The separation of the enantiomers of diquats by capillary electrophoresis using randomly sulfated cyclodextrins as chiral selectors

Abstract

Diquats, derivatives of the widely used herbicide diquat, represent a new class of functional organic molecules. A combination of their special electrochemical properties and axial chirality could potentially result in their important applications in supramolecular chemistry, chiral catalysis, and chiral analysis. However, prior to their practical applications, the diquats have to be prepared in enantiomerically pure forms and the enantiomeric purity of their P- and M-isomers has to be checked. Hence, a chiral capillary electrophoresis (CE) method has been developed and applied for separation of P- and M-enantiomers of 11 new diquats. Fast and better than baseline CE separations of enantiomers of all 11 diquats within a short time 5–7 min were achieved using acidic buffer, 22 mM NaOH, 35 mM H₃PO₄, pH 2.5, as a background electrolyte, and 6 mM randomly sulfated α-, β-, and γ-cyclodextrins as chiral selectors. The most successful selector was sulfated γ-cyclodextrin, which baseline separated the enantiomers of all 11 diquats, followed by sulfated β-cyclodextrin and sulfated α-cyclodextrin, which baseline separated enantiomers of 10 and nine diquats, respectively. Using this method, a high enantiopurity degree of the isolated P- and M-enantiomers of three diquats with a defined absolute configuration was confirmed and their migration order was identified.

Article Related Abbreviations

CD

cyclodextrin

DQ

diquat

DS

degree of substitution

HPC

hydroxypropyl cellulose

S-CD

sulfated CD

1 INTRODUCTION

Diquats (DQs), derivatives of the widely used herbicide diquat, are N-heteroaromatic dications containing the structural motif of 2,2´-bipyridyl (see Figure 1). A DQ in the form of dibromide salt was first presented as a strong and fast acting herbicide in 1958 by the researchers at the Imperial Chemical Industries [1, 2]. The DQ is a non-selective contact herbicide acting as a desiccant and a defoliant, which is applied in the form of a spray only on the green part of a plant. It causes the degradation of cell membranes and decreases the photosynthetic activity of the plant [3]. DQ is no longer approved for use in the European Union, but it is still registered in the United States and in many other countries. The identification and determination of achiral DQ dibromide as a herbicide was performed by voltammetry [4] or HPLC-MS [5].

FIGURE 1

The interconversion of the P- and M-enantiomers of the bipyridinium skeleton of diquat.

DQs represent a new class of functional organic molecules. They are particularly interesting in two aspects. First, the presence of two quaternary nitrogen atoms incorporated into two aromatic cores creates an electron-deficient system with a large electron-acceptor potential and endows these molecules with special electrochemical redox properties [6, 7]. Depending on the oxidation-reduction conditions of the surrounding medium, the bipyridinium scaffold can occur in the form of a dication, a cation-radical or a neutral molecule [8-11]. The second interesting feature of DQs is their axial chirality. They exhibit P and M atropoisomeric conformations due to the noncoplanarity of the two pyridinium rings linked by the axis of chirality (see Figure 1). The DQs have a helical geometry; nevertheless, with the exception of a few derivatives prepared by the Lacour's group [8], their P and M isomers do not have sufficient configurational stability [12]. In spite of that, racemic DQs have already been used for various applications, for example, they have been employed as structural building blocks in supramolecular chemistry [13, 14], electron acceptors in light-harvesting chromophore-quencher systems [15], components for the synthesis of ion-pair charge-transfer complexes [16] and elements of nonlinear optical systems [16].

Recently, a series of new 11 DQ derivatives has been synthesized at our Institute [17]. For their structures, see Figure 2 and Table 1; their systematic chemical names and relative molecular masses are presented in Table S1. A combination of their special electrochemical properties and axial chirality makes them suitable for testing as candidates for important applications, such as the chiral catalysts of electron transfer, chiral selectors, redox indicators, dyes, herbicides, DNA intercalators, and components of molecular devices. However, for these practical applications, the DQs have to be prepared in an enantiomerically pure form; in addition, for the control of the enantiomeric purity of their P and M isomers, a suitable separation method has to be available.

FIGURE 2

The structures of diquats (DQs) and the capillary electrophoresis (CE) separations of their P- and M-enantiomers using 22 mM NaOH, 35 mM H₃PO₄, pH 2.5, as the background electrolyte (BGE) and S-α-, S-β-, and S-γ-CDs, all of which at 6 mM concentrations, as chiral selectors. For CE experimental conditions, see Section 2.3. For the DQ7 in the separation with S-β- and S-γ-CDs, the BGE was composed of 16 mM NaOH, 50 mM H₃PO₄, pH 2.1.

TABLE 1. The structures of the analyzed diquats 1−11 and selected parameters of the CE separations of their enantiomers using 22 mM NaOH, 35 mM H₃PO₄, pH 2.5, as the background electrolyte (BGE) and S-α-, S-β-, and S-γ-CDs, all of which at 6 mM concentration, as chiral selectors: The medians of (i) migration times (t_mig,1), (ii) effective electrophoretic mobilities (m_eff,1), and (iii) the numbers of the theoretical plates (N₁) of the first migrating enantiomers, (iv) resolutions (R_S), and (v) selectivity factors (α).

			Chiral selector
DQ	Structure	Separation parameter	S-α-CD	S-β-CD	S-γ-CD
		tmig,1 [min]	4.13	3.95	4.48
		meff,1 [10^−9 m^2V^−1s^−1]	−39.19	−40.76	−35.20
1		N1 × 10^−3	318.2	212.2	196.9
		RS [−]	7.65	7.59	3.78
		α [−]	1.05	1.05	1.04
		tmig,1 [min]	4.35	4.04	4.69
		meff,1 [10^−9 m^2V^−1s^−1]	−37.14	−39.89	−33.64
2		N1 × 10^−3	396.5	267.8	237.7
		RS [−]	6.30	10.21	4.28
		α [−]	1.05	1.07	1.03
		tmig,1 [min]	4.52	4.28	4.86
		meff,1 [10^−9 m^2V^−1s^−1]	−35.60	−37.55	−32.48
3		N1 × 10^−3	360.3	404.7	248.6
		RS [-]	5.75	7.51	4.09
		α [-]	1.04	1.05	1.04
		tmig,1 [min]	3.89	4.06	4.66
		meff,1 [10^−9 m^2V^−1s^−1]	−41.22	−39.48	−33.85
4		N1 × 10^−3	128.1	233.3	240.9
		RS [−]	3.02	11.04	7.27
		α [−]	1.03	1.08	1.07
		tmig,1 [min]	3.88	4.10	4.58
		meff,1 [10^−9 m^2V^−1s^−1]	−41.55	−39.06	−34.39
5		N1 × 10^−3	299.1	404.6	237.6
		RS [−]	4.21	7.70	2.82
		α [−]	1.03	1.05	1.03
		tmig,1 [min]	3.96	4.12	4.30
		meff,1 [10^−9 m^2V^−1s^−1]	−40.72	−38.84	−36.57
6		N1 × 10^−3	167.0	384.4	218.0
		RS [−]	2.78	10.11	12.47
		α [−]	1.02	1.07	1.12
		tmig,1 [min]	4.32	9.13	5.34
		meff,1 [10^−9 m^2V^−1s^−1]	−37.30	−17.58	−30.06
7^a		N1 × 10^−3	148.5	56.6	45.5
		RS [−]	27.50	34.49	59.86
		α [−]	1.38	1.89	3.24
		tmig,1 [min]	4.00	4.36	4.72
		meff,1 [10^−9 m^2V^−1s^−1]	−39.90	−36.62	−33.11
8		N1 × 10^−3	193.5	372.4	211.3
		RS [−]	0	1.29	3.40
		α [−]	1.00	1.01	1.03
		tmig,1 [min]	3.96	4.08	4.53
		meff,1 [10^−9 m^2V^−1s^−1]	−40.67	−38.93	−34.44
9		N1 × 10^−3	141.7	391.4	218.9
		RS [−]	2.67	15.17	12.10
		α [−]	1.02	1.11	1.12
		tmig,1 [min]	3.96	4.12	4.54
		meff,1 [10^−9 m^2V^−1s^−1]	−40.66	−38.49	−34.21
10		N1 × 10^−3	135.8	353.5	209.9
		RS [−]	2.58	14.30	11.73
		α [−]	1.02	1.11	1.12
		tmig,1 [min]	4.05	4.41	4.81
		meff,1 [10^−9 m^2V^−1s^−1]	−38.75	−36.00	−32.19
11		N1 × 10^−3	136.3	248.5	187.7
		RS [−]	0	7.40	8.95
		α [−]	1.00	1.07	1.13

• ^a The parameters of the DQ 7 were measured in the BGE composed of 16 mM NaOH, 50 mM H₃PO₄, pH 2.1.

For that reason, the aim of this work was to develop an analytical separation method that could resolve DQ enantiomers. One of the suitable methods for such a purpose is capillary electrophoresis (CE). Thanks to its several advantages, such as high separation efficiency, high mass (substance amount) sensitivity, small sample volume, low chemical consumption, and short analysis times, CE is widely applied for the chiral analysis of various types of compounds [18-22]. Enantioseparation is achieved by the addition of a chiral selector to the background electrolyte (BGE) used, in which the CE separation is performed. This selector interacts with the enantiomers of the analyte(s) with different intensity and forms with them transient diastereomeric complexes with different effective electrophoretic mobilities of the enantiomers, which enables their CE separation.

In chiral CE, various chiral selectors are used, including crown ethers, metal complexes, peptides, proteins, and oligo- and polysaccharides [23]. Among them, the native and especially the charged derivatized cyclodextrins (CDs) belong to the most often and the most successfully used chiral selectors [24-26]. CE using randomly sulfated α-, β-, and γ-CDs (S-α-CD, S-β-CD, and S-γ-CD, respectively) as chiral selectors was able to separate helquat enantiomers [27-29]. Because of the structural similarity between helquats [9] and diquats [17] (both of them being dicationic N-heteroaromatic electron-deficient molecules with axial/helical chirality and absorbing radiation in the low UV region), the aim of this work was to develop a CE method for the separation of the enantiomers of 11 DQ derivatives using the chiral selectors of the same class, that is, randomly highly sulfated α-, β-, and γ-CDs. The method should be optimized toward their fast separation with high resolution so that it could be applied for the enantiopurity analysis of preparatively resolved P- and M-isomers. When the absolute configuration of crystallized DQ enantiomers is available, the migration order of P- and M-isomers should be determined.

2 MATERIALS AND METHODS

2.1 Chemicals

All the chemicals used were of analytical grade purity. The chiral selectors, randomly sulfated α-cyclodextrin sodium salt (S-α-CD) (cat. no. 494542, the average degree of substitution (DS) 10.2) and randomly sulfated β-cyclodextrin sodium salt (S-β-CD) (cat. no. 389153, the average DS 9.5), were supplied by Sigma-Aldrich, and randomly sulfated γ-cyclodextrin sodium salt (S-γ-CD) (cat. no. A50924, average DS 11.8) was obtained from Beckman-Coulter as 20% m/v solution in water. The average DSs were not provided by the manufacturers but were determined in our earlier paper [28]. In all CE experiments, the uniform average concentration of the sulfated CDs (S-CDs) in the BGE was 6 mM. Phosphoric acid, acetic acid, dimethyl sulfoxide (DMSO), and hydroxypropyl cellulose (HPC) of the average M_r of 100,000 (cat. no. 191884) were obtained from Sigma-Aldrich. Sodium hydroxide and sulfuric acid were purchased from Penta and Tris was supplied by Serva. Deionized water was prepared using the Millipore Milli-Q Synthesis A10 system (Merck Ltd.).

2.2 Analyzed compounds

Racemic DQs (mixtures of P- and M-enantiomers) were synthesized as dibromide or ditriflate ((bis)trifluoromethanesulfonate) (TfO⁻) salts at our institute as described elsewhere [17]. The chemical structures of the DQs analyzed are shown in Figure 1 and Table 1. For their systematic names and relative molecular masses including counterions, see Table S1. The pure enantiomers P-4, M-7, and P-9 were prepared by the differential crystallization of their diastereomeric salts with an enantiopure anion, an R,R-isomer of dibenzoyl tartrate, as described in [17].

2.3 Capillary electrophoresis

All CE measurements were performed on a 7100 CE analyzer (Agilent Technologies). The fully automated apparatus is equipped with a UV-VIS spectrophotometric diode array detector (UV-VIS DAD) and is controlled by OpenLab CE G7100A software. CE data were evaluated using a Clarity chromatographic and electrophoretic station (DataApex), the data-analysis and graphing SW Origin 8.5 (OriginLab), the electrophoretic evaluation software CEval [30] (for the correction of migration times by the Haarhoff-Van der Linde function), and Excel 365 Pro Plus software (Microsoft).

CE separations took place in fused silica capillaries with the outer polyimide coating of 50/375 μm id/od (cat. no. TSP050375) supplied by Polymicro Technologies. The total capillary length was 385 mm and the effective length (from the injection end to the center of the detection window) was 300 mm. To suppress the analyte sorption, the inner capillary wall was permanently coated with thermally attached HPC using the procedure described elsewhere [31-33]. Briefly, a bare fused silica capillary of 2 m in length was rinsed with 5% (m/v) solution of HPC in distilled water using a purging device by applying a nitrogen pressure of 10 bar. The total volume of approximately 20 μL of HPC solution was flushed through the capillary. Subsequently, the solution was purged out of the capillary by applying a nitrogen pressure of 1 bar and dried overnight. With the constant flow of nitrogen at 1 bar, the capillary was heated in a GC oven. The temperature of 140°C was reached with ramp rate of 5°C/min and held for 40 min. Before use, the capillary was cut into the required lengths (4 × 50 cm) and a detection window was made by dissolving the outer polyimide coating by hot sulfuric acid (100°C).

The capillary cartridge was thermostated by circulating air to the temperature of 22°C. The separation voltage of -12 kV (with the cathode at the injection capillary end) generated the electric field intensity of 312 V/cm, the electric current of 45 μA and the input power (Joule heat) of 0.54 W, which caused a temperature increase inside the capillary by 3°C, as a results of which the mobilities were measured directly at the reference temperature of 25°C. The carousel with samples was thermostated by circulating water to the same temperature.

Before CE measurements, the HPC-coated capillary was rinsed by water for 5 min and by the BGE for another 5 min, in both cases using the pressure of 950 mbar. Between individual analyses, the capillary was rinsed by the BGE at the same pressure for 2 min.

The DQs were dissolved in water at 0.2 mM concentration and were introduced into the capillary hydrodynamically by the pressure of 7 mbar for 10−15 s. Each CE measurement was performed at least in triplicate. After each three analyses of the particular DQ, the electroosmotic flow (EOF) mobility was determined by the Williams and Vigh method [34] using dimethyl sulfoxide (DMSO) at 5 mg/mL concentration as an electroneutral marker of EOF. DQs were detected by a UV-VIS DAD set at the wavelength of 200 nm with 20 Hz frequency of data acquisition.

2.4 Calculations

The apparent electrophoretic mobility of the DQs (μ_app) was calculated using Equation (1):

$$$$\begin{equation}{\mu }_{app} = \frac{{{L}_{{\mathrm{tot}}}{L}_{{\mathrm{eff}}}}}{{{t}_{{\mathrm{mig}}}{U}_{{\mathrm{sep}}}}}\end{equation}$$$$(1)

where L_tot and L_eff are the total and effective capillary lengths, respectively, t_mig is the migration time of the DQ, and U_sep is the separation voltage.

The low EOF mobility of the neutral EOF marker (DMSO), μ_eof, in the HPC-coated capillaries was determined by Equation (2) [34]:

$$$$\begin{equation}{\mu }_{eo{\mathrm{f}}} = \frac{{\left( {{t}_{\mathrm{3}} - 2{t}_{\mathrm{2}} + {t}_{\mathrm{1}}} \right){L}_{{\mathrm{eff}}}{L}_{{\mathrm{tot}}}}}{{{U}_{{\mathrm{eof}}}\left( {{t}_{\mathrm{U}} - \frac{1}{2}{t}_{{\mathrm{ramp}}}} \right)\left( {{t}_{\mathrm{3}} + \frac{1}{2}{t}_{{\mathrm{inj}}}} \right)}}\end{equation}$$$$(2)

where t₁, t₂, and t₃ are the mobilization times of the EOF marker zones, U_eof is the voltage during the measurement of the EOF mobility, t_U is the time of voltage application, t_ramp is the time of the voltage ramp, and t_inj is the injection time of the EOF marker zones.

The effective mobility of the DQs (μ_eff) was obtained as the difference of μ_app and μ_eof:

$$$$\begin{equation}{\mu }_{eff} = {\mu }_{app} - {\mu }_{eof}\end{equation}$$$$(3)

The separation efficiency of the DQ expressed by the plate number (N) was calculated by Equation (4):

$$$$\begin{equation}N = 16{\left( {\frac{{{t}_{{\mathrm{mig}}}}}{w}} \right)}^2\end{equation}$$$$(4)

where t_mig and w are the migration time and the peak width at the baseline of the DQ analyte, respectively.

The resolution (R_S) of the separated P- and M-enantiomers of DQs was calculated using Equation (5):

$$$$\begin{equation}{R}_S = \frac{{2 \times \left| {{t}_{{\mathrm{mig,2}}} - {t}_{{\mathrm{mig,1}}}} \right|}}{{({w}_{\mathrm{1}} + {w}_{\mathrm{2}})}}\end{equation}$$$$(5)

where t_mig,1 and t_mig,2 are the migration times, and w₁ and w₂ are the peak widths at the baseline of the first- and second-migrating enantiomers, respectively.

The selectivity factor (α) was obtained as the ratio of the effective mobilities of the faster (μ_eff,1) and the slower (μ_eff,2) migrating enantiomer, respectively:

$$$$\begin{equation}\alpha = \frac{{{\mu }_{{\mathrm{eff,1}}}}}{{{\mu }_{{\mathrm{eff,2}}}}}\end{equation}$$$$(6)

The enantiomeric excess of enantiomer 1 (ee₁) was calculated using Equation (7):

$$$$\begin{equation}e{e}_1 = \frac{{\left| {{A}_{\mathrm{1}} - {A}_{\mathrm{2}}} \right|}}{{{A}_1 + {A}_2}}\end{equation}$$$$(7)

where A₁ and A₂ are the migration time normalized peak areas of the enantiomers 1 and 2, respectively.

3 RESULTS AND DISCUSSION

3.1 The selection of experimental conditions and the CE separation of DQ enantiomers

The selection of the experimental conditions for the CE separation of DQ enantiomers was based on their physicochemical properties. Thanks to the presence of two quaternary nitrogen atoms and double positive charge, the polar molecules of DQs are positively charged in the whole conventional pH range and well soluble in aqueous buffers. Due to their aromatic structure, the DQs strongly absorb light in the low-UV region and and can be directly detected by UV-absorption at the wavelength of 200 nm. Therefore, DQs are suitable cationic analytes for CE with UV detection in a wide pH range. Moreover, they are similar to helquats—structurally related compounds containing also two quaternary nitrogen atoms in their helical molecules. Helquat enantiomers were successfully separated by CE in the acidic sodium phosphate BGE (22 mM NaOH, 35 mM H₃PO₄, pH 2.5) using randomly sulfated S-α-, S-β-, and S-γ-CDs at 6 mM concentration as chiral stereoselectors [27-29]. For that reason, the same experimental conditions were first tested also for the enantioseparation of DQs. These conditions were also shown to be suitable for the enantioseparation of DQs.

The P- and M-enantiomers of DQs were resolved by the same separation mechanism as helquats. With anionic multiply negatively charged S-α-, S-β-, and S-γ-CDs, dicationic DQ enantiomers formed the transient diastereomeric complexes with the resulting negative charge and anodic mobility. In the HPC-coated FS capillary with close to zero EOF mobility and in the electric field with the reverse polarity of the separation voltage (the cathode at the injection capillary end), they migrated to the detector at the anodic end of the capillary. The more strongly bound enantiomer migrated first and the more weakly bound enantiomer migrated second. The strong electrostatic interactions between doubly positively charged DQs and multiply negatively charged S-CDs with the average DSs of 9.5−11.8 were responsible for the resulting anionic migration of the DQ-CD complexes formed and for the high intensity of the DQ-CD interactions. However, because of their space-nonspecific general way of action, Coulombic forces are not considered major stereo-discriminative interactions. What is crucial for chiral resolution are the spatial structures of the analyte and the selector as well as the weak non-covalent interactions between different groups of their molecules. For CDs as chiral selectors, the decisive forces for chiral resolution of enantiomers of given analyte include the formation of inclusion host–guest complexes, especially the degree of analyte embedding into the hydrophobic cavity, and the hydrogen bond interactions between the functional moieties of the analyte and the hydroxyl groups on the rims of the CD molecules.

All three types of S-CDs exhibited high chiral stereoselectivity for most of the DQ enantiomers. Many of the DQ enantiomers were baseline separated by means of complexation with all S-CDs within a short time of approximately 5 min (see Figure 2). The best selector was S-γ-CD, separating at the baseline level the enantiomers of all 11 DQs (100%), of which six (55%) had a resolution greater than 5. The second best selector was S-β-CD separating at the baseline level the enantiomers of 10 DQs (91%), all of which (100%) with the R_S > 5. The worst but still very good selector was S-α-CD with a relatively high success ratio: enantiomers of nine DQs (82%) were baseline separated, four of which (44%) with the Rs > 5. The high enantioresolution power of S-CDs for diquats was thus similar to their high enantioresolution power for the structurally related helquats. Their enantiomers were also baseline separated at least by one of the three S-CDs tested and they were mostly separated with relatively high R_S values as well. Interestingly, even the order of the success rate of the S-CDs for DQs and helquats was the same, that is, S-γ-CD > S-β-CD > S-α-CD.

For all DQ separations with all three types of S-CDs, the important separation parameters have been calculated and are presented in Table 1. In particular, these include the migration times, the effective mobilities (Equation (3)) and the separation efficiencies (Equation (4)) of the first-migrating enantiomers, and the resolutions (Equation (5)) and the selectivity factors (Equation (6)) of enantioseparations. The complete parameters for the CE separation of both enantiomers of all DQs with S-α-CD are presented in Table S2, with S-β-CD in Table S3, and with S-γ-CD in Table S4.

A comparison of the CE enantioseparations of DQs with three types of S-CDs shows that, with a few exceptions, the enantiomers of the DQs 1−6, and 8−11 migrated with the short migration times in the narrow range of 3.88−4.52 min, relatively high effective electrophoretic mobility values in the narrow range of (35.6−41.6) × 10⁻⁹ m²V⁻¹s⁻¹, high separation efficiency values of (128.1−404.6) × 10³ theoretical plates, high resolutions up to R_S = 15.1, and high selectivity factors up to α = 1.13 (see Table 1).

The enantiomers of the DQs 1−3 free of other aromatic rings than the original 2,2′-bipyridyl skeleton and differing only in the length of the aliphatic chain between quaternary nitrogen atoms migrated with the shortest migration times (the highest effective mobilities, see Table 1) in the BGE containing S-β-CD, thus forming the strongest complexes with only this chiral selector. The DQs 4−6 and 8−11, containing, in addition to the original 2,2′-bipyridyl skeleton, also another aromatic ring (benzene core) attached to the aliphatic chain, migrated with the shortest migration times (the highest effective mobility) in the BGE containing S-α-CD, hence being the most strongly complexed with this chiral selector. The high DS (10.2) and the smaller size of S-α-CD than those of S-β-CD and S-γ-CD could also contribute to the shortest migration times and the highest effective mobilities of the enantiomers of these DQs when complexed with S-α-CD than with S-β-CD or S-γ-CD.

With the exception of the DQ 7, all the other DQs migrated with the longest migration times (the lowest effective mobilities) in the BGE with S-γ-CD, thus forming complexes with the lowest strength but with the best success rate concerning the enantioresolution of all DQs. Apparently, the stereospecificity of DQ-CD interactions was a result of not only the electrostatic attractive forces but also the hydrophobic interactions of DQs with the hydrophobic cavity of CDs and the hydrogen bond interaction of DQs with the hydroxyl groups on the rims of the CD molecules.

The exceptional behavior of the DQ 7 is caused by the presence of two carboxyl groups in its molecule. Their partial dissociation at pH 2.5 and thus the presence of the negative charge significantly reduced the positive charge of the basic skeleton of this DQ. It caused a significant decrease in the strength of its interaction with negatively charged CDs and resulted in extremely long migration times and low effective mobilities of the enantiomers of this DQ. Therefore, for the separation of the enantiomers of this DQ 7, the pH of the BGE was reduced to 2.1. This led to the lower dissociation degree of two carboxyl groups, a reduction of their partial negative charge, an increase in the total positive charge of these enantiomers, and the strength of their interactions with S-CDs and to their reasonable migration times in the BGE containing the S-CDs as chiral selectors.

3.2 The determination of the enantiopurity degree and migration order of isolated P- and M-enantiomers of diquats with known absolute configuration

From the 11 DQs, the absolute configurations of the pure enantiomers were available only for the P-enantiomers of the DQs 4 and 9 and M-enantiomer of the DQ 7. They were obtained by the resolution of racemic DQs based on (i) the anion exchange of an achiral bromide DQ counterion for an enantiopure counterion, an R,R-isomer of dibenzoyl tartrate ([Na][R,R-DBT]), (ii) the formation of the diastereomers [P-DQ][R,R-DBT]₂ and [M-DQ][R,R-DBT]₂, and (iii) different solubilities and/or crystallization abilities of the above two diastereomeric salts formed from the racemic DQ dication and enantiopure anion. The high enantiopurity degree expressed by the enantiomeric excess (ee) of the isolated P- and M-enantiomers of the DQs 4 and 9 was confirmed by their high ee values (96%–98%) determined from the peak areas in their CE analyses shown in Figure 3A, B using Equation (7).

FIGURE 3

Records of the capillary electrophoresis (CE) analyses of racemic mixtures (above), the P-enantiomers (center) and M-enantiomers (below) of the DQ 4 (A) and the DQ 9 (B) using 22 mM NaOH, 35 mM H₃PO₄, pH 2.5, as the background electrolyte (BGE) and 6 mM S-γ-CD as a chiral selector. For the CE experimental conditions, see Section 2.3.

The identification of the P- and M-enantiomers of the three DQs (4, 7, and 9) was performed by the CE analysis of the 1:1 mixed samples of 0.2 mM racemic DQs and a 1 mM solution of their pure P- or M-enantiomers (see Figure 4A–C). Therefore, the peak with greater height and/or a larger area was assigned to the pure enantiomer because it was present in the mixed sample in a higher concentration. As follows from Figure 4A,C, the P-enantiomers of DQs 4 and 9 migrated in the presence of all tree types of S-CDs as the first, that is, the faster (more strongly bound), enantiomer. On the other hand, Figure 4B shows that the P-enantiomer of the DQ 7 migrated first only in the presence of S-α-CD and S-β-CD; in the BGE containing S-γ-CD, the migration order was changed and the M-enantiomer migrated the first. Obviously, the migration order is not identical and depends on the structures of both complexing partners, DQs as well as S-CDs. For the identification of the P- and M-enantiomers of DQ 7, the BGE with pH 2.5 was applied for the CE separation of these isomers with S-α-CD, whereas the BGE with pH 2.1 was applied for their separation via complexation with S-β-CD and S-γ-CD in order to prevent their long migration times as described above.

FIGURE 4

Records of the capillary electrophoresis (CE) analyses for the identification of the P- and M-enantiomers of the DQs 4 (A), 7 (B), and 9 (C) using 22 mM NaOH, 35 mM H₃PO₄, pH 2.5, as the background electrolyte (BGE), and S-α-, S-β-, and S-γ-CDs, all of which at 6 mM concentrations, as chiral selectors. For CE experimental conditions, see Section 2.3. For the DQ7 in the separation with S-β- and S-γ-CDs, the BGE was composed of 16 mM NaOH, 50 mM H₃PO₄, pH 2.1.

4 CONCLUDING REMARKS

The CE using acidic phosphate buffers (22 mM NaOH, 35 mM H₃PO₄, pH 2.5, or 16 mM NaOH, 50 mM H₃PO₄, pH 2.1) as BGEs and randomly sulfated CDs, S-α-, S-β-, and S-γ-CDs, all of which at 6 mM concentrations, as chiral selectors has proven to be a powerful technique for enantioseparation of diquats (DQs). The P- and M-enantiomers of all 11 newly synthesized DQs have been separated with better than baseline resolution within a short time of 5−7 min. For the isolated P- and M-enantiomers of DQ 4 and DQ 9, a high enantiopurity degree has been confirmed by the high values of enantiomeric excess determined by CE. For three DQs (4, 7 and 9) with the determined absolute configuration of the isolated P- or M-enantiomers, the migration order of these enantiomers has been identified. The method is ready to be applied as a control analytical method for the determination of the enantiopurity of the other newly prepared P- and M-enantiomers of the remaining DQs, the preparative resolution of which has not yet been achieved. The advantage of the method is that it needs only very small amounts of DQ analytes (a few microliters of solutions at submillimolar concentration, from which only a few nanoliters are injected per analysis) as well as chiral selectors (a few milliliters of S-CD solutions in the BGE at millimolar concentration).

ACKNOWLEDGMENTS

The work was financially supported by the Czech Science Foundation, project no. 20–03899S, and by the Czech Academy of Sciences, research project no. RVO 61388963.

CONFLICT OF INTEREST STATEMENT

The authors declare no conflicts of interest.

Open Research

DATA AVAILABILITY STATEMENT

Data are available on request from the authors.