The final enzymes that we will consider are the myosins. These enzymes catalyze the hydrolysis of adenosine triphosphate (ATP) to form adenosine diphosphate (ADP) and inorganic phosphate (P_i) and use the energy associated with this thermodynamically favorable reaction to drive the motion of molecules within cells.

For example, when we lift a book, the energy required comes from ATP hydrolysis catalyzed by myosin in our muscles. Myosins are found in all eukaryotes and the human genome encodes more than 40 different myosins. Myosins generally have elongated structures with globular domains that actually carry out ATP hydrolysis (Figure 9.42). In this chapter, we will focus on the globular ATPase domains, particularly the strategies that allow myosins to hydrolyze ATP in a controlled manner and to use the free energy associated with this reaction to promote substantial conformational changes within the myosin molecule. These conformational changes are amplified by other structures in the elongated myosin molecules to transport proteins or other cargo substantial distances within cells. In Chapter 35, we will examine the action of myosins and other molecular-motor proteins in much more detail.

276

As will be discussed in Chapter 15, ATP is used as the major currency of energy inside cells. Many enzymes use ATP hydrolysis to drive other reactions and processes. In almost all cases, an enzyme that hydrolyzed ATP without any such coupled processes would simply drain the energy reserves of a cell without benefit.

In our examination of the mechanism of restriction enzymes, we learned that an activated water molecule performs a nucleophilic attack on phosphorus to cleave the phosphodiester backbone of DNA. The cleavage of ATP by myosins follows an analogous mechanism. To understand the myosin mechanism in more detail, we must first examine the structure of the myosin ATPase domain.

The structures of the ATPase domains of several different myosins have been examined. One such domain, that from the soil-living amoeba Dictyostelium discoideum, an organism that has been extremely useful for studying cell movement and molecular-motor proteins, has been studied in great detail. The crystal structure of this protein fragment in the absence of nucleotides revealed a single globular domain comprising approximately 750 amino acids. A water-filled pocket is present toward the center of the structure, suggesting a possible nucleotide-binding site. Crystals of this protein were soaked in a solution containing ATP and the structure was examined again. Remarkably, this structure revealed intact ATP bound in the active site with very little change in the overall structure and without evidence of significant hydrolysis (Figure 9.43). The ATP is also bound to a Mg²⁺ ion.

Kinetic studies of myosins, as well as many other enzymes having ATP or other nucleoside triphosphates as a substrate, reveal that these enzymes are essentially inactive in the absence of divalent metal ions such as magnesium (Mg²⁺) or manganese (Mn²⁺) but acquire activity on the addition of these ions. In contrast with the enzymes discussed so far, the metal is not a component of the active site. Rather, nucleotides such as ATP bind these ions, and it is the metal ion–nucleotide complex that is the true substrate for the enzymes. The dissociation constant for the ATP–Mg²⁺ complex is approximately 0.1 mM, and thus, given that intracellular Mg²⁺ concentrations are typically in the millimolar range, essentially all nucleoside triphosphates are present as NTP–Mg²⁺ complexes. Magnesium or manganese complexes of nucleoside triphosphates are the true substrates for essentially all NTP-dependent enzymes.

The nucleophilic attack by a water molecule on the γ-phosphoryl group requires some mechanism to activate the water, such as a basic residue or a bound metal ion. Examination of the myosin–ATP complex structure shows no basic residue in an appropriate position and reveals that the bound Mg²⁺ ion is too far away from the phosphoryl group to play this role. These observations suggest why this ATP complex is relatively stable; the enzyme is not in a conformation that is competent to catalyze the reaction. This observation suggests that the domain must undergo a conformational change to catalyze the ATP-hydrolysis reaction.

277

The catalytically competent conformation of the myosin ATPase domain must bind and stabilize the transition state of the reaction. In analogy with restriction enzymes, we expect that ATP hydrolysis includes a pentacoordinate transition state.

Such pentacoordinate structures based on phosphorus are too unstable to be readily observed. However, transition-state analogs in which other atoms replace phosphorus are more stable. The transition metal vanadium, in particular, forms similar structures. The myosin ATPase domain can be crystallized in the presence of ADP and vanadate, . The result is the formation of a complex that closely matches the expected transition-state structure (Figure 9.44). As expected, the vanadium atom is coordinated to five oxygen atoms, including one oxygen atom from ADP diametrically opposite an oxygen atom that is analogous to the attacking water molecule in the transition state. The Mg²⁺ ion is coordinated to one oxygen atom from the vanadate, one oxygen atom from the ADP, two hydroxyl groups from the enzyme, and two water molecules. In this position, this ion does not appear to play any direct role in activating the attacking water. However, an additional residue from the enzyme, Ser 236, is well positioned to play a role in catalysis (Figure 9.44). In the proposed mechanism of ATP hydrolysis based on this structure, the water molecule attacks the γ-phosphoryl group, with the hydroxyl group of Ser 236 facilitating the transfer of a proton from the attacking water to the hydroxyl group of Ser 236, which, in turn, is deprotonated by one of the oxygen atoms of the γ-phosphoryl group (Figure 9.45). Thus, in effect, the ATP serves as a base to promote its own hydrolysis.

278

Comparison of the overall structures of the myosin ATPase domain complexed with ATP and with the ADP–vanadate reveals some remarkable differences. Relatively modest structural changes occur in and around the active site. In particular, a stretch of amino acids moves closer to the nucleotide by approximately 2 Å and interacts with the oxygen atom that corresponds to the attacking water molecule. These changes help facilitate the hydrolysis reaction by stabilizing the transition state. However, examination of the overall structure shows even more striking changes.

A region comprising approximately 60 amino acids at the carboxyl- terminus of the domain adopts a different configuration in the ADP–vanadate complex, displaced by as much as 25 Å from its position in the ATP complex (Figure 9.46). This displacement tremendously amplifies the relatively subtle changes that take place in the active site. The effect of this motion is amplified even more as this carboxyl-terminal domain is connected to other structures within the elongated structures typical of myosin molecules (Figure 9.42). Thus, the conformation that is capable of promoting the ATP hydrolysis reaction is itself substantially different from other conformational changes that take place in the course of the catalytic cycle.

Myosins are slow enzymes, typically turning over approximately once per second. What steps limit the rate of turnover? In an experiment that was particularly revealing, the hydrolysis of ATP was catalyzed by the myosin ATPase domain from mammalian muscle. The reaction took place in water labeled with ¹⁸O to track the incorporation of solvent oxygen into the reaction products. The fraction of oxygen in the phosphate product was analyzed. In the simplest case, the phosphate would be expected to contain one oxygen atom derived from water and three initially present in the terminal phosphoryl group of ATP.

279

Instead, between two and three of the oxygen atoms in the phosphate were found, on average, to be derived from water. These observations indicate that the ATP hydrolysis reaction within the enzyme active site is reversible. Each molecule of ATP is cleaved to ADP and P_i and then re-formed from these products several times before the products are released from the enzyme (Figure 9.47). At first glance, this observation is startling because ATP hydrolysis is a very favorable reaction with an equilibrium constant of approximately 140,000. However, this equilibrium constant applies to the molecules free in solution, not within the active site of an enzyme. Indeed, more-extensive analysis suggests that this equilibrium constant on the enzyme is approximately 10, indicative of a general strategy used by enzymes. Enzymes catalyze reactions by stabilizing the transition state. The structure of this transition state is intermediate between the enzyme-bound reactants and the enzyme-bound products. Many of the interactions that stabilize the transition state will help equalize the stabilities of the reactants and the products. Thus, the equilibrium constant between enzyme-bound reactants and products is often close to 1, regardless of the equilibrium constant for the reactants and products free in solution.

These observations reveal that the hydrolysis of ATP to ADP and P_i is not the rate-limiting step for the reaction catalyzed by myosin. Instead, the release of the products, particularly P_i, from the enzyme is rate limiting. The fact that a conformation of myosin with ATP hydrolyzed but still bound to the enzyme persists for a significant period of time is critical for coupling conformational changes that take place in the course of the reaction to other processes.

Myosin molecules function to use the free energy of hydrolysis of ATP to drive macroscopic motion. Myosin molecules move along a filamentous protein termed actin, as we will discuss in more detail in Chapter 35. Using a variety of physical methods, scientists have been able to watch single myosin in action. For example, a myosin family member termed myosin V can be labelled with fluorescent tags so that it can be localized when fixed on a surface with a precision of less than 15 Å. When this myosin is placed on a surface coated with actin filaments, each molecule remains in a fixed position. However, when ATP is added, each molecule moves along the surface. Tracking individual molecules reveals that each moves in steps of approximately 74 nm as shown in Figure 9.48. The observation of steps of a fixed size as well as the determination of this step size helps reveal details of the mechanism of action of these tiny molecular motors.

280

X-ray crystallography has yielded the three-dimensional structures of a number of different enzymes that share key structural characteristics and, almost certainly, an evolutionary history with myosin. In particular, a conserved NTP-binding core domain is present. This domain consists of a central β sheet, surrounded on both sides by α helices (Figure 9.49). A characteristic feature of this domain is a loop between the first β strand and the first helix. This loop typically has several glycine residues that are often conserved between more closely related members of this large and diverse family. The loop is often referred to as the P-loop because it interacts with phosphoryl groups on the bound nucleotide. P-loop NTPase domains are present in a remarkably wide array of proteins, many of which participate in essential biochemical processes. Examples include ATP synthase, the key enzyme responsible for ATP generation; signal-transduction proteins such as G proteins; proteins essential for translating mRNA into proteins, such as elongation factor Tu; and DNA- and RNA-unwinding helicases. The wide utility of P-loop NTPase domains is perhaps best explained by their ability to undergo substantial conformational changes on nucleoside triphosphate binding and hydrolysis. We shall encounter these domains throughout the book and shall observe how they function as springs, motors, and clocks. To allow easy recognition of these domains in the book, they will be depicted with the inner surfaces of the ribbons in a ribbon diagram shown in purple and the P-loop shown in green (Figure 9.50).

281