| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Articles |
1
Metabolic Screening Laboratory and
2
Pharmaceutical Analysis Laboratory, King Faisal Specialist Hospital and Research Centre, MBC-03, PO Box 3354, Riyadh 11211, Saudi Arabia.
aAuthor for correspondence. Fax 966-1-442-4546; e-mail rashed{at}kfshrc.edu.sa.
 |
Abstract |
|---|
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Methods: We used a narrow bore chiral macrocyclic glycopeptide teicoplanin column for the enantioseparation of D-pipecolic acid (D-PA) and L-PA and interfaced the column directly to the electrospray source of a tandem mass spectrometer. We used phenylalanine-d5 as internal standard added to 50 µL of plasma followed by deproteinization, evaporation, and injection. The analysis was performed in the selected-reaction monitoring mode using two transitions: m/z 130
m/z 84 for PA, and m/z 171m/z 125 for phenylalanine-d5. L-PA eluted at 7 min, and D-PA eluted at 11.7 min, whereas phenylalanine-d5 eluted at 6 min. The turnaround time for the assay was 20 min.
Results: Linear calibration curves were obtained in the range of 0.5–80 µmol/L. At a plasma concentration of 1.0 µmol/L, the signal-to-noise ratio was 50:1. The intra- and interassay variations were 3.1–7.9% and 5.7–13%, respectively, at concentrations of 1–50 µmol/L. Mean recoveries of L-PA added to plasma were 95% (5 µmol/L) and 102% (50 µmol/L). The method clearly distinguished between healthy individuals and peroxisomal disease patients.
Conclusions: The novel LC-MS/MS method is simple, rapid, and stereoselective, and uses only 50 µL of plasma, no derivatizing reagents, and a commercially available internal standard. Sample preparation is not complex and is faster than for all other methods.
|
Introduction |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
-aminoadipic acid (3)(4). On the other hand, it is believed that D-pipecolic acid (D-PA) originates from the metabolism of intestinal bacteria and from dietary sources (5)(6). It was found that both D-PA and L-PA were moderately increased in patients with liver cirrhosis and in patients with chronic hepatic encephalopathy. Although L-PA remained the predominantly circulating form, D-PA was proportionally higher in liver disease patients than in healthy individuals (6). From a diagnostic standpoint, the measurement of PA is considered as a supplementary test for peroxisomal diseases following the analysis of plasma very long chain fatty acids (VLCFAs), bile acids, phytanic acid, and pristanic acid. It is also useful in patients with normal or equivocally increased VLCFA concentrations who clinically are strongly suspected of having a generalized peroxisomal disorder. A plasma PA within the reference interval in Zellweger-like cases with abnormal VLCFA concentrations is strong evidence for a single enzyme defect, such as acyl-CoA oxidase deficiency or bifunctional enzyme deficiency (7). The determination of plasma or urine PA is also becoming increasingly more important because of several recent reports of isolated hyperpipecolic acidemia cases, which suggest that hyperpipecolic acidemia should be classified as a single peroxisomal enzyme deficiency (8)(9)(10).
Initially, total PA (D-PA and L-PA) was determined by ion-exchange chromatography on an amino acid analyzer or by HPLC using the acid ninhydrin method (11). However, this method suffers form severe limitations in sensitivity, specificity, and selectivity. Several more-sensitive single-ion monitoring gas chromatography-mass spectrometry (GC-MS) methods have been published for the determination of total PA in plasma and urine (7)(12). These usually involve solid-phase extraction followed by derivatization and analysis on an achiral GC column. An extremely sensitive method has been developed based on in situ alkaline derivatization of PA to the corresponding N-methylcarbamate followed by solvent extraction and further derivatization of the carboxyl moiety to the pentafluorobenzyl ester. The derivative is then analyzed by electron-capture negative-ion GC-MS or GC with electron-capture detection (13)(14). In addition to the determination of PA in plasma and urine, this method allowed analysis of PA in cerebrospinal fluid, where the concentrations are extremely low.
As far as the stereochemistry of PA is concerned, there are only two reported methods where the PA from biological samples was separated and determined as individual enantiomers. Struys and Jakobs (15) recently published a method for the enantiomeric analysis of D- and L-PA in plasma using a CP Chirasil-Dex CB chiral column and electron-capture negative-ion GC-MS. Plasma samples were first analyzed for total PA by GC with electron-capture detection as mentioned above, and when total PA was increased, the chiral GC-MS procedure was carried out on the same sample.
To our knowledge, the only published studies of the stereochemistry of PA in patients with peroxisomal diseases vs healthy individuals were those of Armstrong and coworkers (16)(17). Their method involved the in situ derivatization of plasma or urine PA to the corresponding carbamate by use of fluorenylmethylchloroformate under basic conditions, followed by acidification and solid-phase extraction. The extracts were analyzed by a double column-switching method using two reversed-phase columns followed by chiral HPLC on a cyclodextrin Cyclobond-I column with fluorescence detection.
In this study, we present a novel and direct chiral liquid chromatography-tandem mass spectrometry (chiral LC-MS/MS) method for the analysis of L-PA in plasma; the method involves no derivatization, simple sample preparation, and a relatively short analysis time.
|
Materials and Methods |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
sample preparation
Samples were prepared by placing 50 µL of patient or control plasma in a microcentrifuge tube (1.5-mL capacity) and adding 10 µL of phenylalanine-d5 solution (2 nmol) as internal standard. Proteins were precipitated by the addition of 200 µL of cold acidified acetonitrile (formic acid-trifluoroacetic acid-acetonitrile; 0.4:0.2:100 by volume). The tubes were then vortex-mixed vigorously for 30 s and centrifuged at 9300g for another 10 min. The clear supernatant was transferred to new microcentrifuge tubes and evaporated to dryness with nitrogen at 35 °C. The residues were stored at 4 °C until analysis. The residues were then reconstituted in 200 µL of methanol-water (60:40 by volume), and 100 µL was transferred to the wells of a 96-well vinyl U-bottomed microplate (200-µL capacity; Costar). The microplate was covered with aluminum foil to prevent solvent evaporation. Of these extracts, 5-µL volumes were injected into the mass spectrometer.
methods
MS/MS and LC-MS/MS.
MS/MS and LC-MS/MS analyses were carried out on a Micromass QuattroLC bench-top triple quadrupole instrument (Micromass) interfaced with a Z-spray electrospray ionization (ESI) source. A Hewlett Packard solvent delivery system equipped with a Model HP 1100 degasser and a Model HP1100 binary pump was used to deliver the mobile phase. A Gilson Model 232XL fitted with a Gilson Model 402 syringe pump and Rheodyne injector Model 7010 was used as an autosampler to inject samples from 96-well microplates. Both the autosampler and the pump were controlled by the Masslynx software of the mass spectrometer. The ESI source was operated in the positive-ion mode at a capillary voltage of 3.0 kV, a cone voltage of 22 V, and extraction voltage of 4 V. Nitrogen was used as the nebulizing gas at a flow rate of 100 L/h and as the desolvation gas at a flow rate of 300 L/h. Argon was used as the collision gas at a pressure of 0.13 Pa. The collision energy was optimized at 14 eV for PA. The ion source temperature was maintained at 125 °C, and the desolvation temperature was optimized to 250 °C. For development work, the product-ion spectra for DL-PA, D-PA, L-PA, and phenylalanine-d5 were acquired in the continuous flow injection mode, using a Harvard Model 22 syringe pump connected directly to the ion source via peak tubing. We used a 0.1 mmol/L solution of DL-PA for signal optimization infused at a rate of 5 µL/min. The resolution of both MS1 and MS2 was maintained at a level that yielded a 10% valley between adjacent ions and 0.75 atomic mass units (amu) peak widths at half height.
The Masslynx software (Ver. 3.4; Micromass) provided with the instrument running under the Windows NT environment was used for instrument control, data acquisition, peak smoothing, integration of peak areas, and signal-to-noise determinations.
Chromatography.
Chiral chromatography was carried out on a Chirobiotic T column [250 x 2.0 mm (i.d.); 5-µm silica gel bonded to the macrocyclic glycopeptide teicoplanin; Advanced Separation Technologies] in the reversed-phase mode at a flow rate of 200 µL/min. The mobile phase consisted of methanol-deionized water (60:40 by volume). The effluent from the injector was passed through an in-line filter (Model A-315; Upchurch Scientific) that was then connected to the column via peek tubing. The effluent from the column was passed directly to the ESI source. The mass spectrometer settings described above were used for chromatography except that the flow of nitrogen was increased to 120 L/h for nebulization and to 400 L/h for desolvation. The source temperature was also increased to 150 °C, and the desolvation temperature was increased to 350 °C. The dwell time was 0.25 s for both PA and phenylalanine-d5.
For selected-reaction monitoring (SRM), the column was interfaced to the ESI source and the signal intensity vs flow rate was optimized using the HP1100 pump as indicated above. The SRM transitions were [M-H]+ at m/z 130 to product ion at m/z 84 [M-H+ - HCOOH] for PA, and at m/z 171 to product ion at m/z 125 [M-H+ - HCOOH] for phenylalanine-d5. The retention time for L-PA was 7 min, and that for D-PA was 11.7 min, whereas the retention time of phenylalanine-d5 was 6 min. The autosampler was programmed to inject a sample every 20 min to allow for the elution of a broad late peak that appeared in the same transition used for PA.
Neutral loss scanning and product-ion scanning were used with the same cone voltage and collision energy to investigate other peaks that appeared in the PA m/z 130m/z 84 transition. Solutions containing 0.1 mmol/L glutamine, lysine, glutamic acid, or pyroglutamic acid were co-injected with L-PA-enriched extracts of pooled plasma samples under the same LC-MS/MS conditions.
Calibration and analysis of biological samples.
The enantiomeric purity of the commercial preparation of L-PA was determined by repeated injections of 1 µmol/L solutions of D-PA and of L-PA (n = 10) and measurement of the peak areas for both compounds. The averaged data indicated that the D-PA preparation contained 4.13% L-PA and that the L-PA preparation contained 4.52% D-PA. This latter factor was used to correct the concentration of L-PA in the stock solutions used for calibration. L-PA stock solutions for enrichment experiments were prepared in deionized water at concentrations of 0.001–1 mmol/L and stored at 4 °C. These solutions were used for preparation of calibration solutions and for recovery experiments. Calibration curves were prepared by the addition of 0, 0.025, 0.075, 0.25, 0.5, 2, and 4 nmol of L-PA to 50 µL of control pooled plasma (0.5–80 µmol/L), followed by sample preparation as described above. A 5-µL volume of each calibrator was injected into the mass spectrometer. A control plasma extract with no additions was injected before and after each calibration curve. The carryover between injections was in the range of 0.6–1.3%. The column was washed between batches with deionized water, followed by mobile phase and 100% methanol, and reequilibrated for a minimum of 2 h with mobile phase.
The contribution of endogenous L-PA in the plasma used for calibration was eliminated by subtracting the peak-area ratio of L-PA to phenylalanine-d5 for the first point in the curve from all the subsequent points. The calibration curves (y = mx + b) were then constructed based on the subtracted peak-area ratios of L-PA to that of the internal standard (y) vs the concentration of added L-PA (x), using Microsoft Excel 2000.
Validation.
For recovery and precision experiments, pooled plasma samples were used; these were divided into three portions, of which two were enriched with L-PA at 5 and 50 µmol/L. Each of these was analyzed six times on the same day to determine the intraday recovery and precision. Additional batches were analyzed on six separate occasions to determine the interday values. In these experiments, L-PA concentrations in control and patient samples were calculated from the resulting peak-area ratios and the regression equation of the calibration curve. We used Microsoft Excel 2000 to carry out the two-tailed t-test of equality for the means of two independent samples with unequal variances to determine the statistical significance of the values obtained for control vs abnormal patient samples.
|
Results |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
. This product ion spectrum was obtained by infusing a 0.1 mmol/L solution of the D- and L-PA mixture. Fig. 1B
shows the product ion spectrum of m/z 171 for the protonated molecular ion of phenylalanine-d5, which was used as internal standard. Under relatively mild CID conditions, both compounds break down to give essentially one major fragment corresponding the loss of the elements of formic acid [M-H+ - HCOOH]. Thus, PA gives a major fragment at m/z 84, and phenylalanine-d5 gives a major fragment at m/z 125. When the individual enantiomers of PA were analyzed under the same conditions, no differences were observed in the CID spectra or in signal intensity (data not shown).
|
chiral lc-ms/ms
The flow injection MS/MS results described above were used to design SRM experiments using the teicoplanin column for both L-PA and phenylalanine-d5 where the mass spectrometer was programmed to cycle between two transitions, m/z 130m/z 84 for PA and m/z 171m/z 125 for phenylalanine-d5. Fig. 2A
shows an extracted SRM chromatogram for the racemic PA mixture, and Fig. 2B
shows an extracted SRM chromatogram for the commercial L-PA. We used methanol-water (60:40 by volume) with no additives as mobile phase. We investigated the addition of acids such as formic acid or trifluoroacetic acid to the mobile phase to promote positive-ion detection. No significant enhancement in signal was observed. The addition of acids such as formic acid (0.25 mL/L) or trifluoroacetic acid (0.1 mL/L) to the mobile phase did improve peak shape and reduced tailing. However, it had a negative impact on the chromatography and led to the appearance of minor extraneous interfering peaks that were otherwise not observed. To overcome slight drifts in retention times, we attempted to use ammonium acetate or triethylamine buffers (pH 4.0) at different concentrations (0.001–0.1 mol/L). However, both had negative effects on the overall signal intensity and on the chromatography. We also evaluated two teicoplanin columns to ensure the reproducibility of the separation and observed essentially no difference.
|
interfering plasma metabolites
The SRM profile in Fig. 3A
shows the m/z 171m/z 125 transition for the phenylalanine-d5 internal standard in a plasma extract. It shows a single peak at 
6 min with no other interfering peaks. On the other hand, the profiles for the m/z 130m/z 84 transition for PA show several peaks other than D- and L-PA at different elution times. Fig. 3, B and C
, shows profiles from plasma extracts for a control and a patient with CHRS, respectively. These plasma peaks, although not interfering with the quantification of L-PA, were investigated by analyzing the plasma extracts by neutral loss scanning for 46 Da (loss of HCOOH), 63 Da (loss of NH3 and HCOOH), and 64 Da (loss of H2O and HCOOH). The early eluting peaks at 3.5 min gave a protonated molecular ion at m/z 148, which corresponded to glutamic acid and cochromatographed with pure L-glutamic acid. The peak at 5 min gave a protonated molecular ion at m/z 147, which corresponded to glutamine and lysine. Protonated glutamine and lysine are isobaric at m/z 147, and both lose ammonia to yield a positive ion at m/z 130, which then undergoes the same transition as PA with the loss of the elements of formic acid to give the ion at m/z 84 (data not shown). However, the peak at 5 min cochromatographed with L-glutamine rather than lysine. Lysine did not elute off the column under the conditions described. The peak at 8 min and a broad late peak eluting at 18.5 min were not identified. Pyroglutamic acid is isobaric with PA and under the same CID conditions yields a major fragment at m/z 84; thus it gives the same m/z 130m/z 84 transition as does PA when analyzed by flow injection MS/MS (data not shown). However, L-pyroglutamic eluted as a somewhat broad peak at <3 min.
|
quantification of plasma l-pa
We obtained linear calibration curves, using phenylalanine-d5 as internal standard, for the determination of plasma L-PA in the range of 0.5–80 µmol/L. The mean (± SD) slope, intercept, and correlation coefficient (r2) were 1.076 (± 0.20), -0.0054 nmol (± 0.01 nmol), and 0.9997 (range, 0.9993–1.0), respectively. A control plasma sample with a measured concentration of 1.0 µmol/L gave a signal-to-noise ratio of 50:1 (5 pmol injected on column). Above 80 µmol/L, the calibration curves tended to show signs of signal saturation with relatively small deviation from linearity.
The intra- and interday accuracy and precision of the method were determined for the controls and two known concentrations (5 and 50 µmol/L) from six experiments run on 6 different days. Accuracy was determined by calculating the mean recovery of the measured concentrations as a percentage of the nominal concentration after subtracting the endogenous L-PA concentration. Precision was determined by calculating the CV of the mean recoveries. The results are summarized in Table 1
and indicate quantitative recovery and good precision for L-PA. The concentrations of L-PA in the control and patient samples are shown in Table 2
. Control values for L-PA were statistically different from those for patients with peroxisomal diseases (P <0.0001). Samples from patients with peroxisomal disease were selected based on previously determined VLCFA values as determined by GC-MS and literature reference intervals (see Table 2
) (18)(19). Samples that gave abnormal L-PA results >80 µmol/L were diluted with control plasma and repeated in the following batch of samples. Under the conditions described, D-PA appeared in some samples, but only as a trace amount.
|
|
It is noteworthy to mention that for the precipitation of proteins, the acidified acetonitrile was far superior to similarly acidified methanol, which gave cloudy extracts on reconstitution in mobile phase. The acetonitrile extracts on reconstitution gave clear supernatants, although they occasionally showed some sedimentation; the sediment, however, was easy to avoid during sample transfer to the microplate for injection. The presence of acids in the precipitating solvent was also necessary for achieving optimum sensitivity for PA.
|
Discussion |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Berthod et al. (25) carried out an extensive study of the use of the macrocyclic antibiotic teicoplanin to resolve a large number of underivatized native amino acids, including PA, using ultraviolet detection. Their work was extremely valuable for the development of the method described here. We essentially used the same mobile phase composition, but we used a narrow bore column, which allowed direct interfacing with the ESI source of the mass spectrometer. The use of MS/MS as the highly specific detection system allowed the direct analysis of plasma L-PA on a single chiral column and substantially simplified sample preparation.
The choice of phenylalanine-d5 as internal standard was made based on its successful use in the LC-MS/MS quantification of several amino acids in plasma (26) and because it showed no interference in the PA transition or in its own SRM transition, as determined by injection of control plasma extracts to which no internal standard was added.
The LC-MS/MS analysis of plasma samples in the SRM mode showed several additional peaks in the transition used for PA. Three of these were identified as pyroglutamic acid, glutamic acid, and glutamine. Fortunately, these three amino acids eluted at different retention times, and thus caused no interference. The basic amino acid lysine, which was found to give the same transition by flow injection MS/MS analysis, was not seen, which agreed with the report by Berthod et al. (25).
Our mean value of 0.96 ± 0.28 µmol/L for L-PA in healthy controls (age >1 week) agrees well with those reported by Armstrong et al. (17), who reported a range for total PA of 0.6–1.3 µmol/L with D-PA constituting <2% of the total PA (age >1 week). Our results also agree with the work by Fujita et al. (6), who reported mean values of 1.20 ± 0.11 µmol/L for total PA and 0.19 ± 0.04 µmol/L for D-PA, and with the data provided by Kok et al. (13), who reported a mean value for total PA of 1.46 µmol/L (range, 0.7–2.46; age >1 week). Plasma samples from a control pediatric population (age <1 week) were not available to us because most samples for these patients are sent to us as dried blood spots. However, the values for L-PA in this particular pediatric population should be studied to determine cutoff values from early-onset peroxisomal disease patients. The abnormal values for L-PA obtained in patients with peroxisomal abnormalities were statistically different from the values for controls (P <0.0001). The clinical features for some of these patients have been already described (27).
The work presented here reflects our interest in the clinical and pharmaceutical applications of chiral chromatography, particularly in combination with MS/MS. Only one study has addressed the issue of the stereochemistry of plasma PA from a small number of peroxisomal deficiency patients (n = 6) (17). We therefore sought to develop a method that may further support previous work with a more specific method and also eliminate the need for the two-pronged approach to sample analysis (15). It became readily apparent from analyzing both control (n = 27) and patient samples (n = 23) that plasma D-PA is indeed very low, appearing as only a trace amount in the chromatograms, and that the we cannot attain enough sensitivity for its determination.
In conclusion, we have developed and validated a novel, simple, rapid, and stereoselective method for the determination of L-PA in plasma. The method requires a small amount of plasma, no derivatizing reagents, and a commercially available internal standard. Sample preparation is not complex and is faster than other methods. The combination of chiral chromatography and MS/MS allowed direct resolution of the enantiomers of PA without the need for any extensive chromatographic separation. The use of MS/MS offered both the specificity and sensitivity necessary for the quantification of L-PA in control and patient plasmas. Although MS/MS remains an expensive technique, many biochemical genetics and screening laboratories currently have this capability, and the very high throughput for routine screening assays, such as the analysis of blood spots for amino acids and acylcarnitines, allows substantial spare time on such instruments that can be used for additional diagnostic work.
|
Acknowledgments |
|---|
|
Footnotes |
|---|
|
References |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
-aminoadipate 
-semialdehyde oxidoreductase. Biochim Biophys 1990;1038:300-305.
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
