J. Lipid Res.
HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
QUICK SEARCH:   [advanced]


     

This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF)
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Right arrow Citation Map
Services
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Li, S.
Right arrow Articles by Schroepfer, G. J.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Li, S.
Right arrow Articles by Schroepfer, G. J., , Jr.
Social Bookmarking
Add to CiteULike   Add to Complore   Add to Connotea   Add to Del.icio.us   Add to Digg   Add to Reddit   Add to Technorati  
What's this?

The Journal of Lipid Research, Vol. 40, 764-772, April 1999
Copyright © 1999 by Lipid Research, Inc.

Methods

New methods for determining the enantiomeric purity of erythro -sphingosine

Shengrong Lia, William K. Wilsona, and George J. Schroepfer, Jr.2,a
a Departments of Biochemistry and Cell Biology and of Chemistry, Rice University, 6100 Main Street, Houston, TX 77005-1892

Correspondence to: William K. Wilson


  ABSTRACT
TOP
 ABSTRACT
INTRODUCTION
 EXPERIMENTAL PROCEDURES AND...
 DISCUSSION
 REFERENCES

The enantiomeric purity of erythro -sphingosine samples can be determined simply, reliably, and accurately from 1H or 19F nuclear magnetic resonance spectra of the {alpha}-methoxy-{alpha}-(trifluoromethyl)phenylacetate (MTPA) derivative. As little as 0.1% of the minor enantiomer could be observed in a 1-mg sample, and detection limits of 1% and 5% were estimated for samples of 100 µg and 10 µg. The two threo -sphingosine enantiomers and four dihydrosphingosine stereoisomers were also differentiated by this technique, which served as an effective method for assessing the purity of sphingosine and dihydrosphingosine samples. Enantiomeric and diastereomeric purities could also be determined by normal-phase high performance liquid chromatographic analysis of the MTPA derivatives.—Li, S., W. K. Wilson, and G. J. Schroepfer, Jr. New methods for determining the enantiomeric purity of erythro -sphingosine. J. Lipid Res. 1999. 40: 764–772.

Supplementary key words: sphingosine, enantiomeric purity, MTPA derivative, NMR, HPLC


  INTRODUCTION
TOP
 ABSTRACT
 INTRODUCTION
EXPERIMENTAL PROCEDURES AND...
 DISCUSSION
 REFERENCES

methods

Sphingosine and other long-chain bases are the defining elements of sphingolipids, a complex and diverse class of lipids of increasing interest in biology and medicine. A large number of studies have demonstrated that the free sphingolipid bases and their 1-phosphate and N -acyl derivatives have important and varied biological activities (1) (2) (3) (4) (5) (6) (7) (8) (9). Investigations of the biological actions of these compounds require samples of defined structure and purity. D-Erythro -sphingosine (1a; Figure 1) has traditionally been isolated from natural sources (10) (11), but total synthesis has become an important alternative ((12) (13) (14) (15) (16) (17) and references therein). A potential problem with total synthesis is the formation of stereoisomeric mixtures. Diastereomeric impurities can generally be detected by thin-layer chromatography (TLC), gas chromatography (GC), or nuclear magnetic resonance (NMR) spectroscopy (10) and removed by chromatographic methods. Enantiomeric impurities, by contrast, are difficult to remove and to detect.



View larger version (24K):
[in this window]
[in a new window]
  		Figure 1. Structures of sphingosine and dihydrosphingosine isomers, their (S)-MTPA derivatives, and related sphingolipid bases described herein.

Optical rotation, the classical method for measuring the enantiomeric purity of sphingosine samples, suffers from low precision and numerous potential sources of error (18). These problems are exacerbated by the low specific rotation of sphingosine and its N -acetate and triacetate derivatives (10). Circular dichroism spectra, from which sphingosine and dihydrosphingosine isomers can be identified (19), would be subject to similar (albeit fewer) sources of error for determining enantiomeric purity. A simple and generally reliable nonoptical method is NMR analysis with chiral lanthanide shift reagents (20) (21), which have been used to measure the enantiomeric purity of an N,O -protected sphingosine (16). However, we did not observe any splitting of 1H NMR signals at 500 MHz upon addition of Pr(hfc)3 to racemic sphingosine (1a and 2a) in CDCl3. Another important method for measuring enantiomeric purity is HPLC on chiral phases (22). The N-benzoyl derivatives of the four dihydrosphingosine isomers 5a–8a have been resolved by connecting two different chiral HPLC columns in tandem (23), and a CHIRALCEL OD column resolved enantiomeric pairs of the N-naphthimido derivatives of sphingosine and dihydrosphingosine (19). However, the promise of good separations is offset by the need to acquire chiral columns and the effort of validating the separation with authentic standards.

Chiral derivatizing agents are the basis of another popular method for measuring enantiomeric purity (21) (24). (S)-{alpha}-Methoxy-{alpha}-(trifluoromethyl)phenylacetate (MTPA) derivatives of chiral alcohols and amines are diastereomeric mixtures that can be analyzed by NMR or by GC or HPLC on achiral columns. NMR analysis of mono-MTPA derivatives has been used to determine the enantiomeric purity of model compounds (15) and synthetic intermediates (14) in total syntheses of sphingosine. For simplicity, we sought to apply this methodology by derivatizing sphingosine as the free base. Despite the potential complications of multiple MTPA groups, we easily converted sphingosine to its tris-MTPA derivative, and NMR analysis furnished accurate enantiomeric purities of erythro -sphingosine samples. Presented herein are the particulars of this simple and robust method and its extension to other sphingosine and dihydrosphingosine stereoisomers and to HPLC analyses.


  EXPERIMENTAL PROCEDURES AND RESULTS
TOP
 ABSTRACT
 INTRODUCTION
 EXPERIMENTAL PROCEDURES AND...
DISCUSSION
 REFERENCES

Materials and methods
D-Erythro -sphingosine (1a), L-threo -sphingosine (3a), and D-erythro -dihydrosphingosine (5a) were isolated from cow brain as described previously (10) (11) and showed high purity by TLC and 1H NMR. L-Erythro -sphingosine (2a) was purchased from Sigma (St. Louis, MO) and showed >=99% purity by NMR as the MTPA derivative 2b. DL-Erythro -sphingosine (1a, 2a; analyzed by NMR as a 40:40:10:10:10:10:1:1 mixture of MTPA derivatives 1b, 2b, 3b, 4b, 11b, 12b, 13b, and 14b), DL-threo -sphingosine (3a, 4a; analyzed by NMR as a 5:5:2:2 mixture of MTPA derivatives 3b, 4b, 13b, and 14b), DL-erythro -dihydrosphingosine (5a, 6a; ~98% purity), and DL-threo -dihydrosphingosine (7a, 8a; ~98% purity) were gifts from H. E. Carter, who received these materials in 1963 from Ciba Ltd. (Basel, Switzerland). This collection of DL mixtures, which included DL-erythro -2-amino-4-octadecyne-1,3-diol (11a, 12a; analyzed by NMR as a 8:8:1:1 mixture of MTPA derivatives 11b, 12b, 13b, and 14b), was evidently synthesized by the methods of Grob and coworkers (25) (26). (2S)-Sphingine (9a, (2S)-2-aminooctadecan-1-ol) and DL-sphingine (9a, 10a) were also gifts from H. E. Carter, who elucidated the structure of sphingine (27). The MTPA derivative of L-threo -dihydrosphingosine (7b) was prepared by hydrogenation of 3b overnight over 10% palladium on carbon in tetrahydrofuran, followed by removal of the catalyst by filtration, evaporation to dryness, and purification of the residue on silica gel (30 x 5 mm column; elution with ethyl acetate–hexane 3:7 (5 ml)); NMR analysis showed a 1:3 mixture of 7b and (2S)-sphingine (9b). Analogous hydrogenation of the MTPA derivative of D-erythro -sphingosine (1b) gave a 2:3 mixture of 5b and 9b. Similar hydrogenolysis during catalytic hydrogenation of D-erythro -sphingosine triacetate has been reported previously (27). (S)- {alpha}-Methoxy- {alpha}-(trifluoromethyl)phenylacetyl chloride (98% ee) and tris(3-(heptafluoropropylhydroxymethylene)-(+)-camphorato) praseodymium(III) (Pr(hfc)3) were purchased from Aldrich (Milwaukee, WI).

NMR spectra were measured on dilute CDCl3 solutions with a Bruker AMX500 or AC250 spectrometer and referenced to internal tetramethylsilane at 0 ppm (1H), CDCl3 at 77.0 ppm (center line; 13C), or CFCl3 at 0 ppm (downfield line; 19F). Spectra on the AMX500 were acquired on an inverse-detection probe at 25°C; spectra on the AC250 were obtained on the inner coil of a normal-geometry probe at ~22°C. Enantiomeric and diastereomeric purities were estimated by integration of spectra obtained with 0.5-Hz line broadening. Most 1D spectra were acquired for 5 min (128 scans) in regular 5-mm tubes, but 1–10 µg samples were acquired for 20–30 min in Shigemi tubes (also used for most 2D spectra). COSY-DQF (1H-1H correlation spectroscopy with double-quantum filtering), HSQC (heteronuclear single-quantum coherence), and HMBC (heteronuclear multiple bond correlation) spectra were used to confirm structures and establish signal assignments; linear prediction was used sparingly to improve resolution in f1. Chemical shifts measured to ±0.001 ppm precision were corrected for effects of strong coupling. No decomposition of the MTPA derivatives in CDCl3 (freshly filtered through basic alumina) was observed, even upon storage for over 2 weeks (mainly at -20°C). Analytical TLC was performed using aluminum-backed silica gel 60 F254 plates (EM Science, Gibbstown, NJ). TLC plates were charred by spraying with 5% ammonium molybdate in 10% sulfuric acid followed by heating for 5 min at 80°C. HPLC was carried out on 5-µm Alltima C18 or silica columns (250 mm x 4.6 mm; 100 Å pore size) from Alltech Associates (Deerfield, IL) with UV detection at 210 nm. Normal-phase HPLC was done with an in-line solvent degasser.

Preparation of (S)-MTPA derivatives of sphingosine and dihydrosphingosine
(S)-{alpha}-Methoxy-{alpha}-(trifluoromethyl)phenylacetyl chloride (30 µl, 160 µmol) was added to a solution of sphingosine or dihydrosphingosine (2 mg, 6.7 µmol) in pyridine (100 µl) followed by shaking for 5 min to mix the reagents. The mixture was left overnight at room temperature, quenched with water (2 ml), and extracted with dichloromethane (3 x 1.5 ml). The combined extracts were washed with water (1 ml) and brine (1 ml), dried over anhydrous sodium sulfate, and evaporated to a residue that was passed through silica gel (30 x 5 mm column; elution with ethyl acetate–hexane 3:7 (5 ml)). The eluate was evaporated to dryness in a stream of nitrogen and further dried in vacuo to a colorless oil. Less MTPA reagent was used for smaller reactions (10 µl for <=100 µg sphingosine; 20 µl for 1 mg sphingosine). TLC on silica gel (ethyl acetate–hexane 3:7) showed a slight separation of some sphingosine and dihydrosphingosine isomers (Rf 0.46–0.52).

Analysis of d- and l-erythro -sphingosine by NMR
(S)-MTPA derivatives of D-erythro -sphingosine (1b), L-erythro -sphingosine (2b), and various mixtures thereof were analyzed by 1H NMR at 250 and 500 MHz and by 19F NMR at 235 MHz. 1H and 19F NMR chemical shifts and 1H-1H coupling constants are summarized in Table 1. Mixtures carefully prepared to have 100:30, 100:10, 100:5, and 100:1 ratios of 1b and 2b showed ratios of 100:30.8, 100: 9.6, 100:4.9, and 100:1.1 by 1H NMR and 100:31.2, 100:9.0, 100:4.4, 100:1.0 by 19F NMR analysis of the MTPA derivatives. Methoxy (1H) and trifluoromethyl (19F) regions of the NMR spectra of the 100:30 and 100:1 mixtures together with spectra of the pure D- and L-erythro -sphingosine isomers are presented in Figure 2. 13C satellites in these spectra were distinguished from trace impurities by their intensities (0.55%) and coupling constants (290 Hz in 19F spectra and 145 Hz for methoxy signals in 1H spectra). Conversion of small quantities (1 mg, 100 µg, 10 µg, and 1 µg) of D-erythro -sphingosine to the MTPA derivative followed by 1H NMR analysis at 500 MHz indicated detection limits for the minor isomer of 0.1%, 1%, 5%, and 20%, respectively.



View larger version (28K):
[in this window]
[in a new window]
  		Figure 2. 1H and 19F NMR spectra of (S)-MTPA derivatives of D-erythro -sphingosine (1b) (panels A and B), L-erythro -sphingosine (2b) (panels G and H), a 100:1 mixture of 1b and 2b (panels C and D), and a 100:30 mixture of 1b and 2b (panels E and F). 1H and 19F NMR spectra were acquired at 500 and 235 MHz, respectively, and processed with 0.5-Hz line broadening (2-Hz line broadening for panel D). Signals of trace enantiomeric impurities, L-threo -sphingosine derivatives, and 13C satellites are marked by arrows, triangles, and closed circles ({downarrow}, {triangledown}, and {bullet}), respectively. The large signal at {delta}H 3.42 in panel C represents a non-sphingolipid byproduct of derivatization coincident with one of the methoxy peaks of 1b and has a correspondingly large 13C satellite at {delta}H 3.28. Other extraneous signals, marked by open circles ({circ}), appear in panels C and D at {delta}F -71.4, -72.2 and {delta}H 3.53.


 
View this table:
[in this window]
[in a new window]
  		Table 1. 1H and 19F NMR chemical shifts and 1H-1H coupling constants of (S)-MTPA derivatives of D-erythro -sphingosine and related sphingolipid basesa,b,c

The 250- and 500-MHz 1H NMR spectra showed chemical shift nonequivalence ({Delta}{delta}) of ~0.05 ppm for the upfield pair of methoxy signals, with peak widths at 2% height of 13 Hz (0.5 Hz line broadening), 10 Hz (no line broadening), or 7 Hz (mild Gaussian apodization). The trifluoromethyl signals of the 235-MHz 19F NMR spectra showed {Delta}{delta} values of ~0.2 ppm and corresponding peak widths of 23, 17, and 12 Hz. The 1H methoxy and 19F trifluoromethyl peaks were broadened by their mutual 1.3-Hz long-range 19F–1H coupling. Even at 250 MHz (or 235 MHz), the methoxy and trifluoromethyl peak separations were adequate for quantitative measurements. Certain 1H NMR signals of the sphingolipid moiety (Table 1) showed good chemical shift nonequivalence ({Delta}{delta} 0.2 ppm), but signal overlap and lower intensities made them less suitable for quantitative measurements.

Analysis of other sphingosine and dihydrosphingosine isomers by NMR
1H and 19F NMR spectra of (S)-MTPA derivatives of L-threo -sphingosine (3b), DL-threo -sphingosine (3b, 4b), D-erythro -dihydrosphingosine (5b), DL-erythro -dihydrosphingosine (5b, 6b), and DL-threo -dihydrosphingosine (7b, 8b), a 1:3 mixture of L-threo-dihydrosphingosine (7b) and (2S)-sphingine (9b), a 2:3 mixture of D-erythro -dihydrosphingosine (5b) and (2S)-sphingine (9b), (2S)-sphingine (9b), and DL-sphingine (9b, 10b) furnished chemical shifts and coupling constants for compounds 3b–8b (Table 1) and 9b and 10b (see below). The methoxy and trifluoromethyl regions of DL mixtures of threo -sphingosine and erythro - and threo -dihydrosphingosine are shown in Figure 3. In each case, at least one diastereomeric pair of methoxy or trifluoromethyl signals of the MTPA derivatives was resolved in an open spectral region. Of the 24 methoxy signals and 24 trifluoromethyl signals of the sphingosine and dihydrosphingosine isomers, the upfield set of 8 methoxy signals and the mid-field set of 8 trifluoromethyl signals showed the best peak dispersion. Isomer ratios could be measured from several 1H NMR signal pairs (Table 1), but the 1H methoxy and 19F trifluoromethyl signals gave the best combination of high sensitivity and good peak dispersion (chemical shift nonequivalence of ~0.05 and ~0.2 ppm, respectively) despite the long-range 19F–1H coupling of 1.3 Hz that broadened the 1H methoxy (W0.5 3 Hz) and 19F trifluoromethyl (W0.5 4 Hz) peaks. Peak dispersion and sensitivity may be improved by increasing the NMR field strength or (more effectively) by use of decoupling to remove the long-range 19F–1H couplings.



View larger version (30K):
[in this window]
[in a new window]
  		Figure 3. 1H and 19F NMR spectra of (S)-MTPA derivatives of L-threo -sphingosine (3b) (panels A and B), DL-threo -sphingosine (3b, 4b) (panels C and D), DL-erythro -dihydrosphingosine (5b, 6b) (panels E and F), and DL-threo -dihydrosphingosine (7b, 8b) (panels G and H). 1H and19F NMR spectra were acquired at 500 and 235 MHz, respectively, and processed with 0.5-Hz line broadening. Minor signals in panels C and D represent 13b and 14b. Broad signals marked by open circles ({circ}) in panels E and F ({delta}H 3.49; {delta}F -71.3) are attributable to a non-sphingolipid MTPA byproduct.

Additional NMR signal assignments for compounds 1b–14b
1H NMR signal assignments3 were made from COSY-DQF spectra and by matching coupling constants from different resonances. Unequivocal assignments for each methoxy (1H) and each trifluoromethyl (19F) signal could not be established from 2D spectra, but chemical shift patterns among 1b–14b, including comparisons with NMR data for MTPA derivatives of amines and alcohols (24) and sphingines 9b and 10b (see below), suggested that the upfield 1H signals ({delta} 3.1–3.3) and downfield 19F signals correspond to the 2–MTPA group and that signals near {delta}F –72 belong to the 1–MTPA. Further comparisons suggested {delta}H 3.52 and 3.43 for the 1–MTPA methoxy signals of the 2S and 2R isomers, respectively. 13C NMR assignments for 1b were determined from HSQC and HMBC spectra and chemical shift comparisons (10): {delta}C 166.28 (1–C=O), 166.18 (2–C=O), 165.15 (3–C=O); 140.54 (C–5); 131.75, 131.68, 129.87, 129.73, 129.60, 128.68, 128.61, 128.47 (aromatic); 127.64, 127.28, 127.20 (aromatic, JCF 1.4 Hz); 122.22 (C–4); 84.8, 84.6, 83.9 (1–MTPA, 3–MTPA, and 2–MTPA quaternary carbons, JCF ca. 30 Hz); 74.88 (C–3), 64.03 (C–1), 55.46 ({delta}H 3.516), 55.41 ({delta}H 3.423), 54.73 ({delta}H 3.186), 50.16 (C–2), 32.21, 31.91, 29.67, 29.66 (2C), 29.64, 29.55, 29.44, 29.34, 29.16, 28.55, 22.68, 14.10. NMR data for 9b: {delta}H 6.69 (d, 8.9 Hz, NH), 4.352 (dd, 11.3, 3.9 Hz, H–1), 4.393 (dd, 11.3, 3.8 Hz, H–1), 4.251 (m, H–2), 3.531 (q, 1.3 Hz, OCH3), 3.285 (q, 1.3 Hz, OCH3), 1.48 (m, H–3); {delta}F -72.00, -69.30. NMR data for 10b: {delta}H 6.84 (d, 8.9 Hz, NH), 4.372 (dd, 11.3, 4.0 Hz, H–1), 4.359 (dd, 11.3, 3.6 Hz, H–1), 4.25 (m, H–2), 3.431 (q, 1.3 Hz, OCH3), 3.295 (q, 1.3 Hz, OCH3), 1.48 (m, H–3); {delta}F -72.11, -69.42. NMR data for 11b, 12b (diastereomeric mixture): {delta}H 7.17 (d, 9.0 Hz, NH), 5.615 (dt, 4.5, 2.0 Hz, H–3), 4.653 (dddd, 9.0, 6.3, 5.3, 4.4 Hz, H–2), 4.524 (dd, 11.6, 5.3 Hz, H–1), 4.405 (dd, 11.6, 6.3 Hz, H–1), 2.14 (m, H–6); 7.05 (d, ~8.7 Hz, NH), 5.644 (m, H–3), 4.62 (m, H–1, H–2), 4.199 (m, H–1), 2.22 (m, H–6); 3.552, 3.514, 3.421, 3.375, 3.238, 3.155 (all q, 1.3 Hz, OCH3); {delta}F -72.21, -72.15, -72.12, -71.89, -69.72, -69.43. NMR data for 13b, 14b (diastereomeric mixture) {delta}H 7.04 (d, 9.0 Hz, NH), 5.624 (dt, 4.0, 2.1 Hz, H–3), 4.608 (dddd, ~9, 7.1, 5.8, 4.0 Hz, H–2), 4.393 (dd, 11.4, 5.9 Hz, H–1), 4.234 (dd, 11.4, 7.1 Hz, H–1), 2.09 (m, H–6); 7.18 (d, 9.0 Hz, NH), 5.598 (dt, 4.1, 2.1 Hz, H–3), 4.634 (dddd, 9.0, 6.9, 5.9, 4.0 Hz, H–2), 4.338 (dd, 11.4, 6.9 Hz, H–1), 4.241 (dd, 11.4, 6.1 Hz, H–1), 2.16 (m, H–6); 3.525, 3.514, 3.489, 3.429, 3.303, 3.203 (all q, 1.3 Hz, OCH3); {delta}F -72.18, -72.02, -72.00, -71.95, -69.67, -69.20.

T1 relaxation measurements
The following T1 relaxation times were measured on nondegassed samples by the inversion–recovery method: 1b, 2b, 0.6 s (H–1), 0.8 s (NH), 0.9 s (H–2), 1.1 s (H–3, 2–MTPA methoxy), 1.2 s (H–4, H–5), 1.3 s (1– and 3–MTPA methoxy); 3b, 4b, 0.6 s (H–1), 0.9 s (NH), 1.0 s (H–2), 1.1 s (H–3, 2–MTPA methoxy), 1.2 s (H–4, H–5), 1.4 s (1– and 3–MTPA methoxy); 5b, 6b, 0.6 s (H–1), 0.9 s (H–2, H–3, NH), 1.0 s (2–MTPA methoxy), 1.3 s (1– and 3–MTPA methoxy); 7b, 8b, 0.5 s (H–1), 0.7 s (H–2, H–3, NH), 0.9 s (2–MTPA methoxy), 1.2 s (1– and 3–MTPA methoxy). A similar 19F NMR experiment for 1b, 2b showed T1 values of 0.6 s (slightly less for the 2–MTPA trifluoromethyl). In 1H and 19F experiments, all pairs of signals corresponding to different diastereomers differed in T1 by <0.1 s. Consequently, relatively short pulse repetition times (e.g., 2–3 s) will not favor one stereoisomer over another. This conclusion was consistent with the observation that methoxy peak pairs in NMR spectra of MTPA derivatives of racemic mixtures had virtually equal intensities. Because of slightly shorter T1 values for 2–MTPA signals relative to 1– and 3–MTPA signals, comparisons should be made between these different sets of MTPA signals only if the NMR data were acquired with a sufficient relaxation delay.

Analysis of sphingosine and dihydrosphingosine isomers by HPLC
Various mixtures of sphingosine and dihydrosphingosine isomers 1b–8b were analyzed by HPLC on a reversed-phase (C18) column with acetonitrile and methanol–water and on a normal-phase (silica) column with methyl tert -butyl ether (MTBE)–hexane. Retention times are given in Table 2, and representative separations on the normal-phase system are shown in Figure 4. The HPLC separations gave ca. 13,000 theoretical plates (acetonitrile), 9,000 plates (methanol–water), and 10,000 plates (normal phase). A resolution factor RS of 1.5 (2% valley between equal-sized symmetrical peaks) at tR 15 min would require a peak separation of 0.8 min (acetonitrile), 0.95 min (methanol–water), and 0.9 min (normal phase); RS of 1.0 (27% valley between peaks) would require peak separations of 0.5–0.6 min. By these criteria, peaks corresponding to the separations of most interest (e.g., 1b and 2b; 1b and 3b; 5b and 6b) would be poorly resolved for quantitative work in either reversed-phase system. The normal-phase system would be capable of providing credible quantitation of mixtures of erythro -sphingosine and erythro -dihydrosphingosine enantiomers but not of mixtures of D-erythro and L-threo diastereomers.



View larger version (16K):
[in this window]
[in a new window]
  		Figure 4. Normal-phase HPLC separations of (S)-MTPA derivatives of sphingosine and dihydrosphingosine isomers: A, mixture of D-erythro -sphingosine (1b), L-erythro -sphingosine (2b), and DL-erythro -dihydrosphingosine derivatives (5b, 6b); B, DL-threo -dihydrosphingosine derivatives (7b, 8b); C, DL-threo -sphingosine derivatives (3b, 4b). In panel C, the peak at tR 11.8 and the additional intensity of the peak at tR 8.7 min are attributable to minor components (13b, 14b) in the DL-threo -sphingosine sample. Conditions: 5-µm Alltima silica column (250 mm x 4.6 mm); MTBE-hexane 1:9 at 1 ml/min; UV detection at 210 nm.


 
View this table:
[in this window]
[in a new window]
  		Table 2. HPLC retention times of (S)-MTPA derivatives of D-erythro -sphingosine and related sphingolipid basesa,b

MTPA byproducts and other components that might interfere with analyses
Some samples of MTPA derivatives of sphingolipid bases showed additional 1H and 19F NMR signals attributable to MTPA byproducts (see Figure 2C, Figure 2D, Figure 3E, and Figure 3F). These signals fell outside the 2–MTPA region ({delta}H 3.1–3.4; {delta}F ca. -69) used for measuring the enantiomeric purity of erythro-sphingosine, but their 13C satellites could be distracting (Figure 2C). The extraneous signals did not match resonances of methyl MTPA ({delta}H 3.91, 3.56; {delta}F -72.3) or the carboxylic acid ({delta}H 3.57; {delta}F -71.5), although the latter may show considerable variation in chemical shift. In one case (Figure 2C, Figure 2D), one set of contaminants ({delta}H 3.53; {delta}F -71.4) was removed by extraction from dichloromethane into aqueous NaHCO3 and another set was not ({delta}H 3.42; {delta}F -72.2). These byproducts, which could interfere with some purity analyses, might be avoided by a TLC purification similar to that described for preparing sphingosine derivatives on a nanogram scale (19). Another potential source of interference is the presence of homologs, such as C20-sphingosine.4 These homologs, easily detected by GC methods, would probably not be distinguished from C18-sphingosine in either the NMR or normal-phase HPLC analyses but might cause confusion in reversed-phase analyses described herein and those on chiral columns.


  DISCUSSION
TOP
 ABSTRACT
 INTRODUCTION
 EXPERIMENTAL PROCEDURES AND...
 DISCUSSION
REFERENCES

We have shown that the enantiomeric purity of sphingosine can be measured simply and accurately by 1H or 19F NMR analysis of the MTPA derivative. The tris-MTPA derivative was easy to prepare and furnished excellent reporter groups for NMR analysis. Methoxy signals corresponding to each enantiomer were cleanly resolved in both 250- and 500-MHz 1H NMR spectra of a racemic mixture, and trifluoromethyl signals in 19F NMR spectra at 235 MHz showed even better chemical shift nonequivalence (Figure 2). Adequate 1H and 19F peak separations on a 250-MHz spectrometer indicated that the enantiomeric purity of sphingosine can be measured on virtually any current NMR instrument. The high accuracy and sensitivity of the present method was demonstrated by analysis of several erythro -sphingosine mixtures of known enantiomeric composition. At 500 MHz with an inverse-detection probe, 0.1% of the minor enantiomer could be detected with 1-mg samples. With a 5% detection limit for the minor enantiomer, analyses required as little as 10 µg of sphingosine.

Similar separations were observed for the enantiomers of threo -sphingosine, erythro -dihydrosphingosine, and threo -dihydrosphingosine. As with the erythro -sphingosine MTPA derivatives, the strong intensity and good separation of the 1H methoxy and 19F trifluoromethyl signals led to accurate and sensitive measurement of enantiomeric purities. The signal dispersion was sufficiently large in both 1H and 19F NMR spectra of sphingosine derivative 1b that any stereoisomeric contaminants and any dihydrosphingosine stereoisomers could be detected by simple inspection of the methoxy or trifluoromethyl region (Figure 2 and Figure 3). An obvious application is the analysis of commercial sphingosine samples, some of which we and others (19) have found to contain significant amounts of stereoisomeric and dihydrosphingosine impurities.

Most 1H NMR chemical shifts of the individual MTPA isomers are given in Table 1 to ±0.001 ppm precision (±0.01 ppm for 19F data), which allows reliable identification of individual isomers and any stereoisomeric contaminants. Unlike chromatographic methods, NMR does not require calibration with authentic standards. Chemical shifts matching our data in Table 1 to ±0.001 ppm (or ca. ±0.01 ppm for 19F) can be obtained by acquiring NMR spectra on a reasonably clean sample in dilute CDCl3 solution (<=10 mM) at approximately 25°C (or ~22°C for 19F) and using tetramethylsilane or CFCl3 as a calibration reference. The 1H NMR chemical shifts of the MTPA derivatives of sphingosine isomers showed much better sample-to-sample reproducibility than those of the free base, which may vary by 0.2 ppm (10). Although triacetate derivatives also give reproducible NMR data and can be used to distinguish threo and erythro isomers (10), the MTPA derivatives provided markedly better NMR analyses because of higher sensitivity, the ability to distinguish enantiomers, and better chemical shift nonequivalence between stereoisomers.

The MTPA derivatives of sphingosine and dihydrosphingosine isomers were also analyzed by HPLC (Table 2). Reversed-phase HPLC with either acetonitrile or methanol–water separated some isomers, but little or no resolution was observed for several isomeric pairs, including D- and L-erythro -sphingosine (1b and 2b). Normal-phase HPLC on silica gel gave much better separations, and all of the sphingosine and dihydrosphingosine stereoisomers were well resolved except those of the D-erythro and L-threo pairs. Relative to NMR, HPLC suffered from lower peak dispersion, the need for careful calibration with authentic standards, and much greater potential for gross errors because quantitation is dependent on measurement of a single peak for each component. However, HPLC provides better sensitivity and the capability for high throughput of samples at low cost.

As with other methods of measuring enantiomeric purity, errors can arise in the present technique by discrimination against one of the enantiomers during derivatization, workup, NMR or HPLC analysis, and data reduction. To avoid partial kinetic resolution of enantiomers, all reactions were carried out for 24 h in the presence of a large excess of reagent with the aim of complete formation the tris-MTPA derivatives. No evidence of bis-MTPA derivatives was observed in NMR spectra. Racemization would seem highly improbable in MTPA derivatizations of sphingosine, and epimerization at the allylic 3-hydroxyl was not observed under the mild reaction conditions described herein. The MTPA derivatives were isolated by filtration of the crude reaction mixture through silica gel with a strong solvent to avoid any possible separation of diastereomers. This necessarily crude procedure sometimes left the analyte contaminated with MTPA byproducts. Interference from the resulting extraneous NMR signals and their 13C satellites was negligible in milligram-scale analyses but adversely affected the sensitivity and reliability in microgram-scale work. Another potential source of error is low enantiomeric purity of the MTPA reagent. The (S)-MTPA chloride used herein appeared to be of high enantiomeric purity and no interfering NMR signals from (R)-MTPA species were observed.5 Measurements of T1 relaxation times in 1H and 19F NMR spectra of representative mixtures showed that long relaxation delays to avoid discrimination among diastereomeric species are generally unnecessary. We calculated enantiomeric ratios from integration of NMR peak areas, but ratios calculated from peak heights were only slightly less accurate despite differences in peak widths owing to differences in long-range 19F–1H couplings constants. Peak areas were also used for HPLC analyses. Perhaps the best evidence that enantiomer discrimination does not occur at any stage of analysis is the agreement of observed NMR measurements with known enantiomeric compositions of samples prepared by mixing various amounts of the individual enantiomers.

In summary, we have presented new methods for determining the stereoisomeric purity of any isomer of sphingosine and dihydrosphingosine. The MTPA derivatives are easily prepared and can be analyzed by NMR or HPLC. Both 1H and 19F NMR give results of high reliability and require no authentic standards and only a few minutes of time on instrumentation that is commonly available in research settings. The NMR methods work best with samples of >=100 µg and should be useful for determining the purity of sphingosine from total synthesis. HPLC analysis, despite the considerable effort required for setup, calibration, and validation and its susceptibility to major errors from adventitious impurities, is the only practical alternative for measuring enantiomeric and diastereomeric purities of small biological samples, which may contain only nanogram levels of sphingosine. The NMR and HPLC methods described herein represent major improvements over previous methodology for determining the stereoisomeric purity of sphingosine and dihydrosphingosine. These new methods should be valuable for assuring the purity and authenticity of sphingosine and dihydrosphingosine samples used in biomedical research.


  FOOTNOTES

3 The validity of the method for measuring enantiomeric and diastereomeric purities is not dependent on NMR signal assignments. Assignments were important in elucidating the composition of diastereomeric mixtures and may be helpful in assigning the stereochemistry of other sphingosine derivatives. Back
4 Samples of sphingolipid bases 1a, 3a, and 5a analyzed herein were obtained by hydrolysis of a mixture of cerebrosides and sphingomyelin from cow brain (11). Unlike gangliosides, cerebrosides and sphingomyelin contain low levels of C20 bases, and GC–MS analysis of 1a showed negligible levels (<0.5%) of the C20 homolog. Back
2 Senior author; deceased December 11, 1998. Back
5 The two chiral centers of sphingosine and the three MTPA chiral centers can produce 16 distinct diastereomers. Designating the chirality of the three MTPA groups as 1S *,2S *,3S * and supposing a 1% level of the (R)-MTPA isomer, we can predict that the 2S, 3R, 1R *, 2R *, 3R * isomer (D-erythro -sphingosine derivatized entirely with the (R)-MTPA contaminant) would have an intensity of 0.0001% of that of the normal 2S, 3R, 1S *, 2S *, 3S * isomer. Similarly, the (R)-MTPA contaminant would not contribute significantly to NMR signal intensities of other sphingosine isomers. In this regard, the tris-MTPA derivative of sphingosine is superior to a mono-MTPA derivative for analysis of enantiomeric purity. However, for each tris-MTPA sphingosine stereoisomer, three artifactual diastereomers of 1% intensity would be produced, i.e., 1S *,2S *,3R *; 1S *,2R *,3S *; and 1R *,2S *,3S *, and their enantiomers would be produced at 0.01% intensity. In the analysis of a single sphingosine isomer, 9 artifactual MTPA NMR signals from 3 artifactual diastereomers would appear at a level of 1% and 9 others at a 0.01% level. If all four sphingosine stereoisomers are present in equal amounts, 36 artifactual NMR signals would be produced at a 1% level. This potential problem is best avoided by using MTPA reagent of high enantiomeric purity. Back


  ACKNOWLEDGMENTS

We gratefully acknowledge the support of the National Institutes of Health (HL-49122) and the Robert A. Welch Foundation (C-583).

Manuscript received September 10, 1998; and in revised form December 4, 1998.

Abbreviations: COSY-DQF, 1H-1H correlation spectroscopy with double-quantum filtering; GC, gas chromatography; HMBC, heteronuclear multiple bond correlation; HPLC, high performance liquid chromatography; HSQC, heteronuclear single-quantum coherence; MTBE, methyl tert-butyl ether; MTPA, {alpha}-methoxy-{alpha}-(trifluoromethyl)phenylacetate; NMR, nuclear magnetic resonance (spectroscopy); Pr(hfc)3, tris(3-(heptafluoropropylhydroxymethylene)-(+)-camphorato) praseodymium(III); TLC, thin-layer chromatography


  REFERENCES
TOP
 ABSTRACT
 INTRODUCTION
 EXPERIMENTAL PROCEDURES AND...
 DISCUSSION
 REFERENCES

  1. Bell, R. M., Hannun, Y. A. 1989. Functions of sphingolipids and sphingolipid breakdown products in cellular regulation. Science. 243:500-507[Abstract/Free Full Text].

  2. Merrill, A. H., Jr., and C. C. Sweeley. 1996. Sphingolipids: metabolism and cell signalling. In Biochemistry of Lipids, Lipoproteins, and Membranes. D. E. Vance and J. E. Vance, editors. Elsevier, Amsterdam. 309–339.

  3. Ariga, T., Jarvis, W. D., Yu, R. K. 1998. Role of sphingolipid-mediated cell death in neurodegenerative diseases. J. Lipid Res. 39:1-16[Abstract/Free Full Text].

  4. Riboni, L., Viani, P., Bassi, R., Prinetti, A., Tettamanti, G. 1997. The role of sphingolipids in the process of signal transduction. Prog. Lipid Res. 36:153-195[Medline].

  5. Kolesnick, R., Krönke, M. 1998. Regulation of ceramide production and apoptosis. Annu. Rev. Physiol. 60:643-665[Medline].

  6. Gómez-Muñoz, A. 1998. Modulation of cell signalling by ceramides. Biochim. Biophys. Acta. 1391:92-109[Medline].

  7. Ghosh, S., Strum, J. C., Bell, R. M. 1997. Lipid biochemistry: functions of glycerolipids and sphingolipids in cellular signaling. FASEB J. 11:45-52[Abstract].

  8. Spiegel, S., Milstien, S. 1995. Sphingolipid metabolites: members of a new class of lipid second messengers. J. Membrane Biol. 146:225-237[Medline].

  9. Merrill, A. H., Jr., Schmelz, E-M., Dillehay, D. L., Spiegel, S., Shayman, J. A., Schroeder, J. J., Riley, R. T., Voss, K. A., Wang, E. 1997. Sphingolipids—the enigmatic lipid class: biochemistry, physiology, and pathophysiology. Toxicol. Appl. Pharmacol. 142:208-225[Medline].

  10. Kisic, A., Tsuda, M., Kulmacz, R. J., Wilson, W. K., Schroepfer, G. J., Jr. 1995. Sphingolipid bases. A revisitation of the O-methyl derivatives of sphingosine. Isolation and characterization of diacetate derivatives, with revised 13C nuclear magnetic resonance assignments for D-erythro -sphingosine. J. Lipid Res. 36:787-803[Abstract].

  11. Carter, H. E., Haines, W. J., Ledyard, W. E., Norris, W. P. 1947. Biochemistry of the sphingolipides. Preparation of sphingolipides from beef brain and spinal cord. J. Biol. Chem. 169:77-82[Free Full Text].

  12. Solladié-Cavallo, A., Koessler, J. L. 1994. A four-step diastereoselective synthesis of D-erythro -sphingosine by an enantioselective aldol reaction using a titanium enolate derived from a chiral iminoglycinate. J. Org. Chem. 59:3240-3242.

  13. Hudlicky, T., Nugent, T., Griffith, W. 1994. Chemoenzymatic synthesis of D-erythro - and L-threo -C18-sphingosines. J. Org. Chem. 59:7944-7946.

  14. Herold, P. 1988. Synthesis of D-erythro - and D-threo -sphingosine derivatives from L-serine. Helv. Chim. Acta. 71:354-362.

  15. Polt, R., Peterson, M. A., DeYoung, L. 1992. Aluminoxy acetals from {alpha}-amino esters: chirality transfer via sequential addition of hydride and C-nucleophiles. 2-Amino alcohols and sphingosines. J. Org. Chem. 57:5469-5480.

  16. Hoffman, R. V., Tao, J. 1998. A synthesis of D-erythro - and D-threo -sphingosine and sphinganine diastereomers via the biomimetic precursor 3-ketosphinganine. J. Org. Chem. 63:3979-3985.

  17. Kobayashi, S., Furuta, T. 1998. Use of heterocycles as chiral ligands and auxiliaries in asymmetric syntheses of sphingosine, sphingofungins B and F. Tetrahedron. 54:10275-10294.

  18. Eliel, E. L., and S. H. Wilen. 1994. Stereochemistry of Organic Compounds. Wiley, New York. 214–265, 1071–1080.

  19. Kawamura, A., Berova, N., Dirsch, V., Mangoni, A., Nakanishi, K., Schwartz, G., Bielawska, A., Hannun, Y., Kitagawa, I. 1996. Picomole scale stereochemical analysis of sphingosines and dihydrosphingosines. Bioorg. Med. Chem. 4:1035-1043[Medline].

  20. Wenzel, T. 1987. NMR Shift Reagents. CRC Press, Boca Raton, FL. 127–154.

  21. Parker, D. 1991. NMR determination of enantiomeric purity. Chem. Rev. 91:1441-1457.

  22. Snyder, L. R., J. J. Kirkland, and J. L. Glajch. 1997. HPLC Method Development. Wiley, New York. 537–615.

  23. Masui, M., Shioiri, T. 1998. Stereoselective synthesis of sphinganine by means of modified asymmetric borane reduction. Tetrahedron Lett. 39:5199-5200.

  24. Dale, J. A., Dull, D. L., Mosher, H. S. 1969. {alpha}-Methoxy-{alpha}-trifluoromethylphenylacetic acid, a versatile reagent for the determination of enantiomeric composition of alcohols and amines. J. Org. Chem. 34:2543-2549.

  25. Grob, C. A., Gadient, F. 1957. Die Synthese des Sphingosins und seiner Stereoisomeren. Helv. Chim. Acta. 40:1145-1157.

  26. Grob, C. A., Jenny, E. F., Utzinger, H. 1951. Die Synthese von rac. 1,3-Dioxy-2-amino-octadecan. Helv. Chim. Acta. 34:2249-2254.

  27. Carter, H. E., Humiston, C. G. 1951. Biochemistry of the sphingolipides V. The structure of sphingine. J. Biol. Chem. 191:727-733[Free Full Text].


Add to CiteULike CiteULike   Add to Complore Complore   Add to Connotea Connotea   Add to Del.icio.us Del.icio.us   Add to Digg Digg   Add to Reddit Reddit   Add to Technorati Technorati    What's this?



This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF)
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Right arrow Citation Map
Services
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Li, S.
Right arrow Articles by Schroepfer, G. J.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Li, S.
Right arrow Articles by Schroepfer, G. J., , Jr.
Social Bookmarking
Add to CiteULike   Add to Complore   Add to Connotea   Add to Del.icio.us   Add to Digg   Add to Reddit   Add to Technorati  
What's this?

HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
 All ASBMB Journals   Journal of Biological Chemistry 
 Molecular and Cellular Proteomics   ASBMB Today