The potential of capillary electrophoresis (CE) with ultraviolet (UV)-excited fluorescence detection for sensitive chiral analysis of amino acids (AAs) was investigated. DL-AAs were derivatized with 9-fluorenylmethoxycarbonyl chloride (FMOC)-Cl to allow their fluorescence detection and enhance enantioseparation. Fluorescence detection was achieved employing optical fibers, leading UVexcitation light (< 300 nm) from a Xe-Hg lamp to the capillary window, and fluorescence emission to a spectrograph equipped with a charge-coupled device (CCD). Signal averaging over time and emission wavelength intervals was carried out to improve the signal-to-noise ratio of the FMOC-AAs. A background electrolyte (BGE) of 40 mM sodium tetraborate (pH 9.5), containing 15% isopropanol (v/v), 30 mM sodium dodecyl sulfate (SDS), and 30 mM β-cyclodextrin (β-CD), was found optimal for AA chemo- and enantioseparation. Enantioresolutions of 1.0 or higher were achieved for 16 proteinogenic DL-AAs. Limits of detection (LODs) were in the 10–100-nM range (injected concentration) for the D-AA enantiomers, except for FMOC-D- tryptophan (536 nM) which showed intramolecular fluorescence quenching. Linearity (R > 0.997) and repeatability for peak height (relative standard deviations (RSDs) < 7.0%; n = 5) and electrophoretic mobility (RSDs < 0.6%; n =5)ofindividualAA enantiomers were established for chiral analysis of DL-AA mixtures. The applicability of the method was investigated by the analysis of cerebrospinal fluid (CSF). Next to L-AAs, endogenous levels of D-glutamine and D-aspartic acid could be measured in CSF revealing enantiomeric ratios of 0.35 and 19.6%, respectively. This indicates the method’s potential for the analysis of low concentrations of D-AAs in presence of abundant L-AAs. . . . . . Keywords Amino acids Chiral separation Capillary electrophoresis FMOC derivatization Fluorescence detection Cerebrospinal fluid Introduction in their L-form. The presence of free D-AAs in the brain of humans and other mammals has been known for many years Amino acids (AAs) play a major role in the physiology of but for long time was thought to originate from bacteria or organisms being building units of proteins but also essential formed by spontaneous racemization of L-AAs . More re- in, e.g., metabolic processes, neurotransmission, and lipid cently, endogenous racemases were shown to be involved in transport. AAs are precursors for the synthesis of hormones D-AA synthesis [5, 6]. Some D-AAs were found in relatively and low-molecular-weight nitrogenous substances [1, 2]. large quantities during the embryonic phase of brain develop- Most AAs in human life are chiral and predominantly occur ment , localized to protoplasmic astrocytes, or closely dis- tributed to NMDA receptors [8, 9], which suggests their func- tion in signaling pathways, neurotransmission, and in brain * Govert W. Somsen biology . D-AAs were found to have essential roles and email@example.com abnormal D-AA levels and AA enantiomeric ratios in cerebro- spinal fluid (CSF) may relate to neurodegeneration. Indeed, D- Division of BioAnalytical Chemistry, Amsterdam Institute for AAs were found to be involved in the pathogenesis of psychi- Molecules, Medicines and Systems, Vrije Universiteit Amsterdam, de Boelelaan 1085, 1081 HVAmsterdam, The Netherlands atric diseases and abnormal levels were associated with hu- man disorders, such as schizophrenia and Alzheimer’sdisease Biomolecular Analysis, Utrecht University, Universiteitsweg 99, 3584 CG Utrecht, The Netherlands [11–15]. For example, Samakashvili et al. showed that Prior A. et al. chiral analysis of AAs in CSF might be useful for the early excitation in the UV region, which is not provided by conven- diagnosis and understanding of metabolism processes related tional CE-Flu systems employing lasers with output in the to neurodegeneration and Alzheimer’s disease. Clearly, anal- visible region. Chan et al.  used a home-built laser-in- ysis of D-AAs and determination of the D/L ratios can be of duced fluorescence (LIF) detector for the CE analysis of importance in clinical and pharmaceutical science but also in FMOC-AAs but did not perform chiral separations. In CE, environmental and food analysis [17–20]. AA enantiomers of UV-excited fluorescence detection of other compounds (e.g., interest often are minor components of multi-component com- vitamins, drugs, phenolic compounds, and proteins) has been plex samples, such as tissues and biological fluids, requiring achieved using a laser source (mostly 266 nm) but also lamp- both chemo- and enantioselective separation with sensitive based excitation using a Xe-Hg light source (for excitation in detection in order to allow their unambiguous assessment. the range of 220–400 nm) has been reported using a dedicated Chiral analysis of AAs can be performed by gas chroma- detection cell [50–55]. The latter showed up to 160-fold sen- tography  and high-performance liquid chromatography sitivity improvement as compared to UV absorbance detec- [22, 23], requiring costly chiral stationary phases and/or rela- tion, whereas sensitivity was comparable as obtained with tively large amounts of (chiral) derivatization agents [24–27]. LIF detection, while providing much higher flexibility in se- In addition, upon repeated analysis of biological samples, life- lection of UV excitation wavelengths. times of chiral columns may be limited. Capillary electropho- In the present paper, we studied chiral CE-Flu of FMOC- resis (CE) has shown to be a versatile alternative tool for the derivatized AAs employing lamp-based UVexcitation. Using enantioselective analysis of AAs requiring minute sample vol- a CE-dedicated fluorescence setup [50–55], the light from a umes and small amounts of solvents and chiral selector mol- Xe-Hg excitation lamp is led to the separation capillary by an ecules. These selectors, such as (derivatized) cyclodextrins optical fiber and focused onto the detection window. The an- (CDs) [28, 29], vancomycin [30, 31], and 18-crown-6- alyte emission light is partly trapped within the fused silica tetracarboxylic acid , can be simply added to the back- capillary and guided along the capillary by total internal re- ground electrolyte (BGE). A limitation of CE is the relatively flection. The fluorescence light is coupled out of the capillary low concentration sensitivity obtained with common ultravi- by an optical cone and directed via a liquid light guide towards olet (UV) absorbance detection, due to the small optical path the detector, which was comprised of a spectrograph with a length provided by the capillary internal diameter. Besides, charge-coupled device (CCD) detector. Fluorescence excita- only the aromatic AAs tryptophan, tyrosine, and phenylala- tion and emission parameters were studied in order to nine show native UV absorbance. Therefore, AAs are often achieve optimal sensitivity of FMOC-AAs. Separation labeled with UV or visible light-absorbing agents and subse- conditions were investigated and optimized for chiral and quently analyzed by CE with UVor fluorescence (Flu) detec- chemical resolution. Analytical aspects of the CE-Flu tion [33–36]. The use of derivatization agents will not only method, such as repeatability, linearity, and detection improve detectability and sensitivity but also detection selec- limits, were evaluated. The method’s applicability was tivity as only analytes with specific reactive groups will be studied by the enantioselective analysis of AAs in CSF. derivatized and thus detected. Furthermore, derivatization of AAs may also enhance enantioseparation . Chiral CE analysis of AAs with fluorescence detection has predominantly been done employing fluorescein isothiocya- Materials and methods nate (FITC) [16, 33, 34, 38–40] but also, e.g., naphthalene- 2,3-dicarboxyaldehyde (NDA) , 4-fluoro-7-nitro-2,1,3- Chemicals benzoxadiazole (NBD-F) , 5-(4,6-dichloro-s-triazin-2- ylamino) fluorescein (DTAF) , and 5-carboxyfluorescein All reagents were of analytical grade. FMOC-Cl, β-CD, pen- succinimidyl ester (CFSE)  have been used for derivatiza- tane, sodium tetraborate, sodium hydroxide, glycine, D- tion. These reagents, however, require relatively long deriva- glutamic acid, D-histidine, D-threonine, L-alanine, L-arginine, tization times (30 min to several hours). More rapid derivati- L-asparagine, L-aspartic acid, L-cysteine, L-glutamic acid, L- zation (few minutes) can be achieved with 9- glutamine, L-histidine, L-isoleucine, L-leucine, L-lysine, L-me- fluorenylmethoxycarbonyl chloride (FMOC-Cl). Under alka- thionine, L-proline, L-serine, L-threonine, L-tryptophan, L-ty- line conditions, FMOC reacts with primary and secondary rosine and L-valine, DL-alanine, DL-arginine, DL-asparagine, amines and allows fast derivatization of all proteinogenic DL-aspartic acid, DL-cysteine, DL-glutamic acid, DL-histidine, AAs . Chiral CE-UVof FMOC-derivatized AAs has been DL-isoleucine, DL-leucine, DL-lysine, DL-phenylalanine, DL- described [45–48]. Although FMOC is fluorescent, and thus proline, DL-serine, DL-tryptophan, and DL-valine were from would allow more sensitive fluorescence detection, chiral CE- Sigma-Aldrich (Steinheim, Germany). Isopropanol, DL-methi- Flu of FMOC-derivatized AAs has not been reported so far. onine, DL-tyrosine, sodium dodecyl sulfate, and acetonitrile This is most probably due to the fact that FMOC requires were supplied by Fluka (Steinheim, Germany). Water was Chiral capillary electrophoresis with UV-excited fluorescence detection for the enantioselective analysis... deionized and purified with a Milli-Q purification system detector cell towards an external outlet vial containing the (Millipore, Belford, NJ, USA). grounding electrode. The external detector adapter guarantees The optimal BGE was 40 mM sodium tetraborate (adjusted undisturbed capillary cooling by facilitating the CE instru- to pH 9.5 with 1 M sodium hydroxide) containing 15% (v/v) ment’s liquid cooling. The emission light optical guide was isopropanol, 30 mM sodium dodecyl sulfate (SDS), and connected to the spectrograph via a home-made fiber holder 30 mM β-CD. The BGE was filtered prior to use through equipped with a back illuminated CCD chip of 256 × 1024 0.45-μm pore size disposable nylon filters from VWR pixels with a pixel size of 26 μm (Andor Technologies). (Amsterdam, The Netherlands). Stock solutions (3 mM) of The spectrograph comprised a grating of 600 lines/mm blazed AAs were prepared in 0.2 M sodium tetraborate (pH 9.5). at 300 nm and a band-pass of 263 nm. The CCD chip was cooled down to − 60 °C. The spectrograph was wavelength- Derivatization calibrated daily using the reference spectral lines of an Hg pen-ray light source (L.O.T.-Oriel, Darmstadt, Germany). The pH of the CSF samples was adjusted by adding 10 μLof Typical detection settings used in CE-Flu experiments were 2 M sodium hydroxide to 990 μLCSF. slit width, 50 μm; exposure time, 3 s; vertical shift speed, Derivatization of AAs with FMOC was carried out as de- 16.25 μs; and horizontal read-out rate, 33 kHz. Acquired spec- scribed earlier . Briefly, 500 μL of 10 mM FMOC in tra were collected using the Full Vertical Binning mode and acetonitrile was added to 500 μLsample (i.e., ≤3mM AA were background corrected. A 300-nm short-pass interference in 0.2 M sodium tetraborate buffer (pH 9.5) or pH-adjusted filter (Asahi Spectra USA Inc., Torrance, CA, USA) was used CSF). This mixture was kept at room temperature for 2 min to select excitation light. Other tested excitation filters were a and then extracted with 1.5 mL pentane to remove excess of 260 (± 10)-nm band-pass filter (Asahi Spectra) and a 240– FMOC reagent. The aqueous phase was diluted ten times with 400-nm broad-pass filter (Flux Instruments). Data acquisition water. The resulting solution was kept at 4 °C until injection. analysis was performed using the software program Andor Solis (Andor Technologies). CE-Flu system Fluorescence spectra CE experiments were carried out with a P/ACE MDQ CE instrument (Beckman Coulter, Brea, CA, USA). CE of AAs Reference excitation and emission spectra of FMOC-AAs was performed using bare-fused silica capillaries (Polymicro (10 μM in water) were recorded using an LS 50B fluorescence Technologies, Phoenix, AZ, USA). The capillaries had an i.d. spectrometer (PerkinElmer, Waltham, MA) at room tempera- of 75 μm, an o.d. of 375 μm, and total/effective lengths of ture using excitation and emission slit widths of 15 and 4 nm, 72.2/55.3 cm. The capillary temperature was set to 22 °C. respectively, and a scan rate of 3 nm/s. New bare-fused silica capillaries were rinsed with 1 M sodium hydroxide for 10 min at 30 psi and deionized water for 10 min Statement of human and animal rights at 30 psi. Between CE analyses, the capillaries were rinsed with BGE for 5 min at 30 psi. Overnight, the capillaries were No human or animal subjects were used in this study. stored in deionized water. Separations were performed in nor- mal polarity mode with a separation voltage of 25 kV. Sample injection was performed hydrodynamically by applying Results and discussion 0.5 psi for 13 s, which corresponds to an injected volume of about 0.8% of the total capillary volume (BGE viscosity rel- Fluorescence detection of FMOC-AAs atively to water = 1.93). Data acquisition was performed using 32 Karat software (Beckman Coulter). Based on previous studies [29, 48, 58], a BGE of 40 mM A previously described wavelength-resolved fluorescence sodium tetraborate (pH 9.5) with 15% (v/v) isopropanol and (wrFlu) detector for CE was used, which was based on an 30 mM β-CD was selected as a starting condition to investi- Argos 250B fluorescence detection cell (Flux Instruments, gate the fluorescence detection of AA enantiomers. A test Basel, Switzerland)  combined with a SR-163 spectro- mixture of the DL-AAs alanine, aspartic acid, glutamic acid, graph equipped with a CCD camera (Andor Technologies, leucine, methionine, and tryptophan was derivatized with Darmstadt, Germany) [56, 57]. The Argos system comprises FMOC-Cl. These AAs represent diverse chemical properties a Xe-Hg lamp for excitation, excitation, and emission optical and exhibit different overall charge after derivatization. A guides and filters and an optical cone detection cell. A capil- bare-fused silica capillary with an i.d. of 75 μm was used in lary cartridge with an external detector adapter (Beckman order to maximize the optical path length for excitation, with- Coulter) was used to guide the CE capillary from the inlet vial out inducing excessive CE current. Preliminary CE-UVexper- out of the CE instrument through the Argos fluorescence iments showed that under these conditions all test AAs were Prior A. et al. separated with an enantioresolutions ranging from 0.7 for al- indeed was achieved; however, it also significantly attenuated anine to 6.4 for glutamic acid. the overall excitation light intensity. Best S/N was obtained by The excitation and emission spectrum of FMOC-DL-phe- using only the 300-nm cutoff short-pass filter for excitation. nylalanine was recorded using a conventional spectrofluorom- Although noise levels significantly increased, absolute signal eter (Fig. 1A and B), clearly indicating that FMOC-AAs re- intensities were nine times higher as obtained with the 260-nm quire excitation in the deep UV region for efficient fluores- band-pass filter, leading to most favorable detection of the cence measurement. In order to achieve appropriate UV exci- FMOC-AA fluorescence. tation, a CE-dedicated fluorescence detector equipped with a The wrFlu detection provides the collection of a series of Xe-Hg source was employed . Previously, this system has emission spectra over time. Using a detected emission wave- shown useful for measuring native protein emission upon UV length of 331 nm, only a fraction of the measured emission is excitation . The system comprises a spectrograph and used to construct an electropherogram. Integration of recorded CCD allowing on-line wavelength-resolved fluorescence emission intensities over a certain wavelength range for every (wrFlu) detection. The emission spectrum of FMOC-DL-phe- measured point in time might be used to increase S/N of the nylalanine recorded with wrFlu differed somewhat from the FMOC-AA signals. To evaluate this option, extracted electro- reference spectrum (Fig. 1B) showing a fluorescence maxi- pherograms were constructed from the CE-Flu data obtained mum at 331 nm. The difference in spectral shape and maxi- for FMOC-D-aspartic acid using the integrated signal of in- mum wavelength is caused by a reduced transmittance for UV creasing wavelength intervals centered around 331 nm wavelengths below 315 nm of the detector optics (optical cone (Fig. 2). The S/N grows steadily with increasing wavelength and emission light fiber) . interval, until it levels off at a width of about 40 nm. For In order to achieve optimum detection of FMOC-AAs, wavelength intervals wider than 40 nm, the integrated signal excitation conditions were varied testing a 260-nm band-pass intensity does not significantly increase, while the integrated filter, a 300-nm cutoff short-pass filter, and a 240–400-nm noise increases proportionally, yielding a loss in S/N. The gain broad-pass filter. A sample of FMOC-DL-aspartic acid obtained with wavelength interval integration is clearly illus- (1 μM) was repeatedly analyzed using the different excitation trated by Fig. 3, showing the analysis of FMOC-DL-aspartic filters and the abovementioned BGE, which provided a reso- acid (1.25 μM for each enantiomer). Using signal integration, lution of 3.7 for the aspartic acid enantiomers. The signal-to- the S/N increased with a factor of 12 with respect to single noise ratio (S/N) obtained for each enantiomer at an emission wavelength detection, leading to limits of detection (LODs) of wavelength of 331 nm was determined (Table 1). The lowest less than 100 nM. S/N was observed using the 260-nm band-pass excitation fil- ter. Noise levels were relatively low with this filter, but abso- Optimization of FMOC-AA separation lute signal intensities were modest, as only a part of the exci- tation spectrum is employed to induce FMOC-aspartic acid In order to achieve enantioselective analysis of multiple AAs fluorescence. Creating a broader band-pass by using the by CE-Flu, both chiral separation (i.e., enantioresolution) and 300-nm cutoff short-pass filter or the 240–400-nm band-pass mutual separation (i.e., chemoresolution) of the different AAs filter, an increase of S/N was observed with respect to the 260- are required. Enhancement of AA separation can be attained nm filter. With the 240–400 nm, much broader excitation by addition of SDS to the BGE, inducing micellar Fig. 1 Excitation (A, C)and emission (B, D) spectra of FMOC-AAs (10 μM) in water recorded with a standalone fluorescence spectrophotometer. (A + B) FMOC-phenylalanine. (C + D) FMOC-tryptophan. Experimental conditions, see section BMaterials and methods^ Chiral capillary electrophoresis with UV-excited fluorescence detection for the enantioselective analysis... Table 1 S/NsofFMOC-DL-aspartic acid (1 μM per enantiomer) obtained during CE-Flu using different excitation filters Enantiomer Excitation filter 260 nm 240–400 nm < 300 nm D-Aspartic acid 67.0 81.0 117.9 L-Aspartic acid 68.7 81.7 117.1 Experimental conditions: emission wavelength, 331 nm; for further con- ditions, see section BMaterials and methods^ electrokinetic chromatography . Therefore, a test mixture of 11 FMOC-DL-AAs (histidine, threonine, alanine, valine, methionine, isoleucine, glutamic acid, aspartic acid, leucine, phenylalanine, and tryptophan) was analyzed using BGEs of Fig. 3 Electropherograms obtained during chiral CE-Flu of FMOC-DL- 40 mM sodium tetraborate (pH 9.5) with 15% (v/v) aspartic acid. (A) extracted electropherogram at an emission wavelength isopropanol, 30 mM β-CD, and 20, 25, or 30 mM SDS. of 331 nm and (B) extracted electropherogram using emission signal Raising the SDS concentration from 20 to 30 mM resulted averaging over 40-nm interval centered around 331 nm. Injected concentrations, 1.25 μM of each enantiomer; D-enantiomer migrates in longer analysis times (40, 47, and 83 min, respectively) before L-enantiomer. For further experimental conditions, see section but also in an up to three times higher resolution of the AAs BMaterials and methods^ and an overall enhancement of enantioresolution (Fig. 4). Indeed, use of SDS in the BGE resulted in significantly in- creased enantioseparation of FMOC-AAs in comparison with isopropanol, 19 out of the 22 FMOC-AA enantiomers were CE employing a buffer with only β-CD, as reported by us mutually resolved, whereas with 15% isopropanol almost full previously . Highest enantioresolutions were observed at separation was achieved with only L-aspartic acid and D-leu- 30 mM SDS for most tested FMOC-AAs, except for aspartic cine co-migrating. The optimum BGE was 40 mM sodium acid that was not significantly affected by the SDS concentra- tetraborate (pH 9.5) with 15% (v/v)isopropanol, 30 mM β- tion and glutamic acid that showed a decrease of CD, and 30 mM SDS. Finally, the effect of the capillary thermostating tempera- enantioresolution with increasing SDS concentration. FMOC-aspartic acid and FMOC-glutamic acid are doubly ture (15–23 °C) on the analysis time and resolution was stud- ied. Increasing the capillary temperature overall resulted in negatively charged and most probably cannot partition into the SDS micelles due to electrostatic repulsion . A BGE shorter migration times. For instance, at 15 °C, the migration concentration of 30 mM SDS was selected for further time of DL-leucine was about 59 min, whereas at 23 °C, the experiments. enantiomers were detected after 44 min. Although for most The isopropanol content in the BGE was varied in the AAs the enantioresolution slightly decreased with raising range of 13–17% in order to further fine-tune the separation capillary temperatures, the chemoresolution of the AAs in- of the 11 test FMOC-DL-AAs. Using 13% isopropanol in the creased. As a compromise between analysis time, BGE, the migration window of the tested FMOC-AAs was chemoresolution, and enantioresolution, a capillary temper- small, and as a result, many FMOC-AAs co-migrated. With ature of 22 °C was selected. Figure 5 shows the CE-Flu analysis of the 11 test DL- 15 and 17% isopropanol in the BGE, the chemo- and enantioresolution improved significantly. With 17% AAs derivatized with FMOC using the optimized meth- od. All analyzed FMOC-AAs show enantioresolution (1.0–8.8) and are almost fully separated mutually. The system peak from the unreacted FMOC reagent is not interfering with the FMOC-AAs. Tryptophan and phe- nylalanine showed relatively long migration times, most probably due to their relatively high hydrophobicity and, therefore, strong interaction with the SDS micelles. Analytical performance Fig. 2 S/N as function of the integrated wavelength interval obtained For the optimized CE-Flu method, precision of migration time during CE-Flu of 1.25 μMFMOC-D-aspartic acid. For experimental con- ditions, see section BMaterials and methods^ and electrophoretic mobility were assessed by analyzing the Prior A. et al. Fig. 4 Effect of SDS concentrationinthe BGE onthe FMOC-AA enantiomeric resolution. BGE, 40 mM sodium tetraborate (pH 9.5) containing 15% isopropanol, 30 mM β-CD and SDS. For further experimental conditions, see section BMaterials and methods^. Asterisk: At 20 mM SDS in the BGE, the histidine enantiomers co-migrated with unreacted FMOC and could not be observed. Double asterisk: At 20 mM SDS in the BGE, the alanine enantiomers were not separated Fig. 5 Electropherogram obtained during chiral CE-Flu of a mixture of 11 DL-AAs. For all FMOC-AAs, the D-form migrates before the L-form. Injected concentrations, 500 nM for each enantiomer, except for tryptophan, 5000 nM. For further experimental conditions, see section BMaterials and methods^ Chiral capillary electrophoresis with UV-excited fluorescence detection for the enantioselective analysis... 11 DL-AAs mixture in five consecutive runs, yielding relative (Fig. 1A, B) for the same concentration. The lower fluo- standard deviations (RSDs) in the range of 1.6–5.8 and 0.1– rescence yield of FMOC-tryptophan is due to intramolec- 0.6%, respectively. RSDs for peak height were in the range of ular quenching of the FMOC emission by the indole moi- 1.6–7.0%. Method linearity was assessed by derivatizing mix- ety of tryptophan . On average, the obtained LODs tures of DL-threonine and DL-leucine, which represent a fast encompass an improvement of the sensitivity of two orders and a slow migrating AA, of different concentration (75– of magnitude when compared with chiral CE-UV methods 3700 nM for each enantiomer). Good linearity was observed for AAs using FMOC derivatization [45–48]. These stud- for the enantiomers of both AAs with coefficients of determi- ies reported LODs in the micromolar range. nation (R ) above 0.997. The chiral performance and sensitivity of the optimized CE-Flu method were evaluated for 19 chiral proteinogenic Application to CSF AAs and glycine (Table 2). Tyrosine, lysine, and cysteine were not detected within 90 min of analysis. Lysine and cys- The feasibility of the developed chiral CE-Flu method for the teine—which carry two FMOC moieties after derivatization detection of D-AAs in biofluids was investigated by the anal- —and tyrosine are quite hydrophobic and show high af- ysis of CSF. CSF was spiked with 13 DL-AAs at levels corre- finity for the SDS micelles, yielding very low mobility. sponding to a concentration of 250 nM in CSF for each enan- Extending the analysis time revealed the enantioseparation tiomer (except for tryptophan, 2500 nM) and analyzed by CE- of these AAs with migration times of up to 3.5 h. The other Flu (Fig. 6A). Assignment of the peaks was performed by FMOC-AAs were successfully enantioseparated, showing spiking CSF with individual FMOC-DL-AAs. All the tested resolution of 1.5 or higher, except for alanine (resolution, FMOC-AAs could be detected in the CSF and each was 1.0). The LODs (injected concentration yielding a S/N of 3) enantioseparated. Nevertheless, for CSF, no chemoresolution were in the range of 14–98 nM (2–15 ng/mL), except for the of histidine and glutamine and of threonine and serine was tryptophan enantiomers, which exhibited a LOD of 536 nM achieved and the L-enantiomers of glutamic acid and aspartic (110 ng/mL). The intensity of the FMOC-tryptophan exci- acid co-migrated. Analysis of blank CSF (Fig. 6B) showed the tation and emission spectra (Fig. 1C, D) indeed was lower natural presence of the L-enantiomers of glutamine, histidine, than the intensity observed for the other FMOC-AAs serine, threonine, alanine, valine, methionine, isoleucine, glutamic acid, aspartic acid, phenylalanine, and tryptophan. More importantly, the sensitivity of the CE-Flu method allowed direct detection of D-aspartic acid in the blank CSF Table 2 Enantiomer resolution and LODs (nM; ng/mL) obtained for 17 (peak at 48 min in Fig. 6B). In addition, the small peak at proteinogenic AAs using chiral CE-Flu 26 min (Fig. 6B) was assigned to D-glutamine as D-histidine a b b Amino acid Enantioresolution LOD (nM) LOD (ng/mL) is not expected to be present in CSF . From the measured peak areas, the D/L-enantiomeric ratio of aspartic acid in CSF Alanine 1.0 22 1.9 was calculated to be 19.6%. For glutamine, the D/L-enantio- Valine 7.4 19 2.2 meric ratio was estimated to be 0.35%. These enantiomeric Methionine 3.9 39 5.8 ratios are within reported ranges for aspartic acid (18–25%) Threonine 2.2 21 2.5 and glutamine (0.1–1.0%) in CSF [60, 61]. Based on the peak Histidine 3.4 98 15.2 areas and the spiked concentrations, endogenous CSF levels Isoleucine 5.4 52 6.8 for D-aspartic acid and D-glutamine of 1365 and 565 nM (182 Glutamic acid 3.8 19 2.8 and 82 ng/mL, respectively) were estimated, which is within Aspartic acid 1.9 27 3.6 reported ranges for these two D-AAs [61, 62]. In order to Leucine 5.4 37 4.8 appreciate the LOD of the CE-Flu method for D-AA analysis Phenylalanine 8.8 28 4.6 in CSF, DL-leucine was selected as this AA was not present in Tryptophan 7.1 536 109.4 the blank CSF analyzed. From the peak area obtained for the Glycine – 27 2.0 spiked CSF, the LOD for D-leucine in CSF was determined to Proline 1.5 38 4.3 be 1050 nM (138 ng/mL) which corresponds to an injected Serine 2.1 16 1.7 concentration of 52 nM (6.8 ng/mL) taking the dilution from Asparagine 2.1 15 1.9 the sample pretreatment into account. This value is similar to Glutamine 1.7 14 2.0 the LOD obtained for leucine in aqueous solution (Table 2), Arginine 3.4 36 6.2 indicating minor effects of the CSF matrix on the analysis of the AAs. Overall, these results indicate the potential of the Injected concentration, 500 nM per enantiomer (except tryptophan, 5000 nM) CE-Flu method to detect D-AAs next to their L-AA enantio- Concentration yielding a S/N of 3 as calculated for the D-enantiomer mers in a biofluid. Prior A. et al. Fig. 6 Electropherograms obtained during chiral CE-Flu of (A)CSF spiked with 13 DL-AAs, and (B) blank CSF. For (A) 5.00 μM per enantiomer was spiked into the CSF, except for tryptophan (50.0 μM), which corresponds to injected concentrations of 250 and 2500 nM, respectively. For further experimental conditions, see section BMaterials and methods^ The developed CE-wrFlu method performance was the LIF-based chiral CE-Flu methods. The only chiral CE- compared to previously reported chiral CE-Flu methods Flu method employing UV excitation published so far  for AAs, which almost all employ visible laser-induced showed more favorable sensitivity but was developed for excitation [16, 38–44](Table 3). Using FMOC, the deriv- one AA only (aspartic acid). atization time (2 min) was significantly shorter than for the earlier applied derivatization agents, in particular for FITC (> 12 h), i.e., the most commonly used agent. The number Conclusion of enantioseparated proteinogenic AAs (16 with chiral res- olution > 1.0) was also considerably higher than for previ- A new chiral CE method for AAs was developed ously reported methods (maximum of 8 AAs), while chiral encompassing fast derivatization with FMOC followed separation of a mixture of 11 AAs could be achieved in a by selective separation employing a BGE with β-CD single run applying the currently presented method. and SDS and sensitive fluorescence detection. Efficient Considering that our method employs regular lamp-based broad-band UV excitation of FMOC-AAs was achieved excitation, the achieved sensitivity is very satisfactory, using a Xe-Hg lamp in combination with a short-pass exhibiting lower sample LODs than reported for most of excitation filter. The optimized CE-Flu method enabled Chiral capillary electrophoresis with UV-excited fluorescence detection for the enantioselective analysis... Table 3 Comparison with previously reported chiral CE-Flu methods for AAs Separation method and BGE Derivatization agent Excitation source AAs with enantioresolution Lowest LOD (injected Application Ref. and time and wavelength(s) ≥ 1.0 concentration/sample concentration; nM) MEKC; 100 mM sodium tetraborate (pH 10.0), FITC; overnight Ar+ laser; 488 nm Arg, Ala, Glu, Asp, Ser, Leu, 0.7/7 CSF  80 mM SDS, 20 mM β-CD Gln, Lys MEKC; 100 mM sodium tetraborate (pH 10.0), FITC; overnight Ar+ laser; 488 nm Arg, Ala, Glu, Asp, Ser 160/3200 Maize  80 mM SDS. 20 mM β-CD MEKC; 100 mM sodium tetraborate (pH 9.4), FITC; overnight Ar+ laser; 488 nm Arg, Asn, Ser, Ala, Glu, Asp 0.3/1200 Orange juice  30 mM SDS. 20 mM β-CD MEKC; 100 mM sodium tetraborate (pH 9.7), FITC; overnight Ar+ laser; 488 nm Arg, Ala, Glu, Asp 16.6/8000 Vinegars  30 mM SDS, 20 mM β-CD MEKC; 150 mM Tris-borate (pH 9.0), NDA; 20 min Violet LED; Asp 0.25/2.5 CSF, soymilk, beer  150 mM SDS with 60 mM HP-β-CD 395–425 nm CE; 100 mM borate (pH 8.0), 8 mM NBD-F; 10 min Ar+ laser; Ex, 488 nm Glu, Asp 50/600 Brain  DM-β-CDand5mMHPA-β-CD MEKC; 10 mM sodium borate (pH 9.1), DTAF; 30 min Ar+ laser; Ex, 488 nm Glu, Asp 0.15/180 Urine  12 mM SC, 1.6% HSA, 10% methanol MEKC; 80 mM sodium tetraborate (pH 9.2), CFSE; 2 h Ar+ laser; Ex, 488 nm Ala, Glu, Asp, His, Ser, 5/5 Water from Mono  30 mM γ-CD, 30 mM STC, 5% acetonitrile Leu, Val Lake, CA MEKC; 40 mM sodium tetraborate (pH 9.5), FMOC; 2 min Xe-Hg lamp; 210–300 nm Ala, Val, Met, Thr, His, Ile, 14/280 CSF this work 30 mM SDS, 30 mM β-CD, 15% isopropanol Glu, Asp, Leu, Phe, Trp, Pro, Ser, Asn, Gln, Arg HP-β-CD hydroxypropyl-β-cyclodextrin, DM-β-CD dimethyl-β-cyclodextrin, HPA-β-CD hydroxylpropylamino-β-cyclodextrin, SC sodium cholate, HSA human serum albumin, DTAF 5-(4,6-dichloro-s- triazin-2-ylamino) fluorescein, STC sodium taurocholate, CFSE 5-carboxyfluorescein succinimidyl ester, Ala alanine, Arg arginine, Asn asparagine, Asp aspartic acid, Glu glutamic acid, Gln glutamine, His histidine, Ile isoleucine, Leu leucine, Lys lysine, Met methionine, Phe phenylalanine, Pro proline, Ser serine, Thr threonine, Trp tryptophan, Val valine Prior A. et al. enantioseparation of 16 FMOC-DL-AAs with a resolution References of 1.0 or higher. Use of wave-guiding principles for emis- sion light collection provided good sensitivity, and further 1. Elango R, Ball RO, Pencharz PB. Amino acid requirements in humans: with a special emphasis on the metabolic availability of improvement of the S/N for individual enantiomers was amino acids. Amino Acids. 2009;37:19–27. accomplished by wavelength-resolved CCD detection uti- 2. Wu G. Amino acids: metabolism, functions, and nutrition. Amino lizing signal averaging along the wavelength axis. Acids. 2009;37:1–17. Resulting LODs for most AA enantiomers were between 3. Bruckner H, Becker D, Lupke M. Chirality of amino acids of mi- croorganisms used in food biotechnology. Chirality. 1993;5:385– 10 and 100 nM (2–15 ng/mL injected concentration), which is significantly better than the micromolar range 4. Friedman M, Gumbmann MR, Masters PM. Protein-alkali reac- LODs reported for similar analyses by chiral CE using tions: chemistry, toxicology, and nutritional consequences. Adv UV absorbance detection [45–48]. The LODs for CE-Flu Exp Med Biol. 1984;177:367–412. 5. Koval D, Jiraskova J, Strisovsky K, Konvalinka J, Kasicka V. are similar to LODs obtained for FMOC-AAs using LIF Capillary electrophoresis method for determination of D-serine detection employing a home-built UV laser . and its application for monitoring of serine racemase activity. Compared to chiral FMOC-AA analysis using CE-MS, Electrophoresis. 2006;27:2558–66. LODs for CE-Flu were better with a factor of 3 for argi- 6. Kim PM, Duan X, Huang AS, Liu CY, Ming GL, Song H, et al. Aspartate racemase, generating neuronal D-aspartate, regulates nine and isoleucine up to a factor of 85 for valine in adult neurogenesis. P Natl Acad Sci USA. 2010;107:3175–9. aqueous solution . The CE-Flu LOD for D-leucine in 7. Errico F, Napolitano F, Nistico R, Centonze D, Usiello A. D-aspar- CSF was 11.9 times lower than the LOD for this AA tate: an atypical amino acid with neuromodulatory activity in mam- obtained with CE coupled with mass spectrometry (CE- mals. Rev Neurosci. 2009;20:429–40. 8. Hashimoto A, Nishikawa T, Oka T, Takahashi K. Endogenous D- MS) . The new chiral CE-Flu method was demon- serine in rat brain: N-methyl-D-aspartate receptor-related distribu- strated to allow separation and detection of AA enantio- tion and aging. J Neurochem. 1993;60:783–6. mers in CSF, providing sufficient resolution and sensitiv- 9. Snyder SH, Kim PM. D-amino acids as putative neurotransmitters: ity to probe low levels of endogenous D-aspartic acid and focus on D-serine. Neurochem Res. 2000;25:553–60. D-glutamine next to excess L-AAs. 10. Yekkala R, Meers C, Van Schepdael A, Hoogmartens J, Lambrichts I, Willems G. Racemization of aspartic acid from human dentin in Compared with several previously reported chiral CE-Flu the estimation of chronological age. Forensic Sci Int. methods for AAs using LIF detection, the presented method 2006;159(Suppl 1):S89–94. has the advantage of fast derivatization time (2 min), number 11. Fisher G, Lorenzo N, Abe H, Fujita E, Frey WH, Emory C, et al. of AAs enantioseparated (16 in total; 11 in one run), and Free D- and L-amino acids in ventricular cerebrospinal fluid from Alzheimer and normal subjects. Amino Acids. 1998;15:263–9. favorable sensitivity while employing straightforward lamp 12. Fuchs SA, de Sain-van der Velden MG, de Barse MM, Roeleveld excitation. As for most other chiral CE-Flu methods, the MW, Hendriks M, Dorland L, et al. Two mass-spectrometric tech- LODs for real samples were compromised by a dilution of niques for quantifying serine enantiomers and glycine in cerebro- the derivatized sample before injection, which was needed to spinal fluid: potential confounders and age-dependent ranges. Clin Chem. 2008;54:1443–50. avoid interferences by matrix components and overloading 13. Luykx JJ, Bakker SC, van Boxmeer L, Vinkers CH, Smeenk HE, of the unreacted FMOC. For the presented method a 10- Visser WF, et al. D-amino acid aberrations in cerebrospinal fluid times dilution was adequate, whereas other methods required and plasma of smokers. Neuropsychopharmacology. 2013;38: much higher dilutions, deteriorating the ultimate LODs [16, 2019–26. 14. Fuchs SA, Berger R, Klomp LW, de Koning TJ. D-amino acids in 38–44]. The intrinsic sensitivity of the new CE-Flu system the central nervous system in health and disease. Mol Genet Metab. would allow D-AA detection down to the low-nanomolar 2005;85:168–80. range, but that would require an improved sample cleanup. 15. Bendikov I, Nadri C, Amar S, Panizzutti R, De Miranda J, Wolosker H, et al. A CSF and postmortem brain study of D- serine metabolic parameters in schizophrenia. Schizophr Res. 2007;90:41–51. Acknowledgments This work was financially supported by the 16. Samakashvili S, Ibanez C, Simo C, Gil-Bea FJ, Winblad B, Netherlands Organization for Scientific Research (ECHO project Cedazo-Minguez A, et al. Analysis of chiral amino acids in cere- 711.011.003). brospinal fluid samples linked to different stages of Alzheimer dis- ease. Electrophoresis. 2011;32:2757–64. Compliance with ethical standards 17. Erny GL, Cifuentes A. Liquid separation techniques coupled with mass spectrometry for chiral analysis of pharmaceuticals com- Conflict of interest The authors declare no conflict of interest. pounds and their metabolites in biological fluids. J Pharmaceut Biomed. 2006;40:509–15. Open Access This article is distributed under the terms of the Creative 18. Hernandez-Borges J, Rodriguez-Delgado MA, Garcia-Montelongo Commons Attribution 4.0 International License (http:// FJ, Cifuentes A. Chiral analysis of pollutants and their metabolites creativecommons.org/licenses/by/4.0/), which permits unrestricted use, by capillary electromigration methods. Electrophoresis. 2005;26: distribution, and reproduction in any medium, provided you give appro- 3799–813. priate credit to the original author(s) and the source, provide a link to the 19. Simo C, Barbas C, Cifuentes A. Chiral electromigration methods in Creative Commons license, and indicate if changes were made. food analysis. Electrophoresis. 2003;24:2431–41. Chiral capillary electrophoresis with UV-excited fluorescence detection for the enantioselective analysis... 20. Chankvetadze B. Recent trends in enantioseparations using capil- 38. Herrero M, Ibanez E, Martin-Alvarez PJ, Cifuentes A. Analysis of chiral amino acids in conventional and transgenic maize. Anal lary electromigration techniques. Trac-Trend Anal Chem. 1999;18: 485–98. Chem. 2007;79:5071–7. 21. Spanik I, Horvathova G, Janacova A, Krupcik J. On the use of solid 39. Simo C, Barbas C, Cifuentes A. Sensitive micellar electrokinetic phase ion exchangers for isolation of amino acids from liquid sam- chromatography-laser-induced fluorescence method to analyze chiral ples and their enantioselective gas chromatographic analysis. J amino acids in orange juices. J Agr Food Chem. 2002;50:5288–93. Chromatogr A. 2007;1150:145–54. 40. Carlavilla D, Moreno-Arribas MV, Fanali S, Cifuentes A. Chiral MEKC-LIF of amino acids in foods: analysis of vinegars. 22. Chen S. The enantioseparation of amino acids on a teicoplanin Electrophoresis. 2006;27:2551–7. chiral stationary phase using non-aqueous mobile phases after 41. Lin KC, Hsieh MM, Chang CW, Lin EP, Wu TH. Stacking and pre-column derivatization with sulfur-containing reagents: the con- separation of aspartic acid enantiomers under discontinuous system siderations of mobile phase composition and analyte structure var- by capillary electrophoresis with light-emitting diode-induced fluo- iation on resolution enhancement. Biomed Chromatogr. 2006;20: rescence detection. Talanta. 2010;82:1912–8. 718–28. 42. Wagner Z, Tabi T, Jako T, Zachar G, Csillag A, Szoko E. Chiral 23. Bhushan R, Nagar H. Indirect enantioseparation of proteinogenic separation and determination of excitatory amino acids in brain amino acids using naproxen-based chiral derivatizing reagent and samples by CE-LIF using dual cyclodextrin system. Anal Bioanal HPLC. Biomed Chromatogr. 2013;27:750–6. Chem. 2012;404:2363–8. 24. Erbe T, Bruckner H. Chiral amino acid analysis of vinegars using 43. Wang S, Fan L, Cui S. CE-LIF chiral separation of aspartic acid and gas chromatography - selected ion monitoring mass spectrometry. Z glutamic acid enantiomers using human serum albumin and sodium Lebensm Unters F A. 1998;207:400–9. cholate as dual selectors. J Sep Sci. 2009;32:3184–90. 25. Erbe T, Bruckner H. Studies on the optical isomerization of dietary 44. Creamer JS, Mora MF, Willis PA. Enhanced resolution of chiral amino acids in vinegar and aqueous acetic acid. Eur Food Res amino acids with capillary electrophoresis for biosignature detec- Technol. 2000;211:6–12. tion in extraterrestrial samples. Anal Chem. 2017;89:1329–37. 26. Blomberg LG, Wan H. Determination of enantiomeric excess by 45. Han Y, Chen Y. On-column labeling technique and chiral CE of capillary electrophoresis. Electrophoresis. 2000;21:1940–52. amino acids with mixed chiral selectors and UV detection. 27. Guan J, Yan F, Shi S, Wang S. Optimization and validation of a new Electrophoresis. 2007;28:2765–70. CE method for the determination of pantoprazole enantiomers. 46. Chan KC, Muschik GM, Issaq HJ. Enantiomeric separation of ami- Electrophoresis. 2012;33:1631–6. no acids using micellar electrokinetic chromatography after pre- 28. Tang W, Ong TT, Ng SC. Chiral separation of dansyl amino acids in column derivatization with the chiral reagent 1-(9-fluorenyl)-ethyl capillary electrophoresis using mono-(3-methyl-imidazolium)-be- chloroformate. Electrophoresis. 1995;16:504–9. ta-cyclodextrin chloride as selector. J Sep Sci. 2007;30:1343–9. 47. Wan H, Andersson PE, Engström A, Blomberg LG. Direct and 29. Fradi I, Servais AC, Lamalle C, Kallel M, Abidi M, Crommen J, et indirect chiral separation of amino acids by capillary electrophore- al. Chemo- and enantio-selective method for the analysis of amino sis. J Chromatogr A. 1995;704:179–93. acids by capillary electrophoresis with in-capillary derivatization. J 48. Wan H, Blomberg LG. Optimized chiral separation of 20 amino Chromatogr A. 2012;1267:121–6. acids derivatized with 9-fluorenylmethyl chloroformate using cy- 30. Sanchez-Hernandez L, Dominguez-Vega E, Montealegre C, clodextrins as chiral selectors in capillary electrophoresis. J Castro-Puyana M, Marina ML, Crego AL. Potential of vancomycin Chromatogr Sci. 1996;34:540–6. for the enantiomeric resolution of FMOC-amino acids by capillary 49. Chan KC, Janini GM, Muschik GM, Issaq HJ. Laser-induced fluo- electrophoresis-ion-trap-mass spectrometry. Electrophoresis. rescence detection of 9-fluorenylmethyl chloroformate derivatized 2014;35:1244–50. amino acids in capillary electrophoresis. J Chromatogr A. 31. Wan H, Blomberg LG. Enantioseparation of amino acids and di- 1993;653:93–7. peptides using vancomycin as chiral selector in capillary electro- 50. Caslavska J, Thormann W. Monitoring of drugs and metabolites in phoresis. Electrophoresis. 1996;17:1938–44. body fluids by capillary electrophoresis with XeHg lamp-based and 32. Verleysen K, Vandijck J, Schelfaut M, Sandra P. Enantiomeric sep- laser-induced fluorescence detection. Electrophoresis. 2004;25: arations in capillary electrophoresis using 18-crown-6- 1623–31. tetracarboxylic acid (18C6H(4)) as buffer additive. Hrc-J High 51. Priego Capote F, Rodriguez JM, Luque de Castro MD. Res Chrom. 1998;21:323–31. Determination of phenolic compounds in grape skin by capillary 33. Singh NS, Paul RK, Sichler M, Moaddel R, Bernier M, Wainer IW. electrophoresis with simultaneous dual fluorescence and diode ar- Capillary electrophoresis-laser-induced fluorescence (CE-LIF) as- ray absorption detection after dynamic superheated liquid leaching. say for measurement of intracellular D-serine and serine racemase J Chromatogr A. 2007;1139:301–7. activity. Anal Biochem. 2012;421:460–6. 52. Priego-Capote F, Luque de Castro MD. Determination of B2 and 34. Nouadje G, Rubie H, Chatelut E, Canal P, Nertz M, Puig P, et al. B6 vitamers in serum by capillary electrophoresis-molecular Child cerebrospinal fluid analysis by capillary electrophoresis and fluorescence-charge coupled detector. Electrophoresis. laser-induced fluorescence detection. J Chromatogr A. 1995;717: 2005;26:2376–83. 293–8. 53. Pobozy E, Michalski A, Sotowska-Brochocka J, Trojanowicz M. 35. Siri N, Lacroix M, Garrigues JC, Poinsot V, Couderc F. HPLC- Determination of melatonin and its precursors and metabolites fluorescence detection and MEKC-LIF detection for the study of using capillary electrophoresis with UVand fluorometric detection. amino acids and catecholamines labelled with naphthalene-2,3- J Sep Sci. 2005;28:2165–72. dicarboxyaldehyde. Electrophoresis. 2006;27:4446–55. 54. Mallampati S, Van Aerschot A, Hoogmartens J, Van Schepdael A. 36. Taga A, Honda S. Derivatization at capillary inlet in high- Analysis of dideoxyadenosine triphosphate by CE with fluores- performance capillary electrophoresis—its reliability in quantifica- cence detection. I. Derivatization through the phosphate group. tion. J Chromatogr A. 1996;742:243–50. Electrophoresis. 2007;28:3948–56. 37. Simo C,RizziA,Barbas C,Cifuentes A. Chiral capillary 55. Radenovic DC, de Kort BJ, Somsen GW. Lamp-based native fluo- electrophoresis-mass spectrometry of amino acids in foods. rescence detection of proteins in capillary electrophoresis. J Electrophoresis. 2005;26:1432–41. Chromatogr A. 2009;1216:4629–32. Prior A. et al. 56. de Kort BJ, ten Kate GA, de Jong GJ, Somsen GW. Capillary 9-fluorenyl methylchloroformate and 1-aminoadamantane. J Chromatogr A. 1990;507:67–77. electrophoresis with lamp-based wavelength-resolved fluorescence detection for the probing of protein conformational changes. Anal 60. Thorsen G, Bergquist J. Chiral separation of amino acids in Chem. 2011;83:6060–7. biological fluids by micellar electrokinetic chromatography 57. de Kort BJ, de Jong GJ, Somsen GW. Lamp-based wavelength- with laser-induced fluorescence detection. J Chromatogr B. resolved fluorescence detection for protein capillary electrophore- 2000;745:389–97. sis: setup and detector performance. Electrophoresis. 2010;31: 61. Prior A, Moldovan RC, Crommen J, Servais AC, Fillet M, de Jong 2861–8. GJ, et al. Enantioselective capillary electrophoresis-mass spectrom- 58. Prior A, Sanchez-Hernandez L, Sastre-Torano J, Marina ML, de etry of amino acids in cerebrospinal fluid using a chiral derivatizing Jong GJ, Somsen GW. Enantioselective analysis of proteinogenic agent and volatile surfactant. Anal Chim Acta. 2016;940:150–8. amino acids in cerebrospinal fluid by capillary electrophoresis-mass 62. Fisher GH, Petrucelli L, Gardner C, Emory C, Frey WH, Amaducci spectrometry. Electrophoresis. 2016;37:2410–9. L, et al. Free D-amino acids in human cerebrospinal fluid of 59. Gustavsson B, Betnér I. Fully automated amino acid analysis for Alzheimer disease, multiple sclerosis and healthy control subjects. protein and peptide hydrolysates by precolumn derivatization with Mol Chem Neuropathol. 1994;23:115–24.
Analytical and Bioanalytical Chemistry
– Springer Journals
Published: May 29, 2018