P–O Bond Destabilization Accelerates Phosphoenzyme Hydrolysis of Sarcoplasmic Reticulum Ca2+-ATPase
2004; Elsevier BV; Volume: 279; Issue: 50 Linguagem: Inglês
10.1074/jbc.m410867200
ISSN1083-351X
AutoresAndreas Barth, Natalya Bezlyepkina,
Tópico(s)Mitochondrial Function and Pathology
ResumoThe phosphate group of the ADP-insensitive phosphoenzyme (E2-P) of sarcoplasmic reticulum Ca2+-ATPase (SERCA1a) was studied with infrared spectroscopy to understand the high hydrolysis rate of E2-P. By monitoring an autocatalyzed isotope exchange reaction, three stretching vibrations of the transiently bound phosphate group were selectively observed against a background of 50,000 protein vibrations. They were found at 1194, 1137, and 1115 cm–1. This information was evaluated using the bond valence model and empirical correlations. Compared with the model compound acetyl phosphate, structure and charge distribution of the E2-P aspartyl phosphate resemble somewhat the transition state in a dissociative phosphate transfer reaction; the aspartyl phosphate of E2-P has 0.02 Å shorter terminal P–O bonds and a 0.09 Å longer bridging P–O bond that is ∼20% weaker, the angle between the terminal P–O bonds is wider, and –0.2 formal charges are shifted from the phosphate group to the aspartyl moiety. The weaker bridging P–O bond of E2-P accounts for a 1011–1015-fold hydrolysis rate enhancement, implying that P–O bond destabilization facilitates phosphoenzyme hydrolysis. P–O bond destabilization is caused by a shift of noncovalent interactions from the phosphate oxygens to the aspartyl oxygens. We suggest that the relative positioning of Mg2+ and Lys684 between phosphate and aspartyl oxygens controls the hydrolysis rate of the ATPase phosphoenzymes and related phosphoproteins. The phosphate group of the ADP-insensitive phosphoenzyme (E2-P) of sarcoplasmic reticulum Ca2+-ATPase (SERCA1a) was studied with infrared spectroscopy to understand the high hydrolysis rate of E2-P. By monitoring an autocatalyzed isotope exchange reaction, three stretching vibrations of the transiently bound phosphate group were selectively observed against a background of 50,000 protein vibrations. They were found at 1194, 1137, and 1115 cm–1. This information was evaluated using the bond valence model and empirical correlations. Compared with the model compound acetyl phosphate, structure and charge distribution of the E2-P aspartyl phosphate resemble somewhat the transition state in a dissociative phosphate transfer reaction; the aspartyl phosphate of E2-P has 0.02 Å shorter terminal P–O bonds and a 0.09 Å longer bridging P–O bond that is ∼20% weaker, the angle between the terminal P–O bonds is wider, and –0.2 formal charges are shifted from the phosphate group to the aspartyl moiety. The weaker bridging P–O bond of E2-P accounts for a 1011–1015-fold hydrolysis rate enhancement, implying that P–O bond destabilization facilitates phosphoenzyme hydrolysis. P–O bond destabilization is caused by a shift of noncovalent interactions from the phosphate oxygens to the aspartyl oxygens. We suggest that the relative positioning of Mg2+ and Lys684 between phosphate and aspartyl oxygens controls the hydrolysis rate of the ATPase phosphoenzymes and related phosphoproteins. Phosphorylation is one of the fundamental regulatory mechanisms in biology (1Mildvan A.S. Proteins. 1997; 29: 401-416Crossref PubMed Scopus (255) Google Scholar). It also governs the catalytic mechanism of the sarcoplasmic reticulum Ca2+-ATPase (SERCA1) (2Hasselbach W. Makinose M. Biochem. Z. 1961; 333: 518-528PubMed Google Scholar), where it controls the ordered sequence of catalytic steps to ensure the efficiency of the overall catalytic reaction. Here we study one of the ATPase phosphoenzyme intermediates by monitoring an enzyme-catalyzed isotopic exchange reaction at the phosphate group with infrared spectroscopy. The result is an infrared spectrum at "atomic resolution" in a crowded spectral region. It observes selectively the transiently bound phosphate group, a group that cannot be studied by site-directed mutagenesis. Our approach can be extended to other phosphoproteins and reveals here the molecular cause of an important property of the phosphoenzyme studied. The Ca2+-ATPase (2Hasselbach W. Makinose M. Biochem. Z. 1961; 333: 518-528PubMed Google Scholar) pumps two Ca2+ ions against a concentration gradient across the sarcoplasmic reticulum membrane, which relaxes a flexed muscle (for reviews see Refs. 3Andersen J.P. Biochim. Biophys. Acta. 1989; 988: 47-72Crossref PubMed Scopus (99) Google Scholar, 4Hasselbach W. Bonting S.L. De Pont J.J.H.H.M. Membrane Transport. Elsevier, Amsterdam1981: 183-208Google Scholar, 5Inesi G. de Meis L. Martonosi A. 2nd Ed. The Enzymes of Biological Membranes. 3. Plenum Press, New York1985: 157-191Google Scholar, 6Martonosi A. Kracke G. Taylor K.A. Dux L. Peracchia C. Soc. Gen. Phys. Ser. 1985; 39: 57-85PubMed Google Scholar, 7Mintz E. Guillain F. Biochim. Biophys. Acta. 1997; 1318: 52-70Crossref PubMed Scopus (90) Google Scholar, 8Lee A. East J. Biochem. J. 2001; 356: 665-683Crossref PubMed Scopus (110) Google Scholar). The energy for this active transport process is provided by the substrate ATP, which phosphorylates the ATPase at Asp351 to form at least two consecutive phosphoenzymes, Ca2E1-P and E2-P. The Ca2+ transport step is associated with the conversion from Ca2E1-P to E2-P. These phosphoenzymes have different catalytic properties; while the first phosphoenzyme intermediate Ca2E1-P dephosphorylates with ADP to reform ATP, the second phosphoenzyme E2-P reacts with water. This switch of catalytic specificity ensures the efficiency of the pump process, because the specificity of Ca2E1-P is such that the energy provided by ATP is not wasted before Ca2+ is transported. After Ca2+ transport, E2-P hydrolyzes remarkably faster than the model compound acetyl phosphate in water. The time constant for E2-P hydrolysis is between 10 and 100 ms near 25 °C depending on the pH (9McIntosh D.B. Boyer P.D. Biochemistry. 1983; 22: 2867-2875Crossref PubMed Scopus (78) Google Scholar, 10Vieyra A. Scofano H.M. Guimaraes-Motta H. Tume R.K. de Meis L. Biochim. Biophys. Acta. 1979; 568: 437-445Crossref PubMed Scopus (15) Google Scholar, 11Chaloub R.M. Guimaraes-Motta H. Verjovski-Almeida S. de Meis L. Inesi G. J. Biol. Chem. 1979; 254: 9464-9468Abstract Full Text PDF PubMed Google Scholar, 12Rauch B. von Chak D. Hasselbach W. Z. Naturforsch. 1977; 32: 828-834Crossref PubMed Scopus (27) Google Scholar), whereas that of acetyl phosphate is 105 s (25 °C, pH 6.8) (13Di Sabato G. Jencks W.P. J. Am. Chem. Soc. 1961; 83: 4400-4405Crossref Scopus (180) Google Scholar). This 106–107-fold acceleration of the reaction by the enzyme is essential for the fast progression of the pump cycle and therefore for the efficient removal of Ca2+ from the cytoplasm. Obviously, the environment of the phosphate group is important in controlling the dephosphorylation properties. It has been found to become more hydrophobic in the transition from Ca2E1-P to E2-P (14De Meis L. Martins O.B. Alves E.W. Biochemistry. 1980; 19: 4253-4261Crossref Scopus (181) Google Scholar, 15Dupont Y. Pougeois R. FEBS Lett. 1983; 156: 93-98Crossref PubMed Scopus (72) Google Scholar), and this seems to account for an increased hydrolysis rate of E2-P as compared with Ca2E1-P (Ref. 16De Meis L. Suzano V.A. FEBS Lett. 1988; 232: 73-77Crossref PubMed Scopus (23) Google Scholar; reviewed in Ref. 17De Meis L. Biochim. Biophys. Acta. 1989; 973: 333-349Crossref PubMed Scopus (140) Google Scholar). The molecular mechanism of dephosphorylation has, however, not been elucidated. We were interested in how the environment of the phosphate group controls its catalytic properties and used infrared spectroscopy to study the phosphate group of E2-P in detail. Infrared spectroscopy has proved valuable for the investigation of protein catalysis (18Zscherp C. Barth A. Biochemistry. 2001; 40: 1875-1883Crossref PubMed Scopus (113) Google Scholar, 19Wharton C.W. Nat. Prod. Rep. 2000; 17: 447-453Crossref PubMed Scopus (36) Google Scholar, 20Mäntele W. Trends Biochem. Sci. 1993; 18: 197-202Abstract Full Text PDF PubMed Scopus (204) Google Scholar, 21Siebert F. Methods Enzymol. 1995; 246: 501-526Crossref PubMed Scopus (134) Google Scholar, 22Jung C. J. Mol. Recognit. 2000; 13: 325-351Crossref PubMed Scopus (216) Google Scholar, 23Gerwert K. Biol. Chem. 1999; 380: 931-935Crossref PubMed Scopus (80) Google Scholar, 24Barth A. Zscherp C. Q. Rev. Biophys. 2002; 35: 369-430Crossref PubMed Scopus (1606) Google Scholar) because it allows measurement of interaction strengths and bond lengths at a level of sensitivity that exceeds even that of x-ray crystallography and NMR (25Deng H. Callender R. Gremlich H.U. Yan B. Infrared and Raman Spectroscopy of Biological Materials. Marcel Dekker Inc., New York2001: 477-515Google Scholar). Since the early days of infrared spectroscopy on proteins, the power of selective isotopic labeling has been exploited to observe a specific group in a large protein (26Belasco J.G. Knowles J.R. Biochemistry. 1980; 19: 472-477Crossref PubMed Scopus (127) Google Scholar, 27Alben J.O. Caughey W.S. Biochemistry. 1968; 7: 175-183Crossref PubMed Scopus (186) Google Scholar). Because of the mass effect on vibrational frequencies, infrared absorption bands of a labeled group are shifted with respect to those of the unlabeled group and can be identified in the spectrum. Specific labeling is usually necessary to identify a specific group in the spectrum, because otherwise the absorption of this group will be hidden under the overwhelming absorption of other groups. For the investigation of protein reactions, another necessity in most cases is the use of reaction-induced infrared difference spectroscopy, where the reaction of interest is started in the infrared cuvette. We release a biological compound of interest from a biologically inactive photosensitive precursor (caged compounds), for example ATP from caged ATP 1The abbreviations used are: caged ATP, P3-1-(2-nitrophenyl)ethyl ATP; vu, valence unit(s). (Ref. 28Barth A. Mäntele W. Kreutz W. FEBS Lett. 1990; 277: 147-150Crossref PubMed Scopus (63) Google Scholar; reviewed in Ref. 29Barth A. Zscherp C. FEBS Lett. 2000; 477: 151-156Crossref PubMed Scopus (49) Google Scholar). By calculating the difference between the infrared absorption before and after start of the reaction, only those groups are detected that participate actively; the absorption of the vast majority of passive groups cancels. Here we combine and extend both approaches to observe selectively the phosphate group of E2-P; [γ-18O3]ATP released from [γ-18O3]caged ATP transfers the labeled γ-phosphate to the ATPase, which then accumulates a labeled E2-P phosphoenzyme under appropriate experimental conditions. This phosphoenzyme catalyzes an isotopic exchange with water at the phosphate oxygens (9McIntosh D.B. Boyer P.D. Biochemistry. 1983; 22: 2867-2875Crossref PubMed Scopus (78) Google Scholar, 30Kanazawa T. Boyer P.D. J. Biol. Chem. 1973; 248: 3163-3172Abstract Full Text PDF PubMed Google Scholar), which can be observed with infrared spectroscopy. Isotopically labeled caged compounds were used before in experiments on the Ca2+-ATPase (31Barth A. Mäntele W. Biophys. J. 1998; 75: 538-544Abstract Full Text Full Text PDF PubMed Scopus (65) Google Scholar, 32Barth A. J. Biol. Chem. 1999; 274: 22170-22175Abstract Full Text Full Text PDF PubMed Scopus (33) Google Scholar) and on Ras (33Cepus V. Scheidig A.J. Goody R.S. Gerwert K. Biochemistry. 1998; 37: 10263-10271Crossref PubMed Scopus (90) Google Scholar, 34Du X. Frei H. Kim S.-H. J. Biol. Chem. 2000; 275: 8492-8500Abstract Full Text Full Text PDF PubMed Scopus (74) Google Scholar, 35Cheng H. Sukal S. Deng H. Leyh T.S. Callender R. Biochemistry. 2001; 40: 4035-4043Crossref PubMed Scopus (45) Google Scholar). They were evaluated by a careful comparison or subtraction of spectra obtained with labeled and unlabeled compounds. This approach has proved to be of limited sensitivity for detecting the E2-P phosphate vibrations (36Barth A. Biopolymers. 2002; 67: 237-241Crossref PubMed Scopus (18) Google Scholar). The approach here is to combine selective labeling of a protein with the induction of an autocatalyzed isotopic exchange (36Barth A. Biopolymers. 2002; 67: 237-241Crossref PubMed Scopus (18) Google Scholar). Thus we obtain directly a difference spectrum between labeled and unlabeled E2-P in a single time-resolved experiment. We observe selectively three stretching vibrations of the phosphate group of a protein with nearly 50,000 normal modes of vibration. The vibrational frequencies identified reveal a shift of electron density from the P–O bond that connects protein and phosphate to the terminal P–O bonds. This weakens the link to the protein and accelerates hydrolysis of E2-P. The approach used here to study one of the Ca2+-ATPase phosphoenzymes can be extended to other phosphoenzymes, if necessary with the help of auxiliary proteins to induce phosphorylation or dephosphorylation. Time-resolved Fourier transform infrared measurements were performed at 10 °C, and the samples were prepared as described previously (37Barth A. von Germar F. Kreutz W. Mäntele W. J. Biol. Chem. 1996; 271: 30637-30646Abstract Full Text Full Text PDF PubMed Scopus (73) Google Scholar). They contained ∼0.9 mm Ca2+-ATPase, 100 mm imidazole buffer, pH 6.5, 5 mm caged ATP, 200 μm Ca2+, 20 mm Mg2+, 5 mm dithiothreitol, 0.5 mg/ml Ca2+ ionophore A23187, 1 mg/ml adenylate kinase, 10% Me2SO. All of the spectra were normalized to an identical protein concentration using the amide II absorbance (31Barth A. Mäntele W. Biophys. J. 1998; 75: 538-544Abstract Full Text Full Text PDF PubMed Scopus (65) Google Scholar). Correlation of Vibrational Frequency with Bond Valence—Deng et al. (38Deng H. Wang J. Callender R. Ray W.J. J. Phys. Chem. B. 1998; 1022: 3617-3623Crossref Scopus (42) Google Scholar) have defined a fundamental frequency for the stretching vibration for phosphates, which for PO32− is ν=[(νs2+2νas2)/3]12, with νs and νas being the symmetric and asymmetric stretching vibrations of the terminal P–O bonds, respectively. The frequency of the asymmetric stretching vibration is weighted by a factor of 2 because this vibration is doubly degenerate. The fundamental frequency is less dependent on the geometry of the phosphate group than the asymmetric and symmetric stretching vibrations of the terminal P–O bonds themselves (38Deng H. Wang J. Callender R. Ray W.J. J. Phys. Chem. B. 1998; 1022: 3617-3623Crossref Scopus (42) Google Scholar). We slightly extend the approach by Deng et al. to phosphate groups in asymmetric environments by defining the fundamental frequency as follows: ν=[(ν12+ν22+ν32)/3]12 where ν1, ν2, and ν3 are the frequencies of the terminal P–O bonds that we have observed for the E2-P phosphoenzyme. In this equation three different frequencies appear for the three terminal P–O vibrations because the degeneracy of the asymmetric stretching vibrations is lifted by the asymmetric environment of the E2-P phosphate group. The fundamental frequency can be used to calculate the bond valence of P–O bonds using the following formula (38Deng H. Wang J. Callender R. Ray W.J. J. Phys. Chem. B. 1998; 1022: 3617-3623Crossref Scopus (42) Google Scholar),s=[0.175×ln(224500cm−1/ν˜)]−4.29(Eq. 1) where s is the bond valence in vu, and ν is the wavenumber that corresponds to the fundamental frequency. In our case, where the phosphate group is in an asymmetric environment, the bond valence s represents the average bond valence of the three terminal P–O bonds. The bond valence of the bridging P–O bond was calculated by subtracting the summed bond valences of the terminal P–O bonds from 5, the atomic valence of phosphorus. The average bond valence of external bonds (noncovalent bonds to the environment) to the terminal phosphate oxygens was calculated by subtracting the average bond valence of the terminal P–O bonds from 2, the atomic valence of oxygen. Correlation of Bond Valence with Bond Length—Bond valences can be derived from bond lengths in such a way that they sum up to the atomic valence within a few percentages (39Brown I.D. Shannon R.D. Acta Crystallogr. Sect. A. 1973; 29: 266-282Crossref Scopus (1326) Google Scholar) for a large number of oxides. We used Equation 2 (40Brown I.D. The Chemical Bond in Inorganic Chemistry: The Bond Valence Model. Oxford University Press, Oxford2002Google Scholar) with the parameters (41Brown I.D. Wu K.K. Acta Crystallogr. Sect. B Struct. Crystallogr. Cryst. Chem. 1976; 32: 1957-1959Crossref Google Scholar) used to derive the frequency versus bond valence correlation (Equation 1) (38Deng H. Wang J. Callender R. Ray W.J. J. Phys. Chem. B. 1998; 1022: 3617-3623Crossref Scopus (42) Google Scholar),s=L1N/LsN(Eq. 2) where s is the bond valence measured in vu, Ls is the bond length of a bond with bond valence s, and L1 is the bond length of a bond with a bond valence of 1 vu. L1 and N are constants for a given type of bond (39Brown I.D. Shannon R.D. Acta Crystallogr. Sect. A. 1973; 29: 266-282Crossref Scopus (1326) Google Scholar, 41Brown I.D. Wu K.K. Acta Crystallogr. Sect. B Struct. Crystallogr. Cryst. Chem. 1976; 32: 1957-1959Crossref Google Scholar): N = 4.29, and L1 = 1.622 Å for P–O bonds (41Brown I.D. Wu K.K. Acta Crystallogr. Sect. B Struct. Crystallogr. Cryst. Chem. 1976; 32: 1957-1959Crossref Google Scholar). Bond lengths determined differed by 0.01 Å (0.6%) or less from those determined with the exponential expression and more recent parameters from Ref. 40Brown I.D. The Chemical Bond in Inorganic Chemistry: The Bond Valence Model. Oxford University Press, Oxford2002Google Scholar. It should be noted that the values given in Table I for bond lengths and bond valences should not be considered to be accurate to three digits after the decimal point. Three digits are given to avoid rounding errors when the differences between acetyl phosphate and E2-P are discussed.Table IProperties of the E2-P phosphate group and acetyl phosphateE2-P phosphate groupPhosphate group of acetyl phosphateTerminal oxygenBridging P-OTerminal P-OBridging P-OWavenumbers of P-O stretching vibrations (cm-1)11941132113711321115982P-O fundamental frequency or wavenumber (cm-1)11491082P-O bond valence (vu)1.4090.7721.3430.972Bond valence of external bonding (formal charge) (vu)0.5910.657P-O bond length (Å)1.4971.7231.5141.633P-O force constant (102 nm-1)7.83.47.44.8 Open table in a new tab To observe isotope exchange at the E2-P phosphate group, E2-P was prepared by the release of ATP from caged ATP in the absence of K+ and the presence of Me2SO and Ca2+ ionophore, which are conditions that accumulate E2-P as discussed and shown in previous work (37Barth A. von Germar F. Kreutz W. Mäntele W. J. Biol. Chem. 1996; 271: 30637-30646Abstract Full Text Full Text PDF PubMed Scopus (73) Google Scholar, 42Barth A. Kreutz W. Mäntele W. Biochim. Biophys. Acta. 1994; 1194: 75-91Crossref PubMed Scopus (42) Google Scholar). This work also established a marker band for E2-P near 1194 cm–1 that was later assigned to the phosphate group (32Barth A. J. Biol. Chem. 1999; 274: 22170-22175Abstract Full Text Full Text PDF PubMed Scopus (33) Google Scholar). In the present work, sample conditions and temperature were fine-tuned so that the E2-P concentration is constant over several minutes. During that time the ATPase dephosphorylates and rephosphorylates several times leading to oxygen exchange between water and phosphate group (30Kanazawa T. Boyer P.D. J. Biol. Chem. 1973; 248: 3163-3172Abstract Full Text PDF PubMed Google Scholar). If water and phosphate oxygen isotopes are different, an isotope exchange at the E2-P phosphate group is the consequence of oxygen exchange that can be followed by infrared spectroscopy (36Barth A. Biopolymers. 2002; 67: 237-241Crossref PubMed Scopus (18) Google Scholar). Several isotope exchange experiments were performed using [γ-16O3]caged ATP in [18O]water, [γ-18O3]caged ATP in [16O]water, [γ-18O1]caged ATP in [16O]water, and [γ-18O1]caged ATP in [18O]water. Control samples contained [γ-16O3]caged ATP in [16O]water or [γ-18O3]caged ATP in [18O]water. Fig. 1 shows infrared difference spectra of E2-P formation from Ca2E1 (Ca2E1 → E2-P) in which the E2-P spectrum was recorded before and after isotope exchange. The bold line spectra were recorded before isotope exchange, and the thin line spectra were recorded afterward. The spectra overlap well above 1300 cm–1 where phosphates do not absorb, which confirms that the E2-P concentration is constant during the experiment. Below 1300 cm–1, spectra before and after exchange differ because of band shifts that are the result of the isotope exchange reaction. The spectra in Fig. 1a were obtained with unlabeled caged ATP in [18O]H2O, and the first, bold spectrum is characteristic of [16O3]E2-P as indicated by the marker band at 1194 cm–1 (32Barth A. J. Biol. Chem. 1999; 274: 22170-22175Abstract Full Text Full Text PDF PubMed Scopus (33) Google Scholar). Within 3 min, this band decays because of isotope exchange at the phosphate group, and the thin line spectrum is obtained. The absorbance changes associated with isotope exchange are very small. This is expected because they are caused by frequency shifts of the terminal P–O stretching vibrations which account for 3 vibrations of 50,000 protein vibrations. The isotope exchange experiment shown in Fig. 1b was carried out in the reverse direction; the ATPase was phosphorylated with [γ-18O3]caged ATP, which produced [18O3]E2-P first and [16O3]E2-P after oxygen exchange with [16O]H2O. Consequently the first spectrum (bold line) shows only a small signal at 1194 cm–1, which has increased in the thin line spectrum after isotope exchange. The comparison of the spectra in panels a and b of Fig. 1 demonstrate the excellent reproducibility of the infrared difference spectra. To show the isotope exchange associated absorbance changes more clearly, the bold and thin line spectra of Fig. 1 were subtracted. The spectra obtained are shown in Fig. 2 and are named isotope exchange spectra. Fig. 2a shows the 16O3 → 18O3 exchange spectrum, with positive bands being caused by [18O3]E2-P and negative bands being caused by [16O3]E2-P. The featureless spectrum above 1300 cm–1 demonstrates again that the E2-P concentration is constant to a very high degree in this experiment. In addition to bands caused by isotope exchange, other reactions might contribute to the spectrum. To distinguish these bands from isotope exchange bands, the exchange experiment was performed in the reverse direction (18O3 → 16O3), which should lead to a mirror image spectrum when compared with that in Fig. 2a. Fig. 2b shows two of these 18O3 → 16O3 exchange spectra. As expected for the isotope exchange spectrum, bands appear at the same positions as in Fig. 2a but with the opposite sign. The spectra show also bands caused by the slow steady state hydrolysis of ATP near 1250 and 1060 cm–1 (28Barth A. Mäntele W. Kreutz W. FEBS Lett. 1990; 277: 147-150Crossref PubMed Scopus (63) Google Scholar, 43Barth A. Mäntele W. Kreutz W. Biochim. Biophys. Acta. 1991; 1057: 115-123Crossref PubMed Scopus (68) Google Scholar). The hydrolysis rate is sensitive to the concentration of Me2SO, which is difficult to control precisely during the preparation of infrared samples. Therefore the hydrolysis rate varied between experiments. The bold line spectrum is the average of all experiments, and the thin line spectrum is the average of experiments with slow hydrolysis. A weighted subtraction of the spectrum of all samples from the spectrum of samples with slow hydrolysis was calculated to subtract the hydrolysis bands from the exchange spectrum (data not shown). The criterion for appropriate subtraction was a flat spectrum around 1250 cm–1 where a strong hydrolysis band is observed. This correction did not abolish the sharp bands between 1200 and 1090 cm–1, which shows that they are not associated with hydrolysis. However, the subtraction reduced the relative amplitudes of the bands below 1105 cm–1. Before discussing the spectra in more detail, further controls are shown in Fig. 2c. Fig. 2c shows several control spectra. For the two full line spectra no isotope exchange is expected because the isotopes of γ-phosphate of ATP and water are the same (18O in the thin line spectrum and 16O in the bold line spectrum). The dashed line spectrum shows an experiment where the ATPase was inhibited by the addition of the Ca2+ chelator EGTA. These spectra demonstrate that the sharp bands between 1200 and 1090 cm–1 in Fig. 2 (a and b) are only observed when the isotopes of γ-phosphate of ATP and water are different and when the ATPase is active. Therefore bands with opposite signs in Fig. 2 (a and b) are attributed to isotope exchange at the phosphate oxygens of E2-P. These bands are easily identified in Fig. 2d where the isotope exchange spectrum of Fig. 2a and that with slow hydrolysis of Fig. 2b are shown. The comparison highlights the mirror image property of the two spectra in the region between 1200 and 1090 cm–1. The mirror image property is not perfect for the amplitudes of positive bands at 1159 and 1091 cm–1 in the bold line spectrum. Correction for hydrolysis of the 18O3 → 16O3 exchange spectrum as discussed above improved the mirror image property by reducing slightly the amplitude of the 1091 cm–1 band. However, we prefer to show the original data here. Another reason for the small deviation from the ideal mirror image property could be incomplete isotope exchange in one of the spectra. We think that isotope exchange is complete for the 18O3 → 16O3 exchange as discussed below. However, it might not be complete for the 16O3 → 18O3 exchange. In these samples, the flat spectrum around 1250 cm–1 indicates very slow steady state hydrolysis of ATP. Therefore, E2-P was more stable in these samples than in samples for the 18O3 → 16O3 exchange with the possible consequence that isotope exchange is not complete during the observation time. To improve further the quality of the exchange spectrum, the two spectra in Fig. 2d were subtracted and divided by two. This cancels contributions to the spectrum that are unrelated to isotope exchange and enhances those related to isotope exchange. The resulting spectrum is shown in Fig. 2e. Negative bands in this spectrum are due to [16O3]E2-P and are found at 1194, 1137, and 1115 cm–1. Identification of [18O3]E2-P bands is less clear because positive bands are less prominent. The largest positive bands are observed at 1157 and 1091 cm–1. We propose that upon 16O3 → 18O3 exchange the band at 1194 cm–1 shifts to 1157 cm–1 and that at 1115 cm–1 shifts to 1091 cm–1. The band at 1137 cm–1 likely shifts to around 1115 cm–1 where it partially cancels the [16O3]E2-P band, which explains the small amplitude of the 1115 cm–1 band compared with the other two [16O3]E2-P bands. Small bands in the spectrum can be caused by substrates of E2-P or by a small percentage of Ca2E1-P. As a further test for the robustness of our isotope exchange spectrum, we conducted experiments with singly labeled caged ATP, [γ-18O1]caged ATP, which are shown in Fig. 3a. The bold line shows the experiment in [16O]water reflecting 18O116O2 → 16O3 exchange (absorbance of [16O3]E2-P minus that of [18O116O2]E2-P), and the thin line shows the experiment in [18O]water reflecting 18O116O2 → 18O3 exchange (absorbance of [18O]E2-P minus that of [18O116O2]E2-P). A discussion of the individual band shifts in the spectra of Fig. 3a is beyond the scope of this work. One band, however, is worth mentioning here. Near 1195 cm–1, a relatively large positive band is observed in the bold line spectrum but only a small band in the thin line spectrum. This indicates that the amplitude of this band is relatively large for [16O3]E2-P but small for [18O116 O2]E2-P, which therefore can be used as a marker band for complete exchange in an 18O3 → 16O3 exchange experiment. The large amplitude of this band in the thin line spectrum of Fig. 2d demonstrates that 18O3 → 16O3 exchange is complete or close to complete in this experiment. The spectra for partial isotope exchange can be used to mathematically obtain a second spectrum for full exchange. By subtracting the two spectra, the contribution of [18O116O2]E2-P cancels, and the absorbance of [18O3]E2-P minus that of [16O3]E2-P is obtained. This spectrum is shown as a thin line in Fig. 3b and compared with the experimental spectrum for full 16O3 → 18O3 exchange from Fig. 2e. The match between these spectra is very good and confirms the above attribution of bands to isotope exchange. The spectral positions of the phosphate bands are unaffected by the presence of Me2SO in our samples; band positions in isotope exchange spectra were the same for 10–30% Me2SO. In experiments not designed to monitor isotope exchange, the 1194 cm–1 band was observed at the same position in the presence and absence of Me2SO in spectra like those shown in Fig. 1. Infrared Bands of the E2-P Phosphate Group—Three bands of the E2-P phosphate group have been identified: those at 1194, 1137, and 1115 cm–1. They are associated with the P–O stretching vibrations of the PO3 moiety. The spectral positions of the phosphate bands are different from those of the model substance acetyl phosphate in water, which shows only two bands at lower wavenumbers: that of the symmetric stretching vibration at 983 cm–1 and that of the degenerated asymmetric vibration at 1132 cm–1. The observation of three bands for E2-P is due to an asymmetric environment around the phosphate group that lifts the degeneracy. The band at 1194 cm–1 has been identified before as a phosphate band (32Barth A. J. Biol. Chem. 1999; 274: 22170-22175Abstract Full Text Full Text PDF PubMed Scopus (33) Google Scholar) by comparing spectra obtained with labeled and unlabeled caged ATP. It was observed at the slightly different position of 1192 cm–1 becaus
Referência(s)