HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS		QUICK SEARCH:   [advanced] Author: Keyword(s): Year:  Vol:  Page:

J. Biol. Chem., Vol. 275, Issue 26, 19505-19512, June 30, 2000

This Article Abstract Full Text (PDF) All Versions of this Article: 275/26/19505    most recent M001784200v1 Alert me when this article is cited Alert me if a correction is posted Citation Map Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Pietri, S. Articles by Culcasi, M. Search for Related Content PubMed PubMed Citation Articles by Pietri, S. Articles by Culcasi, M. Social Bookmarking                      What's this?

- and -Phosphorylated Amines and Pyrrolidines, a New Class of Low Toxic Highly Sensitive ³¹P NMR pH Indicators

MODELING OF pK_a AND CHEMICAL SHIFT VALUES AS A FUNCTION OF SUBSTITUENTS^*

Sylvia Pietri, Malvina Miollan, Sophie Martel, François Le Moigne, Bruno Blaive^§, and Marcel Culcasi

From the Structure et Réactivité des Espèces Paramagnétiques, CNRS-UMR 6517 and the ^§ Ecole Nationale Supérieure de Chimie, Universités d'Aix-Marseille I et III, Faculté des Sciences de Saint-Jérôme, F-13397 Marseille Cedex 20, France

Received for publication, March 3, 2000

	ABSTRACT

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION APPENDIX REFERENCES

Fourteen linear and cyclic - and -aminophosphonates in which the P-atom is substituted by alkoxy groups have been synthesized and evaluated as ³¹P NMR pH markers in Krebs-Henseleit buffer. pK_a values varied with substitution in the range 1.3-9.1, giving potentially access to a wide range of pH. Temperature had a weak influence on pH and a dramatic increase in ionic strength slightly modified the pK_a of the pyrrolidine diethyl(2-methylpyrrolidin-2-yl)phosphonate (DEPMPH).All compounds displayed a 4-fold better NMR sensitivity than inorganic phosphate or other commonly used phosphonates, as assessed by differences _b-_a between the chemical shifts of the protonated and the unprotonated forms. In isolated perfused rat hearts, a non-toxic concentration window of 1.5-15 mM was determined for three representative compounds. Using empirical linear relationships, the experimental values of pK_a, _a, and _b have been correlated with the two-dimensional structure, i.e. the chemical nature of substituents bonded to the secondary amine and P-atom. The data suggest that DEPMPH and its cyclic and linear variants are ideal versatile ³¹P NMR probes for the study of tenuous pH changes in biological processes.

	INTRODUCTION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION APPENDIX REFERENCES

Since the pioneering studies of Moon and Richards (1), ³¹P NMR has become the standard noninvasive technique to determine the intracellular pH in a variety of biological models, based on the dependence of the endogenous inorganic phosphate (P_i) chemical shift on pH. In addition, the ex vivo or in vivo use of ³¹P NMR can concomitantly give access to dynamic processes involving the cellular phosphorylated metabolites. However, the usefulness of P_i as a ³¹P NMR probe for the measurement of pH in biological systems is limited by its low and varying concentration during metabolism and by its sensitivity to ionic strength. Indeed, owing to the large line widths of cellular peaks, and since the mean difference _ab between the ³¹P NMR chemical shifts of the protonated (_a) and the unprotonated (_b) forms of P_i is only 2.6-2.7 ppm, it cannot be used to provide information on differences on pH between intra- and extracellular compartments of less than 0.2 pH units (2).

A number of endogenous and exogenous analogues of phosphates and phosphonic acid metabolites have been tested as alternatives to P_i for the ³¹P NMR measurement of pH but their pK_a and chemical shifts were found to be influenced by ionic strength and Mg²⁺ concentration while their NMR sensitivity _ab was not greatly increased with respect to that of P_i (1-4). Later, it was considered that changing the chemical nature of the substituents around the protonation site of selected phosphates (i.e. the two anionic oxygens) might result in an increase in the NMR sensitivity for pH determinations. Therefore, commercially available alkyl- and aminophosphonates have been extensively studied in test solutions (3-11), cell cultures or preparations (5-9, 12), tumors (10), and isolated organs (13-16). The most commonly used charged phosphonates (e.g. methylphosphonate or phenylphosphonate), which have a similar narrow pK_a value range of 6.5-7.5 as P_i, were often found poorly permeable to cell membranes (13-16), in certain cases relatively toxic (8) and not significantly more sensitive than P_i although they can selectively give access to extracellular spaces (13-16). A marked modulation of the pK_a value range was obtained by introducing an amino function in the  position of the phosphonate structure (7, 10) and examples of such charged aminophosphonates having their pK_a values in a more acidic domain have been recently reported (7). However, because the NMR sensitivity of these compounds was still in the range of 2-3 ppm (7, 10), they have met with limited success as pH probes despite their low toxicity and relative insensitivity to temperature and ionic strength.

Taking into account the results of these previous studies (5-16), the rationale behind the present investigation was that, regarding the property of ³¹P NMR pH marker, there might be a favorable interplay between the phosphorylated and amine functions of a properly designed aminophosphonate if the preferred protonation site is at the nitrogen atom rather than at the anionic oxygen atoms linked to the phosphorus nucleus. Thus, the pK_a would be very sensitive to the environment of the nitrogen atom and the ³¹P NMR behavior (i.e. _ab) would strongly depend on the substituents linked to the phosphorus nucleus and their electric charge.

In this study, we have therefore designed a new series of uncharged - and -aminophosphonates with improved and versatile characteristics as ³¹P NMR pH probes in biological studies, i.e. that allow to investigate the largest range of pK_a values and that may offer the maximum NMR sensitivity at a given pK_a value (i.e. with resonances reasonably different from those of cellular metabolites, and the greatest _ab values). The calibration curves of these phosphorylated pH probes have been obtained in modified Krebs-Henseleit medium and the effects of temperature and ionic composition have been assessed. Semi-empirical models have been established to correlate the experimental pK_a, _a, and _b values with the chemical structure of any of the uncharged and charged aminophosphonates described herein or previously reported in the literature. Finally, to allow a sensitive detection by ³¹P NMR a concentration of at least 0.5 mM for such pH probes is required in biological experiments, and thus the toxicity of three selected uncharged -aminophosphonates has been evaluated in normoxic isolated hearts.

	MATERIALS AND METHODS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION APPENDIX REFERENCES

Synthetic Chemistry

All reagents and commercially available phosphonates were of the highest grade from Aldrich Chimie, Saint-Quentin-Fallavier, France. The general structures of the studied aminophosphonates are indicated in Structure 1. In the case of previously described syntheses, satisfactory elemental (± 0.3% for C, H, and P) and NMR data were obtained. Detailed synthetic procedures and analytical data for the three newly prepared - and -aminophosphonates are described in the "Appendix."

View larger version (31K): [in this window] [in a new window]   Structure 1.   Structures of aminophosphonates. Me, CH3; Et, CH2CH3; iPr, CH(CH3)2; Ph, C6H5.

-Phosphorylated Pyrrolidines (Structure 1A)

The following compounds were prepared according to published procedures: diethyl(2-methylpyrrolidin-2-yl)phosphonate 1 (DEPMPH)¹ (17, 18), diethyl(2-phenylpyrrolidin-2-yl)phosphonate 2 (19), tetraethyl(pyrrolidin-2,2-diyl)bisphosphonate 3, tetraisopropyl(pyrrolidin-2,2-diyl)bisphosphonate 4, and trans-tetraethyl(2,5-dimethylpyrrolidin-2,5-diyl)bisphosphonate 5a (20). The synthesis of cis- tetraethyl(2,5-dimethylpyrrolidin-2,5-diyl)bisphosphonate 5b will be described elsewhere. Ethyl(2-methylpyrrolidin-2-yl)methylphosphinate 6 was obtained by condensing ethyl-methylphosphinate on 2-methylpyrroline following a literature procedure (21) which will be described elsewhere.

Unbranched - and -Aminophosphonates (Structure 1B)

Diethyl(2-propylaminoprop-2-yl)phosphonate 7 and diethyl[1-(4-nitrophenyl)-1-propylaminoethyl]phosphonate 8 were prepared by a one-pot Kabachnik-Fields aminophosphorylation (22, 23) involving diethylphosphite and a mixture of the appropriate ketone and amine. Diethyl(2-butylaminoprop-1-yl)phosphonate 9 was obtained by a two-step synthesis adapted from a published procedure (24, 25) and involving the condensation of chloroacetone with triethylphosphite yielding diethyl(propan-2-one)phosphonate, which was then reacted with butylamine to afford 9.

Branched -Aminophosphonates (Structure 1C)

Diethyl(1-tert-butylamino-2-methylprop-1-yl)phosphonate 10 and diethyl(1-methylethylamino-1-cyclohexyl)phosphonate 11 were obtained by a procedure described in Ref. 26 involving the reaction of diethylphosphite with the imine formed by in situ condensation of the appropriate alkylamine and aldehyde (see Ref. 21). Tetraethyl(N-tert-butylaminomethylene)bisphosphonate 12 was prepared by addition of POCl₃ to a mixture of triethylphosphite and N-tert-butylacetamide as reported (20, 23).

³¹P NMR Spectroscopy: Temperature and Ionic Strength Studies

³¹P NMR spectra were acquired at 161.9 MHz with proton decoupling using a AMX400 Bruker spectrometer (Karlsruhe, Germany) equipped with a 10-mm broadband probe and a temperature control unit. The chemical shifts () were expressed relative to external 85% H₃PO₄ in D₂O. Fourier-transformed spectra were obtained with a pulse duration of 10 µs (90°) and a pulse interval of 1 s; the acquisition time was 2 s and 64 scans were typically averaged. As test compounds were intended to serve as pH markers in biological milieu, NMR titrations were performed in a standard Krebs-Henseleit buffer containing: KH₂PO₄ (1.2 mM), NaCl (118.5 mM), KCl (4.8 mM), MgSO₄ (1.2 mM), NaHCO₃ (25 mM), and EDTA (0.55 mM) in doubly distilled deionized water. Solutions of each test compound (5 mM) were freshly prepared and filtered through a 0.2-µm Millipore filter prior to use. For each test compound, the temperature of the solution was maintained at 20 or 37 °C using a circulating water bath, the pH was adjusted (Tacussel pH meter) to a selected value in the range 1.0-12.0 (10-30 values) using 6 M HCl or KOH and the final solution was transferred into a 10-mm sample tube for NMR assay. The pK_a value and the limiting chemical shifts _a and _b were calculated by iteratively fitting the  versus pH data (n = three independent experiments for each compound) to the Henderson-Hasselbalch equation,

(Eq. 1)

using a nonlinear regression.

To examine the effect of increasing ionic strength on titration curves, additional experiments were performed on 1 in which ³¹P NMR spectra were recorded under the same conditions as stated above except that buffer contained 0.3 M of either NaCl or KCl. These experiments also allowed the measurement of the effects of high ionic strengths on the KH₂PO₄ resonance peak (i.e. of P_i). Since the presence of calcium and glucose is required in many organ perfusion buffers (e.g. in isolated heart studies), their effect on titration curves was assessed in additional experiments using a CaCl₂ (2.5 mM)- and glucose (11.1 mM)-supplemented Krebs buffer.

Toxicity Experiments on Isolated Perfused Hearts

Perfusion Conditions-- Male Wistar rats (350-370 g) were anesthetized with sodium pentobarbital (50 mg/kg) by intraperitoneal injection. Hearts were quickly excised and retrogradely perfused at 37 °C via the aorta in the nonrecirculating Langendorff mode. Constant perfusion pressure of 100 cm H₂O was used. The perfusate was the CaCl₂- and glucose-supplemented Krebs buffer prepared as described above, gassed with 95% O₂, 5% CO₂ (pH 7.35) and passed through a 0.2-µm Millipore filter before use. A saline-filled latex balloon connected to a pressure transducer (Gould Statham P23) was inserted into the left ventricle, allowing measurement of contractile function. The filling pressure was adjusted until left ventricular end-diastolic pressure was in the range 8-12 mm Hg. The heart was enclosed into a water-jacketed chamber at a thermostatically controlled temperature of 37 °C. Throughout perfusion, heart rate, left ventricular end-diastolic pressure, left ventricular developed pressure, and its first derivative with time, dp/dt, were monitored at 5-min intervals via a Gould 8000S recorder. Cardiac performance was estimated by rate pressure product, calculated as the product of heart rate and left ventricular developed pressure. Coronary flow was measured by timely collecting the coronary effluent. Animal care and handling conformed to the Guide for the Care and Use of Laboratory Animals (40).

Experimental Groups and Protocols-- The perfusion protocol consisted of a 30-min normothermic control period with plain buffer followed by three subsequent 30-min periods where increasing concentrations of the tested aminophosphonate were added to the perfusion medium, and a final 30-min stabilization period with plain buffer. Three groups of hearts (n = 6) receiving the following compounds: 1 (at 5, 10, and 15 mM), 2 (at 0.5, 1.5, and 2.5 mM), and 3 (at 1, 5, and 10 mM) were studied and hemodynamic data were compared with that of an untreated group (n = 6) perfused with Krebs buffer throughout.

Data Analysis and Statistics

In titration experiments, data represent the means of three separate experiments and are expressed as mean ± S.D. In hemodynamic experiments, data as absolute values are expressed as mean ± S.D. Evaluation of statistical significance was conducted by two-way analysis of variance (ANOVA), with repeated measures over the different phases of the experimental protocol.

	RESULTS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION APPENDIX REFERENCES

pH-Dependent ³¹P NMR Behavior of Aminophosphonates-- The ³¹P NMR chemical shift of each aminophosphonate was determined as a function of pH at 20 °C in standard Krebs buffer containing physiological concentrations of NaCl and KCl (i.e. 118.5 and 4.8 mM, respectively). The ionic strength of this buffer being 0.207, measured pK_a values must therefore be considered as apparent pK_a values as obtained in intracellular medium (ionic strength range 0.15-0.20). For all compounds monophasic titration curves were obtained which could be fitted to the equation of a single sigmoidal curve, likely suggesting that protonation always occurred at the nitrogen atom. The fits were characterized by the sets of data presented in Table I and representative titration curves are shown in Fig. 1. Since compound 6 has two chiral centers (see Structure 1A), its couples of diastereoisomers 6a and 6b can be distinguished by ³¹P NMR, yielding two distinct resonance peaks that allowed calculation of their respective pK_a values (Table I).

View larger version (18K): [in this window] [in a new window]   Fig. 1.   pH dependence of the 31P NMR chemical shifts of Pi and selected aminophosphonates. The pH titrations were performed at 22 °C with 5 mM aminophosphonate solutions in a standard Krebs-Henseleit buffer containing 1.2 mM Pi as KH2PO4, 118.5 mM NaCl, 4.8 mM KCl, 1.2 mM MgSO4, 25 mM NaHCO3, and 0.55 mM EDTA. pH was adjusted with HCl or KOH. Each curve was obtained by a nonlinear fitting of three independent measurements on 10-30 pH values.

When compared with commonly used pH markers such as P_i and methylphosphonate, all -aminophosphonates showed a 2-4-fold increase in sensitivity, as assessed by _ab = _a  b measurements (Table I). This increase in _ab was also observed for the -aminophosphonate 9 but in a lesser extent (Table I). Titration curves were also obtained in the same buffer at 37 °C for P_i and six selected compounds and, as expected, a general decrease in pK_a with increasing temperature was observed (Table I). From temperature studies a change of 0.004 pH unit/°C was calculated for P_i, similar to that reported by DeFronzo and Gillies (6) for methylphosphonate. Interestingly, the decrease rate of pK_a with increasing temperature was found higher for -monophosphorylated compounds such as 1 and 7 (i.e. 0.013 and 0.011 pH unit/°C, respectively) than for -bisphosphorylated compounds such as 3, 4, and 5 (i.e. 0.008, 0.005, and 0.005 pH unit/°C, respectively). Of the tested compounds, -aminophosphonate 9 showed the most sensitive pK_a dependence with temperature (0.027 pH unit/°C).

It is known that any change in ionic strength dramatically affects pK_a values for P_i, methylphosphonate, and other phosphorylated metabolites (4, 6, 9). However, when titrations were repeated for 1 at 20 °C under nonphysiological ionic strength conditions (i.e. in Krebs buffer containing KH₂PO₄ and supplemented with 0.3 M Na⁺ or K⁺, yielding a ionic strength of 0.388), the observed changes in pK_a were even lower (i.e. +0.04 pH unit/0.1 M NaCl and +0.01 pH unit/0.1 M KCl; Table I) than that reported for methylphosphonate (6) while the pK_a of P_i was strongly decreased (i.e. 0.13 pH unit/0.1 M NaCl and 0.09 pH unit/0.1 M KCl; Table I). Consistently, when Krebs buffer was converted into a standard organ perfusion buffer by adding CaCl₂ (2.5 mM) and glucose (11.1 mM), the pK_a value of 1 remained unchanged at 7.01.

Modeling pK_a, _a, and _b Values as a Function of Substituents-- Since titration experiments showed that aminophosphonates have an enhanced potential as ³¹P NMR pH indicators with respect to P_i and other commonly used phosphonates (4-8), it was considered pertinent to devise a predictive model that could account for the large pK_a values and chemical shift ranges as a function of the chemical structure (Table I).

According to the hypothesis that the nitrogen atom of tested compounds is the protonation site, the pK_a should therefore mainly depend upon the substituents which are in the vicinity of the secondary amine function, i.e. those R₁  R₆ connected to the two sp³ carbon atoms bonded to the nitrogen atom (Fig. 2A). Assuming that the effect of substituents on pK_a was additive, the validity of the linear model given by Equation 2 was tested,

(Eq. 2)

where each substituent was classified into the seven types defined in Table II. In Equation 2, the coefficients n_i and a_i represent the number of each substituent of type i (with n_i = 6) and its corresponding electronic effect on the pK_a, respectively. Given that most of the studied compounds belong to the pyrrolidine series (termed as "cyclic alkyl" in Table II), the cyclic alkyl substituent was chosen as reference substituent with its coefficient a_i arbitrary taken as 0. The coefficients a₀-a₆ were computed by linear regression of the pK_a values at 20 °C of the 14 tested pyrrolidines and aminophosphonates of Table I, together with additional data on six relevant nonphosphorylated secondary amines taken from the literature (Refs. 27 and 28, see Table III). A satisfactory fit to the experimental pK_a values was obtained with Equation 2 (standard deviation  = 0.46 pH unit), which was further improved in the final Equation 3 ( = 0.17 pH unit) by introducing a new coefficient termed as a_gem·n_gem accounting for the geometric nonlinear effect induced when a second phosphorylated substituent is linked in gem position to a carbon substituted by a phosphorylated group (e.g. in compounds 3, 4, and 12).

(Eq. 3)

Tables II and III show the computed a_i values and the calculated pK_a values at 20 °C obtained from Equation 3, respectively.

View larger version (5K): [in this window] [in a new window]   Fig. 2.   Coordinate models used in the calculation of pKa values (A) and 31P NMR chemical shifts of the acidic and basic forms (B) for studied aminophosphonates.

View this table: [in this window] [in a new window]   Table II Calculated effects of substituents on pKa values at 20 °C drawn from : pKa = a0 + (ai·ni) + agem · ngem The contribution ai of each substituent Ri (i = 1-6; defined from the general structure represented in Fig. 2A) and the constant a0 were calculated by linear regression analysis of the experimental pKa values at 20 °C of (a) the 14 compounds (- and -aminophosphonates) shown in Table I and (b) the pKa values in water at 20-25 °C (27, 28) of the six nonphosphorylated amines listed in Table III. In pyrrolidines, the cyclic alkyl substituent was chosen as the reference and assigned a null coefficient. The abbreviations used are: P(O)XY, phosphonyl substituent; agem, incremental value to be added when a gem phosphorylated group is present. As typical examples, the respective values for n1 to n6 are (1, 2, 0, 0, 0, 1) for DEPMPH (1); (2, 2, 0, 1, 0, 1) for 8; and (0, 2, 0, 0, 0, 2) for the pyrrolidine 3 which contains a gem phosphorylated moiety.

As the sensitivity for the pH measurement by ³¹PNMR clearly depends upon the _a and _b values, it should therefore be related to the electronic influence of the substituents in the close environment of the tetra coordinated phosphorus atom of amino phosphonates, i.e. R₁-R₅ represented in Fig. 2B. It was then assumed that these substituents had additive effects on _a and _b, and therefore the linear model given by the system of Equations 4 and 5 was tested.

(Eq. 4)

(Eq. 5)

In Equations 4 and 5, n_j represents the number of each substituent of type j as defined in Table IV and b_ja and b_jb are their respective contributions to _a and _b. As alkyl substituents generally induce low displacing effects on ³¹P NMR chemical shifts, they were chosen as reference substituents and their coefficients b_a and b_b were arbitrary taken at zero. The set of contributions b_ja and b_jb of Table IV were finally calculated by linear regression from the experimental chemical shifts of aminophosphonates of Table I with additional data from 10 phosphonates determined either experimentally or found in the literature (7, 8, 29). As shown in Table III, this method gave a satisfactory fit to the experimental chemical shift values regarding the large chemical shift window (about 50 ppm) which was considered ( = 1.31 and 1.54 ppm for _a and _b, respectively). The mean percentage of absolute deviation of the calculated _a and _b values was only of 5% when compared with the experimental values. For each substituent of type j, the difference (b_jb  b_ja) (Table IV) is therefore believed to reflect the importance of the structural changes occurring at the level of the substituent due to the protonation of the molecule. As expected, the most important (b_jb  b_ja) differences were obtained for protonation sites, i.e. the hydroxyl groups of phosphates and the amino function of - and -aminophosphonates (Table IV).

View this table: [in this window] [in a new window]   Table IV Calculated effects of substituents on 31P NMR acidic (a) and basic (b) chemical shifts at 20 °C drawn from: a = b0a + bja · nj and b = b0b + bjb · nj The contribution bja or bjb of each substituent R1-R5 (defined from the general structure represented in Fig. 2B) was obtained by linear regression analysis of the experimental a or b values at 20 °C of (a) methylphosphonate and the 14 - and -aminophosphonates shown in Table I, and (b) the 10 phosphonates listed in the bottom of Table III. The contributions of a methyl group linked to a P atom and of a sp3-carbon linked to an -C atom are taken as zero. The abbreviations used are: P(O)XY, phosphonyl substituent; NHX, amino substituent linked to a phosphate group; NH, amino substituent linked to a dialkoxyphosphonyl group.

Toxicity of Selected Aminophosphonates on the Normoxic Isolated Rat Heart-- To determine a general non-toxic concentration window for the use of aminophosphonates as ³¹P NMR pH markers in biological studies, the effect of three selected compounds was evaluated on the baseline hemodynamic function of aerobically perfused rat hearts. Since cellular acidosis is a common feature in many pathological situations, compounds 2 and 3 were chosen due to their acidic pK_a values, and were compared with DEPMPH (1) whose pK_a is at the physiological pH (see Table I). Table V shows that perfusing increasing concentrations of either 1 or the bisphosphonate 3 up to 15 and 10 mM, respectively, caused no significant changes in baseline hemodynamic variables during a 150-min perfusion sequence, as assessed by rate pressure product and coronary flow measurements. The same pattern was observed for 1 and 3 when considering the effects on left ventricular developed pressure, left ventricular end-diastolic pressure, and dp/dt (not shown). Although hearts receiving 2.5 mM of the phenyl substituted phosphonate 2 demonstrated a significantly altered cardiac function, decreasing the concentration to 1.5 mM restored hemodynamic values similar to that of the untreated group (Table V).

View this table: [in this window] [in a new window]   Table V Dose-dependent effect of aminophosphonates pH markers on selected hemodynamic parameters of normoxic rat isolated hearts Data are presented as mean ± S.E. (n = 6 in each group). The abbreviations used are: RPP, rate pressure product (in mm Hg · beats · min1); CF; coronary flow (in ml · min1 · g1).

	DISCUSSION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION APPENDIX REFERENCES

Among the methods that are available to measure the intracellular pH in living cells (e.g. use of fluorescent dyes or micro-electrodes), NMR remains unique since it can be performed noninvasively. Although sensitive ¹⁹F NMR techniques associated with the use of fluorinated pH markers have been developed (30, 31), ³¹P NMR has been most widely utilized in vivo since it can also provide information on important phosphorylated metabolites such as ATP or creatine phosphate. Historically, naturally occurring ³¹P NMR pH indicators such as P_i or other phosphorus-containing metabolites (ATP, 2,3-phosphoglycerate, and glucose 6-phosphate) have been used extensively (3, 4) but as metabolites, their concentrations are prone to vary during a metabolic pathway and, in some cases, their NMR detection becomes quite difficult. Nevertheless, P_i is usually considered as the best endogenous indicator of the cytosolic pH (with ± 0.05 pH unit accuracy). However, the main limitation of P_i as ³¹P NMR pH probe relies on the small difference existing between the resonances of the intracellular and extracellular peaks (2) that precludes any accurate study of transsarcolemmal proton movements. Another important limitation of P_i as a ³¹P NMR probe is encountered when high concentrations of KH₂PO₄ are required in perfusion medium such as that used in organ preservation studies.

From the above considerations, it has become clear that exogenous ³¹P NMR probes had to be developed to improve the information provided by endogenous phosphorylated compounds and many efforts toward this aim have been performed in recent years (4-15). The basic requirements (32) for an endogenous pH marker are to display (a) a pK_a value in the physiological range, (b) a large effect of pH on _ab, (c) a resonance frequency clearly separable with that of other phosphorylated metabolites, (d) a low ionic strength dependence of the chemical shift of the resonance peak, (e) the lowest toxicity, and (f) a reasonable cellular permeability. This report promotes the concept that it is possible to cover a very large range of pH values (with 1.31 < pK_a < 9.05, see Table I) by using molecules belonging to the same family, i.e. - and -aminophosphonates. Besides the well investigated pH ranges occurring during normoxia (i.e. 7.0-7.4) or ischemic conditions (i.e. as low as pH 6), there is now considerable interest in the study of very acidic cellular compartments (i.e. 4 < pH < 6) such as those found in certain exocytic and endocytic vacuolar systems or in hepatocytes using other methods than NMR (33, 34). Recently, Brénot et al. (7) reported the first ³¹P NMR assessment of an acidic endosomal compartment in Dictyostelium discoideum amoebae using commercially available aminomethylphosphonate (pK_a 5.5) and 2-aminoethylphosphonate (pK_a 6.3) but, since the amplitude of their chemical shift variation _ab with pH was in the narrow range of 2.0-2.5 ppm, the accuracy of pH measurement using these two probes remains similar to that obtained with P_i at physiological pH. Table I and Fig. 1 show that an important feature of the aminophosphonates developed in this study is that their _ab values are in the range of 8-11 ppm while the _ab range of alkylphosphonic acid derivatives reported in the literature is only 2-3 ppm (2-16). Owing to the sigmoidal shape of titration curves, the pH value is best determined within a range where the slope is maximum, i.e. generally within ±1 pH unit of the pK_a value. This sensitivity is in the range of 1.0-1.5 ppm/pH unit for the best phosphorylated probes reported to date (2-16), while it is raised to 4-5 ppm/pH unit for the aminophosphonates described in the present study. Furthermore, outside this larger window, calibration curves for all aminophosphonates described herein show that rather accurate pH measurement can be performed at an extra ± 1 pH unit of the pK_a. In the case of 1 (pK_a 7.01), the chemical shift variation, which is as great as 5 ppm in the pH range 6-7, is still of 1.5 ppm in the more acidic domain of pH 5-6 and a measurable change in the chemical shift of 0.3 ppm can be still observed in the pH range of 4-5 (Fig. 1).

The results shown in Tables I-IV confirm that when the protonation occurs at the nitrogen atom of an - or -aminophosphonate instead at phosphorus as in alkylphosphonic acids, then both the pK_a and _ab values strongly vary upon the nature and position of substituents. From the results of the modeling studies (Tables II and III), it can be suggested that the electronic effects of substituents ("inductive model") play an important role on pK_a and _ab. Indeed, it is not unexpected that increasing the inductive electron withdrawing effect in the vicinity of the nitrogen atom led to a decrease of the basicity and thus to a decrease of the pK_a value. This was observed in the cyclic series when the methyl group of 1 was replaced by inductive attractors such as phenyl or dialkoxyphosphoryl groups in 2 and 3-5, respectively. Conversely, any structural factor that can increase the nucleophilicity of the nitrogen lone pair (e.g. removal of an electron withdrawing group by intercalating a methylene) increases the basicity of the compound raising its pK_a value, as was seen for linear compounds 7 and 9 (Table I). The calculations shown in Table II also demonstrate that, in the case of bisphosphorylated pyrrolidines, a nonlinear geometrical increment n_gem·a_gem must be added to the inductive model to predict more accurately the pK_a values of pyrrolidines 3 and 4 in which the two phosphorylated groups are linked to the same -carbon. Because such a strong steric effect in gem position to the amine function is able to induce a marked change in the conformational equilibrium of the five-membered ring, it can slightly increase the basicity of the pyrrolidine by destabilizing the nitrogen lone pair. Such an effect of gem steric crowding is supported by the finding that the pK_a values of cis- and trans-5, which lack this gem interaction, are well modeled by the non-corrected inductive model (Table II) although some deviation occurred for cis-5 (Table III).

Less straightforward is the interpretation of the dependence of ³¹P NMR chemical shifts _a and _b upon the nature of substituents around the phosphorus nucleus, even if the linear model of Equations 4 and 5 gave satisfactory results (Tables III and IV). A widely accepted theoretical model for the effect of substitution on ³¹P chemical shifts involves a linear combination of the "inductive effect" of substituents, the changes in the pyramidalization at the phosphorus atom and the p-d bonding with adjacent atoms (35). If one considers the general trend for _a, _b, and _ab values in Table I, the "inductive" effect does not appear to prevail as in the case of pK_a since the linkage of an electron withdrawing substituent to the phosphorus atom induces an apparent shielding whereas electron density is expected to decrease. Interestingly, semi-empirical molecular orbital calculations demonstrated that electron density and pK_a decreased with increasing electron withdrawing effect of substituents in a series of alkylphosphonic acids but no correlation with the ³¹P chemical shifts was reported (36). The data presented herein therefore suggest that at least the changes in the pyramidalization at the phosphorus atom (i.e. bond angles) that are likely to occur between the acidic and basic forms of aminophosphonates could play an important role in the observed variations of chemical shifts, i.e. _ab (Table I). A deshielding with decreasing bond angles at the phosphorus atom has been established in a large series of alkylphosphates (35). For - and -aminophosphonates, a general stereochemical process can be hypothesized in which bond angles _a and _a in the acidic form increase with respect to angles _b and _b in the basic form, as a consequence of a stabilizing intramolecular hydrogen bonding between the ammonium ion and one phosphoryl oxygen lone pair (Reaction 1).

View larger version (12K): [in this window] [in a new window]   Reaction 1.   Hypothetical stereochemical scheme showing the increase in bond angles  and  occurring during protonation of - (n = 1) and -aminophosphonates (n = 2). This increase of a and a could result from a stabilizing hydrogen bonding and explain (see Ref. 35) the observed shielding in the acidic form with respect to the basic form observed in the 31P NMR spectra.

A third important feature of - and -aminophosphonates as ³¹P NMR pH markers is that although their characteristic resonance peak is highly pH-sensitive, it is always located in the downfield position relative to other endogenous phosphorus-containing metabolites, making integral determinations straightforward. This relative high-field resonance peak of dialkoxyphosphoryl-containing compounds has recently allowed the concomitant monitoring by ³¹P NMR of the metabolic behavior of high-energy phosphates and a perfused nitrone spin-trap derived from pyrrolidine 1 during myocardial ischemia/reperfusion (37).

Another great advantage of using ³¹P NMR of aminophosphonates to determine pH values in biological systems results from their relatively low sensitivity to ionic strength conditions (Table I). This situation strongly contrasts with that of P_i for which an increase in ionic strength of 0.2 caused a decrease of the apparent pK_a of more than 0.4 pH unit (3, 38, 39) while the apparent pK_a value of 1 was not significantly affected (Table I). The ionic strength-dependent decrease of the chemical shift of methylphosphonate was reported comparable in magnitude to that of P_i (3, 6, 9). According to the Debye-Hückel equation, the apparent pK_a depends on the ratio of the activity coefficients of the acidic and basic forms which are functions of ionic strength (38). Thus, the observed differences between phosphates and aminophosphonates toward ionic strength sensitivity likely reflect the variations in activity coefficients occurring during the ionization of the phosphonate group while protonation of the non-dissociable amino group does not affect the activity coefficient to a large extent.

Based on the present toxicity studies in isolated perfused rat hearts (Table V), concentrations of aminophosphonates in the range of 1-10 mM can be used safely. This concentration range, which is similar to that reported with phosphonates in several model systems (13-16), is compatible with a high sensitivity of the NMR detection.

In summary, the data presented herein suggest that - and -aminophosphonates in which alkoxyl groups are linked to the phosphorus atom represent valuable general tools for the ³¹P NMR study of pH modifications during biological processes. Semi-empirical correlations between the chemical structure and experimental data have allowed probes to be designed that encompass a wide range of pK_a values. Provided the low toxicity found for DEPMPH is a common feature of this class of compounds, aminophosphonates could be of great use to enhance our understanding of ischemic cell injury at the subcellular level.

	FOOTNOTES

^* The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

To whom correspondence should be addressed: SREP-CNRS UMR 6517 (Case 521), Universités d'Aix-Marseille I et III, Faculté des Sciences de Saint-Jérôme, Avenue Escadrille Normandie Niemen, 13397 Marseille Cedex 20, France. Tel.: 33-4-91-28-85-79; Fax: 33-4-91-98-85-12; E-mail: pietri@srepir1.univ-mrs.fr.

Published, JBC Papers in Press, March 21, 2000, DOI 10.1074/jbc.M001784200

	ABBREVIATIONS

The abbreviations used are: DEPMPH, diethyl(2-methylpyrrolidin-2-yl)phosphonate; _a, limiting chemical shift in acidic medium; _b, limiting chemical shift in basic medium; P_i, inorganic phosphate.

	Appendix

Synthetic Procedures for Linear Unbranched Aminophosphonates 7, 8, and 9 (see Structure 1)

¹H and ³¹P NMR spectra were acquired on a Bruker AMX 400 spectrometer at 400 and 161.9 MHz with proton decoupling, respectively. Chemical shifts () are reported in ppm from internal Me₄Si (¹H NMR) and from external 85% H₃PO₄ (³¹P NMR). Downfield shifts are noted positive in all cases.

Diethyl(2-propylaminoprop-2-yl)phosphonate 7Diethylphosphite (28 g, 0.20 mol) and n-propylamine (25 g, 0.42 mol) were dissolved in acetone (24 g, 0.40 mol). The reaction mixture was stirred for 3 days at room temperature under nitrogen bubbling and treated with 37% aqueous HCl to pH 1 and extracted with diethyl ether (4 × 50 ml). Aqueous NaOH was added to the aqueous layer to pH 10, and the mixture was then extracted with methylene chloride (3 × 50 ml). The organic phase was washed with brine and dried over MgSO₄. Removal of the solvent under reduced pressure afforded 7 as a colorless oil (31.3 g, 66%), bp 65-70 °C at 0.06 mbar. ¹H NMR (CDCl₃, 400 MHz)  4.10 (m, 4 H), 2.63 (td, 2 H, ³J_HH 7.3 Hz, ³J_HH 1.0 Hz), 1.55 (br s, 1 H), 1.41 (tq, 2 H, ³J_HH 7.3 Hz, ³J_HH 7.2 Hz), 1.28 (t, 6 H, ³J_HH 7.0 Hz), 1.25 (d, 6 H, ³J_PH 15.8 Hz), 0.88 (t, 3 H, ³J_HH 7.3 Hz); ³¹P (¹H) NMR (CDCl₃, 161.9 MHz)  30.42. 


Diethyl[1-(4-nitrophenyl)-1-propylaminoethyl]phosphonate 8A mixture of 4-nitrobenzophenone (9.45 g, 58 mmol), propylamine (5.49 g, 93 mmol), and Na₂SO₄ (8 g, 58 mmol) and 0.5 ml of 37% HCl were stirred for 2 days at room temperature. Diethylphosphite (9.6 g, 70 mmol) was added, and the reaction mixture was stirred for 2 days at room temperature under nitrogen bubbling, treated with 37% aqueous HCl until pH 1, and extracted with diethyl ether (4 × 50 ml). Aqueous NaOH was added to the aqueous layer until pH 10, and the mixture was then extracted with methylene chloride (3 × 50 ml). The organic layer was washed with brine and dried over MgSO₄. Removal of the solvent under reduced pressure afforded 8 as a colorless oil (5.0 g, 25%), bp 120 °C at 0.05 mbar. ¹H NMR (CDCl₃, 400 MHz)  8.15 (d, 2 H, J_HH 10.4 Hz), 7.20 (d, 2H, J_HH 10.0 Hz), 4.05 (m, 4 H), 2.46 (m, 2 H), 1.78 (d, 3 H, ³J_PH 15.6 Hz), 1.45 (m, 2 H), 1.21 (m, 6 H), 0.82 (t, 3 H, ³J_HH 7.3 Hz); ³¹P (¹H) NMR (CDCl₃, 161.9 MHz)  24.06. 


Diethyl(2-butylaminoprop-1-yl)phosphonate 9A mixture of triethylphosphite (52.1 g, 0.314 mol) and freshly distilled chloroacetone (29 g, 0.314 mol) was refluxed for 4 h at 160-165 °C to yield diethyl acetylmethanephosphonate (6 g, 10%), bp 98 °C at 1 mm Hg, as described in Ref. 24. A mixture of diethyl acetylmethanephosphonate (1 g, 5.2 mmol), n-butylamine (0.40 g, 5.5 mmol), NaBH(OAc)₃ (1.64 g, 7.7 mmol), and acetone (0.33 g, 5.7 mmol) was stirred in 10 ml of 1,2-dichloroethane for 8 h at room temperature under nitrogen bubbling. Water was added to the mixture and concentrated HCl (37%) was added until pH 1. The aqueous layer was separated and washed with methylene chloride (4 × 50 ml). Aqueous NaOH was added to the aqueous layer until pH 10, and the mixture was then extracted with methylene chloride (3 × 10 ml). The organic layer was washed with brine and dried over MgSO₄. Removal of the solvent under reduced pressure afforded 9 as a colorless oil (0.7 g, 55%), bp 65 °C at 0.065 mbar. ¹H NMR (CDCl₃, 400 MHz)  4.08 (m, 4 H), 3.04 (qdd, 1 H, ³J_HH 6.3 Hz, ³J_HH 7.0 Hz, ³J_HH 5.7 Hz), 2.60 (dt, 2 H, ¹J_HH 11.0 Hz, ³J_HH 7.0 Hz), 2.45 (dt, 2 H, ¹J_HH 11.0 Hz, ³J_HH 7.2 Hz), 1.93 (ddd, 1 H, ²J_PH 18.5 Hz, ¹J_HH 15.3 Hz, ³J_HH 5.7 Hz), 1.62 (brs, 1 H), 1.74 (ddd, 1 H, ²J_PH 17.8 Hz, ¹J_HH 15.3 Hz, ³J_HH 7.0 Hz), 1.42 (m, 2 H), 1.33 (m, 1 H), 1.29 (t, 6 H, ³J_HH 7.0 Hz), 1.15 (dd, 3 H, ³J_HH 6.3 Hz), 0.88 (t, 3 H, ³J_HH 7.3 Hz); ³¹P (¹H) NMR (CDCl₃, 161.9 MHz)  29.39. 

	REFERENCES

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION APPENDIX REFERENCES

1.	Moon, R. B., and Richards, J. H. (1973) J. Biol. Chem. 248, 7276-7278
2.	Gillies, R. J., Alger, J. R., den Hollander, J. A., and Shulman, R. G. (1982) in Intracellular pH: Its Measurement, Regulation, and Utilization in Cellular Functions (Nuccitelli, R. , and Deamer, D. W., eds) , pp. 79-104, Alan R. Liss, New York
3.	Schlisefeld, L. H., Burt, C. T., and Labotka, R. J. (1982) Biochemistry 21, 317-320
4.	Robitaille, P. M. L., Robitaille, P. A., Brown, G. G., Jr., and Brown, G. G. (1991) J. Magn. Reson. 92, 73-84
5.	Labotka, R. J., and Kleps, R. A. (1983) Biochemistry 22, 6089-6095
6.	DeFronzo, M., and Gillies, R. J. (1987) J. Biol. Chem. 262, 11032-11037
7.	Brénot, F., Aubry, L., Martin, J. B., Satre, M., and Klein, G. (1992) Biochimie (Paris) 74, 883-895
8.	Satre, M., Martin, J. B., and Klein, G. (1989) Biochimie (Paris) 71, 941-948
9.	Stewart, I. M., Chapman, B. E., Kirk, K., Kuchel, P. W., Lovric, V. A., and Raftos, J. E. (1986) Biochim. Biophys. Acta 885, 23-33
10.	Gillies, R. J., Liu, Z., and Bhujwalla, Z. (1994) Am. J. Physiol. 267, C195-C203
11.	Szwergold, B. S., Brown, T. R., and Freed, J. J. (1989) J. Cell. Physiol. 138, 227-235
12.	Thoma, W. J., Steiert, J. G., Crawford, R. L., and Ugurbil, K. (1986) Biochem. Biophys. Res. Commun. 138, 1106-1109
13.	Fisher, M. J., and Dillon, P. F. (1987) Circ. Res. 60, 472-477
14.	Clarke, K., Stewart, L. C., Neubauer, S., Balschi, J. A., Smith, T. W., Ingwall, J. S., Nédelec, J. S., Humphrey, S. M., Kléber, A. G., and Springer, C. S. (1993) NMR Biomed. 6, 278-286
15.	Clarke, K., Anderson, R. E., Nédélec, J. F., Foster, D. O., and Ally, A. (1994) Magn. Res. Med. 32, 181-188
16.	Bruynseels, K., Gillis, N., Van Hecke, P., and Vanstapel, F. (1997) NMR Biomed. 10, 263-270
17.	Fréjaville, C., Karoui, H., Tuccio, B., Le Moigne, F., Culcasi, M., Pietri, S., Lauricella, R., and Tordo, P. (1995) J. Med. Chem. 38, 258-265
18.	Barbati, S., Clément, J. L., Olive, G., Roubaud, V., Tuccio, B., and Tordo, P. (1997) NATO ASI Ser. Ser. A Life Sci. 27, 39-47
19.	Karoui, H., Nsanzumuhire, C., Le Moigne, F., and Tordo, P. (1999) J. Org. Chem. 64, 1471-1477
20.	Pietri, S., Miollan, M., Le Moigne, F., and Culcasi, M. (March 18, 1998) French Patent CNRS 98,03317
21.	Afarinkia, K., Rees, C. W., and Cadogan, J. I. G. (1990) Tetrahedron 46, 7175-7196
22.	Medved, T. Y., and Kabachnik, M. I. (1954) Izv. Akad. Nauk. SSSR Otd. Khim. Nauk. 255-261; Engl. Transl. (1954) Bull. Acad. Sci. USSR Div. Chem. Sci. 314-320
23.	Fields, E. K. (1952) J. Am. Chem. Soc. 74, 1528-1531
24.	Dawson, N. D., and Burger, A. (1952) J. Am. Chem. Soc. 74, 5312-5314
25.	Abdel-Magid, A. F., Maryanoff, C. A., and Carson, K. G. (1990) Tetrahedron Lett. 31, 5595-5598
26.	Grimaldi, S. (1997) Ph. D. Thesis , Université d'Aix-Marseille III, France
27.	Gordon, A. J., and Ford, R. A. (1972) The Chemist's Companion: A Handbook of Practical Data, Techniques, and References , pp. 55-81, Wiley-Interscience, New York
28.	Sosnovsky, G., and Bell, P. (1998) Life Sci. 62, 639-648
29.	Grabenstetter, R. J., Quimby, O. T., and Flautt, T. J. (1967) J. Phys. Chem. 71, 4194-4202
30.	Deutsch, C. J., and Taylor, J. S. (1987) Ann. N. Y. Acad. Sci. 508, 33-47
31.	Deutsch, C. J., and Taylor, J. S. (1989) Biophys. J. 55, 799-804
32.	Szwergold, B. S. (1992) Annu. Rev. Physiol. 54, 775-798
33.	Anderson, R. G. W., and Orci, L. (1988) J. Cell Biol. 106, 539-543
34.	Wadsworth, S. J., and van Rossum, G. D. V. (1994) J. Membr. Biol. 142, 21-34
35.	Gorenstein, D. G. (1984) in Phosphorus-31 NMR: Principles and Applications (Gorenstein, D. G., ed) , pp. 7-36, Academic Press, Orlando, FL
36.	Ohia, K. (1992) Bull. Chem. Soc. Jpn 65, 2543-2545
37.	Maurelli, E., Culcasi, M., Delmas-Beauvieux, M. C., Miollan, M., Gallis, J. L., Tron, T., and Pietri, S. (1999) Free Radic. Biol. & Med. 27, 34-41
38.	Seo, Y., Murakami, M., Watari, H., Imai, Y., Yoshizaki, K., Nishikawa, H., and Morimoto, T. (1983) J. Biochem. 94, 729-734
39.	Roberts, J. K. M., Wade-Jardetsky, N., and Jardetzky, O. (1981) Biochemistry 20, 5389-5394
40.	U.S. Department of Health and Human Services (1985) NIH publication 85-23