Molecular Mechanisms of Enzyme Activation by Monovalent Cations*

Regulation of enzymes through metal ion complexation is widespread in biology and underscores a physiological need for stability and high catalytic activity that likely predated proteins in the RNA world. In addition to divalent metals such as Ca2+, Mg2+, and Zn2+, monovalent cations often function as efficient and selective promoters of catalysis. Advances in structural biology unravel a rich repertoire of molecular mechanisms for enzyme activation by Na+ and K+. Strategies range from short-range effects mediated by direct participation in substrate binding, to more distributed effects that propagate long-range to catalytic residues. This review addresses general considerations and examples.


General Considerations
M ϩ activation manifests itself as a hyperbolic increase in the rate of substrate hydrolysis that obeys the law of mass action (6). Replacement of M ϩ with a bulky alkyl ammonium such as choline serves as a necessary control to establish that activation is specific and not due to changes in ionic strength. The independent Michaelis-Menten parameters k cat and k cat /K m become of interest in the context of available structural information. In general, the midpoint of the hyperbolic dependence of k cat on [M ϩ ] measures the equilibrium dissociation constant for M ϩ binding to the enzyme-substrate complex. The midpoint of the dependence of k cat /K m on [M ϩ ] yields an approximate measure of the equilibrium dissociation constant for M ϩ binding to the free enzyme. The extent of activation and comparison of the midpoints of k cat and k cat /K m define the thermodynamic coordinates that link M ϩ and substrate recognition.
M ϩ complexation benefits enzyme-substrate interaction and catalysis in some general ways, independent of the specific mechanism of activation. M ϩ binding is typically associated with a large entropy cost required for ordering the site of complexation, as shown by Na ϩ binding to thrombin (15). The effect contributes a more favorable entropy balance when substrate binds to the M ϩ -bound form as compared with the M ϩ -free form. M ϩ complexation also propagates entropic benefits to the entire structure of the enzyme by selecting more ordered and catalytically active conformations from an ensemble dominated by disordered and poorly active conformers. This is documented in clotting proteases (16,17), inosine monophosphate dehydrogenase (18), several ␣-amylases (19 -21), and kinases (22,23), and is a determining factor of ion selectivity in the Streptomyces lividans K ϩ channel (10). ATPdriven sequential switching between Na ϩ -specific and K ϩ -specific conformations drives ion transport in the Na/K-ATPase (11,24).
Specific components of the mechanism of M ϩ activation may be identified from structural analysis. The locale for M ϩ binding pinpoints the origin of a transduction pathway that eventually influences residues of the active site and produces enhanced catalytic activity. The coordination shell of the bound M ϩ is composed mainly of O atoms from the protein backbone and water molecules. Six ligands are common for Na ϩ with average Na ϩ -O distances of 2.4 Ϯ 0.2 Å, but K ϩ prefers six or seven ligands with average K ϩ -O distances of 2.8 Ϯ 0.3 and 2.9 Ϯ 0.3 Å, respectively, due its larger ionic radius (Fig. 1). M ϩ -O distances and coordination optimize the ion-specific valence of the binding site (25).
Activation is defined as Type I when M ϩ is in direct contact with substrate, or Type II when M ϩ binds to a separate site (5). Although M ϩ binding is necessary for activation, it is certainly not sufficient. The initial M ϩ binding event must be transduced into enhanced catalytic activity to produce a biological effect. Binding and transduction are mediated by the same locale in Type I activation where the bound M ϩ is a key determinant of substrate recognition. Diol dehydratase (26) and pyruvate kinase (27) offer two relevant examples. Separation of M ϩ and substrate binding sites in Type II activation poses the additional challenge of identifying the pathway of transduction. Communication between M ϩ and the active site may be traced to specific residues from inspection of the crystal structure, as in dialkylglycine decarboxylase (28), or may be long-range and more difficult to dissect as in thrombin (29).

Type I Activation
In Type I activation, the M ϩ anchors substrate to the active site of the enzyme, often acting in tandem with divalent cations such as Mn 2ϩ , Mg 2ϩ , or Zn 2ϩ . Numerous examples of this synergism as a staple of M ϩ activation cover structural effects that are local on the active site or extend to other regions of the enzyme. Typically, the requirement for M ϩ is absolute. Two subgroups should be distinguished based on whether M ϩ is needed for substrate binding (Type Ia) or hydrolysis (Type Ib). Type Ia activation is associated with a value of k cat independent of [M ϩ ] and a value of k cat /K m that increases hyperbolically with [M ϩ ] (6). Diol and glycerol dehydratases are the simplest examples of Type Ia activation (26,30). The absolute requirement for K ϩ (31) is explained by the crystal structure bound to propanediol ( Fig. 2A): K ϩ is coordinated by five ligands from the protein and functions as bait for two hydroxyl O atoms of substrate (26,30,32). The activity of ␤-galactosidase is enhanced preferentially by Na ϩ over K ϩ (14), and synergy with Mg 2ϩ secures substrate binding to the active site (33,34). Changes in k cat /K m and k cat for the hydrolysis of para-nitrophenyl-␤-D-galactopyranoside are 16-fold and only 2-fold, respectively (35), suggestive of Type Ia activation. Far more common is Type Ib activation, where both k cat and k cat /K m increase hyperbolically with [M ϩ ] (6). This is the case reported originally for pyruvate kinase (36) and encountered in many other enzymes.

Kinases
Pyruvate kinase, an allosteric tetrameric enzyme of the glycolytic pathway catalyzing the conversion of phosphoenolpyruvate and ADP to pyruvate and ATP, was the first reported enzyme to require K ϩ for catalytic activity (12), in addition to two Mn 2ϩ or Mg 2ϩ . Structural biology reveals how the cations cooperate and enable substrate binding to the active site (27,37). Mn 2ϩ anchors substrate via its carboxylate and phosphoester O atoms to the carboxylate O atoms of Glu 242 and Asp 266 , whereas K ϩ increases electrostatic coupling of the phosphate group by screening the carboxylate O of Asp 84 . The critical role of K ϩ creates very favorable conditions for the transfer of the phosphate group from substrate to ATP. Although a convincing case can be made for the key role of K ϩ in the activation of pyruvate kinase, the strong preference of K ϩ over Na ϩ as an activator remains puzzling, given that the Na ϩ -bound structure shows no significant changes in the architecture of the active site (38).
Coordination of substrate coupled to ordering of the structure explains K ϩ activation in branched-chain ␣-ketoacid dehydrogenase (BCKD) kinase (22) and pyruvate dehydrogenase kinase (23). These mitochondrial serine protein kinases belong to the GHKL ATPases that also comprise gyrase, Hsp90, bacterial histidine kinase CheA ,and the DNA mismatch repair protein MutL involved in DNA metabolism, protein folding, and signal transduction. All crystal structures available for this class of enzymes show Mg 2ϩ bound to the O␦1 atom of a conserved Asn and to the triphosphate moiety of ATP. K ϩ sits on the opposite side of ATP relative to Mg 2ϩ , rather than in cis as in pyruvate kinase, and bridges one O atom of the phosphate moiety of TP to typically four protein atoms (Fig. 2B). Comparison of the structures of BCKD kinase in the apo form and bound to ATP reveals how K ϩ stabilizes an entire segment of the protein from His 302 to Phe 336 that is completely disordered in the apo form. Similar observations have been reported for pyruvate dehydrogenase kinase (23).  (47,892). For coordination numbers 4 and 5, there is a secondary peak around 2.75 Å, suggesting possible misidentification of water molecules as Na ϩ . B, K ϩ -O coordination. The average bond distances for coordination numbers 6 and 7 are 2.8 Ϯ 0.2 and 2.9 Ϯ 0.3 Å, respectively, over the entire range (20,494). C, water-O coordination. The coordinations of all crystallographic waters in the data set used for Na ϩ and K ϩ analysis (1,067,258) were calculated for comparison. For coordination number 4, shoulders around the peak correspond approximately to distances of 2.7 and 3.2 Å. The 2.7 Å shoulder is consistent with the secondary Na ϩ -O peaks observed for coordination numbers 4 and 5, suggesting that stronger peaks in electron density maps and longer bond distances may refer to water molecules.
Pyridoxal kinase is a member of the ribokinase superfamily involved in the ATP-dependent phosphorylation of pyridoxal to provide pyridoxal 5Ј-phosphate (PLP), a widely used coenzyme. The enzyme requires K ϩ and Zn 2ϩ as absolute cofactors (39), and the crystal structure reveals how K ϩ assists formation of the enzyme-substrate complex through interactions with a negatively charged phosphate moiety (40).

GroEL and Hsc70
GroEL is an allosteric tetradecameric protein composed of two stacked heptamers that define a large central cavity when in complex with GroES (41). The activity of GroEL is influenced by Mg 2ϩ and has an absolute requirement for K ϩ (9), as seen in pyruvate kinase. NH 4 ϩ and Rb ϩ partially substitute for K ϩ , but Li ϩ , Na ϩ , or Cs ϩ are poor activators. The crystal structure of GroEL bound to ATP reveals Mg 2ϩ and K ϩ acting in tandem to assist binding of ATP to the protein (Fig. 2C).
Two K ϩ in tandem with Mg 2ϩ influence catalysis in the molecular chaperone Hsc70 (42), a member of the heat shock family of proteins involved in the binding and release of polypeptides linked to ATP hydrolysis (43). Similar to GroEL and pyruvate kinase, the ATPase activity of Hsc70 is optimal in the presence of K ϩ and is minimal in Na ϩ (44). Crystal structures of a fragment of Hsc70 retaining M ϩ activation are available in the presence of Na ϩ (45) and K ϩ (42) and explain the functional difference between these M ϩ , unlike the case of pyruvate kinase discussed above. K ϩ provides optimal electrostatic coupling for the phosphate moiety of substrate and optimizes the register for docking in the enzyme active site and formation of the transition state. The function is assisted by a divalent cation, Mg 2ϩ in this case, that forms a ␤,␥-bidentate complex with ATP favoring nucleophilic attack on the P␥. The phosphate moiety of ADP is forced to clash within the active site when K ϩ is replaced by Na ϩ (45,46).

Phosphatases and Aldolases
Synergism between K ϩ and Mg 2ϩ is also observed in fructose-1,6-bisphosphatase (47) and S-adenosylmethionine synthase (48). Fructose-1,6-bisphosphatase is a key enzyme of glu-coneogenesis and catalyzes the conversion of ␣-D-fructose 1,6bisphosphate to ␣-D-fructose 6-phosphate. The enzyme has an absolute requirement for Mg 2ϩ , but the activity is further enhanced by K ϩ and inhibited by Li ϩ (49). K ϩ anchors the substrate to the active site and assists the role of two neighboring Mg 2ϩ located in cis, as for other enzymes catalyzing phosphoryl transfer. K ϩ replaces the guanidinium group of Arg 276 and polarizes the phosphate group for nucleophilic attack. The inhibitory role of Li ϩ is due to replacement of one of the two Mg 2ϩ (47).
S-Adenosylmethionine synthase catalyzes the formation of S-adenosylmethionine from ATP and Met and provides the most widely used methyl donor in biology. The enzyme has an absolute requirement for Mg 2ϩ and K ϩ (50). Substantial crystallographic work has been carried out on this enzyme in complex with various substrates, cofactors, and inhibitors (51). In the presence of an ATP analog and the substrate Met, the structure reveals two Mg 2ϩ and K ϩ in the active site anchoring the phosphate moiety of the cofactor (48). The architecture is similar to that of pyruvate kinase and BCDK and explains the absolute requirement of K ϩ for activation.

Recombinases
Repair of double-stranded DNA breaks or stalled replication forks and homologous gene recombination involve several M ϩ activated members of the recombinase superfamily. The activity of bacterial RecA, archaeal RadA, or archaeal and eukaryal Rad51 depends on ATP and Mg 2ϩ , but also requires K ϩ in the case of human and yeast. In the Rad51 homolog from Methanococcus voltae, the requirement for K ϩ is absolute (13). The crystal structure of MvRadA has been solved in the presence of an ATP analog and Mg 2ϩ , with and without K ϩ (13, 54). The structures reveal a typical arrangement of Mg 2ϩ and two K ϩ in the active site that polarize the P␥ of ATP. Each K ϩ bridges an O atom from the P␥ and a carboxylate from the protein, but also makes extensive contacts at the dimer interface formed upon assembly of the MvRadA filament that explain the absolute requirement for K ϩ . Notably, binding of K ϩ in the active site produces long-range conformational ordering of the putative single-stranded DNA binding domain, establishing a link between M ϩ binding and selection of functionally active conformations.

Type II Activation
In Type II activation, M ϩ binds to a site not in direct contact with substrate and enhances enzyme activity through conformational transitions. Unlike Type I activation, the requirement for M ϩ is less stringent. Measurements of k cat and k cat /K m as a function of [M ϩ ] document a hyperbolic increase in both parameters that is difficult to distinguish from Type Ib activation without independent insight from structural biology (6). A relevant example is the large (Ͼ100-fold) K ϩ -induced increase in activity from a minuscule baseline level reported for inosine monophosphate dehydrogenase from Tritrichomonas foetus (55). In general, Type II activation poses challenges to structural interpretation as the underlying mechanism becomes more difficult to resolve the greater the spatial separation between M ϩ and residues of the active site.

Ribokinase and a Path to Type II Activation
Ribokinase breaks the typical K ϩ -Mg 2ϩ tandem of kinases obeying Type I activation and sequesters K ϩ in a ␤-turn adjacent to the active site, but separated from the solvent and substrate (56) (Fig. 3A). The same arrangement is used by aminoimidazole riboside kinase (57). Lack of information on the structure of the apo form makes it difficult to identify the structural determinants responsible for enhanced catalytic activity. A recent structural analysis of ribokinase from Vibrio cholerae shows significant conformational changes induced by Na ϩ binding that also acts as a preferred activator (58). It is tempting to speculate that these enzymes may represent end points of an evolutionary pathway where coordination of the phosphate moiety of ATP transitions from direct binding to M ϩ (Type I activation) to complete separation from the M ϩ (Type II activation). A possible intermediate in this transition is MutL, an enzyme whose broad M ϩ specificity has so far eluded structural interpretation (59).

Dialkylglycine Decarboxylase
Dialkylglycine decarboxylase is a PLP-dependent enzyme capable of both decarboxylation and transamination. Activity depends on K ϩ , with Na ϩ producing modest enhancement (60). The enzyme is composed of four identical subunits, each containing a PLP binding domain and N-terminal and C-terminal domains. Active sites in the tetramer are close to each other and formed by residues from both monomers of a tightly assembled dimer. The resulting tetramer is formed by two such dimers associated symmetrically. Crystal structures of the enzyme solved in the presence of K ϩ and Na ϩ , with PLP bound to the active site, reveal the mechanism of M ϩ activation and the need for K ϩ over Na ϩ (28,61). K ϩ binds to O␦1 of Asp 307 , O␥ of Ser 80 , and the carbonyl O atoms of Leu 78 , Thr 303 , and Val 305 near the dimer interface where PLP binds. A water molecule completes the octahedral coordination shell (Fig. 3B). When Na ϩ is bound to this site, a water molecule replaces Thr 303 and Ser 80 in the coordination shell around the smaller M ϩ . The O␥ of Ser 80 relocates and causes the phenyl ring of the active site residue Tyr 301 to adopt a conformation no longer favorable for substrate binding. This is a simple and elegant example of how the architecture of the M ϩ binding site is precisely tailored for K ϩ and not Na ϩ to optimize communication with nearby residues of the active site. A similar strategy is used by other PLP-dependent enzymes such as Ser dehydratase (62), tryptophanase, and tyrosinase (63,64), for which K ϩ is absolutely required for activity and Na ϩ acts as a poor activator.

Dehydrogenases
Two K ϩ binding sites have been identified in BCKD. One site controls binding of thiamine diphosphate, and the other, also FIGURE 3. Type II activated enzymes. A, ribokinase (PDB ID 1GQT) shown with substrate, relevant residues, and Cs ϩ (yellow sphere) that plays a functional role analogous to K ϩ . The bound M ϩ is sequestered from solvent and contact with substrate, the ATP analog phosphomethylphosphonic acid adenylate ester (ACP). B, dialkylglycine decarboxylase (PDB ID 1DKA) shown with substrate, relevant protein residues, and K ϩ (yellow sphere). When Na ϩ replaces K ϩ in the site, a structural rearrangement brings the O␥ of Ser 80 in conflict with the phenyl ring of Tyr 301 that adopts a new conformation incompatible with substrate binding. C, Trp synthase (PDB ID 1BKS) shown with relevant protein residues and Na ϩ (yellow sphere) that binds to the ␤ subunit, away from substrate and PLP, but near the tunnel that shuttles the indole for complexation with L-Ser.
found in pyruvate dehydrogenase (65), stabilizes the quaternary structure (66). The BCKD catalytic machine is a member of the highly conserved mitochondrial ␣-ketoacid dehydrogenase complexes including the BCKD complex (BCKDC), the pyruvate dehydrogenase complex (PDC), and the ␣-ketoglutarate dehydrogenase complex (67). The BCKDC contains multiple copies of BCKD, as well as a dihydrolipoyl transacylase, the BCKD kinase, and phosphatase. The activity of BCKD and BCKDC is abolished by phosphorylation of Ser 292 , which promotes an order-disorder transition in the phosphorylation loop of BCKD (68). BCKDC utilizes the entire repertoire of K ϩ binding sites found in BCKD (66) and its kinase (22). The crystal structure of BCKD bound to thiamine diphosphate shows two K ϩ binding sites in crucial positions, with one separated from cofactor and substrate that is most likely responsible for enzyme activation. The second K ϩ site has a structural role and maintains the tetrameric assembly of BCKD (66). This second site is also found in pyruvate dehydrogenase, where the first K ϩ is constitutively replaced by a pair of H-bonds (65).

Trp Synthase
Among the enzymes utilizing PLP-mediated catalysis, Trp synthase has been studied in great detail both structurally and kinetically (69). Trp synthase is a tetramer with the subunits arranged in a linear ␣␤␤␣ fashion. The ␣ subunit catalyzes cleavage of indole 3-glycerol phosphate (IGP) to glyceraldehyde 3-phosphate (G3P) and indole, which is then tunneled to a neighboring ␤ subunit that catalyzes condensation of indole with L-Ser to give L-Trp. The enzyme requires Na ϩ or K ϩ for optimal catalysis (70,71). M ϩ coordination increases catalytic activity 30-fold by affecting the distribution of intermediates along the ground and transition state. Crystal structures of Trp synthase bound to Na ϩ or K ϩ show that the M ϩ does not contact substrate or PLP and binds the ␤ subunit near the tunnel that shuttles the indole for complexation with L-Ser (72). Binding to the active site in the ␣ subunit displaces Na ϩ from its site in the ␤ subunit through an allosteric communication involving the salt bridge between Asp 56 in the ␣ subunit and Lys 167 in the ␤ subunit (73). When Na ϩ is bound (Fig. 3C), Asp 305 in the ␤ subunit assumes two possible orientations, one in contact with Lys 167 and the other rotated away from this residue. In the K ϩ structure, Lys 167 flips 180°and engages Asp 56 in the ␣ subunit, thereby establishing a critical communication within the ␣␤ dimer. Changes are propagated to the tunnel that is partially blocked by the bulky side chains of Phe 280 and Tyr 279 in the Na ϩ form, but is more open in the K ϩ form. Significant changes are confirmed by more recent structures solved in the presence of Na ϩ and Cs ϩ (74).

Thrombin
The stimulatory effect of Na ϩ on the activity of some clotting factors has been known for a long time (75)(76)(77). A simple structure-function link identifies the presence of Tyr 225 near the Na ϩ binding site (29) as a necessary determinant of Na ϩ activation in the entire family of trypsin-like proteases (78). Na ϩ binding shifts the pre-existing equilibrium of the trypsin-fold between active and inactive conformers and produces specific changes that promote substrate binding and catalysis (16,17), rigidify the structure (79), and increase thermal stability (80). The Na ϩ binding site is located Ͼ15 Å away from residues of the catalytic triad within loops that control the primary specificity of the enzyme (Fig. 4). The structural determinants of this long-range communication offer an instructive example of allosteric control that has eluded x-ray structural biology (81) and even NMR measurements (79), unlike the cases of dialkylglycine decarboxylase or Trp synthase. A clear separation of roles exists between residues responsible for Na ϩ binding and those transducing this event into enhanced catalytic activity in thrombin and other clotting proteases. In this case, site-directed mutagenesis and linkage analysis are uniquely suited to arrive at the mechanism of activation.
The bound Na ϩ in thrombin is octahedrally coordinated by two backbone O atoms from Arg 221a and Lys 224 and four buried water molecules anchored to the side chains of Asp 189 , Asp 221 , and the backbone O atoms of Gly 223 and Tyr 184a (Fig. 4A). Mutagenesis of residues in immediate proximity to the site, such Asp 189 (Fig. 4B), results in significantly (Ͼ10-fold) reduced Na ϩ affinity and weakened activation (81). Other mutations do not affect Na ϩ binding, yet abrogate Na ϩ activation (Fig. 4B). They involve residues strategically positioned along the corridors of communication between the Na ϩ site and the active site: Asp 221 supports one of the waters in the coordination shell (81), the backbone N atom of Asn 143 makes an important H-bond interaction with the backbone O atom of Glu 192 that ensures a correct architecture of the Glu 192 -Gly 193 peptide bond organizing the oxyanion hole (82), and Ser 195 is a member of the catalytic triad (83). Asp 221 functions as the initial reporter of the bound Na ϩ and transmits information to the neighbor Cys 191 -Cys 220 disulfide bond that splits the signal toward the active site along the Cys 191 -Asp 194 and Ser 214 -Cys 220 corridors. Additional positive contributions to the Na ϩ effect come from Asp 194 , which stabilizes the fold by H-bonding to the new N terminus generated after zymogen activation, and Trp 215 , which is involved in the pre-existing equilibrium between active and inactive forms of the enzyme. A negative contribution comes from Ser 214 that H-bonds to the catalytic Asp 102 . Removal of the side chain of Ser 214 significantly enhances the Na ϩ effect. The end point of transduction along the two corridors is the rotamer of the catalytic Ser 195 itself, as assessed by the complete loss of Na ϩ activation in the S195T mutant of thrombin and other clotting proteases such as activated protein C and factor Xa (83). The Thr replacement constrains mobility of the O␥ nucleophile within the active site. The Na ϩ effect of thrombin is the result of direct communications and more distributed dynamic components.

Na/K-ATPase
A similar interplay between long-range conformational transitions and direct pathways of communication is observed in the Na/K-ATPase, a ubiquitous ATP-driven ion pump within the family of P-type ATPases (11,24,84). Na ϩ and K ϩ bind at sites separate from ATP binding and phosphorylation, which require Mg 2ϩ as a cofactor. During a catalytic cycle of ATP hydrolysis, the pump switches from the K ϩ -specific E 2 form to the Na ϩ -specific E 1 form by adjusting the ligation distances and coordination of the M ϩ at the same set of sites and changing their orientation rather than moving the M ϩ from one site to another. In the E 2 form, two dehydrated K ϩ are bound and orientation is toward the extracellular phase. In the alternative E 1 form bound to ATP, the geometry is suitable for coordination of three dehydrated or partially dehydrated Na ϩ ions and the sites are oriented toward the cytoplasm. Communication between the M ϩ binding sites and the site of phosphorylation in the P domain is mediated by long-range conformational transitions the involve the transmembrane domain of the ATPase.

Evolutionary Origins
Structural biology also provides a framework to understand the evolutionary origin of M ϩ activation. Widespread occurrence of enzymes activated by M ϩ in plants and the animal world underscores a physiological need for stability and high catalytic activity that likely predated proteins in the RNA world. Mg 2ϩ stabilizes tRNA structures and assists phosphoryl transfer reactions in ribozymes (85,86). However, RNA catalysis may have required M ϩ to broaden its chemical repertoire (87), as suggested by the architecture of the rRNA of the large ribosomal subunit from the archaeon Haloarcula marismortui (88). In small ribozymes, such as the hammerhead, hairpin, and Varkud satellite, M ϩ are sufficient to stimulate catalysis even in the absence of divalent metal ions and stabilize a catalytically competent conformation (89 -91). With the emergence of proteins, stability in high temperatures or salinity became key to extremophiles and revealed the thermodynamic benefit of a more ordered structure for catalysis. The formyltransferase of the archaeon Methanopyrus kandleri utilizes high concentra-tions of K ϩ for activity and thermostability (92). Na ϩ binding sites have been reported in archaeal dehydrogenases (93,94) and aldehyde ferredoxin oxidoreductase of the hyperthermophile Pyrococcus furiosus (95). The architecture of these sites has been retained during evolution (6). Carbonic anhydrase of the halophile alga Dunaliella salina carries an added loop for specific Na ϩ binding that confers stability and resistance to high salinity (96). The loop is strikingly similar to the Na ϩ binding loop of thrombin (6), an enzyme that emerged much later from the deuterostome lineage and that utilizes Na ϩ not only for stability but also for optimal physiological function.

Conclusion
Much has been learned about the structural determinants of enzyme activation by M ϩ . We currently understand several mechanisms to promote catalysis and how the biological abundance of Na ϩ and K ϩ has been strategically utilized during evolution. Much remains to be learned from the subtlety of M ϩ activation in some systems. Structural biology pinpoints likely players, but their interconnectedness may be complex and often involves more distributed, dynamic properties of the protein. That explains why switching M ϩ specificity (15,97) or engineering M ϩ activation de novo in protein scaffolds devoid of such a property (98) requires a large number of amino acid substitutions. Achieving high activity by mimicry M ϩ activation is no simpler, as shown by nature's unique success with actin (99) and murine thrombin (100). Engineering proteins for optimal catalysis may benefit a great deal from increased attention to the structural determinants of M ϩ activation. . Molecular mechanism of Na ؉ activation in thrombin. A, structural determinants of Na ϩ activation in thrombin (PDB ID 1SG8). Shown are the Na ϩ (yellow sphere) coordination shell with water (red spheres) and relevant protein residues. Na ϩ binding is detected by Asp 189 and Asp 221 and then channeled through the corridors Cys 191 -Asp 194 and Ser 214 -Cys 220 to the catalytic residues Asp 102 and Ser 195 . The rotamer of Ser 195 is the end point of the Na ϩ effect, as demonstrated by the properties of the S195T mutant in panel B. The spatial separation of key residues responsible for transduction of the Na ϩ effect (arrows) underscores the contribution of backbone dynamics and overall conformational changes. B, contribution to Na ϩ activation of thrombin from residues in the two corridors Cys 191 -Asp 194 and Ser 214 -Cys 220 connecting the Na ϩ site to the catalytic residues Asp 102 and Ser 195 (see also panel A; thrombin has no residue 218). Three residues are of particular importance, as their mutation has no effect on Na ϩ affinity but abrogates Na ϩ activation (81)(82)(83): Asp 221 supports one of the waters in the coordination shell, Asn 143 stabilizes the functional conformation of the backbone N atom of Gly 193 in the oxyanion hole, and Ser 195 is a member of the catalytic triad.