2 Wrong = Right

Double mutants & Insights therefrom

Righting Wrongs: double mutants restore motor function!

he essential idea behind our work is that we can deduce structural states of the motor by observing how mutants derange the motor's cycle and discovering which further mutations can "re-balance" the motor thus restoring it to function. The basic thinking behind this approach is detailed in the Philosophy section; I'll re-state it briefly here:
  • The myosin motor is a complex machine with many states, corrresponding to the many states of a car engine (e.g. injecting fuel into cylinders -> igniting fuel -> driving revolution -> venting exhaust...)
  • Ideal mutants will disrupt the cycle by specifically perturbing one state of the motor--rendering it too stable (instead of cycling as it should, the motor 'sticks' at one point. Keeping to the car analogy, it would be as if one of the pistons got stuck in one part of its cylinder, holding up everyting)
  • We can learn a lot about the motor by seeing what further changes can restore function when the original defect is still present (e.g., in the case above, a mutation that widened the cylinder at the point where the damaged piston was sticking would reverse the problem even though the oddly shaped cylinder was still present)
  • By capturing and studying many such fixes (known in genetic parlance as suppressors since they 'suppress' the defective behavior) we hope to learn what the motor looks like when it is doing parts of the cycle that are currently impenetrable by other means. Again, the analogy would be that if we get many mutations that widen the cylinder, we will eventually "map out" the shape of a cylinder and hopefully deduce its relationship to the piston.

  • f course, it's much trickier than that in practice! Unfortunately, straightforward situations like the one described above have proven to be hard to come by; generally the repairing event is more distant and subtle--more like the case where a car with a "tire mutation" that pulled to the right (such as a flat on the right side) could be fixed by a "steering wheel mutation" that favored a leftward-biased position. The two changes would be distant from each other in space even though they were related in function. This example, however, highlights an important aspect of our double-mutant approach. If we were to fix the tire on the double-mutant (tire + steering wheel) car, we would notice that the steering wheel mutation alone caused the car to drift to the left. We could then know that the "fix-it" mutation was indeed a functionally relevant compensating change, and we could begin trying to deduce the relationship between the imbalance caused by the tire and the opposing imbalance introduced by the alteration of the steering wheel.

    s always, painting a lovely picture of what we hope is happening isn't compelling evidence that things will work out as we hope. However, we are now deep enough into the approach to have garnered some striking evidence that we're barking up the right tree. To wit: our search for suppressors of the G680V mutation has netted us a variety of changes that seem to not only have common properties, but properties that are sensible counters to the defect introduced by the G680V mutation. Recall that the G680V mutant motor appears to suffer from a reluctance to open the "gateway" and release Pi. Since it executes this step poorly, it spends more time in a "moderately strong" actin binding state. Using the linear cycle introduced in the MotorCycle, the G680V mutant cycle looks like this:

    Where the green rectangle surround the state that is "overoccupied" in the G680V mutant; note that it corresponds to an actin binding state whose strength (indicated by the width of the brown bar) is greater than the weak binding states (ATP, ADP + Pi) but weaker than the later states (ADP, no nucleotide).

    o, what do we expect to learn by restoring myosin motors bearing the G680V change to near-normal functioning via further mutations? Several kinds of things, potentially! The different types of suppressors we might expect are outlined below
      Fixing the G680V motor
    • Class I The most obvious way to fix the problem is at the source. The G680V mutation introduces a bulky residue (valine) where a small, flexible one (glycine) is normally found. A trivial solution would be to shrink some nearby residue to accommodate the increased space required by a valine at position 680.
    • Class II If the defect truly arises from an excessively stiff quot;gateway" then the most obvious remedy would be to make a leakier or weaker gateway. Of course, this may be impossible! Given that there are only 20 amino acid "words" to choose from, there may be no substitution that would make a weak gateway; our choices may be a strong one or none at all!
    • Class III An alternative would be to create myosins whose "mood" or preference is to open the gateway. Such mutants would not be in the doorway itself, but rather in amino acids that influence the shape that the motor adopts. These would be analogous to the steering wheel mutants mentioned above--they don't directly alter the tire, but change the "state" of the steering apparatus so that the car "prefers" to try to turn to the left, compensating for the fact that the wheels now cause a rightward bias.

    f course, the three types of fixes would look different when lit up on a map of myosin. The first set, those that simply shrink nearby residues, would be easily spotted since they would be very close to position 680 in three dimensions. One candidate for this type of suppression occurs at position 483. The N483 residue (asparagine 483) is spatially close to position 680, and thus the N483S mutation (serine is significantly smaller than asparagine) is a candidate for this type of change. In the figure below, G680 is shown in dark blue and N483 in orange.

    The second class of suppressors, those that alter the gateway itself, would be expected to spatially define the gateway. Since we have a candidate gateway, residues R238 and E459 (labeled "gateway" above), we'd expect to recover changes in these residues or adjacent ones. Note that the Pi analog is shown in black. Intriguingly, one suppressor changes residue 240 (cyan in figure above) from a glycine (G) to a valine (V), i.e. G240V, while a second changes residue 235 (magenta) from asparagine to aspartic acid (N235D). However, it is important to note that at this point we're just guessing, but that's half the fun of science. If only I knew what the other half was....

    he third class of suppressors is the hardest to make sense of, but paradoxically, they're potentially the most informative. These would be changes that alter myosin's preference for different shapes. Recall that, to paraphrase Shakespeare, All the world's a stage, and all the myosins merely motors: they have their exits and their entrances; and one myosin in its cycle takes many shapes. To be a good motor, the different shapes must be balanced such that the motor neither skips a step nor gets stuck overlong in any one state. By learning what structural alterations influence myosin's preference for or against a certain state we hope to deduce just what those states look like!

    potting mutants that have this effect is difficult; indeed, to date we guess that they're members of this class partly because they don't obviously belong to either of the first two classes and partly because such a guess is pleasing to us. Fortunately, in the case of G680V suppressors at least, the suppressors in this class array themselves in a very interesting pattern in the myosin structure and demonstrate some extremely suggestive biochemical properties! Several of the residues belonging to this group are termed "the Cluster" throughout this site. The residues are shown below; their wild type shapes are the ones in green, while the blue and cyan forms depict the mutant residues that can individually suppress changes at position 680

    G680A and G680V suppressors: wild type residues are shown in green. Mutant residues capable of suppressing G680V are shown in cyan, those capable of suppressing G680A in blue.
    Note how this group of mutants nicely define a small region of the myosin molecule. Of course, we also have several OTHER suppressors that are not members of Class 1 or Class 2 and which do not fit into this neatly packed group (for a complete showing of G680V suppressors, click here).

    o far so good: the putative class III suppressors (those that we hope are influencing the state or shape of the motor rather than interacting directly with the original mutant or defining a functional element such as the gateway) include a bunch that are close to each other. But how do we know if this placement is a significant "message from myosin" as opposed to a a cruel jest by an uncaring universe? Recall our other prediction about Class III suppressors: if they are all causing the molecule to prefer a similar shape, they should all alter the properties of the molecule in a similar way. What's more, they should ideally cause an alteration opposing that induced by the original mutation (G680V in this case).

    emember how the G680V mutant suffered from a reluctance to release Pi? This generates a trivial prediction for the behavior of any mutant that would restore a balance to myosin: the second mutation should favor Pi-releasing states of the motor. Fortunately, this type of behavior is much easier to spot than Pi retention. This is because Pi release is normally a very slow step for a myosin motor running in the absence of actin. Indeed, it is so slow that this step is the primary even in determining how fast the motor runs. Thus the rate of this step determines the rate of the cycle as a whole. As such, it is technically termed the 'rate-limiting step'.

    ince Pi release "sets the pace" of the motor, we can easily detect enhanced Pi release since it will manifest as a faster motor. As you may or may not've learned in the Smoke and Mirrors section, the motor's rate of cycling is reasonably easy to measure. So, do the mutations that identify the Cluster also give rise to faster cycling motors? For a look at the actual data, follow this link to the DataDungeon, but for the credulous and/or lazy: YES! Briefly, the G680V mutation lowers the basal ATPase (= Pi release rate) while its suppressors enhance Pi release! Thus the Cluster residues satisfy our criteria for Class III suppressors: they aren't close to the primary mutation, they don't identify a known functional element, they form a definable group, and they confer defects similar to one another but opposite of the mutation they counteract.

    o, what do we think is going on? You'll have to brave the Sea of Hypotheses to get the whole story, but in brief it runs like this: G680V renders the myosin motor reluctant to open the Gateway and let out Pi. Further mutants that restore balanced function to the motor act by 1) reversing the original structural defect (potentially N483S), or 2) enhancing the motor's "desire" to open the gateway. Residues in the Cluster may be acting in this way, and we are currently trying to create an explicit model of how they do so.

    You're in: 2 Wrong = Right Places to go:


    How to capture
    altered Myosins

    One Bad

    Dissecting a
    mutant's Defects




    Data behind
    the Claims
    NavTable Comments Help!

    Bruce Patterson