Determination of the pKa Values of some Biologically Active and Inactive Hydroxyquinones

As constantes de dissociação aparentes (pK a ) de quatro 2-hidroxinaftoquinonas, diferentemente substituídas em C-3, foram determinadas em meio aquoso-etanólico (1:1, v/v), utilizando titulações pH-métricas e espectrofotométricas. O isolapachol (pK a <6) mostrou ser mais ácido que seu análogo natural, o lapachol, (pK a >6). Os derivados 3-metilaminados apresentam dois valores de pK a , um relacionado ao grupo enólico e o outro ao sal de amônio, e são zwitteriônicos, em larga extensão, em pH fisiológico. As possíveis conseqüências desses parâmetros frente a suas atividades biológicas são discutidas.


Introduction
Lapachol (1) possesses anti-tumour, antibiotic, antimalarial, anti-inflammatory and anti-ulcer properties, 1 whilst both 1 [2][3][4][5][6] and isolapachol (2) 4,6 exhibit significant activities against the etiological agents of a number of tropical diseases.Additionally, the 2-hydroxy-1,4-naphthoquinones 1 and 2 are potent molluscicides (Table 1) with activities against both mature adults and egg masses of Biomphalaria glabrata, 2,3,7 the intermediate host of the causative agent of schistosomiasis (Schistosoma mansoni). 8,9In contrast, a group of fourteen 2-hydroxy-3-methylnitrogenated-1,4naphthoquinones were inactive in molluscicidal assays, 7 and the representative compounds 3 and 4 showed no activity against Artemia salina, 10 an organism used for cytotoxicity screening. 4,11 has been suggested that the mechanism of molluscicidal action of the 2-hydroxy-1,4-naphthoquinones may be related to their redox parameters, 7 which in turn are strongly associated with their acid-base properties. 12However, it is well-established that biological activity towards a live host is a complex outcome that is not usually dominated by one parameter.Generally, various physico-chemical properties impact on the pharmacokinetic and metabolic fate of a biologically active compound and, hence, a thorough understanding of such characteristics is crucial in understanding the mode of action. 13][15][16] Of particular interest is the role of pK a on absorption, since this is often related to its effect on lipophilicity and solubility. 14In order to measure pK a values, it is necessary to expose the analyte to an environment of changing pH and to monitor a specific property that varies as a function of the ionisation state of the molecule.Several methods for conducting such an analysis are available including the rapid techniques of pH-metric titration and hybrid pHmetric/UV titration.The pH-metric titration is an important reference method because it can be used to measure pK a values between 2 and 12 in the presence or absence of a chromophore, provided that the sample can be dissolved in water or water and a co-solvent over the pH range of interest.In this method, acid-base titration of a sample solution is monitored with a glass pH electrode, and pK a is calculated from the change in shape of the titration curve in comparison with that of a blank titration carried out in the absence of the sample. 14The hybrid method, also known as spectrophotometric pK a determination, is an attractive alternative provided that the sample is soluble in water to the extent of 10 -6 mol L -1 and also contains chromophore(s) in proximity to the ionisation centre(s) such that the protonated and deprotonated forms exhibit sufficient spectral dissimilarities. 17In this method, an acid-base titration is conducted across a pH range that includes the pK a , and UV spectra are recorded for each pH value. 14ith the aim of elucidating the spectra of biological activities of the 2-hydroxy-3-methylamino-1,4-naphthoquinones, we have determined the pK a values of representative compounds 3 and 4, together with those of the active 2-hydroxy-1,4-naphthoquinones 1 and 2, and have evaluated these data in terms of the known bioactivities of the respective analytes.The pK a values were determined in water: ethanol (1:1, v/v) solutions, in order to allow the complete dissolution of the samples over the entire pH range, using both the pH-metric and hybrid pH-metric/UV titration methods.

Potentiometric titrations
Potentiometric titrations were carried out using a Cerko Lab System microtitration unit, consisting of a microprocessor-controlled dosing unit, a precise 18-bit pH-meter and a host PC, together with a Mettler Toledo InLab 423 pH microelectrode.For measurements in water solutions, the pH electrode was calibrated against at least five buffers as recommended by IUPAC. 20For the water: ethanol system, the electrode was calibrated using potassium hydrogen phthalate as described in the reference value pH standard (RVS) method previously published. 21,22In a typical experiment, an aliquot (2 mL) of a solution of quinone and perchloric acid (ca. 5 x 10 -4 mol L -1 ) was titrated against sodium hydroxide (ca. 10 -2 mol L -1 ).The resolution of the titration step was < 0.001 mL (using a 0.5 mL Hamilton syringe), and the resolution of the voltage measurement was < 0.1 mV.Equilibrium constants were calculated with the aid of the STOICHIO program, which employs a non-linear least-squares Gauss-Newton-Marquardt algorithm. 23

Spectrophotometric titrations
Compounds 1-4 were dissolved in water:ethanol (1:1), acidified to low pH values (~ 3) with 0.02 mol L -1 hydrochloric acid, and aliquots (typically 1.5 mL) titrated against sodium hydroxide (8.944 x 10 -3 mol L -1 ) in steps of 0.002 mL or less to a final pH of 12. 24 Initially, sample solutions contained the protonated hydroxyquinones at concentrations of ca. 5 10 -4 mol L -1 , the exact values being c 1 = 5.32 10 -4 mol L -1 , c 2 = 3.92 10 -4 mol L -1 , c 3 = 5.39 10 -4 mol L -1 and c 4 = 7.43 10 -4 mol L -1 .Samples (typically 1.5 mL) were titrated in steps of 0.002 mL or less with sodium hydroxide (8.944 x 10 -3 mol L -1 ) to a final pH of 12. 24 The UV-Vis spectra were recorded, after the addition of gradually increasing amounts of base, using a Perkin Elmer Lambda 40 spectrophotometer and a 1cm quartz cuvette.All measurements were carried out at 298.1 K. Plots of absorption at different wavelengths versus apparent pH values of sample the solution generated sigmoid curves.The pK a constants were determined for samples 1-4 on the basis of Henderson-Hasselbach equations. 25,26

Results and Discussion
Under acidic conditions, the spectrum in the range 300-600 nm of a water:ethanol (1:1) solution of lapachol (1) showed two absorption bands with maxima at 337 and 390 nm, whereas in alkaline solution the spectrum exhibited a single absorption band with its maximum located at 490 nm (Figure 1A).The same features could be observed upon titration of an acidic solution of 1 with sodium hydroxide (Figure 1B), and the changes in UV absorbance monitored at 337, 370, 390 and 490 nm throughout the spectrophotometric titration showed a similar pattern (Figure 2).A pK a value for 1 of 6.15 ± 0.01 was calculated at a wavelength of 490 nm on the basis of Henderson-Hasselbach equations.
Potentiometric titration of (1) performed in a water:ethanol (1:1) solution containing perchloric acid showed a pattern characteristic of a mixture of more and less strong acids.Two steps were observed in the potentiometric curve (Figure 3), the first relating to the neutralization of perchloric acid and the second to the formation of the enolate from the enol.From the potentiometric data, the equilibrium constant for 1 was calculated on the basis of the Gauss-Newton-Marquardt algorithm 23 and gave a value for pK a of 6.31 ± 0.03.The data obtained for 1 from the spectroscopic and potentiometric titrations are in good agreement and support the model of the equilibrium in water:ethanol solution.
The pH-dependent changes in the UV spectrum of isolapachol (2) were comparable with those observed for 1 monitored under similar conditions.At pH 3.2, compound 2 exhibited absorptions at 337 nm (shoulder I) and 445 nm (band II), but in alkaline solution a bathochromic shift was observed and a broad band at 540 nm appeared (Figure 4A).The same features were displayed upon titration of an acidic solution of 2 with sodium hydroxide (Figure 4B).Slight alterations in the spectrum of 2 could be observed between pH 3.2 and 4.7, with an increase in absorbance at shoulder I and a concomitant decrease of band II.This could be related to possible tautomerism and protonation of the double bond at lower pH values (Figure 5), a phenomenon that did not affect the pK a .A value of 5.75 ± 0.01 for the pK a of 2 was calculated at a wavelength of 540 nm (Figure 6) on the basis of Henderson-Hasselbach equations. 25,26The pK a of 2 was also determined by the potentiometric method as described for compound 1, and a value of 5.98 ± 0.03 was obtained (data not shown).
With respect to the 2-hydroxy-3-methylaminonaphthoquinones 3 and 4, more complex features were, as expected, observed in the plots of UV absorption versus pH.In the case of compound 3, which carries a tertiary cyclic amine (methylmorpholine), the UV-Vis spectrum measured at pH 1.2 exhibited a maximum absorbance at 330 nm (Figure 7A).Upon the addition of base, an absorbance maximum was observed at 323 nm and a second broad band appeared at 452 nm, the latter undergoing a bathochromic shift towards 470 nm as the pH approached 12 (Figure 7B).After 3h under alkaline conditions, this absorbance pattern changed slightly indicating a certain degree of instability of the quinone.The changes in UV absorbance of 3 were monitored at 330, 366, 440 and 470 nm throughout the spectrophotometric titration (Figure 8), and the plots obtained indicated the presence of two points of inflection.Although  it was not possible to obtain a complete curve at lower pH values, the first ionisation value could be estimated as < 3.5 at a wavelength of 470 nm.The second ionisation constant, obtained at higher pH, could calculated precisely from the second inflection point at 470 nm and gave a pK a2 for compound 3 of 8.59 ± 0.02.
The spectroscopic behaviour of 4 resembled that of compound 3 and revealed the presence of two pK a values.For compound 4, however, the absorption at 450 nm increased upon base addition and attained an almost constant value in the pH range 4-9.5 (Figure 9).Moreover, a further increase in pH resulted in an increase in absorption, such that the data only allowed the conclusions that pK a1 could be lower than 3.50 and that pK a2 could be higher than 10.
More precise information was obtained from potentiometric investigations of water: ethanol solutions of compounds 3 and 4. In the potentiometric curve of 3 (Figure 10), only one inflection point was visible.However, after fitting of the data using the STOICHIO program 23,24 the existence of two equilibria became apparent, one with pK a1 (OH) at 3.36 0.13 and the second with pK a2 (NH + ) at 8.81 0.08.For compound 4, potentiometric measurements (Figure 11) and fitting of the data using the STOICHIO program 23,24 revealed values for pK a1 (OH) of 3.36 0.13 and pK a2 (NH + ) of 10.01 0.07. 24The pI values of compounds 3 and 4 were estimated to be 6.08 and 6.68, respectively.
The deprotonation steps for the aminohydroxyquinones, as exemplified by compound 3, are depicted in Figure 12 showing        that pK a1 is related to the ionisation of the enol and that pK a2 relates to the removal of proton from the salt.
Comparison of the potentiometric and spectroscopic data for compounds 1-4 lead to the conclusion that the proposed equilibrium models were valid and generate the expected values for pK a1 and pK a2 .Generally, the potentiometric method gave more precise pK a values than the spectrometric method, and could also be employed over a broader range of pK.The main disadvantage of the method relates to the difficulty in finding suitable models for the equilibria.
The profiles of the distributions of base (AO -), neutral (AOH) and acidic (AOH 2 + ) species of compounds 1-4 are shown in Figure 13.In terms of pK a1 the order of acidity is 3 = 4 > 2 > 1, and an enolate ion is generated in all cases.For compounds 3 and 4, intramolecular hydrogen bonding   explains the higher acidity of these quinones (Figure 14B), whilst for isolapachol (2), the extension of conjugation to the double bond of the substituent explains its higher acidity in comparison with 1.
Generally, the variation in basicity of the aliphatic amines (primary, secondary and tertiary) is associated with stereoelectronic and solvation effects of their conjugate acids.For compounds 3 and 4, the differences in pK a2 must be related to solution 27 and substituent electronwithdrawing effects, the latter referring to the morpholine group of 3. Solvation through hydrogen bonding tends to increase the apparent strength of all amines, since a positively charged ammonium ion is more effectively solvated than an uncharged amine.In comparison with 3, the extent of solvation of the ammonium salt of 4 is enhanced since increased hydrogen bonding stabilizes the conjugate acid (Figure 14A) and consequently leads to a higher pK a2 value. 27n terms of biological properties, compounds 1 and 2 exhibit a broad spectrum of activities, whilst 3 and 4, as representatives of a class of fourteen 2-hydroxy-3methylnitrogenated-1,4-naphthoquinones, are inactive in various biological assays (Table 1).
The great majority of drugs are passively absorbed and need to cross the lipid bilayers that constitute the hydrophobic environment of biological membranes.][16] In this context, the hydrophobic nature of a molecule is fundamental to the facilitation of its transport across cell membranes and can be important in its interactions with receptors. 16A zwitterionic species possesses a large intramolecular multipole moment owing to its multiplicity of oppositely charged groups. 28Consequently, most zwitterionic compounds exhibit low solubility in polar and non-polar media, and this results in low membrane permeability. 29As measured in the present study, the pI values of 3 and 4 (6.08 and 6.68, respectively) indicate a strong contribution of the zwitterionic forms of these molecules at physiological pH values, and this could be the reason for their lack of biological activity.However, organisms possess various mechanisms that can circumvent such problems of transportation involving, for example, ion-pairing or the participation of L-carnitine.Hence, in order to obtain unequivocal proof of the inactivity of the methylaminohydroxyquinones, in vivo or in situ assays need to be carried out. 29,30

Conclusions
At physiological pH, isolapachol (2) is more acidic than lapachol (1), and the biologically inactive methylaminohydroxyquinones 3 and 4 are, to a large extent, zwitterionic.In all four compounds, the first pK a is related to the ionisation of the enol, and the differences in acidity of the -OH group may be explained in terms of the stabilisation of the enolate ion by hydrogen bonding.Compound 3 has a tertiary amino functional group, the conjugate acid of which is stronger than the acid derived from the secondary amine present in 4. The results obtained can be considered as starting points in the understanding of the mechanism of biological activity, however, additional work is needed in order to determine whether these parameters are relevant under physiological conditions and in vivo.

Figure 2 .
Figure 2. Absorption at different wavelengths of a water:ethanol (1:1) solution of 1 as a function of pH.The point of inflection on the 490 nm plot gave a pK a value of 6.15.

Figure 6 .
Figure 6.Absorption at 490 nm of a water:ethanol (1:1) solution of 2 as a function of pH.The point of inflection gave a pK a value of 5.75.

Figure 8 .
Figure 8. Absorption at different wavelengths of a water:ethanol (1:1) solution of 3 as a function of pH.The point of inflection on the 470 nm plot gave a pK a value of 8.59.

Figure 14 .
Figure 14.A) Possible hydrogen-bonding interactions between solvent and the enolate ion of (a) compound 4, and (b) compound 3. B) Stabilization by intramolecular hydrogen bonding of the enolate ion of (a) compound 4, and (b) compound 3.