G Model
ARTICLE IN PRESS
CHROMA-355267; No. of Pages 8
Journal of Chromatography A, xxx (2014) xxx–xxx
Contents lists available at ScienceDirect
Journal of Chromatography A journal homepage: www.elsevier.com/locate/chroma
Direct enantioseparation of underivatized aliphatic 3-hydroxyalkanoic acids with a quinine-based zwitterionic chiral stationary phase Federica Ianni a,b , Zoltán Pataj b,c , Harald Gross b , Roccaldo Sardella a , Benedetto Natalini a , Wolfgang Lindner d , Michael Lämmerhofer b,∗ a
Department of Pharmaceutical Sciences, University of Perugia, Via del Liceo 1, 06123 Perugia, Italy Institute of Pharmaceutical Sciences, University of Tübingen, Auf der Morgenstelle 8, 72076 Tübingen, Germany c Regional Centre of Advanced Technologies and Materials, Department of Analytical Chemistry, Faculty of Science, Palack´ y University, 17. listopadu 12, 771 46 Olomouc, Czech Republic d Department of Analytical Chemistry, University of Vienna, Währinger Strasse 38, 1090 Vienna, Austria b
a r t i c l e
i n f o
Article history: Received 25 January 2014 Received in revised form 19 March 2014 Accepted 21 March 2014 Available online xxx Keywords: Enantiomer separation Underivatized 3-hydroxycarboxylic acids 3-Hydroxyalkanoic acids, liquid chromatography Cinchona-based zwitterionic chiral stationary phase Natural compounds, polar-organic mode
a b s t r a c t While aliphatic 2-hydroxyalkanoic acids have been more or less successfully enantioseparated with various chiral stationary phases by HPLC and GC, analogous applications on underivatized aliphatic 3hydroxyalkanoic acids are completely absent in the scientific literature. With the aim of closing this gap, the enantioseparation of 3-hydroxybutyric acid, 3-hydroxydecanoic acid and 3-hydroxymyristic acid has been performed with two ion-exchange type chiral stationary phases (CSPs): one containing the anion-exchange type tert-butyl carbamoyl quinine chiral selector motif (Chiralpak QN-AX), and the other carrying the new zwitterionic variant based on trans-(S,S)-2-aminocyclohexanesulfonic acid-derivatized quinine carbamate (Chiralpak ZWIX(+)) as the chiral selector and enantiodiscriminating element, respectively. The zwitterionic enantiorecognition material provided better results in terms of enantioselectivity and resolution compared to the anion-exchanger CSP at reduced retention times due to the intramolecular counterion effect imposed by the sulfonic acid moiety and its competition with the 3-hydroxyalkanoic acid analyte for ionic interaction at the quininium-anion exchanger site. It is thus recommended as the CSP of first choice for enantioseparations of the class of aliphatic 3-hydroxyalkanoic acids. With use of polar organic eluent composed of ACN/MeOH/AcOH – 95/5/0.05 (v/v/v), a good compromise in terms of analysis time and enantioresolution quality was accomplished. The major experimental variables have been investigated for optimization of the resolution and allowed to derive information on the enantiorecognition mechanism. Corresponding Chiralpak ZWIX(−), based on pseudo-enantiomeric selector derived from quinidine and trans-(R,R)-2-aminocyclohexanesulfonic acid with opposite configurations provided reversed enantiomer elution orders. It has further to be stressed that these separations can be obtained with mass spectrometry compatible mobile phases. © 2014 Elsevier B.V. All rights reserved.
1. Introduction 3-Hydroxyalkanoic acids are along with their 2hydroxyalkanoic acid counterparts important disease-related biomarkers, constituents of natural compounds, substrates for chemical synthesis and biochemical transformations, and many more applications [1–8]. In spite of the advanced state in chromatographic enantiomer separation, the direct chromatographic
∗ Corresponding author. Tel.: +49 7071 29 78793; fax: +49 7071 29 4565. E-mail address: [email protected] (M. Lämmerhofer).
separation of enantiomers of chiral aliphatic 3-hydroxyalkanoic acids (devoid of any aromatic group) without chiral or achiral derivatization is still one of the most challenging, hitherto unsolved problems. On contrary, for 2-hydroxyalkanoic acids a number of methods exist for their enantiomer separation, in particular for aromatic but also aliphatic ones. Aromatic 2-hydroxyalkanoic acids have been resolved without derivatization for instance by capillary electrophoresis (CE) with cyclodextrin derivatives as chiral additives to the background electrolyte (BGE) [9,10], by chiral ligand-exchange CE [11] or capillary electrochromatography (CEC) with CSPs modified with chelating selectors and metal ions in the BGE [12], and by various liquid chromatography methods,
http://dx.doi.org/10.1016/j.chroma.2014.03.060 0021-9673/© 2014 Elsevier B.V. All rights reserved.
Please cite this article in press as: F. Ianni, et al., J. Chromatogr. A (2014), http://dx.doi.org/10.1016/j.chroma.2014.03.060
G Model CHROMA-355267; No. of Pages 8
ARTICLE IN PRESS F. Ianni et al. / J. Chromatogr. A xxx (2014) xxx–xxx
2
O HO
OH R
R= -CH3, 3-hydroxybutyric acid, 1 -C7H15, 3-hydroxydecanoic acid, 2 -C11H23, 3-hydroxymyristic acid, 3
Fig. 1. Analytes investigated in this study.
for instance after derivatization of the hydroxyl group with FLEC (1-(9-fluorenyl)ethyl chloroformate) by achiral HPLC with ODS [13], or chiral ligand exchange chromatography (CLEC) using chiral additives such as phenylalanine amide and Cu(II)-salts in the mobile phase as well as achiral ODS stationary phase [14], or directly with various chiral stationary phases (CSPs) such as based on antibody- [15], polysaccharide-[16], cyclodextrin[17], -electron donor–acceptor (Pirkle)-type [18], and Cinchona alkaloid [19] selectors. Aliphatic 2-hydroxyalkanoic acids can be separated into enantiomers by achiral gas chromatography (GC) but it needs a double-derivatization step to form diastereomeric O-trifluoroacetylated (S)-(+)-3-methyl-2-butyl esters prior to injection [20]. Further, enantioselective CE methods with cyclodextrin derivative [21] or ligand exchange selectors [22,23] as BGE additives have been proposed for aliphatic 2-hydroxyalkanoic acid enantiomer separations. As far as the LC-based enantioseparation of aliphatic 2-hydroxyalkanoic acids is concerned, precolumn derivatization with FLEC and separation on ODS column [13] as well as various direct HPLC methods based on Cinchona alkaloids[24], antibiotic- [25] and ligand-exchange-based [26,27] CSPs have been reported. However, to the best of our knowledge, direct applications of enantioselective HPLC to aliphatic 3-hydroxyalkanoic acids without derivatization are completely missing in the panorama of scientific literature. There are only a few methods reported so far for HPLC enantiomer separation of 3-hydroxyalkanoic acids [28,29]. However, most of them require achiral precolumn derivatization of the hydroxyl group to carbamates [30,31] or ester derivatives [32] before separation on CSPs. The present study deals with the challenging topic of direct liquid chromatographic enantiomer separation of this class of compounds. Three aliphatic 3-hydroxyalkanoic acids (Fig. 1) were selected as test compounds. The separations were performed with two ion-exchange-based CSPs: one containing the tert-butyl carbamoyl quinine selector motif (Chiralpak QN-AX, CSP 1, Fig. 2) [33–35], and the other carrying the relatively new zwitterionic variant based on trans-2-aminocyclohexanesulfonic acid-decorated
quinine carbamate (Chiralpak ZWIX(+), CSP 2, Fig. 2) [36–38]. The effect of the primary experimental variables has been investigated to achieve optimal separations and allow to derive information on the retention mechanism. 2. Experimental 2.1. Materials All the reagents used were of analytical grade. Acetonitrile (ACN), methanol (MeOH), ethanol (EtOH) and 2-propanol (iPrOH) were purchased from Sigma–Aldrich (Munich, Germany). Acetic acid (AcOH) was purchased from Honeywell Riedel-de Haën (Hannover, Germany). The racemic mixtures of 3-hydroxybutyric acid, 3hydroxydecanoic acid and 3-hydroxymyristic acid, and the enantiomerically pure compound (R)-3-hydroxybutyric acid, were purchased from Sigma–Aldrich (Munich, Germany). Enantiomerically pure (R)-3-hydroxymyristic acid was obtained from Alberta Research Chemicals Inc. (Edmonton, Canada). The preparation of (R)-3-hydroxydecanoic acid was described elsewhere [30]. The enantiomeric elution order was assessed by injection of single enantiomers and non-racemic mixtures obtained by fortification of the racemate with single enantiomer. The analytes were dissolved in methanol at the approximate concentration of 0.5 mg/mL. 2.2. Instrumentation and chromatographic method The analytical HPLC measurements were made on a Merck Hitachi LaChrom 7000-series HPLC System (Tokyo, Japan) equipped with a D-7000 Interface, a diode array detector L-7455, a L-7200 Autosampler, and a L-6200A Intelligent Pump system. High Performance Liquid Chromatography System Manager (HSM) software from Hitachi allowed performing the acquisition and processing of the chromatographic data. Because of absence of suitable chromophoric moieties for sensitive UV detection in the sample structures, the HPLC system was connected with the Corona® Charged Aerosol Detector, CAD® (ESA Biosciences Inc., Chelmsford, U.S.A.). The nitrogen flow of the CAD was adjusted to 35 psi. The CHIRALPAK® QN-AX (150 mm × 4.6 mm ID, 120 A˚ pore size, 5 m particle diameter) column, and CHIRALPAK® ZWIX(+) and ZWIX(−) (150 mm × 4.0 mm ID, 120 A˚ pore size, 3 m particle diameter) columns were from Chiral Technologies Europe (Illkirch, France). Further, a prototype of CHIRALPAK® ZWIX(+) (250 mm × 4.0 mm ID, 120 A˚ pore size, 5 m particle diameter) was used for initial optimization experiments. Before use, all the employed mobile phases were degassed through sonication. The column was conditioned with the selected mobile phase at a flow rate of 1.0 mL/min for at least 20 min, before running the analysis. If not otherwise stated, the column temperature was fixed at 20 ◦ C with a column thermostat from W.O. Electronics (Langenzersdorf, Austria). 3. Results and discussion 3.1. Column selection
Fig. 2. Chiral stationary phases employed in this study.
Recently, we reported the direct enantiomer separation of underivatized lactic acid on chiral anion-exchange CSP 1 (Chiralpak QN-AX) [24] and of 2-hydroxyalkanoic acids including lactic acid, 2-hydroxybutyric acid, and glyceric acid on a noncommercial structural analog of CSP 1 with 2,6-diisopropylphenyl carbamoyl residue [39] (see also supplementary material). 3Hydroxyalkanoic acids, in contrast, were not separated with the employed conditions. Further, retention times were long on the
Please cite this article in press as: F. Ianni, et al., J. Chromatogr. A (2014), http://dx.doi.org/10.1016/j.chroma.2014.03.060
G Model CHROMA-355267; No. of Pages 8
ARTICLE IN PRESS F. Ianni et al. / J. Chromatogr. A xxx (2014) xxx–xxx
(b) Detector signal
anion-exchange type CSP 1 and acceleration by increase of counterion concentrations was precluded by its negative effect on UV detection sensitivity. It is to note that in ion exchange chromatography retention is controlled by the concentration of suitable counter ions. Thus, herein we compare the performance provided by two different ion-exchange based CSPs: besides the anionexchange CSP 1 (Fig. 2) [33–35], also a recently commercialized zwitterionic variant (CSP 2) in which both a cation-exchange (CX) and an anion-exchange (AX) motif are hybridized into a new selector (SO) unit (Fig. 2) [36–38]. In this case, the sulfonic acid moiety can serve as intramolecular counterion which competes with the 3-hydroxcarboxylic acid solutes for ionic interaction at the quininium anion-exchange site (possibly by a long range electrostatic interaction) with the consequence of reducing retention on anionic species. This intramolecular counterion effect allows for working with eluents having low counterion concentrations which is favourable for UV, CAD and MS detection sensitivity. For this comparison, a polar-organic (PO) elution mode was selected which was preferred over reversed-phase (RP) and normal-phase (NP) chromatographic elution conditions [40,41]. Previous experiments with 2-hydroxyalkanoic acids revealed that aqueous conditions such as in an RP elution mode are detrimental. In fact, the screening conditions with 2% water suggested by the column supplier as initial test did not give any resolution for current 3-hydroxyalkanoic acids. The problem may be that the hydroxyl group at the stereogenic centre is strongly solvated and thus cannot participate in a concerted simultaneous anionexchange driven multi-point interaction with the chiral selector. On the other hand, hydroxyalkanoic acids are very polar and poorly soluble in NP eluents. An important benefit of PO eluent systems resides in their ability of preserving strong electrostatic forces between complementary functional groups of SO and selectand (SA), here the 3-hydroxyalkanoic acids, such as H-bonding and long-range Coulomb attractions [42,43] while they are still balanced so that retention does not get too strong. In the current chromatographic systems, ionic interactions are supposed to be the dominant driving forces for binding of hydroxyalkanoic acids on the quinine carbamate anion-exchanger moiety competing with the counterions in the eluent for the sorption sites. On the other hand, H-bonds between the hydroxyl group of the selectand and complementary sites at the selector (carbamate moiety) are supposed to be active in case of hydroxyalkanoic acid enantiomer separation in PO mode as well and supportive for chiral distinction, like hydrophobic and/or van der Waals type interactions in case of the long chain 3-hydroxy fatty acids (compounds 2 and 3). The role of steric interactions is less obvious but they can also favourably contribute to stereoselectivity in PO media [44]. On the other hand, the PO mode turns out to be effective in simultaneously suppressing both hydrophobic and hydrophilic non-specific interactions, e.g. with support material and linker, which can have a negative effect on the observed enantioselectivity. Methanol and other alcohols were deemed to be too strong competitors, although much weaker than water, for hydrogen bonding of the hydroxyl group at the stereogenic centre of the analytes with the chiral selector and hence acetonitrile was chosen for the initial column screening as solvent system [37,40,45]. AcOH was used as the exchangeable counter-ion and displacer, respectively, to attenuate the otherwise too strong retention caused by the anion-exchange contribution [37,40,45]. Fig. 3 shows the chromatograms obtained for compound 2 (Fig. 1) with the screening eluent composed of ACN/AcOH-100/0.1 (v/v) (corresponding to a total concentration of 17.5 mM). As evident from the chromatographic traces, the zwitterionic CSP 2 (Fig. 3b) revealed greatly reduced retention compared to CSP 1 (in spite of the longer column) (Fig. 3a) due to the intramolecular counter-ion effect imposed by the sulfonic acid in the ZWIX
3
(a)
0
10
20
30
40
50
60
70
Time (min) Fig. 3. Chromatographic traces of compound 2, obtained with the two CSPs (a) anion-exchange CSP 1 (Chiralpak QN-AX, 5 m, 150 mm × 4 mm ID) and (b) zwitterionic CSP 2 (Chiralpak ZWIX(+), 5 m, 250 mm × 4 mm ID). Eluent: ACN/AcOH – 100/0.1 (v/v) (corresponding to a 17.5 mM total acid concentration); flow rate 1.0 mL/min; column temperature 20 ◦ C.
selector. It also slightly outperformed the anion-exchanger CSP 1 in terms of resolution. 3.2. Optimization of separation conditions Above separation indicated the potential of the ZWIX CSP 2 but for baseline separation conditions needed to be optimized. Compound 2 was selected for optimization by an OVAT (onevariable-at-time) approach and this tuning of the experimental conditions can be also useful to gain mechanistic information [34]. 3.2.1. Counterion concentration AcOH represents the counterion and displacer in this chromatographic system. When its concentration was reduced from 17.5 mM down to 4.4 mM, retention factors significantly increased in accordance with an anion-exchange process [34,37,40] (Fig. 4a). At the same time, enantioselectivity was nearly unchanged between 8 and 20 mM and increased slightly at very low concentrations. This behaviour strongly suggests a negligible effect of counterion on other tentative binding increments mainly responsible for stereoselectivity (namely H-bond and van der Waals/steric interactions) [46]. Likewise, resolution factors (R) remain more or less constant within the investigated eluent ionic strength (Fig. 4b). For CAD and MS detection, the more volatile formic acid was deemed to be advantageous and tested, but clearly revealed worse enantioselectivity and resolution. 3.2.2. MeOH content In order to study the role of H-bonds with the hydroxyl group for complex stabilization besides the primary ion-pairing mechanism, the impact of the MeOH content in the ACN-based eluent on the chromatographic performance was evaluated. MeOH which is a polar protic solvent is a stronger H-bond competitor than ACN, and thus, by increasing its concentration in ACN, retention of 2 decreased for both enantiomers. As expected, this clearly reveals that H-bonding plays a major role for retention of 3-hydroxyalkanoic acids on the ZWIX phase (Fig. 4c) [40,47]. Obviously, the higher the ACN content in the eluent the more significant this retention contribution, while addition of MeOH leads to stereoselective disruption of such interactions and an increase of enantioselectivity with the MeOH content (Fig. 4c). Further increase from 5 to 10% did not lead to improved resolution, but stronger baseline noise. At 50% MeOH resolution declined and vanished completely at 100% MeOH. Lower column efficiencies with ACNricher mobile phases along with smaller separation factors explain
Please cite this article in press as: F. Ianni, et al., J. Chromatogr. A (2014), http://dx.doi.org/10.1016/j.chroma.2014.03.060
G Model
ARTICLE IN PRESS
CHROMA-355267; No. of Pages 8
F. Ianni et al. / J. Chromatogr. A xxx (2014) xxx–xxx
4
k(R)
a)
k(S)
α
0.75
0.08
0.65
0.07
0.55 0.35
0.05
0.25
0.04
0.15
1.20 1.15 1.10
0.06
0.45 Log k
b)
1.05 R 1.00
Log α
0.95
0.03
0.90
-0.05
0.02
0.85
-0.15
0.01
0.80
0.05
0.64
0.94
k(R)
k(S)
α
1.25
0.07
1.05
0.06
0.85
d)
1.00 0.80 R
0.04 Log α
0.45
0.03
0.25 0.05
0.02
-0.15
0.01 0
2
1.40 1.20
0.05
0.65 Log k
0.60 0.40 0.20 0.00
5
0
% MeOH
e)
k(R)
k(S)
17 .50
Conc. AcOH (mM)
Log. Conc. AcOH (mM)
c)
8.75
4.40
1.24
2
5
% MeOH
α
0.80
0.060
f)
R
α
1.60
1.50 1.45 1.40 1.35 1.30 1.25 α 1.20 1.15 1.10 1.05 1.00
1.40
0.70 0.60
0.055
1.20 1.00
0.50 Log k 0.40
0.050 Log α
0.30
R 0.80 0.60
0.20
0.045
0.10
0.40 0.20
0.040
0.00 MeOH
EtOH
IPA
0.00 0.5
0.3
Type of alcoholic modifier
1.0
Eluent flow rate (mL/min)
Fig. 4. Influence of some experimental variables on retention, selectivity, and resolution of compound 2 on Chiralpak ZWIX(+) (5 m; 250 mm × 4 mm ID). Experimental conditions: (a) and (b) mobile phase, ACN/MeOH – 95/5 (v/v), total AcOH concentration varied in the range between 4.40 and 17.50 mM; flow rate, 1.0 mL/min; column temperature, 20 ◦ C; (c) and (d) mobile phase, ACN without and with MeOH (2% and 5%, v/v), total AcOH concentration 17.50 mM; flow rate, 1.0 mL/min; column temperature, 20 ◦ C; (e) mobile phase, ACN/type of alcoholic modifier – 98/2 (v/v), total AcOH concentration 17.50 mM; flow rate, 1.0 mL/min; column temperature 20 ◦ C; (f) mobile phase, ACN/MeOH – 95/5 (v/v), total AcOH concentration 8.75 mM; flow rate varied in the range between 0.3 and 1.0 mL/min; column temperature, 20 ◦ C.
the trend of decreasing R values with decrease of MeOH content (Fig. 4d).
3.2.3. Alcohol-type The replacement of MeOH with EtOH or iPrOH was then evaluated. The three alcohols were always used at 2% (v/v). As can be seen from Fig. 4e, retention factors get slightly reduced with the higher alcohols. Besides Coulombic attraction and H-bonding, hydrophobic interactions may contribute to retention to some extent. Accordingly, due to its ability to better disrupt hydrophobic interactions, the iPrOH containing eluent produced the highest elution strength [48]. The ␣-value was largest for EtOH, then MeOH followed by iPrOH [48] (Fig. 4e). Further experiments were made with MeOH as polar protic modifier because it provided the best resolution.
3.2.4. Temperature Temperature has multiple effects on the chromatographic behaviour of solutes in HPLC enantiomer separation. In spite of that, it is often neglected as variable to optimize separations. In current case, it turned out worthwhile to examine this parameter in order to achieve full baseline resolution. To do so, temperature was studied in the range between 10 and 40 ◦ C in steps of 10 ◦ C. The results are summarized in Tables 1 and 2. Adsorption thermodynamics can be analyzed by van’t Hoff equations (Eqs. (1) and (2))
ln k = −
S ◦ H ◦ + + ln RT R
Please cite this article in press as: F. Ianni, et al., J. Chromatogr. A (2014), http://dx.doi.org/10.1016/j.chroma.2014.03.060
(1)
G Model
ARTICLE IN PRESS
CHROMA-355267; No. of Pages 8
F. Ianni et al. / J. Chromatogr. A xxx (2014) xxx–xxx Table 1 Temperature dependence of retention factor of first eluting enantiomer (k1 ), separation factor (˛) and resolution (R) of 3-hydroxyalkanoic acids on Chiralpak ZWIX(+). Compound
k, ˛, R
k1 ˛ RS k1 ˛ RS k1 ˛ RS
1
2
3
Temperature (◦ C)
Elution order
10
20
30
40
1.53 1.07 0.61 1.71 1.18 1.40 1.80 1.16 1.23
1.32 1.07 0.58 1.46 1.16 1.34 1.52 1.15 1.16
1.12 1.07 0.56 1.21 1.14 1.21 1.22 1.14 1.14
0.91 1.07 0.59 0.99 1.12 1.08 0.98 1.12 1.06
R
ARTICLE IN PRESS
CHROMA-355267; No. of Pages 8
Journal of Chromatography A, xxx (2014) xxx–xxx
Contents lists available at ScienceDirect
Journal of Chromatography A journal homepage: www.elsevier.com/locate/chroma
Direct enantioseparation of underivatized aliphatic 3-hydroxyalkanoic acids with a quinine-based zwitterionic chiral stationary phase Federica Ianni a,b , Zoltán Pataj b,c , Harald Gross b , Roccaldo Sardella a , Benedetto Natalini a , Wolfgang Lindner d , Michael Lämmerhofer b,∗ a
Department of Pharmaceutical Sciences, University of Perugia, Via del Liceo 1, 06123 Perugia, Italy Institute of Pharmaceutical Sciences, University of Tübingen, Auf der Morgenstelle 8, 72076 Tübingen, Germany c Regional Centre of Advanced Technologies and Materials, Department of Analytical Chemistry, Faculty of Science, Palack´ y University, 17. listopadu 12, 771 46 Olomouc, Czech Republic d Department of Analytical Chemistry, University of Vienna, Währinger Strasse 38, 1090 Vienna, Austria b
a r t i c l e
i n f o
Article history: Received 25 January 2014 Received in revised form 19 March 2014 Accepted 21 March 2014 Available online xxx Keywords: Enantiomer separation Underivatized 3-hydroxycarboxylic acids 3-Hydroxyalkanoic acids, liquid chromatography Cinchona-based zwitterionic chiral stationary phase Natural compounds, polar-organic mode
a b s t r a c t While aliphatic 2-hydroxyalkanoic acids have been more or less successfully enantioseparated with various chiral stationary phases by HPLC and GC, analogous applications on underivatized aliphatic 3hydroxyalkanoic acids are completely absent in the scientific literature. With the aim of closing this gap, the enantioseparation of 3-hydroxybutyric acid, 3-hydroxydecanoic acid and 3-hydroxymyristic acid has been performed with two ion-exchange type chiral stationary phases (CSPs): one containing the anion-exchange type tert-butyl carbamoyl quinine chiral selector motif (Chiralpak QN-AX), and the other carrying the new zwitterionic variant based on trans-(S,S)-2-aminocyclohexanesulfonic acid-derivatized quinine carbamate (Chiralpak ZWIX(+)) as the chiral selector and enantiodiscriminating element, respectively. The zwitterionic enantiorecognition material provided better results in terms of enantioselectivity and resolution compared to the anion-exchanger CSP at reduced retention times due to the intramolecular counterion effect imposed by the sulfonic acid moiety and its competition with the 3-hydroxyalkanoic acid analyte for ionic interaction at the quininium-anion exchanger site. It is thus recommended as the CSP of first choice for enantioseparations of the class of aliphatic 3-hydroxyalkanoic acids. With use of polar organic eluent composed of ACN/MeOH/AcOH – 95/5/0.05 (v/v/v), a good compromise in terms of analysis time and enantioresolution quality was accomplished. The major experimental variables have been investigated for optimization of the resolution and allowed to derive information on the enantiorecognition mechanism. Corresponding Chiralpak ZWIX(−), based on pseudo-enantiomeric selector derived from quinidine and trans-(R,R)-2-aminocyclohexanesulfonic acid with opposite configurations provided reversed enantiomer elution orders. It has further to be stressed that these separations can be obtained with mass spectrometry compatible mobile phases. © 2014 Elsevier B.V. All rights reserved.
1. Introduction 3-Hydroxyalkanoic acids are along with their 2hydroxyalkanoic acid counterparts important disease-related biomarkers, constituents of natural compounds, substrates for chemical synthesis and biochemical transformations, and many more applications [1–8]. In spite of the advanced state in chromatographic enantiomer separation, the direct chromatographic
∗ Corresponding author. Tel.: +49 7071 29 78793; fax: +49 7071 29 4565. E-mail address: [email protected] (M. Lämmerhofer).
separation of enantiomers of chiral aliphatic 3-hydroxyalkanoic acids (devoid of any aromatic group) without chiral or achiral derivatization is still one of the most challenging, hitherto unsolved problems. On contrary, for 2-hydroxyalkanoic acids a number of methods exist for their enantiomer separation, in particular for aromatic but also aliphatic ones. Aromatic 2-hydroxyalkanoic acids have been resolved without derivatization for instance by capillary electrophoresis (CE) with cyclodextrin derivatives as chiral additives to the background electrolyte (BGE) [9,10], by chiral ligand-exchange CE [11] or capillary electrochromatography (CEC) with CSPs modified with chelating selectors and metal ions in the BGE [12], and by various liquid chromatography methods,
http://dx.doi.org/10.1016/j.chroma.2014.03.060 0021-9673/© 2014 Elsevier B.V. All rights reserved.
Please cite this article in press as: F. Ianni, et al., J. Chromatogr. A (2014), http://dx.doi.org/10.1016/j.chroma.2014.03.060
G Model CHROMA-355267; No. of Pages 8
ARTICLE IN PRESS F. Ianni et al. / J. Chromatogr. A xxx (2014) xxx–xxx
2
O HO
OH R
R= -CH3, 3-hydroxybutyric acid, 1 -C7H15, 3-hydroxydecanoic acid, 2 -C11H23, 3-hydroxymyristic acid, 3
Fig. 1. Analytes investigated in this study.
for instance after derivatization of the hydroxyl group with FLEC (1-(9-fluorenyl)ethyl chloroformate) by achiral HPLC with ODS [13], or chiral ligand exchange chromatography (CLEC) using chiral additives such as phenylalanine amide and Cu(II)-salts in the mobile phase as well as achiral ODS stationary phase [14], or directly with various chiral stationary phases (CSPs) such as based on antibody- [15], polysaccharide-[16], cyclodextrin[17], -electron donor–acceptor (Pirkle)-type [18], and Cinchona alkaloid [19] selectors. Aliphatic 2-hydroxyalkanoic acids can be separated into enantiomers by achiral gas chromatography (GC) but it needs a double-derivatization step to form diastereomeric O-trifluoroacetylated (S)-(+)-3-methyl-2-butyl esters prior to injection [20]. Further, enantioselective CE methods with cyclodextrin derivative [21] or ligand exchange selectors [22,23] as BGE additives have been proposed for aliphatic 2-hydroxyalkanoic acid enantiomer separations. As far as the LC-based enantioseparation of aliphatic 2-hydroxyalkanoic acids is concerned, precolumn derivatization with FLEC and separation on ODS column [13] as well as various direct HPLC methods based on Cinchona alkaloids[24], antibiotic- [25] and ligand-exchange-based [26,27] CSPs have been reported. However, to the best of our knowledge, direct applications of enantioselective HPLC to aliphatic 3-hydroxyalkanoic acids without derivatization are completely missing in the panorama of scientific literature. There are only a few methods reported so far for HPLC enantiomer separation of 3-hydroxyalkanoic acids [28,29]. However, most of them require achiral precolumn derivatization of the hydroxyl group to carbamates [30,31] or ester derivatives [32] before separation on CSPs. The present study deals with the challenging topic of direct liquid chromatographic enantiomer separation of this class of compounds. Three aliphatic 3-hydroxyalkanoic acids (Fig. 1) were selected as test compounds. The separations were performed with two ion-exchange-based CSPs: one containing the tert-butyl carbamoyl quinine selector motif (Chiralpak QN-AX, CSP 1, Fig. 2) [33–35], and the other carrying the relatively new zwitterionic variant based on trans-2-aminocyclohexanesulfonic acid-decorated
quinine carbamate (Chiralpak ZWIX(+), CSP 2, Fig. 2) [36–38]. The effect of the primary experimental variables has been investigated to achieve optimal separations and allow to derive information on the retention mechanism. 2. Experimental 2.1. Materials All the reagents used were of analytical grade. Acetonitrile (ACN), methanol (MeOH), ethanol (EtOH) and 2-propanol (iPrOH) were purchased from Sigma–Aldrich (Munich, Germany). Acetic acid (AcOH) was purchased from Honeywell Riedel-de Haën (Hannover, Germany). The racemic mixtures of 3-hydroxybutyric acid, 3hydroxydecanoic acid and 3-hydroxymyristic acid, and the enantiomerically pure compound (R)-3-hydroxybutyric acid, were purchased from Sigma–Aldrich (Munich, Germany). Enantiomerically pure (R)-3-hydroxymyristic acid was obtained from Alberta Research Chemicals Inc. (Edmonton, Canada). The preparation of (R)-3-hydroxydecanoic acid was described elsewhere [30]. The enantiomeric elution order was assessed by injection of single enantiomers and non-racemic mixtures obtained by fortification of the racemate with single enantiomer. The analytes were dissolved in methanol at the approximate concentration of 0.5 mg/mL. 2.2. Instrumentation and chromatographic method The analytical HPLC measurements were made on a Merck Hitachi LaChrom 7000-series HPLC System (Tokyo, Japan) equipped with a D-7000 Interface, a diode array detector L-7455, a L-7200 Autosampler, and a L-6200A Intelligent Pump system. High Performance Liquid Chromatography System Manager (HSM) software from Hitachi allowed performing the acquisition and processing of the chromatographic data. Because of absence of suitable chromophoric moieties for sensitive UV detection in the sample structures, the HPLC system was connected with the Corona® Charged Aerosol Detector, CAD® (ESA Biosciences Inc., Chelmsford, U.S.A.). The nitrogen flow of the CAD was adjusted to 35 psi. The CHIRALPAK® QN-AX (150 mm × 4.6 mm ID, 120 A˚ pore size, 5 m particle diameter) column, and CHIRALPAK® ZWIX(+) and ZWIX(−) (150 mm × 4.0 mm ID, 120 A˚ pore size, 3 m particle diameter) columns were from Chiral Technologies Europe (Illkirch, France). Further, a prototype of CHIRALPAK® ZWIX(+) (250 mm × 4.0 mm ID, 120 A˚ pore size, 5 m particle diameter) was used for initial optimization experiments. Before use, all the employed mobile phases were degassed through sonication. The column was conditioned with the selected mobile phase at a flow rate of 1.0 mL/min for at least 20 min, before running the analysis. If not otherwise stated, the column temperature was fixed at 20 ◦ C with a column thermostat from W.O. Electronics (Langenzersdorf, Austria). 3. Results and discussion 3.1. Column selection
Fig. 2. Chiral stationary phases employed in this study.
Recently, we reported the direct enantiomer separation of underivatized lactic acid on chiral anion-exchange CSP 1 (Chiralpak QN-AX) [24] and of 2-hydroxyalkanoic acids including lactic acid, 2-hydroxybutyric acid, and glyceric acid on a noncommercial structural analog of CSP 1 with 2,6-diisopropylphenyl carbamoyl residue [39] (see also supplementary material). 3Hydroxyalkanoic acids, in contrast, were not separated with the employed conditions. Further, retention times were long on the
Please cite this article in press as: F. Ianni, et al., J. Chromatogr. A (2014), http://dx.doi.org/10.1016/j.chroma.2014.03.060
G Model CHROMA-355267; No. of Pages 8
ARTICLE IN PRESS F. Ianni et al. / J. Chromatogr. A xxx (2014) xxx–xxx
(b) Detector signal
anion-exchange type CSP 1 and acceleration by increase of counterion concentrations was precluded by its negative effect on UV detection sensitivity. It is to note that in ion exchange chromatography retention is controlled by the concentration of suitable counter ions. Thus, herein we compare the performance provided by two different ion-exchange based CSPs: besides the anionexchange CSP 1 (Fig. 2) [33–35], also a recently commercialized zwitterionic variant (CSP 2) in which both a cation-exchange (CX) and an anion-exchange (AX) motif are hybridized into a new selector (SO) unit (Fig. 2) [36–38]. In this case, the sulfonic acid moiety can serve as intramolecular counterion which competes with the 3-hydroxcarboxylic acid solutes for ionic interaction at the quininium anion-exchange site (possibly by a long range electrostatic interaction) with the consequence of reducing retention on anionic species. This intramolecular counterion effect allows for working with eluents having low counterion concentrations which is favourable for UV, CAD and MS detection sensitivity. For this comparison, a polar-organic (PO) elution mode was selected which was preferred over reversed-phase (RP) and normal-phase (NP) chromatographic elution conditions [40,41]. Previous experiments with 2-hydroxyalkanoic acids revealed that aqueous conditions such as in an RP elution mode are detrimental. In fact, the screening conditions with 2% water suggested by the column supplier as initial test did not give any resolution for current 3-hydroxyalkanoic acids. The problem may be that the hydroxyl group at the stereogenic centre is strongly solvated and thus cannot participate in a concerted simultaneous anionexchange driven multi-point interaction with the chiral selector. On the other hand, hydroxyalkanoic acids are very polar and poorly soluble in NP eluents. An important benefit of PO eluent systems resides in their ability of preserving strong electrostatic forces between complementary functional groups of SO and selectand (SA), here the 3-hydroxyalkanoic acids, such as H-bonding and long-range Coulomb attractions [42,43] while they are still balanced so that retention does not get too strong. In the current chromatographic systems, ionic interactions are supposed to be the dominant driving forces for binding of hydroxyalkanoic acids on the quinine carbamate anion-exchanger moiety competing with the counterions in the eluent for the sorption sites. On the other hand, H-bonds between the hydroxyl group of the selectand and complementary sites at the selector (carbamate moiety) are supposed to be active in case of hydroxyalkanoic acid enantiomer separation in PO mode as well and supportive for chiral distinction, like hydrophobic and/or van der Waals type interactions in case of the long chain 3-hydroxy fatty acids (compounds 2 and 3). The role of steric interactions is less obvious but they can also favourably contribute to stereoselectivity in PO media [44]. On the other hand, the PO mode turns out to be effective in simultaneously suppressing both hydrophobic and hydrophilic non-specific interactions, e.g. with support material and linker, which can have a negative effect on the observed enantioselectivity. Methanol and other alcohols were deemed to be too strong competitors, although much weaker than water, for hydrogen bonding of the hydroxyl group at the stereogenic centre of the analytes with the chiral selector and hence acetonitrile was chosen for the initial column screening as solvent system [37,40,45]. AcOH was used as the exchangeable counter-ion and displacer, respectively, to attenuate the otherwise too strong retention caused by the anion-exchange contribution [37,40,45]. Fig. 3 shows the chromatograms obtained for compound 2 (Fig. 1) with the screening eluent composed of ACN/AcOH-100/0.1 (v/v) (corresponding to a total concentration of 17.5 mM). As evident from the chromatographic traces, the zwitterionic CSP 2 (Fig. 3b) revealed greatly reduced retention compared to CSP 1 (in spite of the longer column) (Fig. 3a) due to the intramolecular counter-ion effect imposed by the sulfonic acid in the ZWIX
3
(a)
0
10
20
30
40
50
60
70
Time (min) Fig. 3. Chromatographic traces of compound 2, obtained with the two CSPs (a) anion-exchange CSP 1 (Chiralpak QN-AX, 5 m, 150 mm × 4 mm ID) and (b) zwitterionic CSP 2 (Chiralpak ZWIX(+), 5 m, 250 mm × 4 mm ID). Eluent: ACN/AcOH – 100/0.1 (v/v) (corresponding to a 17.5 mM total acid concentration); flow rate 1.0 mL/min; column temperature 20 ◦ C.
selector. It also slightly outperformed the anion-exchanger CSP 1 in terms of resolution. 3.2. Optimization of separation conditions Above separation indicated the potential of the ZWIX CSP 2 but for baseline separation conditions needed to be optimized. Compound 2 was selected for optimization by an OVAT (onevariable-at-time) approach and this tuning of the experimental conditions can be also useful to gain mechanistic information [34]. 3.2.1. Counterion concentration AcOH represents the counterion and displacer in this chromatographic system. When its concentration was reduced from 17.5 mM down to 4.4 mM, retention factors significantly increased in accordance with an anion-exchange process [34,37,40] (Fig. 4a). At the same time, enantioselectivity was nearly unchanged between 8 and 20 mM and increased slightly at very low concentrations. This behaviour strongly suggests a negligible effect of counterion on other tentative binding increments mainly responsible for stereoselectivity (namely H-bond and van der Waals/steric interactions) [46]. Likewise, resolution factors (R) remain more or less constant within the investigated eluent ionic strength (Fig. 4b). For CAD and MS detection, the more volatile formic acid was deemed to be advantageous and tested, but clearly revealed worse enantioselectivity and resolution. 3.2.2. MeOH content In order to study the role of H-bonds with the hydroxyl group for complex stabilization besides the primary ion-pairing mechanism, the impact of the MeOH content in the ACN-based eluent on the chromatographic performance was evaluated. MeOH which is a polar protic solvent is a stronger H-bond competitor than ACN, and thus, by increasing its concentration in ACN, retention of 2 decreased for both enantiomers. As expected, this clearly reveals that H-bonding plays a major role for retention of 3-hydroxyalkanoic acids on the ZWIX phase (Fig. 4c) [40,47]. Obviously, the higher the ACN content in the eluent the more significant this retention contribution, while addition of MeOH leads to stereoselective disruption of such interactions and an increase of enantioselectivity with the MeOH content (Fig. 4c). Further increase from 5 to 10% did not lead to improved resolution, but stronger baseline noise. At 50% MeOH resolution declined and vanished completely at 100% MeOH. Lower column efficiencies with ACNricher mobile phases along with smaller separation factors explain
Please cite this article in press as: F. Ianni, et al., J. Chromatogr. A (2014), http://dx.doi.org/10.1016/j.chroma.2014.03.060
G Model
ARTICLE IN PRESS
CHROMA-355267; No. of Pages 8
F. Ianni et al. / J. Chromatogr. A xxx (2014) xxx–xxx
4
k(R)
a)
k(S)
α
0.75
0.08
0.65
0.07
0.55 0.35
0.05
0.25
0.04
0.15
1.20 1.15 1.10
0.06
0.45 Log k
b)
1.05 R 1.00
Log α
0.95
0.03
0.90
-0.05
0.02
0.85
-0.15
0.01
0.80
0.05
0.64
0.94
k(R)
k(S)
α
1.25
0.07
1.05
0.06
0.85
d)
1.00 0.80 R
0.04 Log α
0.45
0.03
0.25 0.05
0.02
-0.15
0.01 0
2
1.40 1.20
0.05
0.65 Log k
0.60 0.40 0.20 0.00
5
0
% MeOH
e)
k(R)
k(S)
17 .50
Conc. AcOH (mM)
Log. Conc. AcOH (mM)
c)
8.75
4.40
1.24
2
5
% MeOH
α
0.80
0.060
f)
R
α
1.60
1.50 1.45 1.40 1.35 1.30 1.25 α 1.20 1.15 1.10 1.05 1.00
1.40
0.70 0.60
0.055
1.20 1.00
0.50 Log k 0.40
0.050 Log α
0.30
R 0.80 0.60
0.20
0.045
0.10
0.40 0.20
0.040
0.00 MeOH
EtOH
IPA
0.00 0.5
0.3
Type of alcoholic modifier
1.0
Eluent flow rate (mL/min)
Fig. 4. Influence of some experimental variables on retention, selectivity, and resolution of compound 2 on Chiralpak ZWIX(+) (5 m; 250 mm × 4 mm ID). Experimental conditions: (a) and (b) mobile phase, ACN/MeOH – 95/5 (v/v), total AcOH concentration varied in the range between 4.40 and 17.50 mM; flow rate, 1.0 mL/min; column temperature, 20 ◦ C; (c) and (d) mobile phase, ACN without and with MeOH (2% and 5%, v/v), total AcOH concentration 17.50 mM; flow rate, 1.0 mL/min; column temperature, 20 ◦ C; (e) mobile phase, ACN/type of alcoholic modifier – 98/2 (v/v), total AcOH concentration 17.50 mM; flow rate, 1.0 mL/min; column temperature 20 ◦ C; (f) mobile phase, ACN/MeOH – 95/5 (v/v), total AcOH concentration 8.75 mM; flow rate varied in the range between 0.3 and 1.0 mL/min; column temperature, 20 ◦ C.
the trend of decreasing R values with decrease of MeOH content (Fig. 4d).
3.2.3. Alcohol-type The replacement of MeOH with EtOH or iPrOH was then evaluated. The three alcohols were always used at 2% (v/v). As can be seen from Fig. 4e, retention factors get slightly reduced with the higher alcohols. Besides Coulombic attraction and H-bonding, hydrophobic interactions may contribute to retention to some extent. Accordingly, due to its ability to better disrupt hydrophobic interactions, the iPrOH containing eluent produced the highest elution strength [48]. The ␣-value was largest for EtOH, then MeOH followed by iPrOH [48] (Fig. 4e). Further experiments were made with MeOH as polar protic modifier because it provided the best resolution.
3.2.4. Temperature Temperature has multiple effects on the chromatographic behaviour of solutes in HPLC enantiomer separation. In spite of that, it is often neglected as variable to optimize separations. In current case, it turned out worthwhile to examine this parameter in order to achieve full baseline resolution. To do so, temperature was studied in the range between 10 and 40 ◦ C in steps of 10 ◦ C. The results are summarized in Tables 1 and 2. Adsorption thermodynamics can be analyzed by van’t Hoff equations (Eqs. (1) and (2))
ln k = −
S ◦ H ◦ + + ln RT R
Please cite this article in press as: F. Ianni, et al., J. Chromatogr. A (2014), http://dx.doi.org/10.1016/j.chroma.2014.03.060
(1)
G Model
ARTICLE IN PRESS
CHROMA-355267; No. of Pages 8
F. Ianni et al. / J. Chromatogr. A xxx (2014) xxx–xxx Table 1 Temperature dependence of retention factor of first eluting enantiomer (k1 ), separation factor (˛) and resolution (R) of 3-hydroxyalkanoic acids on Chiralpak ZWIX(+). Compound
k, ˛, R
k1 ˛ RS k1 ˛ RS k1 ˛ RS
1
2
3
Temperature (◦ C)
Elution order
10
20
30
40
1.53 1.07 0.61 1.71 1.18 1.40 1.80 1.16 1.23
1.32 1.07 0.58 1.46 1.16 1.34 1.52 1.15 1.16
1.12 1.07 0.56 1.21 1.14 1.21 1.22 1.14 1.14
0.91 1.07 0.59 0.99 1.12 1.08 0.98 1.12 1.06
R
