|Phylogenetics Intro: using diversity left behind by Evolution|
o some extent, our approach is incredibly unoriginal--it's been going on since well before living creatures first crawled out of the ocean. How can this be? Simply put, mutations happen. We all know that sunlight isn't the greatest friend of accurate maintenance of the genes. Other processes also contribute to DNA damage and the potential that it will not be properly repaired. The consequence? Mutation. While we greatly accelerate the process in lab by exposing cells to tremendous quantities of UV light, there's nothing truly different about that process and the one occurring in nature. Mutation provides the grist for the mill of Evolution.
y acknowledging this, we stand to benefit from the outcome of millions of cycles of mutation followed by natural selection. While only successful designs will be perpetuated down through evolutionary history, the fact of mutation results in continued "experimentation" in terms of novel designs. If a new design is roughly as good or better than the original, then both may co-exist, or the new design may come to predominate. The final outcome is that while there was presumably some ancestor that contained the primordial mysoin gene, its descendants have altered and perhaps improved upon the original design. Thus as with a family tree, a single ancestor can give rise to diverse descendants. Studying this diversity can lead us to insights into the myosin motor.
ecall that our business is the discovery and characterization of functionally related positions in the myosin motor. We introduce defects and then search for further, compensating changes that restore the motor to function. This same process occurs in nature. UV light, radiation, mutagens and just plain old bad luck can give rise to a change in the DNA. This can create an intolerable situation--a bum protein! If the gene is to be perpetuated, it must rely upon another, more fortunate accident that brings the protein it encodes back to life. Should this occur, the gene may yet have a long and prosperous evolutionary future. More commonly, the organism will die, or its inferior descendants will rapidly be eliminated. The bottom line: changes are usually deleterious, and are only allowed to continue if they are made good--an iffy proposition, but given the number of times the dice are rolled by organisms through Time, every now and again long odds will come out OK. Two factors that makes things more likely are 1) gene duplication events intermittently gives rise to "spare" copies of genes that are free to change without deleterious consequence to the organism, and 2) the happy state of diploidly--the presence of two copies of each chromosome provides a backup which may permit a dud copy of a given gene to be persist for some time. Either of these situations "buys time" during which a dud gene may undergo a compensating change.
o return to the march of the genes through time, scenarios like the one above will be occurring rarely, but occurring nonetheless. When mutation-repair events occur, they give rise to new alleles. These may fortuitously find themselves in organisms whose descendants give rise to new species or whose myosins diversify to take on new functions. In either case event, diversity creeps into a protein's sequence. A couple billion years of this kind of thing and all myosins would no longer be created equal. Each organism would contain functional myosin(s), but they would have diverged from the ancestral myosin that gave rise to them all. By surveying the myosin family, we behold a variety of answers to the question "what amino acid sequences constitute a molecular motor?". The differences between the end product myosins represent several things:
ur goal is to harvest this vast treasury of information. The question is, how do we tell the third group (compensating changes) from the noise of the others? If we can do so, we stand to inherit the vast amounts of information accumulated through evolutionary time. Fortunately, the groundwork has already been laid for us--other researchers have determined the DNA and protein sequences of more than 80 myosin genes from different organisms or different functional families within a given organism. A summary of these sequences and their evolutionary relationships can be found at the MRC's Myosin Homepage. The deductions and hypotheses presented below derive from sequence alignments made available by Tony Hodge and Jamie Cope.
o, what kinds of clues can we find by looking at the variability extant in 80 myosin sequences? Recall that our goal is to determine unknown conformations of the myosin motor by learning what amino acids interact or influence each other. Our primary method is to study mutations that represent defect-repair partnerships. Since Mother Nature has been playing the same game, it behooves us to capitalize on work already done. However, our task is to deduce which changes between proteins represent related, functional interactions as opposed to random changes or property-altering changes. We work from the following simple prediction: if two positions in the myosin molecule have a meaningful functional relationship, then a significant change in the structure of the amino acid at one position would have to be accompanied by a significant, complementary change at the other. A simplistic representation of this idea is the "lock and key" model--if an important interaction is simulated by a lack-and-key relationship between two amino acids, then changes in the lock must be answered by complementary changes in the key.
y looking at 80 myosins simultaneously, we can hope to spot correlated changes within myosin sequences. For example, suppose we "aligned" the 80 myosin sequences, i.e. put them in register such that column A showed the corresponding amino acid from myosin #1, myosin #2, myosin #3, etc. Similarly, column B showed the next amino acid from myosins 1, 2, 3,.... and so on until we had a table of all the amino acids in all 80 myosins. We could then look for pairwise relationships. Returning to our example, suppose amino acids 1 and 26 had a lock-and-key relationship. If this were true, we would find that any myosin with amino acid X in column A would have amino acid Q in column Z. However, a myosin with amino acid B in column A would have amino acid O in column Z. In the 80 myosins, we might observe a variety of pairings between column A and column Z, but by and large there would be a correspondence between columns A and Z. To make this relationship more concrete, observe the following alignment showing amino acid positions 484 and 509 (according to the Dictyostelium myosin II numbering). I have colored the amino acids based on their chemical properties. Note that when the amino acid in the top row (position 484) changes, there is often a corresponding change in the amino acid in the bottom row (509).
The relationship becomes even more striking when you consider the actual properties of the amino acids. Positively charged amino acids are shown in Red, while negatively charged ones are indicated in Blue. Magenta amino acids are "polar" ones that have both positive and negative characteristics. Note that ALL pairings allow a compatible charge distribution at the two positions (excepting the black and green residues, which we'll ignore for now). Kind of striking, isn't it? In the Dictyostelium sequence, residue 484 is a histidine (positive charge), while residue 509 is an aspartic acid (negative charge)
o, what does this signify for the structure of the motor? I would like to believe it reflects an interaction that takes place at some point. However, I'm not paid to have opinions, what do the facts indicate? The two residues that are showing correlated changes exist in two adjacent helices in the structure--the Camshaft and the ActBind helix (511-536). In the figure at right, the Camshaft is shown in pink while the ActBind helix is shown in cyan. Residues 484 and 509 are shown in "normal colors" (carbon in white, nitrogens in blue and oxygens in red). In histidine, the blue nitrogen is the site of the positive charge, while the two red oxygens in aspartic acid are both negatively charged. Note that while the two residues aren't touching, they are pretty close, and both are "pointed" in the same direction (out of the screen). A small rotation and sliding of the Camshaft or Actbind helices could juxtapose the two residues. On the basis of the co-variation shown above, I propose that just such a rotation happens at some point in the myosin cycle.
his is all very well and good, but it's still just talk, though hopefully somewhat persuasive. What do I indtend to do about it? Employ a classic genetic test for interactions. IF the hypothesis is true that the residue at position 484 indeed must be compatible with the residue at 509, then if I make both positively charged or both negatively charged, then the resulting myosins will fail to function. However, a myosin with a negative charge at position 484 will work if it is paired with a positive charge at position 509, despite the fact that this represents a mutation at BOTH positions! In short, the following table is my prediction for what will happen if a positive, negative, or ambivalent (dual specificity) residue is present at each position:
bviously, the upper left hand corner of the table is the most interesting, where "GOOD" and "BAD" interactions are predicted; "OK" hopefully means GOOD, but if it is less than good I won't recant the theory. Presently, I have constructed several myosin mutants with the characteristics called for by the table, and will soon have some preliminary indication of how myosin feels about each arrangement.
here to go from here? Two places. First, just knowing that two amino acids interact is very nice, but the key to our success is figuring out WHEN in the motorCycle an interaction takes place, and what the importance of that interaction is. To move forward from the mutant data depicted in the table, we will biochemically characterize the mutant myosins and try to determine what biochemical features of the mutant are aberrant. For example, I might guess that a [484(-)/509(-)] myosin would be defective in strong binding to actin, a prediction that is easily verified by studying the tenacity with which the mutant holds onto actin in the absence of ATP.
he second direction to go is to use our projected success with this somewhat non-earthshattering pairing to go hunting for pairings that would shed light on more dramatic and unanticipated relationships. At present, I am looking at several pairings that are not as convincing as the 484-509 pair shown above, but which stand to support the radical "Pouting Myosin Model", which I haven't even had the guts to come forth on these pages with yet. Briefly, I'm positing that the actin binding site of myosin undergoes a much more dramatic re-arrangement than anybody has suggested so far. If I can get some supporting data from these or other experiments, you can bet it'll make a big splash in the Sea of Hypotheses!