Studies of the competing rates of catechol oxidation and suicide inactivation of tyrosinase

Tyrosinase oxidation of catechols to ortho -quinones is accompanied by suicide inactivation of the enzyme. The rates of these competing processes vary and depend on the nature of ring substituents. For a series of 4-substituted catechols the relationships between structure and reaction rates have been examined using multiple regression. Significant but different structure-rate relationships were found for each process. The oxidation rate (k 1 ) is greatest for short hydrophobic substituents; there is an optimum substituent hydrophobicity (  ~ 0.7) for the rate of inactivation (k 2 ).


Introduction
The enzyme tyrosinase (EC 1.14.18.1) occurs widely in nature and one of its functions is the oxidation of tyrosine to dopaquinone in the biosynthetic pathway to the melanin pigments. 1,2In addition to this mono-oxygenase activity (phenols → ortho-quinones), tyrosinase is unusual in that it can also function as an oxidase and oxidises catechols 1 to ortho-quinones 2 by a subtly different mechanism. 2Because tyrosinase is uniquely expressed in pigment-generating cells and is up-regulated in the majority of pigment cell tumours it has been proposed as a targeting strategy for selective chemotherapy for malignant melanoma. 3However, although catecholic substrates are generally more readily oxidised by tyrosinase, the situation is complicated by the fact that oxidase activity of tyrosinase (catechols → ortho-quinones) is accompanied by enzyme inactivation (suicide inactivation) which would seriously compromise any therapeutic action.
The suicide inactivation associated with oxidation of catechols has been known for many years 4,5 but the mechanism remained unknown.We have recently provided evidence that this inactivation occurs when the catecholic substrate presents itself to the active site as a phenolic substrate, [6][7][8][9] leading to irreversible reductive elimination of copper 10 from the enzyme.During the course of these studies we have examined a number of 4-substituted catecholic substrates and have measured rate constants for their oxidation (k1) and for the rates at which they inactivate the enzyme (k2) (Scheme 1).In this paper we describe quantitative relationships between catechol structure and the rate constants k1 and k2, and demonstrate how the nature of the substituent influences the relative rates of the competing processes shown in Scheme 1.

Results and Discussion
We examined the oxidation by tyrosinase from Agaricus bisporus of a series of catecholic substrates and measured the initial oxidation rates and the total oxygen utilization by polarimetry.The kinetics of oxygen utilization by the catechols 1a-g (see Table 1), adjusted to give equivalent initial oxidation rates, are shown in Figure 1(a) and demonstrate significant variation in the relative inactivation rates resulting in different total oxygen uptake.A plot of the k1 and k2 values (Figure 1(b)) failed to indicate any correlation between the oxidation rate and the inactivation rate.The linear correlation coefficient (r 2 = 0.381) is nonsignificant suggesting that different reactions are involved.
To investigate the effect of substitutent groups on the oxidation and inactivation rates a multiple regression analysis was carried out.Values of log k1 and log k2 for catechols 1a-m are shown in Table 1 together with the substituent parameters used in the analysis.The catechol substituents (R) were chosen to give a good variation of steric, electronic and hydrophobic properties. 11(2) n = 9, r = 0.949, s = 0.140, F = 27.22,p = 0.001 We interpret Equation 1 as indicating that for rapid catechol oxidation by mushroom tyrosinase a short (L) and hydrophobic () substituent R is required.This provides a quantitative of the qualitative observation that catechol 1e (R = H), 4-fluorocatechol 1b (R = F) and 4-methylcatechol 1c (R = CH3) have the fastest oxidation rates and larger substituents or hydrophilic substituents result in slower oxidation.
If  and L were the only substituent properties that determined the rate of catechol oxidation by tyrosinase, the catechols 1k-m (R = NO2, CN, CF3) would have log k1 values in the range 0.9-1.7 based on Equation 1. Clearly there is an additional effect in these derivatives, which it is reasonable to suppose is an electronic effect, since NO2, CN and CF3 are strongly electronwithdrawing (p 0.5-0.8).It is well established that strongly electron-withdrawing substituents make catechols more difficult to oxidize and we have encountered similar effects in our studies of the autoactivation mechanism of tyrosinase. 12rosinase inactivation (log k2) An analysis of the ten catechols that inhibit tyrosinase (1a-j, Table 1) identified a significant parabolic relationship between log k2 and  (Equation 3).This relationship suggests an optimum substituent hydrophobicity (opt ~ 0.7).Without the  2 term a significant correlation is not observed.We have systematically searched for steric parameters, including L, B1, B4 and MR, that give a correlation without the use of  2 but no other significant correlation was found.Leaving out catechol 1j (R = OH), for the reasons discussed above, gave Equation 4 Apart from the hydroxy derivative 1j discussed above, the only other outlier in Equation 3 is the chloro derivative 1g [log k2(obs.)-0.149] which is significantly less active as an inhibitor than predicted [log k2(calc.)0.463].This low activity is surprising since this substituent has optimal hydrophobicity based on Equation 3 [opt 0.7; Cl 0.71].A possible reason for this observation is that an electronic effect is becoming significant.The electron-withdrawing power of Cl (measured by p) is much greater than that of the other substituents in the set 1a-j.However, Cl (p 0.23) is not as electron withdrawing as NO2 (p 0.78), CN (p 0.66) and CF3 (p 0.54), and for this reason it may not show complete lack of activity like the NO2, CN and CF3 derivatives 1k-m.To further investigate this apparent electronic effect, we made the assumption that the NO2, CN and CF3 derivatives are very weakly active and that log k2 = -3.0,which is below the lower limit of measurement.Using these values for compounds 1k-m, multiple regression gave the correlation shown in Equation 5.In this relationship the  and  2 terms are essentially the same as in Equations 3 but now the p term is also significant and the Cl derivative is no longer an outlier.Leaving out catechol 1j improves the correlation (Equation 6).log k2 = 1.129 (0.416) -0.956(0.481)  2 -3.000 (0.638) p -0.239 (5) n = 13, r = 0.873, s = 0.786, F = 9.58, p = 0.004 log k2 = 0.677 (0.330) -0.707(0.353)  2 -3.634 (0.499) p -0.076 (6) n = 12, r = 0.942, s = 0.562, F = 21.07,p = 0.0004

Conclusions
The analysis described above shows that the substituent properties that influence the rate of tyrosinase oxidation of catechols to ortho-quinones (Equation 1) are different from the properties that influence the rate of inactivation of tyrosinase by catechols (Equation 3).This is in agreement with our proposal that during oxidation (oxidase activity) both catechol oxygen atoms bind to the coppers in the active site whereas for inactivation (mono-oxygenase activity) only one catechol oxygen atom binds to copper. 8The ring substituents therefore present to different regions of the enzyme during the competing transformations.

Figure 1 .
Figure 1.(a) Comparative oxygen utilization kinetics for the catechols 1a-g adjusted for equivalent initial oxidation rates; (b) Scatter plot of initial oxidation rate (k1) versus the inactivation rate (k2).

Table 1 .
Rate constants (log k1 and log k2) and structural parameters for catechols 1

Oxidase activity (log k1) Three
revealed the significant relationship summarised in Equation1, where  is the substituent hydrophobic constant and L is the Verloop STERIMOL substituent length parameter.No evidence of correlation with the other substituent parameters shown in Table1was found.Using Equation 1 the oxidation rate for catechol 1j is high [log k1(calc.)1.311].Since the rate constants for catechol 1j were obtained at pH 3.5, whereas all other compounds were measured at pH 6.75, we investigated the exclusion of this derivative.Equation2shows the relationship without catechol 1j.Although the correlation is slightly improved, leaving out this derivative does not significantly change the relationship.= 0.876, s = 0.216, F = 11.52,p = 0.006 log k1 = 0.288 (0.067) -0.336(0.051)L + 2.578 of the catechols 1k-m were not oxidised by tyrosinase and had no inhibitory effect as shown by subsequent oxidation of 4-methylcatechol.They were not included in the analysis of oxidase activity.A preliminary analysis of the other ten compounds 1a-j