this modest innovation would be.
They began by looking at a large "superfamily" known as the pyridoxal-5'-phosphate (PLP) dependent transferases. This is a well-characterized family of enzymes that share a common fold (shape) but catalyze distinctly different reactions. They identified a pair within that superfamily with very close structural similarity but no functional overlap. Kbl2 is involved in threonine (a type of amino acid) metabolism, and BioF2 is part of the biotin biosynthesis pathway. They then used a three-stage process to identify the sequences mostly likely to confer a functional change.
The experimental question is: How many mutations are required to convert Kbl to BioF function?
There are about 250 different amino-acid differences between Kbl and BioF. This is a huge number, and probably many more than the minimum number of amino acids that are needed to convert one enzyme's function to the other's. In order to determine the minimum number of amino acid changes necessary for functional innovation to occur, Gauger and Axe followed a three stage process. First they used sequence and structural comparisons of the two enzymes to identify candidate amino acids most likely to be significant for function. Second, they mutated those amino acids in BioF, making them like Kbl, and checked for loss of BioF activity. They identified three groups of amino acids, each consisting of six or seven individual amino acids, and one single amino acid, H152, that were essential for BioF function. Finally, they tested whether changing these groups in Kbl to look like BioF would enable the mutated Kbl to substitute for BioF.
The experiment ended up showing that no functional conversion could be achieved, even when all identified changes were made, including every amino acid in the enzyme's active site (the place where the enzyme's chemistry is carried out). Gauger and Axe estimate that seven or more mutations would be required to convert Kbl to BioF function.
So what does this all mean?
Two major implications need to be noted from the results of this experiment. In a second post, we will have a further discussion on implications of this research for neo-Darwinism.
The first finding was that H152 was vital to the functionality of the BioF. Perhaps what is most interesting about this finding is that H152 is not within the active site but is on the enzyme surface away from the active site. It is generally believed that the active site is the area of interest for enzymes within a family and the rest of the enzyme (the "scaffold") just holds the active site. However, these experimental findings seem to indicate that the non-active site differences, however minimal they may be, need to be considered, and that these differences may be more important than the apparent similarities.
The second implication from this failure to convert functionality is the question of whether a neo-Darwinian process of step-by-step conversion from one enzyme to another is actually feasible. The two enzymes in this study were very similar enzymes, yet even with generous estimates for mutation rate, gene duplication rate, and no fitness cost for carrying the extra gene, there does not seem to be enough time for mutations of this sort to occur:
"...seven is a reasonable lower-bound estimate of the specific nucleotide substitutions required for conversion...this places the Kbl [to] BioF conversion outside the bounds of what can be achieved by the Darwinian mechanism." When using the established mechanisms and estimates, it would require 10^30 or more generations to elapse before any type of BioF-like conversion could be established. There is not enough time to accomplish this relatively small innovation! As Axe and Gauger aptly summarize:
This places the innovation well beyond what can be expected within the time that life has existed on earth, under favorable assumptions. In fact, even the unrealistically favorable assumption that kbl duplicates carry no fitness cost leaves the conversion just beyond the limits of feasibility.
These are not large leaps or large-scale changes, but small-scale changes. And other research, cited in this paper, have shown the same difficulty in achieving enzyme conversions. This calls into question a fundamental assumption in the neo-Darwinian paradigm, that similarity of structure or form means ease of conversion, and implies that a different paradigm is necessary to account for enzymatic functional conversion.
In a second post, we will unpack some of the implications of these findings for neo-Darwinian theory.
(For Doug Axe's post on some of the implications of their research, see here)