Monovalent Cation Activation of Tryptophanase*

The interaction of monovalent cations with holotryptophanase has been examined by spectral and kinetic methods. Using S-orthonitrophenyl-L-cysteine as a substrate, activation by the following monovalent cations was demonstrated; values of K A (mM, in italics) and V ,,, a x (pm01 min-’ mg-I) are given in parentheses: It was demonstrated by circular dichroic spectra that the competitive inhibitor, ethionine, interacts with the holoenzyme in the absence of activating monovalent cations, although it does not undergo labilization of the (Y proton. On addition of monovalent cation to the holoenzyme.ethionine complex, a marked increase occurs in absorption at 508 nm resulting from labili- zation of the (Y proton with formation of the quinoid form of the pyridoxal phosphate moiety of the enzyme. substrate complex at the catalytic center (Merino,

The interaction of monovalent cations with holotryptophanase has been examined by spectral and kinetic methods. was linearly related to the maximum velocity observed with each cation except NH,+, which was anomalously active. When measured at 500 nm, the change in absorption ranged from a AA = 0.45 rng-' of tryptophanase for NH4+ to 0.06 rng-' for Li+. Two moles of thallium (I) were bound per mole of subunit. The data are most consistent with the interaction of monovalent cation at or near the catalytic center in such a way that it either participates directly in the reaction or is required for the critical alignment of one or more functional groups necessary for catalysis.
Monovalent cations such as K+, Na+, or NH,+ participate in biological processes in a variety of ways. For example, they act as charge carriers in nerve impulses, cofactors in transport, cofactors inaintaining osmotic balance, important elements of the environment of halophiles and aquatic organisms, and activators of enzymes. Our interest in these cations is as activators of several enzymes (1,2). Enzymes subject to this activation fall into two main classes, those catalyzing phosphoryl transfer reactions and those catalyzing elimination reactions. The latter group includes a subgroup of pyridoxal ' The abbreviations used are: Hepps, 4-(2.hydroxyethyl)-l-piperazinepropanesulfonic acid; SOPC, S-orthonitrophenyl-L-cysteine.
Kinetics -Kinetic constants at 25" were determined from initial velocities at SOPC concentrations ranging from 0.02 to 0.6 mM in 0.025 M (CH,),N/Hepps, pH 8.0, or by analysis of complete reaction progress curves starting at an initial concentration of 0.12 mM SOPC (11). The data were analyzed according to Wilkinson (12).
Spectral Titrations-Dissociation constants for interaction of each cation with tryptophanase in 0.025 M (CH,),N Hepps, pH 8.0, containing 10 mM ethionine were determined by following changes in absorption at 500 nm after addition of I-to 50-~1 aliquots of concentrated solution of cation in 0.025 M (CH,),N/Hepps. After correction of each AA,,on, for dilution, the data were plotted as l/AA,,,n, versus l/M+. The data were analyzed according to Wilkinson (12).
Visible absorption and circular dichroic spectra were obtained with the Cary 14 and Cary 60 spectrophotometers, respectively.

RESULTS
To evaluate the effect of monovalent cations on an enzyme, it is essential that the effect of each cation on the kinetic parameters, K, and V,,,,,, for that enzyme be examined. Hence the first experiments involved a determination of the K,,, for SOPC at various concentrations of K+. Fig. 1 Fig. 1). This latter behavior allowed the determination of activation constants (KA) for each cation at 0.6 mM SOPC and the K, and V,,,,, values for SOPC at saturating levels of each cation.
[SOPC]'l mM -' acid, or piperidine HCl were tested as nonactivating salts for maintenance of ionic strength, however, each was inhibitory at concentrations below those required for this purpose (50 to 300 mM). For this reason, no attempt was made to control ionic strength in the kinetic or equilibrium experiments. Table I lists the activating constant, KA, for each cation obtained at 0.6 mM SOPC, and also the kinetic constants for SOPC at saturating concentrations of each cation. The concentration of each cation was, at least, 10 times the K,. Several points can be made. First, each cation activates, lithium being the least effective and ammonium the best. Second, the K, for SOPC is not greatly affected by ionic strength. Third, the K,,, for SOPC is not affected significantly by varying the activating cation, except for Na+, whereas the KA for each cation varies from 54 mM for Li+ to 0.2 mM for NH,+. The K, for SOPC in the absence of monovalent cation was not determined, since the activity in the absence of monovalent cations is very small and is affected by the ammonia released during the reaction.
As demonstrated by Morino and Snell(41, when holoenzyme is allowed to interact with a competitive inhibitor such as alanine, a marked change in the visible absorption spectrum occurs, especially at 500 nm. This is confirmed by the spectra of Fig. 2, showing spectral changes arising from interaction of ethionine with holotryptophanase in the presence of K+. Spectrum 1 is holoenzyme alone, showing a major peak at 420 nm. Addition of K+ results in a decrease in absorption at 420 mn with an increase at 337 mn (Spectrum 2). When the competi- b The values given in parentheses were obtained by titrating the ethionineholoenzyme complex with cation, and following the change in absorption at 500 nm.
( These data were obtained from initial velocities at various substrate concentrations, rather than from analysis of complete reaction progress curves (11).
tive inhibitor ethionine is added in addition to K+, a reduction in absorption at 420 nm and 337 nm is noted with a marked increase in absorption at 508 nm (Spectrum 3).
The extent of the spectral perturbation at 508 nm in the presence of 10 mM ethionine (K, = 0.52 mM) (13) was next measured as a function of increasing concentration of monovalent cation. The K, for ethionine in the presence of each monovalent cation was not determined, since the K,,, for SOPC was not affected significantly by each cation. These data were plotted as a function of l/M+ versus l/AA,,,,, and extrapolated to infinite concentration of monovalent cation. A linear plot was observed in each case. The dissociation constants obtained from these data are given in Table I. The AASoo ,,,,, at infinite concentration of M+ was then plotted as a function of the maximum velocity observed with each cation, using SOPC as the substrate (Fig. 3). For all the cations except ammonia, a nearly direct relationship exists between AA,,,, and V,,,. That is, the maximum velocity for a given cation is directly proportional to the extent to which it elicits formation of the 500 nm absorbing species with ethionine.
The next question of interest is whether or not the substrate binds in the absence of cations. Fig, 4A shows that the absorption spectrum of holoenzyme in the absence of monovalent cation, plus or minus the competitive inhibitor, ethionine, are nearly identical. On the other hand, circular dichroic spectra of holotryptophanase, plus or minus the competitive inhibitor ethionine, in the absence of activating monovalent cations (Fig. 4B), are very different and are consistent with a complex between holoenzyme and ethionine. The difference between spectra 1 and 2 in the 360 to 440 nm range cannot be due to contamination with a monovalent cation such as ammonia, since the change in circular dichroism at 420 nm is much larger than that at 508 nm, where the maximum change occurs following interaction of monovalent cation (Spectrum 3). The formation of a complex between ethionine and holoenzyme in the absence of monovalent cations is consistent with the observation that saturating SOPC allows Michaelis-Menten kinetics for the activation with K+ (Fig. 1).
Finally, stoichiometry of binding of M+ to tryptophanase was examined in an effort to evaluate the specificity of its interaction with monovalent cations. The data of Fig. 5 indicate the binding of 2 mol of Tl+ per tryptophanase subunit with KD equivalent to that measured by kinetics or spectral titration. DISCUSSION The sigmoid kinetics for activation of tryptophanase by K+ at subsaturating concentrations of substrate, first reported by Hogberg-Raibaud et al. (51, were confirmed in this study. Fig.  1 indicates nonlinear Michaelis-Menten kinetics for the activation by K+ at low SOPC concentrations and for the reaction with SOPC at low K+ concentrations. However, at saturating concentrations of either K+ or SOPC, we observed linear Michaelis-Menten kinetics. A similar behavior was noted with tryptophan as the substrate (data not given).
Several models may explain these data. Hogberg-Raibaud et al. (5) suggested for one model that a minimum of 3 sites per subunit of enzyme, consistent with an observed Hill slope of 2.7, must be complexed with monovalent cation to achieve the activated enzyme form. No interactions between protomers would be necessary. Alternatively, they suggested cooperative binding of potassium. We suggest a third possibility, namely, that the K,) for the pyridoxal phosphate.apoenzyme complex is sufficiently large in the absence of monovalent cations and  substrate that the holoenzyme partially dissociates in the assay mix at low concentrations of enzyme, substrate, and metal ion activator. Presumably, K+ or substrate brings about an increased affinity of enzyme for pyridoxal-P, which results in an increase in active enzyme at higher substrate or K+ concentrations, resulting in the nonlinear Michaelis-Menten kinetics. The binding studies with Tl+ indicate 2 mol of Tl+ bound per protomer, and if this is true also for K+, the first model suggested by Hogberg-Raibaud et al. (5), as an explanation of nonlinear K+ saturation at subsaturating substrate, could not be true. Sufficient data are not available to distinguish between the other alternatives.
However, addition of pyridoxal-P to an assay at low K+ and SOPC concentrations eliminates much of the curvature of the Lineweaver-Burk plot2 This observation favors dissociation of pyridoxal-P from the holoenzyme in the assay mix as an explanation of the sigmoid kinetics.
The data of Table I indicate that Na+ elicits nearly 30% of the activity observed with NH,+. Further, a reaction with Na+ at 1 M, 0.17 M S-methylcysteine (K,,, = 10 mM) (15) 0.025 M (CH,,),N/Hepps, pH 8.0, 25", gave a rate equal to 16.5% of that observed with 0.05 M NH,+." These observations are in direct contrast to earlier reports with other substrates, in which Na+ was shown to elicit less than 5% of the activity observed with NH,+ (3)(4)(5)16). The earlier results are consistent, however, with our findings (cf Table I) that much greater concentrations of cation, substrate, or both are required to demonstrate activity of Na+, and that the V,,,,, observed with this cation is much lower than that observed with K+. For example, in the experiment reported by Hogberg-Raibaud et al. (5), S-methylcysteine (K,,, = 10 mM (15)) and Na+ were employed at concentrations of only 70 and <50%, respectively, of those required for enzyme saturation.
Under these conditions, the enzyme would show much lower activity than the maximum expected assuming Michaelis-Menten kinetics. Thus, failure to note activity with Na+ results primarily from the fact that insufficient Na+ and substrate were added to the assay mixture.
The circular dichroic data of Fig. 4B provide convincing evidence that the competitive inhibitor, ethionine, interacts, with the holoenzyme in the absence of monovalent cation, but does so in such a way that the 508 nm absorbing species is not formed. The structure of the former adduct is not known; one possibility (since, as shown in Fig. 4A, it has the same absorption spectrum as enzyme alone) is that it has Structure ZZ (  (13). The exact ionic forms of the coenzyme and substrate are unknown. 6). The addition of cations shifts the reaction fromZZ toZZZ, a or b (Fig. 6), which for ethionine absorbs maximally at 508 nm (Fig. 2).
As indicated in Fig. 3, the maximum absorption at 508 nm, and thus the extent of formation of the quinoid intermediate III elicited by each monovalent cation in the presence of the inhibitor, ethionine, is directly proportional to the I',,,, observed with each cation and SOPC except for NH,. The activity with ammonia is much greater than expected, perhaps due to the stereochemical nature of the interaction with its ligands. Since the extent of formation of the quinoid intermediate reflects the labilization of the amino acid LY proton (41, the data are consistent with the view that labilization of the LY proton is the rate-limiting step in the reaction with SOPC. This is in contrast to the data obtained for a variety of other substrates of tryptophanase, including L-tryptophan, serine, S-ethyl-L-cysteine, and S-methyl-L-cysteine (4). When reactions with these substrates are allowed to proceed to partial completion in tritiated water, the unreacted substrate contained tritium. This is consistent with the argument that the rate-limiting step in the reaction occurs after the quinoid intermediate III, Fig. 6. Morino and Snell (4) supported this argument by showing the existence of appreciable intermediate near 500 nm during the steady state reaction with several substrates, such as tryptophan, S-methyl-, and S-ethyl-L-cysmine. However, the V,,,,, for SOPC is substantially higher than that for any of the previously tested substrates, and this higher rate could well be associated with a change in the ratelimiting step. In accordance with this, experiments similar to those reported previously (41, but with SOPC as substrate in "H,O containing 0.05 M NH,+, 0.1 M K+, or 1 M Na+, showed no incorporation of tritium into the unreacted SOPC.' In addition, no 500 nm absorbing species could be detected by a scanning stopped flow (17) spectrophotometric study of the reaction with SOPC in 0.1 M KCl; appreciable 500 nm absorption was detected when tryptophan, S-ethyl-, and S-benzyl-Lcysteine were used as substrates in the tryptophanase reac-tion4 Thus, the rate-limiting step with SOPC occurs before elimination of the /3 substituent; labilization of the LY proton as the rate-limiting step is consistent with these observations. The extent of the 508 nm perturbation, and therefore, the  Based on crystallographic structures of monovalent cation complexes such as nonactin (18) and enniatin B (19), it is expected that the monovalent cation would be complexed without its water of hydration. Therefore, the interaction of a cation with an enzyme must be considered in terms of the following scheme: Assuming that the hydration energy of the enzyme cation binding site is negligible, the affinity for a cation is primarily controlled by the difference between the free energy of hydration of the cation and the free energy of interaction of the naked cation with its site on the enzyme surface (20). For example, the free energy of hydration ranges from approximately 96 kcal/mol for Li+ to 47 kcal/mol for Cs+. The free energy of interaction of monovalent cation with nonactin, when referred to a vacuum, is essentially of the same magnitude (20). The differences between these values gives the net free energy of interaction (ranging from 1 to 4 kcal/mol) defining the order of affinities of nonactin with M+ in aqueous solution, K+ > Rb+ > Cs+ > Na+ > Li+, which is identical to that observed with tryptophanase disregarding NH,+ and Tl+. The binding constants for NH,+ and Tl+ with nonactin are anomously high (20). Based on these arguments, it is expected that the affinities of the naked cation for the enzyme are approximately the same as those with nonactin, and that their sequence follows the same order; namely, Li+ > Na+ > K+ > Rb+ > Cs+. Therefore, once bound, the actual attractive force between the cation and the interacting groups on an enzyme surface bears little resemblance to the order of maximum velocities. Furthermore, this attractive force is much larger than that indicated by the dissociation constant of M+ with enzyme and may provide important clues regarding the role of monovalent cations in catalysis. A similar argument was made by Jencks (21) for other systems.
Another interesting feature of these cations is their size. Cations with ionic radii smaller than 1.3 A, or larger than 1.5 A, are poor activators. This relationship between the size of a cation, including its volume and surface area, and the maximum velocity are presented in Fig. 7. The plot is given as a function of optimum size assuming that the size of NH,' is optimum, but that the rate with NH,+ is anomalous. The parameters are presented in terms of a percentage of the optimum. When the parameter was larger than the optimum, the excess over 100 was subtracted from 100. The data suggest that the forces of attraction between the naked cation and the enzyme, which, in fact, differ little, bring about the alignment of the peptide backbone into an optimum configuration.
An optimum configuration requires an optimum size or volume for the monovalent cation, so that the proper alignment of reacting groups needed for efficient catalysis can be achieved. The anomalous behavior of NH,+ suggests that the directional aspect of the attractive force may be important in the critical alignment of the interacting groups. Does the cation exert its effect by interaction directly with the reactants at the catalytic site, or at some site distant from the catalytic site? Data of Fig. 2 show that the cation binds in the absence of substrate, and those of Fig. 4 show that the  Fig. 6). Since proteins are flexible molecules, it would seem unlikely that the relatively small differential effect of each cation could be mediated if it were bound at some site on the protein surface distant from the active site. The most likely possibilities are that the cation is bound at the catalytic site and participates directly in the reaction, or alternatively is bound near the catalytic center to bring about the critical alignment of one or more functional groups. The fact that most pyridoxal-P-dependent enzymes that catalyze CX, /? elimination reactions require monovalent cations for maximum activity (1,13,22), whereas those catalyzing transamination reactions do not, also supports the view that these cations are involved directly or indirectly in the catalytic process.
A stoichiometry of 2 Tl+ per subunit is also consistent with a specific role for the cation, rather than a nonspecific conformational effect. The hyperbolic saturation curve for monovalent cation activation when excess substrate is present is consistent with participation of 1 cation per catalytic center. Whether or not the second mole has mechanistic significance will have to await additional data; for example, x-ray crystallographic evidence.