Enantioselective discrimination and analysis of chiral compounds is prerequisite in separation sciences and has become an attractive and critically important topic especially in pharmaceutical field because of the importance of chirality in living organisms.1,2) Pharmaceuticals constitute of large varieties of compounds where more than 50% in use are chiral compounds often marketed as drugs in racemic forms.3) Thus, the safety and quality of chiral drugs requires a convenient and efficient analytical platform that can prepare enantiopure compounds so that the administered enantiomerically pure drugs fit properly to the target or receptor site for optimal biological activities.1) Chiral amines, α-amino acids, and their ester derivatives are very common and important chiral intermediates employed in the pharmaceutical, agrochemical and chemical industry as chiral building blocks that act as scaffolds in the synthesis of new pharmaceuticals and biologically active molecules.4,5) Enantiomer resolution of chiral compounds by high performance liquid chromatography (HPLC) employing chiral stationary phases (CSPs) is recognized as one of the most effective technique to determine enantiomeric composition and to obtain optically pure chiral molecules.6,7) Also, the critical factor for the successful enantiomeric separation and resolution is the right choice of CSP. Among numerous commercially available CSPs, phenylcarbamates derivatives of amylose or cellulose polysaccharides are one of the most versatile and powerful chiral selectors because of their high chiral recognition ability and reproducibility and extensively used for chiral resolution of various structurally different chiral molecules.8,9) Previously, in our group, several researches were performed for enantiomeric resolution of chiral amines, amino acids as well as their esters as 7-nitro-1,2,3-benzoxadiazole, benzophenone imine, 9-anthraldimine and fluorenylmethoxycarbonyl derivatives under normal phase HPLC using polysaccharide based CSPs.10-13) A fluorogenic derivatizing agent, 4-chloro-7-nitro-1,2,3-benzoxadiazole (NBDCl), was employed to obtain good interaction sites as well as to increase the detection sensitivity for investigated analytes.14) This paper describes development of selective and validated normal phase HPLC enantiomer separation method of chiral amines and α-amino acid esters as NBD derivatives on polysaccharide based CSPs under simultaneous ultraviolet (UV) and fluorescence (FL) detection.
Enantioselective separations were performed on 1100 series HPLC system from Agilent Technologies (Palo Alto, CA, USA). The chromatographic system was equipped with a solvent degasser, a G1310A isocratic pump, an autosampler, a thermostatic column compartment, a simultaneous multiwavelength G1315A ultraviolet, and a HP1046A fluorescence detector. Hewlett-Packard ChemStation chromatographic data software was used for instrument control, data acquisition and analysis. Ten amylose or cellulose phenylcarbamates derived chiral columns categorized as two types (covalently bonded and coated type) were used in the present study. The six covalently bonded Chiralpak IA, Chiralpak IB, Chiralpak IC, Chiralpak ID, Chiralpak IE and Chiralpak IF columns were supplied by Daicel Company (Tokyo, Japan). The four coated type columns, Chiralcel OD-H and Chiralpak AD-H were obtained from Daicel Company (Tokyo, Japan) whereas Lux Cellulose-1 and Lux Amylose-1 were purchased from Phenomenex (Torrance, CA, USA) respectively. All chiral columns used were 4.6×250 mm in size and packed with 5μm particles.
Four studied chiral amines and (S)-amino acid esters as well as derivatizing agent, 4-chloro-7-nitro-2,1,3-benzooxadiazole (NBDCl), and sodium bicarbonate were purchased from Across Organics (Fair Lawn, NJ, USA), Alfa Aesar (Haverhill, MA, USA) and Sigma-Aldrich (St. Louis, MO, USA). The three racemic acid esters were prepared using conventional method.15) HPLC grade hexane, 2-propanol, ethanol, triethylamine and dimethylformamide (DMF) for the preparation of mobile phase and sample were obtained from Burdick & Jackson (Morristown, NJ, USA) or Daejung Chemical and Metals Co. Ltd. (Gyeonggido, South Korea).
For chiral amines, a new preparation method was employed by stirring chiral amines and NBD-Cl (1 equivalent) with excess triethylamine (10 equivalents) in DMF (5mL) at room temperature for 6 h.16) The NBD derivatives of all α-amino acid esters were prepared with excess of sodium bicarbonate (5 equivalents) and NBD-Cl (2 equivalents) in ethanol (5 mL) at room temperature for 6 h and sonicated at 50°C for 30 to 60 min. Mild reaction conditions were preferred to check the racemization of the derivatized samples. The resultant sample solution was then filtered and was further diluted to a proper concentration for the injection in normal HPLC.
Separation of enantiomer using chiral HPLC was performed at room temperature with a flow rate set at 1 mL/min. The employed isocratic mobile phase prepared with HPLC grade solvents consists of 10~30% 2-propanol/hexane (V/V). Simultaneous detection was performed at UV 310 nm and FL (excitation 470 nm and emission 530 nm). Developed liquid chromatographic method was validated in accordance with International Council for Harmonisation of Technical Requirements for Pharmaceuticals for Human Use (ICH) guidelines.17) Limit of detection (LOD) and limit of quantification (LOQ) at signal-to-noise ratios of 3:1 and 10:1, intra- and interday accuracy and precision as well as stability tests were performed to assess the analytical method.
As shown in Table 1 and 2, the enantioseparation of selected four chiral amines as NBD derivatives was performed by normal phase HPLC using ten CSPs of different nature, selector backbone and the position of substituents in tris phenylcarbamate moiety. For four chiral amines (Table 1 and 2), the performance of the covalently bonded Chiralpak IE derived from amylose tris(3,5-dichlorophenylcarbamate) and coated type Chiralcel OD-H derived from cellulose tris(3,5-dimethylphenylcarbamate) was greater than those of the other CSPs used and were able to resolve all four chiral amines to baseline separation. Interestingly, for three aliphatic amines (entries 1-3), the enantiomeric discrimination was highly successful on Chiralpak IE, Chiralcel OD-H and Lux Cellulose-1, while that of aromatic amine (entry 4) on Chiralpak IC and Chiralpak IF. The aromatic moiety might have impact on different chiral interaction with chiral selector of CSPs and found to favor Chiralpak IC and Chiralpak IF. Especially, amongst the examined six covalently bonded CSPs (Table 1), the chlorine substituted covalently bonded CSPs (Chiralpak IC, Chiralpak ID, Chiralpak IE and Chiralpak IF) exhibited good enantiomer separation and resolution of three NBD derivatized chiral amine analytes. On contrary, resolution on amylose tris(3,5-dimethylphenylcarbamate) derived CSPs (both coated and covalently bonded) (Chiralpak IA, Chiralpak AD-H and Lux Amylose-1) were not satisfactory except for one analyte (entry 2, Table 1 and 2). The resolving powers are significantly affected by the substituents of phenylcarbamate moiety which are electron withdrawing groups (chlorine) for Chiralpak IC, Chiralpak ID, Chiralpak IE and Chiralpak IF.6,18) Presumably, phenylcarbamate chiral selector derivatives having electron-withdrawing substituents such as chlorine could possess a higher resolving power and have favored the enantioselective interactions likely through enhanced π-π interactions with greater retention times (especially on Chiralpak IC and Chiralpak IE).6,18,19) Also, as noted in Table 1 and 2, the coated type CSPs showed better chiral discrimination ability than the covalently bonded type CSPs for the same chiral selectors. This chromatographic behavior could be caused by different accessibility of enantioselective interaction sites of the polymer backbone structure of the coated CSPs than of immobilized ones.20) Especially, the analyte of 1,3-dimethylbutylamine (entry 2, Table 2) showed the best enantioseparation (α=1.17~2.45) on four coated type CSPs. Interestingly, the elution order reversal of enantiomers of two analytes, the aliphatic amine of 3,3-dimethyl-2-butylamine (entry 1) and the aromatic amine of α-methylbenzylamine (entry 4) was shown on amylose tris(3,5-dichlorophenylcarbamate) derived Chiralpak IE and cellulose tris(3,5-dimethylphenylcarbamate) derived Chiralcel OD-H, which provided the great enantioselectivities in Table 1 and 2.19,21) It is considered that the kind of analyte, backbone differences (amylose or cellulose) and position of substituents of the selector moieties were involved in the chiral recognition pattern toward analytes.18,19) Table 3 and 4 show the enantiomeric discrimination results of three α-amino acids ethyl esters as NBD derivatives on six covalently bonded and four coated type CSPs. Compared to the chiral amines, the enhanced performance in terms of enantioselectivity and resolution of these α-amino acids ethyl esters was observed on ten phenylcarbamates derived CSPs. Almost all the analytes investigated in this study were resolved. Unlikely the chiral amines, for resolution of α-amino acids ethyl esters, hydrogen bonding interactions could occur between the carbonyl group of analytes and chiral selectors for the better chiral recognition during the separation process. Among six covalently bonded CSPs for these three analytes, Chiralpak IA having amylose tris(3,5-dimethylphenylcarbamate) as chiral selector showed the greatest separation and followed by the other covalently bonded CSPs (Chiralpak ID, Chiralpak IE and Chiralpak IF). On contrary, cellulose derived Chiralpak IC showed comparatively worst separation among the CSPs used. In case of the four coated type CSPs, the performance of Chiralpak AD-H and Lux Amylose-1 with the chiral selector of amylose tris(3,5-dimethylphenylcarbamate) was superior to that of Chiralcel OD-H and Lux Cellulose-1 with only difference in polymer backbone as cellulose tris(3,5-dimethylphenylcarbamate) chiral selector. Consequently, the greatest enantioselectivities were observed on Chiralpak AD-H and Lux Amylose-1 CSPs derived from amylose tris(3,5-dimethylphenylcarbamate) as chiral selector. For example, the highest enantioselectivities were observed on the Lux Amylose-1 CSP for leucine (α=2.87, Rs=12.90) and phenylglycine (α=2.48, Rs=13.50) analytes in Table 4. There is no doubt that the nature and structural differences between helical type amylose and sheet type cellulose based CSPs have an impact in the chiral recognition pattern toward analytes and it appears that the investigated analytes fit better to the amylose type CSPs than to the cellulose type CSPs.7,18) The separation performances of complementary CSPs having same chiral selector but different nature (covalently bonded and coated type) have been observed and were also compared. Comparison between covalently bonded (Chiralpak IA) in Table 3 and coated type (Chiralpak AD-H and Lux Amylose-1) CSPs in Table 4, which contains the same chiral selectors of amylose tris(3,5-dimethylphenylcarbamate) revealed higher separation factor and resolution in all cases on two coated CSPs than on the immobilized CSP. Similarly, the same achievements of two coated type CSPs Chiralcel OD-H and Lux Cellulose-1 (Table 4) in compare to covalently bonded Chiralpak IB (Table 3) with same chiral selector of cellulose tris(3,5-dimethylphenylcarbamate) were observed as these two CSPs exhibited better enantioseparation of α-amino acids ethyl esters as NBD derivatives. Between the effectiveness of immobilized and coated polysaccharide phases, in most cases, the employment of coated phases seems more advantageous but in general the covalent phases are preferred due to their high robustness.22) Identical elution order patterns were observed for all three analytes as the (S)-isomers were preferentially retained in seven CSPs. However, opposite trend was observed for chlorine substituted amylose based CSPs (Chiralpak ID, Chiralpak IE and Chiralpak IF, Table 3) as (R)-isomers were secondly eluted. It is very difficult to rationalize such reversal of elution orders rather it can be assumed that the introduction of electron withdrawing chloro group in phenylcarbamate moiety might lead to the different chiral interaction and recognition for enantiomeric separation and results in different elution order.18,19) Figure 2 shows the typical HPLC chromatograms for the enantiomeric separation of chiral amine and α-amino acid ester on the covalently bonded CSPs. The unreacted derivatizing agent (NBDCl) as starting material peak appeared under UV 310 nm (Fig. 2B), but was not found under fluorescence detection because NBDCl is fluorescence inactive. Thus, simple analytical chromatograms under high sensitive fluorescence detection show an advantage of this present analytical method.
Intra- and interday tests for accuracy and precision as well as LOD and LOQ determination were performed using (S)-3,3-dimethyl-2-butylamine and (S)-leucine ethyl ester as NBD derivatives on Chiralpak IE and Lux Amylose-1 with a mobile phase of 20% 2-propanol/hexane (V/V) to provide evidence that the developed method will show acceptable reproducibility and accuracy to be applicable for further purposes.17) Table 5 shows the validated data of three (S)-3,3-dimethyl-2-butylamine and (S)-leucine ethyl ester samples with enantiomeric purities of 98.6, 96.6 and 94.6% of the accuracy and precision experiment of the developed analytical method under FL detection. The intra- and interday accuracy for (S)-3,3-dimethyl-2-butylamine was found to be 99.92-101.46 and 100.61-101.78%, respectively, while intraand interday precision calculated in terms of relative standard deviation (% RSD) was 0.96-1.57 and 0.81-1.91%, respectively. On the other hand, intra- and interday accuracy assay of (S)-leucine ethyl ester were 100.05-102.12 and 101.25-102.73%, respectively, while the intra- and interday precision of the developed normal phase HPLC method were found to be 0.24-1.01% and 0.65-1.35%, respectively. The amount of (S)-3,3-dimethyl-2-butylamine as NBD derivative on Chiralpak IE for LOD and LOQ determination was found to be 2.5 and 7.00 ng, respectively. Similarly, the amount of (S)-leucine ethyl ester for LOD and LOQ determination on Lux Amylose-1 was 1.38 and 4.62 ng, respectively. Based on these obtained chromatographic results, this developed analytical method demonstrated to be quite sensitive, selective and suitable for the enantiomeric separation of chiral amines and α-amino acid ethyl esters. Moreover, stability test for chiral amine as NBD derivative stored at 4°C was performed. The optical purity data for the stability test determined on Chiralpak IC for (S)-α-methylbenzylamine as NBD derivative is shown in Table 6. During storage of 25 days at 4°C, no significant changes in its optical purity were found in this study.
As an application of the validated analytical method, enantiomeric purities of two commercially available analytes were measured by this developed analytical method. The enantiomeric impurities of (S)-α-methylbenzylamine and (S)-leucine ethyl ester were found to be 0.35 and 0.05%, respectively. Figure 3 shows the representative simultaneous UV and FL chromatograms of two analytes as NBD derivatives for the determination of enantiomeric purities.
Enhanced discrimination, selectivity and sensitivity of the NBD derivatives with the ease of their preparation provide chromatographers with a general and effective alternative for the separation of chiral amines and α-amino acid ethyl esters as NBD derivatives. Ten commercially available amylose and cellulose phenylcarbamate derived CSPs have been studied under the simultaneous UV and FL detection. Sometimes, only subtle changes in the structure of a chiral selector (phenylcarbamate moiety) may lead to a difference in the affinity for a particular pair of enantiomers and ultimately separation. The separation of the chiral amine enantiomers was more dependent on chlorine substituted phenylcarbamates derived CSPs whereas Chiralpak AD-H and Lux Amylose-1 (Amylose based) CSPs was found to be the more appropriate for the separation of α-amino acid ethyl esters in this study. Also, with validation and application of the present study, it clearly demonstrated that present analytical method to be highly feasible and reproducible and is suitable for determination of chiral purity of chiral amine and α-amino acid esters.
This above research work was supported by Basic Science Research Program through the National Research Foundation of Korea (NRF) funded by the Ministry of Education (2019R1I1A3A01059479).
The authors declare that there is no conflict of interest regarding any issues for the publication of this research work.