摘要:

PhysChemComm, 2000, 9 Matthias O. Schmitt and Siegfried Schneider Institut für Physikalische und Theoretische Chemie, Friedrich-Alexander-Universität Erlangen-Nürnberg, Egerlandstraße 3, 91058 Erlangen, Germany. E-mail: schneider@chemie.uni-erlangen.de Received 17th July 2000, Accepted 25th July 2000, Published 10th August 2000 The complexation of various "normal" tetracyclines and anhydrotetracycline with Mg2+ and Ca2+ was investigated under physiological conditions (aqueous Tris buffer at pH 7.0 and 8.5) by spectrometric titration. The generated sets of UV-Vis absorption and fluorescence spectra were analysed by two techniques, namely (i) classical evaluation which relies on the occurrence of isosbestic points and (ii) multivariate fitting employing the program Specfit.The latter yields not only the conditional binding constants, but also, for each tested kinetic scheme, the absorption or emission spectra of the individual metal– tetracycline complexes. It was found that absolute values of binding constants and deduced spectra are in all cases in favour of a consecutive metal ion complexation yielding 1 : 1 and 2 : 1 metal–tetracycline ion complexes. Tetracycline and oxytetracycline with Mg2+ at both pH 7.0 and 8.5 form the so-called Mg-type complexes which were already described in the literature (first metal ion at C11– O{Mg}O–C12, second at C4–Me2N{Mg}O–C3). Doxycycline and anhydrotetracycline bind the first Mg2+ like tetracycline, but the binding site of the second ion is still unclear. In the case of Ca2+ two ions are bound consecutively as Mg-type complexes only at pH 7.0.At pH 8.5, so-called Ca-type complexes are formed, the first Ca2+ being bound via C12–O{Ca}O–C1 and the second via C10–OH{Ca}O–C11. In most cases binding constants determined from absorption data coincided with those derived from fluorescence data. An exception is anhydrotetracycline where excited state reactions obviously lead to an emitting state with a different conformation. R Spectroscopic investigation of complexation between various tetracyclines and Mg2+ or Ca2+ The existing literature does not provide a sufficient and unequivocal image of Tc complexation. A variety of stoichiometries are proposed for the complexes, depending on the type of measurement and the solvent used.Table 2 gives a short overview. Table 1 Substitution pattern for the tetracyclines investigated Derivative X2 Abbreviation X1 H OH H CH3 OH CH3 OH 5a, 6 dehydrated Tetracycline Tc 6-Deoxy-5- Doxy oxytetracycline 5-Oxytetracycline Oxy Anhydrotetracycline Atc OH CH3 CH3 1 Introduction The interaction of tetracyclines (Table 1) with protons and various metal ions was the subject of numerous publications during the past decades. The main reason for this interest was the fact that the understanding of protonation1–15 and complexation of tetracyclines with calcium and magnesium ions (e.g. references in Table 2) as present in blood and cell plasma seemed to be important to rationalize their pharmacokinetical properties as, e.g., membrane permeability, their partition in different tissues and the incorporation in the bones during bone formation.Although still unexplored in detail, it is known that tetracyclines are not only transported as metal complexes after their application, but also that the antibiotic action of tetracyclines necessitates magnesium ions. This means that the active antibiotic species is a magnesium complex.16,17 Even the Tn10 encoded tetracycline resistence in bacteria is triggered by a tetracycline magnesium complex.18,19 In recent times there have been reports of tetracyclines that inactivate metallo matrix proteinases by complexing the metal ions of their active centres and the inactivation of the HIV integrase in the same manner.20,21 DOI: 10.1039/b005722nTable 2 Stoichiometry of Mg2+ and Ca2+ complexes as proposed in various publications.Protonation of the complexes is disregarded Derivative M2+ Stoichiometries M2+–Tc Literature Tc Oxy Doxy ATc The variety of different stoichiometries is expanded by the proposal of different binding sites of the M2+ ions to tetracyclines within a complex of given composition and, in case of binding more than one metal ion, also by the sequence of binding. Furthermore, there is some controversy about the conformation such complexes are said to adopt. The general line of arguments used when drawing conclusions on protonation and metal complexation behaviour is as follows: acidity of a functional group correlates with the basicity of the non-protonated unit.High basicity, in turn, is a prerequisite for strong metal ion coordination, despite the fact that the affinity of a deprotonated functional group towards Mg2+ and Ca2+ does not necessarily parallel exactly its affinity to the proton. Consequently, the acidic groups should be the main anchors for complexation. We assume that the dominant interaction between the metal cation and tetracyclines is coulombic in nature and that the changes induced in the absorption spectra by M2+ or proton complexation are approximately the same. From the above said it is quite reasonable to assume that in any metal ion binding to tetracyclines at least one of the molecules’ acidic groups is involved.The hydrochloride salts of the tetracyclines are considered here to be tri-protic acids, even though a fourth deprotonation has been postulated recently.15 The tetracyclines’ most acidic functional group is the tricarbonyl system centered at C2 with pKA1 ß 3.2 (C3–OH group),3 which absorbs at ~270 nm.35,36 The molar extinction coefficient of the protonated form with e the deprotonated one with e272 nm ß15 400 M–1 cm–1 (values for tetracycline, publication in preparation). By convention, this group is named the A-ring chromophore and here it is assumed to be non-fluorescent,7,35 even though A-ring fluorescence is discussed in literature.37 Mg2+ 1 : 1 Mg2+ 1 : 2, 1 : 1, 2 : 1 Mg2+ 1 : 1, 2 : 1 Ca2+ 1 : 1 Ca2+ 2 : 1, 2 : 1 Ca2+ 1 : 1, 2 : 1 Mg2+ 1 : 1 Mg2+ 1 : 1, 2 : 1 Mg2+ 1 : 2, 1 : 1, 2 : 1 Ca2+ 1 : 1 Ca2+ 1 : 1, 2 : 1 Ca2+ 1 : 2, 1 : 1, 2 : 1 Mg2+ 1 : 1 Mg2+ 1 : 1, 2:1 Mg2+ 1 : 2, 1 : 1, 2 : 1 Ca2+ 1 : 1 Ca2+ 1 : 1, 2 : 1 Ca2+ 1 : 2, 1 : 1, 2 : 1 Mg2+ 1 : 1 Mg2+ 1 : 1, 2 : 1 Ca2+ 1 : 1 Ca2+ 1 : 1, 1 : 2 268 nm ß18 000 M–1 cm–1 is higher than that of Ref.9, 14, 22–29 Ref. 30, 31 Ref. 32, this work Ref. 9, 22–25, 27, 29 Ref. 30, 31 Ref. 32, this work Ref. 1, 9, 23, 24, 33 Ref. 34, this work Ref. 30, 31 Ref. 9, 23, 33 Ref. 34, this work Ref. 30, 31 Ref. 9, 27 This work Ref. 30, 31 Ref. 27 This work Ref. 30, 31 Ref. 32 Ref. 5, this work Ref. 32 This work The protonated C4 dimethylamino group (pKA3 ß9.6 , ref.3, 6) has the lowest acidity and does not contribute directly to the UV absorption of the A-ring. Depending on the protonation state, it forms a hydrogen bond with C3–O or C12–O (ref. 36) and thus stabilizes different conformations of the molecule. Since through this hydrogen bond the A as well as the BCD chromophore are altered in their geometry, the UV-Vis absorption of both chromophores and also the fluorescence of the BCD chromophore could be influenced. max) The b-hydroxyketo system at C11, C11a and C12 (pKA2 ß 7.6, ref. 3) is part of the BCD chromophore which is made responsible for the long wavelength absorption band of the tetracyclines, and also contributes to the absorption below 325 nm. The long wavelength absorption maximum ( l of the protonated form at about 365 nm has e365 nm ß15 000 M–1 cm–1 and that of the deprotonated form e375 nm ß17 000 M–1 cm–1 (values for tetracycline).The BCD chromophore represents the fluorophore of the tetracyclines,7, 35 and consequently the emission is dependent on the protonation state of the b-hydroxyketo system. The fluorescence yield of the protonated fluorophore is low, and that of the deprotonated system at least ten times higher. In view of the above mentioned biological aspects, the aim of this work was to investigate the complexation behaviour of tetracyclines under experimental conditions which are close enough to those being present in living organisms and, on the other hand, are appropriate to produce results which can be unequivocally interpreted. Consequently, spectroscopic titrations of the tetracyclines with metal ions in aqueous Tris buffer were performed at two fixed pH values, namely at pH 7.0 and 8.5.Consistent sets of UVVis absorption and fluorescence spectra were recorded at various times during the titration. Because of the spectral changes discussed above, quantitative measurements allow not only for the determination of the association constants, but also—by means of certain algorithms—the determination of the absorption and emission spectra of the different moieties differing either by the number of metal ions complexed or the position of complexation. Consequently qualitative analysis of the absorption spectra supplies additional information on the binding sites of the metal ions and is supported by the comparison with the information obtained from fluorescence spectra.Nevertheless unique statements on the conformation of the complexes can not be deduced from absorption or fluorescence measurements and we therefore refer to the numerous CD 5,23,25,30,32,36,38–46 and NMR47–49 measurements aiming at this topic and to recent MO-calculations.15,50,51 Probable binding sites and the characteristic spectral changes related to M2+ binding to these sites will however be discussed. 2 2 Materials and methods 2.1 Materials Tetracycline hydrochloride (Tc·HCl), oxytetracycline hydrochloride (Oxy · HCl) and doxycycline hydrochloride (Doxy·HCl) (Fluka, Deisenhofen, Germany) and anhydrotetracycline (ATc) (Acros, Geel, Belgium) were used without further purification.Mg2+ and Ca2+ solutions were prepared by diluting 1 M stock solutions of MgCl and CaCl2 that were made by dissolving MgCl2hexahydrate (Aldrich, Steinheim, Germany) and CaCl2 dihydrate (Merck, Darmstadt, Germany). Deionised water from a Millipore-Q unit was used throughout to prepare solutions. All solutions were buffered by 50 mM Trisbuffer (ICN Biomedicals, Aurora, Ohio, USA). The desired pH was obtained by adding dilute hydrochloric acid. 2.2 Steady state spectroscopy All experiments were carried out at room temperature (20 ± 1 °C) using freshly prepared solutions of the tetracyclines. The total concentrations cT = 3 × 10–5 M for Tc, Oxy and Doxy and 2 × 10–5 M for ATc were kept constant during each titration by adding the tetracycline in the same concentration to all the metal ion solutions used.This procedure allowed permanent examination of the quality of the titration during the experiment. The procedure necessitates reversibility of the complexation, which was found to be valid. The total volume at the beginning of a titration was 100 ml, samples of 2 ml were used to record spectra after each addition of metal ions. UV-Vis absorption spectra were recorded with a Perkin- Elmer Lambda-2 spectrometer at a wavelength resolution of D l = ± 1 nm. Hellma 10 × 10 mm QS cuvettes were used throughout. Fluorescence emission spectra were recorded on a Perkin-Elmer LS-50B spectrometer.The experimental parameters are given in Table 3. The excitation wavelength was chosen as the wavelength of the isosbestic point found for the first complexation step in the 350 nm region. 2.3 Numerical analysis Due to the better signal-to-noise ratio, the higher content of information (vide supra) and direct linearity between optical density and concentration, this analysis relies mainly on absorption data although fluorescence data have also been evaluated. Analysis of the data obtained during a titration with respect to binding constants was carried out using two methods. A preliminary analysis made use of the various isosbestic points occurring in a series of spectra. For an equilibrium of absorbing species A and B, A B, at least one isosbestic point can be found.If this first equilibrium is followed by a A B iso second one, B C, measurements performed at l can be treated as a single equilibrium and yield a simple sigmoidal titration curve. The same applies for the B C iso B by measurements at l evaluation of A . Binding constants were obtained by fitting these curves with a sigmoidal function. For detailed analysis we made use of the program Specfit.52–55 The program performs a single value decomposition Y = USVT of the experimental data matrix Y which is built up of the set of wavelength scans. In this algorithm, U and V are sets of orthogonal evolutionary and spectral eigenvectors, S is a set of weighting factors.This procedure provides the number of eigenvectors n (equivalent to the number of distinquishable species) that contribute significantly to the dataset. Then, a set of equilibria between n assumed species (model) of the actual system must be inputted into the program. Taking into account the relations which must exist between the concentrations of the various species due to the postulated equilibria, the program calculates the conditional binding constants b� and the spectra of the n incorporated species based on a nonlinear least squares (NLSQ) fit. In addition, the concentration profiles of the various species resulting from the calculated b� values are provided. Various models (some of them are mentioned in Table 2) have been tested and were ranked with respect to the quality of the fitting parameters and the physical significance of the calculated spectra (see below).Table 3 Experimental parameters for the titrations applying absorption and fluorescence spectroscopy M2+ cT/M Derivati ve Tc Oxy Doxy ATc 3 × 10–5 Mg2+ 3 × 10–5 Ca2+ 3 × 10–5 Mg2+ 3 × 10–5 Ca2+ 3 × 10–5 Mg2+ 3 × 10–5 Ca2+ 2 × 10–5 Mg2+ 2 × 10–5 Ca2+ Equilibrium A B A B A B A B B C B C B C B C Equilibrium 3 Results 3.1 Titrations of tetracycline and oxytetracycline with Mg2+ and Ca2+ Tetracycline and oxytetracycline show completely analogous behaviour with respect to absorption and fluorescence spectra when titrated with Mg2+ and Ca2+.Only subtle differences are found in position and intensity of the bands. For both ions, two successive complexational equilibria can be found that are confirmed by several isosbestic points per equilibrium (see Table 4 for details). Table 4 Equilibria and corresponding isosbestic points for the titrations of tetracycline and oxytetracycline with Mg2+ and Ca2+ Tetracycline M2+ pH 7.0 Mg2+ 8.5 Ca2+ 7.0 8.5 Oxytetracycline pH M2+ 7.0 Mg2+ 8.5 Ca2+ 7.0 8.5 A B A B A B A B B C B C B C B C lex/nm pH = 8.5 pH = 7.0 350 366 349 354 354 —442 445 353 349 354 354 350 356 440 440 liso/nm 226, 272, 353 (283), 349 229, 270, 353, 409 280, 353, [390–440] 227, 270, 356 285,340 222, 266, 302, 320, 366 280 liso/nm 226, 272, 351, 420 282, 354 224, 273, 348, 407 282, 334, 396 228, 270, 354 283,346 248, 323, 364 291(a) Fig.1 Absorption spectra of tetracycline (3 × 10–5 M) in aqueous Tris buffer with increasing concentrations of Mg2+ at (a) pH 7.0 and (b) pH 8.5; cT(Mg2+) steps between the two spectra are 0.2 log units. Selected concentrations are given in the insets. (a) Fig. 2 Absorption spectra of tetracycline (3 × 10–5 M) in aqueous Tris buffer with increasing concentrations of Ca2+ at (a) pH 7.0 and (b) pH 8.5; cT(Ca2+) steps between the two spectra are 0.2 log units. Selected concentrations are given in the insets. Fig. 1 and 2 show the absorption spectra for the titration with Mg2+ and Ca2+ at pH 7.0 and 8.5, respectively. The absorption spectra of the titration with Ca2+ at pH 8.5 differ noticeably from those of all the other combinations of tetracycline and Mg2+.During the titration with Mg2+, at low Mg2+ concentrations for both pH values one finds a bathochromic shift of the long wavelength absorption band (the kink at 380 nm is an experimental artefact due to changes of filters). This suggests binding of the first M2+ ion to the BCD chromophore. The spectra show identical changes when tetracycline itrated with Ca2+ at pH 7.0 over the range of low Ca2+ concentrations. The second step of complexation has only a small effect on the position of the long wavelength band, but gives rise to a sharp band at about 270 nm.This indicates the involvement of the A chromophore in binding the second metal ion for these three combinations of metal ions and pH. (b) (b) Equally in two steps, two Ca2+ ions are bound to tetracycline at pH 8.5. Here, both steps cause a shift of the long wavelength absorption. The sharp A ring band is absent, whereas a broad band with a maximum at 296 nm (Tc), which has not yet been seen in any of the other titrations, appears at low Ca2+ concentrations. Due to its wavelength this band must be caused by the BCD chromophore. The fluorescence intensity during this titration initially rises with the Ca2+ concentration but drops again for higher concentrations [see Fig. 3 (b)]. This points to an involvement of the BCD chromophore in both complexation steps.Fig. 3(a) shows, exemplary for the three equally behaving systems mentioned above, the fluorescence spectra for the complex formation between tetracycline and Mg2+ at pH 7.0. Only a steady rise of fluorescence intensity is found for these three systems, which matches perfectly with the rise of absorbance during binding of the first metal ion (compare log K�1 absand log K�1 flin Table 7).(a) Fig. 3 Fluorescence spectra (for lex see Table 3) of tetracycline (3 × 10–5 M) in aqueous Tris buffer with increasing concentrations of (a) Mg2+ at pH 7.0 and (b) Ca2+ at pH 8.5; cT(M2+) steps between the two spectra are 0.2 log units. Selected concentrations are given in the insets 3.2 Titrations of doxycycline with Ca2+ and Mg2+ Absorption data (see Fig.4) for all combinations of doxycycline, M2+ and pH under investigation reveal two consecutive equilibria of complexation, as was found for tetracycline and oxytetracycline. Here, as well, any of the equilibria are clearly characterized by several isosbestic points (see Table 5). The absorption spectra of doxycycline complexes under the conditions applied for investigation (Ca2+-containing solutions were unstable at pH 8.5 and not further examined) showed a bathochromic shift of the long-wavelength absorption band for binding the first metal ion [see Fig. 4(a)]. This agrees exactly with the behaviour of the (a) Fig. 4 Complexation of doxycycline (3 × 10–5 M)/Mg2+/pH 8.5 in aqueous Tris buffer with increasing concentrations of Mg2+; (a) absorption spectra and (b) fluorescence spectra.The cT(M2+) steps between the two spectra are 0.2 log units. Selected concentrations are given in the insets. (b) tetracyclines discussed above. As opposed to the behaviour of those, binding the second metal ion does not generate the sharp 270 nm band observed there. A hypsochromic shift of the long-wavelength absorption in the second step is clearely noticeable and this was not observed for tetracycline or oxytetracycline. The absorption spectra thus indicate an involvement of the BCD chromophore in binding either of the two metal ions. For doxycycline, binding both metal ions affects the fluorescence intensity and the wavelength of the emission maximum [see Fig.4(b)]. This also indicates binding of both metal ions in a way that the BCD chromophore is altered. (b)Table 5 Equilibria and corresponding isosbestic points for the titrations of doxycycline with Mg2+ and Ca2+ Doxycycline Equilibrium pH M2+ Mg2+ 7.0 B C 8.5 Ca2+ 7.0 A B A B A B B C B C 3.3 Titration of anhydrotetracycline with Ca2+ and Mg2+ The spectrometric titrations for anhydrotetracycline with Ca2+ and Mg2+ (see Fig. 5) also yield one set of isosbestic points in the low M2+ concentration domain and another for high concentrations (see Table 6). This means that anhydrotetracycline is also successively complexed by two of the metal ions. Changes in absorbance are very small compared to those of the "normal" tetracyclines.Fig. 5 shows, exemplary for all of the four combinations of M2+ and pH, the UV-Vis absorption spectra recorded during the titration with Mg2+ at pH 7.0. In contrast to tetracycline and oxytetracycline, the complexation of anhydrotetracycline with Ca2+ at pH 8.5 does not show an exceptional behaviour. For any of the combinations investigated, M2+ complexation leads to a loss in intensity and concomitant Table 6 Equilibria and isosbestic points for the titration of anhydrotetracycline with Mg2+ and Ca2+ Anhydrotetracycline M2+ Equilibrium pH 7.0 Mg2+ A B B C 8.5 A B C Ca2+ 7.0 B C 8.5 B A B A B B C liso/nm 226, 270, 350 252, 262, 281, 396 227, 260, 282, 294, 354 279, 311, 390 226, 271, 349 283 liso/nm 276, 317, 441 244, 269, 290, 473 235, 247, 276, 315, 442 241, 259, 290, 310, 334 325 No clear point 285, 447 290, 465 bathochromic shift of the 270 nm band, diminishing of the 335 nm band and a shift of the redmost band towards longer wavelengths. Binding of the second metal ion causes a gain in the intensity of the latter and the 270 nm band.The fluorescence emission (integrated intensity) of the Mg2+-containing solutions rises barely during the first complexation. For Mg2+ concentrations greater than 10–2 M the intensity rises however substantially [see Fig. 6(a)] The sets of spectra recorded with Ca2+ ions in solution do not show the isoemissive point that one finds for Mg2+ complexation of ATc.The emission maximum shifts only bathochromically [see Fig. 6(b)]. 4 Discussion 4.1 Determination of conditional binding constants Conditional binding constants [eqn. (1) and (2)] have been evaluated from both absorption and fluorescence data, whenever there was an evolution of spectra as a consequence of the first and/or second complexation step. K (1) 1 2+ = ¢ K2 = ¢ + [ ] {tcM} [ ] tc[ ] M (2) 2+ + {tcM} 2+ [ ] {tcM } [ ]2[M ] A first graphical determination at selected wavelengths was followed by numerical analysis with the program Specfit as described in the section Numerical analysis. All reported constants refer to the consecutive formation of a 1 : 1 and a 2 : 1 M2+–tetracycline complex under the conditions given in Table 7.Error estimations given there are the statistical standard errors s( b�) in the case of the globally fitted constants and the maximal deviation of the arithmetic mean value in the case of graphical determination at several wavelengths. In the latter case, the number of determinations was restricted to the number of isosbestic points in the preceeding or following equilibrium. Since pH was not varied in these experiments, they do not allow the determination of the protonation state of the complexes. Therefore complexational constants are given as conditional constants and protons are disregarded in their definition. Some plausible arguments about protonation will be presented in the conclusion.For a tetracycline complex of given stoichiometry, different sites of complexation are realised depending on the combination of Tc, M2+ and pH.Similarities in the absorption and emission spectra allow one to define groups of magnesium-type and calcium-type complexes for the tetracyclines, as shown below.(a) (b) (c) (d) (e) Fig. 5 Absorption spectra of anhydrotetracycline (2 × 10–5 M) in aqueous Tris buffer with increasing concentrations of Mg2+ at pH 7.0 (a), (b), (c), (d) and (e) show details of the spectra; cT(M2+) steps between the two spectra are 0.2 log units. Selected concentrations are given in the insets.(a) Fig. 6 Fluorescence spectra of anhydrotetracycline (2 × 10–5 M) in aqueous Tris buffer with increasing concentrations of Mg2+ at pH 7.0 (a) and Ca2+ at pH 7.0 (b); cT(M2+) steps between the two spectra are 0.2 log units.Selected concentrations are given in the insets. 4.2 Complexation of tetracycline and oxytetracycline with Mg2+ and Ca2+ If one compares the evolution of the absorption and fluorescence spectra within the titrations of tetracycline or oxytetracycline with Mg2+ and Ca2+, one must conclude that binding of the metal ions occurs in the same way in Table 7 Conditional binding constants and types of complexes evaluated from the spectroscopic titrations. Numbers in bold print wtained from a global fit with Specfit, the others from graphical evaluation Derivative Tc Oxy Doxy ATc pH M2+ Mg2+ 7.0 8.5 7.0 Ca2+ 8.5 7.0 Mg2+ 8.5 7.0 Ca2+ 8.5 7.0 Mg2+ 8.5 7.0 Ca2+ 7.0 Mg2+ 8.5 7.0 Ca2+ 8.5 (b) both derivatives under equivalent conditions. In the following, only the complexation of tetracycline will be discussed in detail, but the same conclusions apply to oxytetracycline.log[K�1 fl · M] log[K�1 abs· M] 2.98 3.95 2.94 ± 0.02 2.88 ± 0.01 3.91 ± 0.03 4.06 ± 0.01 2.48 4.16 2.56 ± 0.05 2.61 ± 0.02 4.41 ± 0.03 4.46 ± 0.04 3.09 4.19 3.03 ± 0.02 3.07 ± 0.01 4.26 ± 0.02 4.13 ± 0.03 2.54 3.90 2.47 ± 0.01 2.61 ± 0.02 3.96 ± 0.10 3.71 ± 0.06 3.07 4.23 3.07 ± 0.03 2.98 ± 0.01 4.33 ± 0.02 4.47 ± 0.02 2.57 2.51 ± 0.01 2.40 ± 0.03 —— 3.34 ± 0.04 3.16 ± 0.03 4.00 ± 0.01 3.99 ± 0.02 —— 3.19 ± 0.07 2.92 ± 0.05 3.12 ± 0.04 3.30 ± 0.05 log[K�2 abs · M] log[K�2 fl· M] —— 1.44 ± 0.01 1.47 ± 0.04 1.62 ± 0.07 1.79 ± 0.03 —2.10 0.39 ± 0.19 0.87 ± 0.12 2.20 ± 0.05 2.11 ± 0.14 —— 1.51 ± 0.02 1.56 ± 0.04 1.45 ± 0.04 1.55 ± 0.10 —2.27 1.56 ± 0.43 1.00 ± 0.11 2.31 ± 0.07 2.15 ± 0.17 1.34 1.44 1.28 ± 0.02 1.13 ± 0.06 1.22 ± 0.01 1.01 ± 0.04 — 1.10 0.93 ± 0.15 1.06 1.27 0.99 ± 0.07 0.79 ± 0.09 1.09 ± 0.01 0.97 ± 0.05 —— 1.37 ± 0.29 1.17 ± 0.03 1.38 ± 0.07 0.77 ± 0.12 Complex Mg-type Mg-type Mg-type Ca-type Mg-type Mg-type Mg-type Ca-type 1:1 Mg-type 1:1 Mg-type Not determ.Mg-type Mg-type Not determ. Not determ.Under the assumption of successive formation of 1 : 1 and 2 : 1 complexes (M2+–Tc) from tetracycline and M2+ and with the help of the program Specfit, we calculated the conditional binding constants (see Table 7), the spectra of the corresponding complexes and their fraction at given total M2+ concentration cT(M2+).From Fig. 7(a) it is evident that Mg2+ and Ca2+ binding at pH 7.0 leads to pairs of nearly identical spectra for the 1 : 1 and 2 : 1 complexes. The calculated spectra for the two types of Mg2+ complexes [Fig. 8(a)] from the titration data at pH 8.5 are practically identical with those derived from the data at pH 7.0. For the Ca2+ complexes, the titration data at pH 8.5 yield completely different spectra. The similarity in the spectra for the 1 : 1 and the 2 : 1 complexes is however conserved.From the calculated spectra one can conclude without doubt that Mg2+ binding at pH 7.0 and 8.5, as well as Ca2+ binding at pH 7.0, must follow the same pattern of complexation whereas Ca2+ at pH 8.5 follows a different pattern. The fact that Mg2+ and Ca2+ bind tetracycline in the same manner was already stated in the past, although no attention was paid to the influence of the pH.23,27 The binding sites within these complexes can be determined as follows: from the above said it is justified to compare the conditional binding constants determined for the Mg2+ binding at each pH because binding obviously occurs at the same functional groups of tetracycline. The difference in the pK�1 abs values determined from the absorption titrations at pH 7.0 and 8.5 is DpK�1 abs = 0.97 ± 0.05. Given that M2+ can displace a proton when binding tetracycline, then the first metal ion binds to an acidic group, which shows a significant difference in the degree of protonation within this pH interval.The appropriate group is the b-hydroxyketo system on the BCD chromophore. This conclusion is underlined by the influence of this first metal ion binding on fluorescence emission. Binding of the first Mg2+ ion displaces the proton at C11–O{H}O–C12 and eliminates the possibility of excited-state intramolecular proton transfer (ESIPT). The latter is made responsible for the large Stokes shift of the tetracycline fluorescence.56 On binding the first Mg2+ ion to form the 1 : 1 complex at pH 7.0, the emission maximum is shifted hypsochromically from lemmax 7.0 = 585 nm to lemmax 7.0 = 535 nm.The latter value is also found for the fluorescence maximum of the free tetracycline at pH 8.5. The constants K'1 abs and K'1 fl determined from absorption and fluorescence measurements, respectively, are practically identical (see Table 7). Further evidence for the binding site within the BCD chromophore will be given after the Ca2+ binding at pH 8.5 has been discussed. Binding of the second metal ion gives rise to the sharp 270 nm absorption band in the three equivalent cases and practically does not influence the absorption assigned to the BCD chromophore. With a DpK'2 abs = 0.18 ± 0.08, binding of the second metal ion is effectively independent of pH.The binding site thus must be either equally protonated (C4–NMe2H+) or deprotonated (C3–O–) at both pH. The fluorescence intensity and emission maximum of the complexes are not altered by this reaction. The second M2+ therefore binds to the A chromophore. The observation of the subtle pH dependence of this binding indicates binding via C4– Me2N{M}O–C3 and disfavours the alternative binding site C3–O{M}NH2–Camid. The spectra deduced from the titration of tetracycline with Ca2+ at pH 8.5 differ significantly from those discussed so far. The absorption spectra show a bathochromic shift of the long wavelength absorption maximum on binding of the first Ca2+ and an increase in e( l) when binding the second.Thus, both reactions influence the absorption due to the BCD chromophore. Also, the broad 296 nm absorption band appearing in both complexes can only be interpreted as being due to BCD chromophore. Distinct from the situation in the Mg-type complexes, here both complexational reactions have an influence on the fluorescence emission providing further evidence for binding both metal ions to the BCD chromophore. 1 abs = It is important to note that at pH 8.5 already the first Ca2+ ion is complexed to a different site on the BCD chromophore than in the Mg-type complexes. This is confirmed by the following observations: appearance of the 296 nm absorption band, the bathochromic shift of the emission maximum (no shift for the Mg-type complexes at pH 8.5) and the absence of evidence for A ring binding (no sharp 270 nm band) even for very high Ca2+ excess.The first Ca2+ ion bound seems to fix the molecule in a conformation that is not suited for further A ring binding. Since it is clear for reasons of repulsive coulomb interactions that both Ca2+ ions cannot bind to the same functional group, we suggest that the first Ca2+ (pK� 4.41 ± 0.03) binds through C12–O{Ca}O–C1 under partcipation of C12a–OH. Then only C10–O(H){Ca}O– C11 remains as a possible site for the binding of a second Ca2+ [see Fig. 9(b)]. The phenolic C10–O(H) group is indeed a less potent donor site and thus accounts for the relatively low value of the binding constant (pK'2 abs = 2.20 ± 0.05).As a consequence of the above reasoning the only possible binding site of the first metal ion in the Mg-type complexes must be C11–O{M}O–C12 which coincides perfectly with the position of the Mg2+ in the ternary complex with tetracycline and the Tet repressor TetR [see Fig. 9(a)].57 From Fig. 7(a) it should be evident that Mg2+ and Ca2+ complexation at pH 7.0 leads to pairs of nearly identical spectra. Only the assumption of formation of equivalent complexes with both metal ions can account for this similarity. Fig. 7(b) visualizes the differences in the binding constants for Mg2+ and Ca2+ at that pH. When binding to the quite rigid C11–O/C12–O moiety, the different radii of the two ions , 66 pm (Mg2+) and 99 pm (Ca2+), result in different binding constants.Binding to the more flexible A chromophore in the saccomplished with about the same constant for both ions. It was claimed that the concentration of tetracyclines reached in body fluids during treatment with the antibiotics is in the order of 10–3 M (ref. 10). Therefore the partition curves displayed in Fig. 7(b) have medical relevance. It is seen that in blood plasma at pH ß7.4 and Mg2+ and Ca2+ concentrations of 9 × 10–4 M and 1.3 × 10–3 M, respectively, there will be significant formation of the 1 : 1 complexes with both metal ions. If one compares the distribution of species displayed in Fig. 7(b) and 8(b) one can conclude that complex formation occurs via displacement of protons since complex formation yet occurs at lower p(M) levels at higher pH in the case of Mg2+.(b) (a) Fig.7 Absorption spectra (a) and fraction F (%) (b) of the individual species for the Mg-type complexes formed between Mg2+ and Ca2+, respectively, and tetracycline (3 × 10–5 M) at pH 7.0 in aqueous Tris buffer. (a) (b) Fig. 8 Absorption spectra (a) and fraction F (%) of the individual species (b) for the Mg-type and Ca-type complex formed between Mg2+, respectively, Ca2+ and tetracycline (3 × 10–5 M) at pH 8.5 in aqueous Tris buffer. (a) (b) Fig. 9 Binding sites in (a) Mg-type (Mg2+, Ca2+) and (b) Ca-type (Ca2+ pH 8.5) complexes of tetracycline and oxytetracycline as derived from absorption and fluorescence measurements (only one mesomeric structure is given).4.3 Complexation of doxycycline with Mg2+ and Ca2+ The complexation of doxycycline with Mg2+ proceeds also in two steps.Both binding reactions affect the fluorescence as well as the absorption of the molecules, whereby the constants evaluated with each of the two analysis techniques agree very well (see Table 7). With the arguments developed in the discussion of tetracycline data, it follows that functional groups of the BCD chromophore are involved in each of the two reactions. The pH dependence of the first conditional binding constant is in favour of binding of the first Mg2+ at the b- hydroxyketo system. This is suggested by the bathochromic shift of the long-wavelength absorption and other changes in the absorption spectra, that correspond to those found for the formation of the 1 : 1 Mg-type complexes (compare Fig.7, 8 and 9). Taking into account that the characteristic 296 nm band of the Ca-type complexes is missing and that A ring binding can be excluded due to the absence of the 270 nm band, C11–O(H)/O–C12 remains the only possible binding site for the first Mg2+ ion. Therefore, doxycycline also forms the 1 : 1 Mg-type complex. The influence of pH on the first complexation is easily recognized in Fig. 10(b) and the similarities in the shape of the absorption spectra of the complex at both pH's become obvious in Fig. 10(a). Binding of the second Mg2+ involves the BCD chromophore, shows practically no pH dependence and does not generate the absorption spectrum typical of the 2 : 1 Mg-type complexes.On the contrary, the longwavelength absorption is noticeably shifted to a shorter wavelength [see Fig. 10(a)]. Mitscher et al. propose binding of Mg2+ through C5–OH and C12a–OH for 5- hydroxytetracyclines. It is claimed that doxycycline can adopt the suited conformation more easily than tetracyclines bearing a C6–OH group because the repulsive interaction of this group and C4–H is missing.23 Even if this chelation remains debatable due to the bad donor properties of alcoholic groups, it would undoubtedly explain all the spectral characteristics as well as the missing pH dependence of the binding constant. (a) Fig. 10 Absorption spectra (a) and fraction F (%) (b) of the individual species for the Mg2+ complexes of doxycycline (3 × 10–5 M) at pH 7.0 and 8.5, respectively, in aqueous Tris buffer.Ca2+ binding to doxycycline could only be investigated at pH 7.0 because solutions at higher pH decomposed when placed in the measuring light beam. For low concentrations cT(Ca2+) at pH 7.0 the absorption spectra are similar to those of the Mg2+ complexes. The same is true for the shift of the emission maximum, the Stokes shift and the form of the fluorescence band. Therefore we suggest the formation of a 1 : 1 Mg-type complex. For the binding of the second Ca2+ ion the formation of a 2 : 1 Mg-type complex can be excluded because the 270 nm band is absent, but the fluorescence intensity does not decrease as in the case of Mg2+ binding.Therefore, the binding site of the second Ca2+ cannot be deduced. 4.4 Complexation of anhydrotetracycline with Mg2+ and Ca2+ As for the normal tetracyclines, we find a two-step complexation at both pH for anhydrotetracycline. For the complexation of the first Mg2+ ion, the bathochromic shift of the long-wavelength absorption band (see Fig. 11) as well as the pH dependence of the binding constant (see Table 7) strongly suggest binding to the BCD unit. A prerequisite for the validity of the last argument is again the fact that binding takes place with the same complexation pattern, as suggested by the similar characteristic changes in the absorption and fluorescence spectra at both pH's. With DpK�1 abs = 0.83 ± 0.05 the difference in the conditional binding constants is not only below the expected value of 1.5 units (assuming the same site of complexation) but even smaller than for tetracycline with DpK�1 abs = 1.18 ± 0.02.This can be explained as a consequence of the lower pKA2 and pKA3 of anhydrotetracycline as compared to tetracycline and the resulting different extending of the protonation of the C4– Me2N and the b-hydroxyketo system. The absolute values of pK�1 abs (see Table 7) are in accordance with binding the Mg2+ ion to the C11–O{Mg}O–C12 site, which produces the 1 : 1 Mg-type complex as suggested in the older literature.5,39 The ground state binding constants determined from absorption measurements deviate, however, from those derived from fluorescence spectra.Obviously, (b)reactions occur in the excited state with the effect that the number of fluorescing molecules is not any more directly proportional to the number of light absorbing molecules.58 Complexation of the second Mg2+ has a considerably lower binding constant than the first one. It results in a growth of the long-wavelength absorption band [see Fig. 11(a)] and of fluorescence intensity with approximately the same dependence on cT(Mg2+). The absolute value of the binding constant and its independence of pH as well as the increase of the 270 nm band suggest binding to the A chromophore, as was already postulated by other authors.5,39 Ca2+ binding leads to comparable changes in the absorption spectra as seen for Mg2+ binding except that the 335 nm band remains unchanged, whereas it was altered on Mg2+ binding.Also, fluorescence spectra change differently in that the emission maximum shifts continuously bathochromically with rising cT(Ca2+). Furthermore, the values of pK’1 abs and pK’2 abs are not in the same range as for Mg2+ binding and thus prevent to transfer the ideas developed there to Ca2+ binding. An unequivocal analysis of the Ca2+ binding sites is is not possible on the basis of our data. (a) (c) Fig. 11 Spectra and fraction F (%) of the individual species for the Mg2+ complexes of anhydrotetracycline (2 × 10–5 M) at pH 7.0 and 8.5, respectively, in aqueous Tris buffer; (a) ATc and complexes at pH 7.0 (for legend see (b)), (b) detail with ATc and the complexes at both pH and (c) fraction F (%) at both pH.5 Summary and conclusion The existence of two complexational equilibria was proven by graphical and numerical analysis (program Specfit) on the basis of UV-Vis absorption spectra recorded during titrations of the tetracyclines with Mg2+ and Ca2+ ions. When metal binding occurs on the BCD chromophore, fluorescence titration could also be used for the determination of binding constants. Our results provide evidence for 1 : 1 and 2 : 1 (M2+–tetracycline) complexes and thereby oppose all models mentioned in the older For the sake of completeness and because anhydrotetracycline is othe most important and toxic degradation products of tetracycline, the distribution of species is shown in Fig.11(b). At the Mg2+ levels present in body fluids 1 : 1 complexes of tetracycline and anhydrotetracycline form with about the same constants and so their partition will mirror their relative amounts present. Formation of 1 : 1 Ca2+ complexes at physiological pH will be favoured compared to that of the corresponding complexes tetracycline forms with Ca2+. (b)literature that postulate only one step complexation (see Table 2). This discrepancy is namely due to the fact that only at higher M2+ excess the contributions of the 2 : 1 complexes to the spectral changes become significant. However, considering the spectral changes related to the binding of the second metal ion also facilitate the allocation of binding sites for the first ion.In our opinion the presented results contradict the existence of 1 : 2 M2+–Tc complexes that were often proposed for the low metal ion concentration domain. Additional experiments showed that no alteration of the Tc spectra occurred for M2+ concentrations below the reported titration range. Analyzing the titration data with equilibrium models that employ such complexes led to worse fits as compared to these with the models discussed above. This conclusion is at variance with models applied by Berthon et al. to fit their potentiometric titration data.10–12,30,31,43,59 It should be emphasized here that our data do not allow the derivation of the protonation state of the complexes as is possible for potentiometric investigations.Nevertheless, plausible arguments can be found to guess the extent of protonation of the complexes proposed in this work. At pH 7.0 C3–OH will be deprotonated in all the tetracyclines investigated. As outlined above the first M2+ always complexes to the BCD unit thereby replacing its acidic proton. Therefore all 1 : 1 Tc–M2+ complexes most probably have the [Tc–MH]+ composition. The third proton will to some extent dissociate at pH 8.5, thus leading to a mixture of [TcM]0 and [TcMH]+ in the case of the Mg-type complexes. The slight pH dependence of complexation favours the idea that the C4–NMe2H+ proton is replaced in the second binding reaction for the Mg-type complexes. This leads to complexes with [TcM2]2+ composition. At Mg2+ and Ca2+ concentrations present in blood plasma only 1 : 1 Tc–M2+ complexes are likely to be formed, a low percentage of which may be present as [TcM]0 in equilibrium with [TcMH]+.The neutral complexes should be able to diffuse through membranes and enter body tissues. Despite these uncertainties with respect to the protonation state, our data allows one to conclude that tetracycline and oxytetracycline bind a first Mg2+ ion via C11–O{Mg}O– C12 and a second one via C4–Me2N{Mg}O–C3. Even Ca2+ at pH 7.0 binds to the same donor groups as Mg2+. At pH 8.5 binding the first Ca2+ ion is via C12–O{Ca}O–C1 and that of the second via C10–O(H){Ca}O–C11, which results in different absorption and fluorescence spectra. Doxycycline and anhydrotetracycline also bind the first Mg2+ via C11–O{Mg}O–C12 but the binding site of the second ion could not be determined for these tetracyclines. For anhydrotetracycline K1' could not be determined from emission spectra although binding the M2+ ion alters the spectra.The reason for this failure could be that anhydrotetracycline undergoes photoinduced conformational changes. Acknowledgement We wish to thank Meryl Dean for proofreading the manuscript. Financial support by Deutsche Forschungsgemeinschaft and Fonds der Chemischen Industrie is also gratefully acknowledged. References 1 A. Albert and C. W. Rees, Nature, 1956, 177, 433. 2 N. Pligler, S. Bag, D. Leyden, J. Sudmeier and C. Reilley, Anal. Chem., 1965, 37,872. 3 L Leeson, J. Krueger and R.Nash, Tetrahedron Lett., 1963, 18, 1155. 4 C. Stephens, K. Murai, K. Brunings and R. Woodward, J. Am. Chem. Soc., 1956, 78, 4155. 5 F. Machado, G. Demicheli, A. Garnier-Suillerot and H. Beraldo, J. Inorg. Biochem., 1995, 60, 163. 6 B. Martin, in Metal ions in biological systems, ed. B. R. Martin, Marcel Dekker, New York, 1985, vol. 19. 7 V. Bhatt and R. Jee, Anal. Chim. Acta, 1985, 167, 233. 8 J. Siqueira, S. Carvalho, E. Paniago, L. Tosi and H. Beraldo, J. Pharm. Sci., 1994, 83, 291. 9 D. J. Silva and M. Dias, Rev. Port. Quim., 1972, 14, 159. 10 M. Brion, G. Berthon and J. Fourtillan, Inorg. Chim. Acta, 1981, 55, 47. 11 L. Lambs, M. Brion and G. Berthon, Inorg. Chim. Acta, 1985, 106, 151. 12 L. Lambs and G.Berthon, Inorg. Chim. Acta, 1988, 151, 33. 13 M. Ghandour, H. Azab and A. Ali, Monatsh. Chem., 1992, 123, 51, 14 M. Ghandour, H. Azab, A. Hassan and A. Ali, Polyhedron, 1989, 8, 189. 15 H. Duarte, S. Carvalho, E. Paniago and A. Simas, J. Pharm. Sci., 1999, 88, 111. 16 R. Connamacher and G. Mandel, Biochem. Biophys. Acta, 1968, 166, 475. 17 J. Rausto da Silva and H. Mendonca Dias, Rev. Port. Quim., 1973, 15, 1. 18 W. Hillen and C. Berens, Ann. Rev. Microbiol., 1994, 48, 345. 19 D. Schnappinger and W. Hillen, Arch. Mikrobiol., 1996, 165, 359. 20 B. Schneider, J. Maimon, L. Golub, N. Ramamurthy and R. Greenwald, Biochem. Biophys. Res. Commun., 21 N. Neamati, H. H. S. Sunder, G. Milne and Y. 22 D. Williamson and G. Everett, J. Am.Chem. Soc. 1975, 23 L. Mitscher, B. Slater-Eng and T. Sokoloski, 24 T. Ohyama and J. Cowan, Inorg. Chem., 1995, 34, 25 E. Newman and C. Rank, J. Pharm. Sci., 1976, 65, 26 S. Day, W. Crouthamel, L. Martinelli and J. Ma, J. 27 A. Caswell and J. Hutchison, Biochem. Biophys. Res. 28 M. Takahashi, J. Degenkolb and W. Hillen, Anal. 29 R. Sharma and S. Joseph, Indian J. Chem., 1996, 35, 30 L. Lambs, B. Decock Le Reverend, H. Kozlowski and 31 G. Berthon, M. Brion and L. Lambs, J. Inorg. Biochem., 32 J. Wessels, W. Ford, W. Szymczak and S. Schneider, J. 33 M. Riaz and N. Pilpel, J. Pharm. Pharmacol., 1984, 36, 34 K. Ibsen and M. Urist, Proc. Soc. Exp. Biol. Med., 35 J. McCormick, E. Jensen, R. Miller and A. Doerschuk, 1992, 188,. Pommier, Mol.Pharmacol., 1997, 52. 97, 2397. Antimicrob. Agents Chemother., 1972, 2, 66. 1083. 1728. Pharm. Sci., 1978, 67, 1518. Commun., 1971, 43, 625. Biochem., 1991, 199, 197. 639. G. Berthon, Inorg. Chem., 1988, 27, 3001. 1983, 19, 1. Phys. Chem., 1998, 102, 9323. 153. 1962, 109, 797. J. Am. Chem. Soc., 1960, 82, 3381.36 L. Mitscher, A. Bonacci and T. Sokoloski, in Antimicrobial Agents and Chemotherapy, ed. G. L. Hobby, American Society for Microbiology, Bethesda, MD, 1968, p. 78. 37 M. Mathew and P. J. Balaxam, Inorg. Biochem., 1980, 13, 339. 38 L. Mitscher, A. Bonacci and T. Sokoloski, Tetrahedron Lett., 1968, 51, 5361. 39 L. Stoel, E. Newman, G. Asleson and C. Rank, J. Pharm. Sci., 1976, 65, 1794. 40 L. Mitscher, A. Bonacci, B. Slater-Eng, A. Hacker and T. Sokoloski,. in Antimicrobial Agents and Chemotherapy, ed. G. L. Hobby, American Society for Microbiology, Bethesda 1969, p. 111. 41 J.-B. Malgorzata and L. Lambs, Inorg. Chem., 1993, 32, 428. 42 M. Jezowska Boiczuk, L. Lambs, E. Kozlowski and G. Berthon, Inorg. Chem., 1993, 32, 428. 43 L. Lambs, M. Venturini, B. Pecock Le Reverend, H. Kozlowski and G. Berthon, J. Inorg. Biochem., 1988, 44 S. Matos and H. Beraldo, J. Braz. Chem. Soc., 1995, 6, 33, 193. 405. 45 V. Grosheva and V. Zolin, Russ. J. Coord. Chem., 1994, 20, 395. 46 E. Newman, PhD Thesis, University of Iowa, 1975. 47 G. Coibion and P. Laszlo, Biochem. Pharmacol., 1978, 28, 1367. 48 R. Curtis and R. Wasylishen, Can. J. Chem., 1991, 69, 834. 49 M. Celotti and G. Fazakerly, J. Chem. Soc., Perkin Trans. 2, 1977, 1319. 50 H. Lanig, M. Gottschalk, S. Schneider and T. Clark, J. Mol. Mod., 1999, 5, 46. 51 M. Gottschalk, H. Lanig and T. Clark, J. Mol. Mod., accepted for publication. 52 H. Grampp, M. Maeder, C. Meyer and A. Zuberbühler, Talanta, 1985, 32, 95. 53 H. Grampp, M. Maeder, C. Meyer and A. Zuberbühler, Talanta, 1985, 32, 257. 54 H. Grampp, M. Maeder, C. Meyer and A. Zuberbühler, Talanta, 1985, 32, 1133. 55 H. Grampp, M. Maeder, C. Meyer and A. Zuberbühler, Talanta, 1986, 33, 943. 56 H. Morrison and G. Olack, J. Am. Chem. Soc., 1991, 113, 8110. 57 W. Hinrichs, C. Kisker, M. Duvel, A. Müller, K. Tovar, W. Hillen and W. Saenger, Science, 1994, 264, 418. 58 A. Kowalczyk, N. Boens and M. Ameloot, in Fluorescence Spectroscopy, ed. L. Brand and M. Johnson, Methods in Enzymology, Academic Press, New York, 1997. 59 L. Lambs, M. Brion and G. Berthon, Agents Acttons, 1984, 743. PhysChemComm © The Royal Society of Chemistry 2000

ISSN:1460-2733    

DOI:10.1039/b005722n

出版商:RSC    

年代:2000

数据来源: RSC