Comparison of performance of Chirobiotic T, T2 and TAG columns in the separation of ß²- and ß³-homoamino acids

Abstract

The enantiomers of eight unusual β²- and ß³-homoamino acids were separated on chiral stationary phases containing the macrocyclic glycopeptide antibiotic teicoplanin (Chirobiotic T or T2) or teicoplanin aglycone (Chirobiotic TAG) as chiral selectors. The effects of the organic modifier, the mobile phase composition and temperature on the separations were investigated. Linear van’t Hoff plots were observed in the studied temperature range, 280–318 K, and the changes in enthalpy, Δ(ΔH°), entropy, Δ(ΔS°), and free energy, Δ(ΔG°), were calculated. The values of the thermodynamic parameters depended on the nature of the selectors, the structures of the analytes, and especially the positions of the substituents on the analytes. A comparison of the separation performances of the macrocyclic glycopeptide stationary phases revealed that the Chirobiotic TAG column exhibited much better selectivity for ß²-homoamino acids, while the separation of β³-homoamino acid enantiomers was better on Chirobiotic T or T2. The elution sequence was determined in some cases and was observed to be R < S.

1. Introduction

ß-Amino acids are key components of numerous bioactive molecules, developmental pharmaceuticals, ß-peptides and peptidomimetics [1–3]. The wide-ranging utility of these compounds has led to increased attention being paid to their enantioselective syntheses [4,5], which require analytical methods for a check on the enantiopurity of the final products.

In the past decade, new types of chiral derivatizing agents and chiral stationary phases (CSPs) have been applied by D’Acquarica et al., Hyun et al. and Péter et al. for the high-performance liquid chromatographic (HPLC) enantioseparation of ß-amino acids. [6–23].

In all chromatographic modes, the selectivity and retention factors are controlled mainly by the concentrations and nature of the mobile phase components, together with other variables, such as the flow rate, and the pH of the mobile phase. Enantioselective retention mechanisms are sometimes influenced by temperature to a greater extent than are ordinary separations. This was noted first in chiral gas chromatography [24]. Additionally, it is known that there are achiral and chiral contributions to retention that can vary with a wide variety of experimental parameters [25–28]. Accordingly, the column temperature has often been optimized in chiral HPLC separations [29–35]. The dependence of the retention of an analyte on the temperature can be expressed by the van’t Hoff equation, which may be interpreted in terms of mechanistic aspects of chiral recognition:

$$\mathit{ln}k\prime =-\frac{\mathrm{\Delta }{H}^{°}}{\mathit{RT}}+\frac{\mathrm{\Delta }{S}^{°}}{R}+\ln \varphi$$ (1)

where ΔH° is the enthalpy of transfer of the solute from the mobile phase to the CSP, ΔS° is the entropy of transfer of the solute from the mobile phase to the CSP, R is the gas constant, T is the temperature and ϕ is the phase ratio of the column (the volume of the stationary phase divided by the volume of the mobile phase). This equation reveals that a plot of ln k′ vs 1/T is linear, with a slope of −ΔH°/R and an intercept of ΔS°/R + ln ϕ, if ΔH° is invariant with temperature. Since the value of ϕ is often not known, the ΔS°^* values [ΔS°^* = ΔS° + R ln ϕ] are generally used. Any uncertainty in the phase ratio affects the ΔS°^* values virtually equally.

The corresponding Δ(ΔH°) and Δ(ΔS°) values for the separated enantiomers can be determined from a modification of Eq. 1:

$$\mathit{ln}\alpha =-\frac{\mathrm{\Delta }(\mathrm{\Delta }{H}^{°})}{\mathit{RT}}+\frac{\mathrm{\Delta }(\mathrm{\Delta }{S}^{°})}{R}$$ (2)

where α is the selectivity factor (α = k₂′/k₁′).

The aim of the present work was to investigate the effectiveness of different macrocyclic glycopeptide-based CSPs for the separation of ß²- and ß³-homoamino acids. The direct HPLC enantioresolution of some ß³-homoamino acid analogs on Chirobiotic T and TAG CSPs was reported earlier by Péter et al. [9,11,12]. For comparison purposes, most of the separations were carried out at constant mobile phase compositions at different temperatures. The influence of specific structural features of the analytes and selectors and the effects of temperature on the retention will be discussed on the basis of the experimental data. The elution sequence was in some cases determined by spiking the racemate with an enantiomer with known absolute configuration.

2. Experimental

2.1. Chemicals and reagents

Racemic 3-amino-2-methylpropanoic acid (ß²-1), 2-aminomethyl-butanoic acid (ß²-2), 2-aminomethyl-3-methyl-butanoic acid (ß²-3) and 3-amino-2-benzylpropanoic acid (ß²-4) (Fig. 1) were prepared in two-step syntheses from methyl cyanoacetate, which in the first step was either alkylated [36] or condensed with aromatic aldehydes [37,38] and in the next step reduced. (S)-ß²-1 and (S)-ß²-3 were generous gifts from Prof. D. Tourwé (Vrije Universiteit Brussels, Belgium). The addition of benzylamine to ethyl ethacrylate furnished the corresponding N-benzylamino ester, which was transformed to the racemic ethyl ester of ß²-2 by catalytic hydrogenolysis. (±)-ß²-2 was resolved via Candida antarctica lipase-A-catalyzed N-acylation in tert-amyl alcohol with ethyl butanoate, yielding (R)-ß²-2 [39]. Racemic 3-aminobutanoic acid (ß³-1) and 3-aminopentanoic acid (ß³-2) (Fig. 1) were prepared from the corresponding α, β-unsaturated acids by benzylamine addition and subsequent debenzylation of the products with 20% metallic palladium on charcoal in a hydrogen atmosphere [40,41]. (R)-ß³-1) was prepared by the same method, but (R)-(+)-α-methylbenzylamine was used in the addition step instead of benzylamine [42]. 3-Amino-4-methylpentanoic acid (ß³-3) and 3-amino-3-phenylpropanoic acid (ß³-4) were synthetized from the corresponding aldehydes by a modification of the procedure of Rodionov and Malivinskaia [43]: the aldehydes were condensed with an equimolar amount of malonic acid in refluxing 96% ethanol in the presence of two equivalents of ammonium acetate [44,45].

Figure 1.

3-Amino-2-methylpropanoic acid (ß²-1), 2-aminomethylbutanoic acid (ß²-2), 2-aminomethyl-3-methylbutanoic acid (ß²-3), 3-amino-2-benzylpropanoic acid (ß²-4), 3-aminobutanoic acid (ß³-1), 3-aminopentanoic acid (ß³-2), 3-amino-4-methylpentanoic acid (ß³-3) and 3-amino-3-phenylpropanoic acid (ß³-4)

Acetonitrile (MeCN) and methanol (MeOH) of HPLC grade were purchased from Merck (Darmstadt, Germany). Triethylamine (TEA), glacial acetic acid (AcOH) and other reagents of analytical reagent grade were from Sigma-Aldrich (St. Louis, MO, USA). The Milli-Q water was further purified by filtration on a 0.45-μm filter, type HV, Millipore (Molsheim, France).

2.2. Apparatus and Chromatography

The HPLC measurements were carried out on a Waters HPLC system consisting of an M-600 low-pressure gradient pump, an M-996 photodiode-array detector and a Millenium³² Chromatography Manager data system; the alternative Waters Breeze system consisted of a 1525 binary pump, a 487 dual-channel absorbance detector, a 717 plus autosampler and Breeze data manager software (both systems from Waters Chromatography, Milford, MA, USA). Both chromatographic systems were equipped with Rheodyne Model 7125 injectors (Cotati, CA, USA) with 20-μl loops.

The macrocyclic glycopeptide-based stationary phases [46] used for analytical separation were teicoplanin-containing Chirobiotic T and T2 and teicoplanin aglycone-containing Chirobiotic TAG columns, 250 mm×4.6 mm I.D., 5-μm particle size (for each column) (Astec, Whippany, NJ, USA). The differences between the Chirobiotic T and T2 columns are that they are both on 5-μm particle size silica gel, but the Chirobiotic T uses a 120 Ängstrom pore size material and Chirobiotic T2 uses a 200 Ängstrom pore material. Also, the linkage chain on Chirobiotic T2 is approximately twice as long as that on Chirobiotic T. Hence, the coverage and spacing are different for the two phases. This manifests itself mainly in the form of steric interaction differences between the two columns.

The columns were thermostated in a Spark Mistral column thermostat (Spark Holland, Emmen, The Netherlands). The precision of the temperature adjustment was ± 0.1 °C.

3. Results and discussion

3.1. Separation of enantiomers on Chirobiotic T, T2 and TAG columns

The analytes in this study (Fig. 1) can be arranged into two classes. The ß²- and ß³-homoamino acids differ in the positions of their substituents. Compounds 1 (ß²-1 and ß³-1) can be considered as analogs of Ala, while compounds 3 (ß²-3 and ß³-3) as analogs of Val. In both classes, analogs 1, 2 and 3 bear alkyl groups, which may have different steric effects. This influences the hydrophobicity, bulkiness and rigidity of the molecules. Compounds 4 (ß²-4 and ß³-4; analogs of Phe) bear aromatic rings, potentially able to make steric/rigid or other interactions with the numerous aromatic rings of the CSPs. The relevant chromatographic data on these two classes of compounds are given in Table 1. All compounds in Table 1 were evaluated with a minimum of three mobile phases, the eluent composition varying in the interval 0.1% triethylammonium acetate (TEAA, pH 4.1)/MeOH = 10/90 – 90/10 (v/v). For comparison purposes, Table 1 lists the chromatographic results obtained when enantiomeric separation was achieved at the 0.1% TEAA (pH 4.1)/MeOH = 30/70 (v/v) mobile phase composition on the three different Chirobiotic columns. With the same mobile phase composition, the retention factors of the first-eluted enantiomers are lower on Chirobiotic T and T2 than on Chirobiotic TAG. Similar trends of higher k₁′ values on a Chirobiotic TAG than on a Chirobiotic T column were observed by Berthod et al. [47], D’Acquarica et al. [8] and Péter et al. [11,12,14,48,49]. Comparison of the Chirobiotic T and T2 CSPs revealed that the retention factors of the first-eluting enantiomers on Chirobiotic T were somewhat larger (Table 1). Slightly higher k₁′ values on Chirobiotic T2 than on Chirobiotic T were observed by Péter et al. [48] for ß³-homoamino acids possessing aryl or heteroaryl side-chains. ß³-Homoamino acids always exhibited lower k₁′ values than ß²-homoamino acids (the only exception was ß³-2 on Chirobiotic T).

Table 1.

Chromatographic data, retention factor of first-eluting enantiomer (k₁′), separation factor (α) and resolution (R_S) for the direct separation of the stereoisomers of ß²- and ß³-homoamino acids on macrocyclic glycopeptide CSPs at a mobile phase composition of 0.1%TEAA (pH 4.1)/MeOH = 30/70 (v/v)

Column	Analyte	k1′	α	RS
T	ß2-1 ß3-1	2.58 2.00	1.14 1.00	0.90 0.00
T2	ß2-1 ß3-1	2.26 1.52	1.00 1.00	0.00 0.00
TAG	ß2-1 ß3-1	4.66 3.30	1.07 1.00	0.30 0.00
T	ß2-2 ß3-2	2.34 2.50	1.15 1.04	1.20 0.50
T2	ß2-2 ß3-2	2.00 1.44	1.08 1.06	0.70 0.60
TAG	ß2-2 ß3-2	3.71 3.32	1.16 1.00	1.65 0.00
T	ß2-3 ß3-3	2.10 1.97	1.14 1.07	1.13 0.90
T2	ß2-3 ß3-3	1.63 1.29	1.14 1.06	0.95 0.55
TAG	ß2-3 ß3-3	3.36 3.23	1.20 1.00	1.30 0.00
T	ß2-4 ß3-4	2.35 2.10	1.32 1.00	3.50 0.00
T2	ß2-4 ß3-4	1.95 1.66	1.19 1.07	1.70 0.45
TAG	ß2-4 ß3-4	4.64 3.83	1.40 1.00	2.60 0.00

Columns, T, Chirobiotic T, T2, Chirobiotic T2, TAG, Chirobiotic TAG; mobile phase, 0.1% TEAA (pH 4.1)/MeOH = 30/70 (v/v); flow rate, 0.5 ml min⁻¹; detection, 205 nm; temperature, ambient

The effect of the MeOH content on the enantiomeric separation was investigated for ß²-1, ß³-1, ß²-3, ß³-3, ß²-4 and ß³-4 on the three stationary phases. For analogs with an alkyl side-chain (ß²-1 and ß³-1), when the MeOH content of the mobile phase was increased, the retention factor progressively increased. This was due to the reduced solubility of the polar amino acids in the MeOH-rich, more apolar mobile phase (Fig. 2). Similar results were obtained for analytes with a branched alkyl side-chain: ß²-3 and ß³-3 (data not shown). This behavior was observed on all three stationary phases. For analogs with an aromatic side-chain (ß²-4 and ß³-4), in most cases a U-shaped retention curve was observed. At higher water content, the retention factor increased again with increasing water content; this was due to enhanced hydrophobic interactions between the analyte and the CSP in the water-rich mobile phases (Fig. 3). In this study, as in an earlier one [48], the position of minimum of k′ differed somewhat for each compound. In the cases, when separations were occurred the α values increased very slightly, while the R_S values progressively increased with increasing MeOH content on all three CSPs (Figs 2 and 3).

Figure 2.

Effects of the MeOH content on the retention factor of the first-eluting enantiomer (k₁′), the separation factor (α) and the resolution (R_S) for ß²-1 and ß³-1 on Chirobiotic T, T2 and TAG CSPs.

Chromatographic conditions: mobile phase, 0.1% TEAA(pH 4.1)/MeOH = 30/70 (v/v); flow rate, 0.5 ml min⁻¹; detection, 205 nm, (■) k′; (●) α; (▲) R_S

Figure 3.

Effects of the MeOH content on the retention factor of the first-eluting enantiomer (k₁′), the separation factor (α) and the resolution (R_S) for ß²-4 and ß³-4 on Chirobiotic T, T2 and TAG CSPs.

Chromatographic conditions: mobile phase, 0.1% TEAA(pH 4.1)/MeOH = 30/70 (v/v); flow rate, 0.5 ml min⁻¹; detection, 205 nm; (■) k′; (●) α; (▲) R_S

A deeper analysis of the results reported in Table 1 shows that at a mobile phase composition of 0.1% TEAA(pH 4.1)/MeOH = 30/70 (v/v) on a given column, the k₁′ values for the same type of analyte (ß²- or ß³-homoamino acid) differed slightly. Overall, the k₁′ values were smallest on the Chirobiotic T2 column and largest on the Chirobiotic TAG column. For the appropriate proteinogenic α-amino acids, Ala, Val, Phe, also a similarity in k₁′ values were observed [49,50]. A possible explanation for this an enantioselective recognition mechanism in which the first-eluted enantiomers do not experience selective interactions and accordingly are eluted at similar retention times (the slight variances in k₁′ may originate from structural differences between the analytes).

The corresponding α and R_S values on the three CSPs for the ß²-homoamino acids were larger than those for the ß³-homoamino acids. However, the α and R_S values were lower than that of for the proteinogenic α-amino acids [49,50]. For the ß²-homoamino acids, the Chirobiotic TAG exhibited higher selectivities (the only exception was ß²-1), while for the ß³-homoamino acids the Chirobiotic T or T2 column did so (the only exception was ß³-1; Table 1). The enantiomers of ß²-1, ß³-1, ß²-2 and ß³-2 with short alkyl side-chains were retained on the CSPs, but they exhibited small α and R_S values. In reversed-phase mode one of the most important interaction between analyte and CSP is the hydrophobic interaction inside the “basket” of glycopeptide. The interactions of the short alkyl side-chains with the hydrophobic wall of the cavity of glycopeptides are probably too weak for effective chiral recognition, small α and R_S values were obtained. From the k₁′ values for ß²-3 and ß³-3 in Table 1, it is evident that the 2-propyl group has a negative role in the stabilization of the molecule–CSP complex, causing a slight decrease in k₁′, but α (and R_S) in most cases slightly improved relative to ß²-1 and ß³-1. A comparison of α and R_S values for Ala and Val also showed the negative role of the presence of 2-propyl group in Val; both α and R_S values were lower for Val comparing to Ala [49].

ß²-4 and ß³-4 possess aromatic rings capable of steric/rigid or other interactions with the CSP. The data in Table 1 for the ß²- and ß³-homoamino acids with aromatic side-chains indicate that (i) on a given column the k₁′ values did not differ considerably from those for amino acids with alkyl side-chains, and (ii) the α and R_S values were in some cases higher (ß³-4 was separable only on the Chirobiotic T2 column). Similarly one of the highest k′, α and R_S values were obtained for Phe in the case of proteinogenic α-amino acids [49,50].

A comparison of the results obtained on the Chirobiotic TAG column relative to Chirobiotic T may contribute to an understanding of the role of the pendant sugar moieties. To quantify the effects of the sugar units, the differences in enantioselective free energies between the two CSPs, Δ(ΔG°)_TAG – Δ(ΔG°)_T, were used. Δ(ΔG°) values were taken from Tables 4 and 5. The Δ(ΔG°)_TAG – Δ(ΔG°)_T, values were plotted as shown in Fig. 4. A negative number means that the stereoisomers are better separated on the aglycone CSP, and a positive number means that the stereoisomers are better separated on the native teicoplanin CSP, which contains the carbohydrate units. As can be seen in Fig. 4, the ß²-homoamino acid enantiomers are much better separated by the aglycone CSP (an exception was ß²-1), while for the ß³-homoamino acids the native teicoplanin CSP was favorable. It also indicates that the aglycone basket of the teicoplanin molecule is solely responsible for the enantiorecognition of most of the ß²-homoamino acid enantiomers. From the aspect of enantiomeric separations, the sugar moieties of the native teicoplanin may intervene in the chiral recognition process in at least three ways [47]: (i) the sugar units occupy space inside the basket; (ii) they block the possible interaction sites on the aglycone (phenolic hydroxy groups and an alcohol moiety); and (iii) they offer competing interaction sites, since the three sugars are themselves chiral and have hydroxy, ether and amido functional groups.

Table 4.

Thermodynamic parameters, ΔH°, ΔS°^*, ΔG°, Δ(ΔH°), Δ(ΔS°), Δ(ΔG°) and correlation coefficient (R²) of analytes measured on Chirobiotic T and T2 columns

Analyte		−ΔH (kJ mol−1)	−ΔS°* (J mol−1K−1)	Correlation coefficient, R2	−ΔG°*298K (kJ mol−1)	−Δ(ΔH°) (kJ mol−1)	−Δ(S°) (J mol−1K−1)	−Δ(ΔG) (kJ mol−1)
	Stereoisomer
Chirobiotic T
ß2-1	1	8.2	19.4	0.9977	2.4	2.2	6.3	0.35
	2	10.4	25.7	0.9974	2.7
ß3-1	1	8.7	17.6	0.9988	3.5	0.9	2.6	0.11
	2	9.6	20.2	0.9988	3.6
ß2-3	1	5.7	13.0	0.9963	1.8	1.6	4.3	0.32
	2	7.3	17.3	0.9971	2.2
ß3-3	1	8.5	18.7	0.9992	2.9	1.5	4.3	0.22
	2	10.0	23.0	0.9993	3.2
ß2-4	1	9.0	23.2	0.9988	2.1	4.2	12.0	0.67
	2	13.2	35.1	0.9993	2.8
ß3-4	1	10.3	25.2	0.9943	2.8	0.9	2.8	0.05
	2	11.2	28.2	0.9989	2.8
Chirobiotic T2
ß2-4	1	12.9	37.8	0.9989	1.7	1.4	3.2	0.47
	2	14.4	40.9	0.9991	2.2
ß3-4	1	10.8	28.5	0.9994	2.3	1.4	3.8	0.24
	2	12.1	32.2	0.9995	2.5

Column, Chirobiotic T; mobile phase, 0.1% TEAA (pH 4.1)/MeOH = 30/70 (v/v); ΔS°^* = ΔS° + ln Φ, where Φ is the phase ratio; R², correlation coefficient of van’t Hoff plot, ln k – 1/T curves

Table 5.

Thermodynamic parameters, ΔH°, ΔS°^*, ΔG°, Δ(ΔH°), Δ(ΔS°), Δ(ΔG°) and correlation coefficient (R²) of analytes measured on Chirobiotic TAG column

Analyte		−ΔH (kJ mol−1)	−ΔS°* (J mol−1K−1)	Correlation coefficient, R2	−ΔG*298K (kJ mol−1)	−Δ(ΔH°) (kJ mol−1)	−Δ(S°) (J mol−1K−1)	−Δ(ΔG) (kJ mol−1)
	Stereoisomer
ß2-1	1	11.5	25.7	0.9980	3.8	1.1	3.3	0.17
	2	12.6	29.0	0.9982	4.0
ß3-1	1	10.9	20.8	0.9953	4.7	0.7	2.0	0.12
	2	12.4	25.6	0.9919	4.8
ß2-3	1	7.7	15.9	0.9969	3.0	1.9	4.6	0.49
	2	9.6	20.5	0.9979	3.5
ß3-3	1	10.9	23.5	0.9966	3.9	0.0	0.0	0
	2	10.9	23.5	0.9966	3.9
ß2-4	1	13.0	30.9	0.9933	3.8	4.5	12.5	0.8
	2	17.5	43.3	0.9949	4.6
ß3-4	1	10.1	20.7	0.9954	4.0	0.0	0.0	0
	2	10.1	20.7	0.9954	4.0

Column, Chirobiotic TAG; mobile phase, 0.1% TEAA (pH 4.1)/MeOH = 30/70 (v/v); ΔS°^* = ΔS° + ln Φ, where Φ is the phase ratio; R², correlation coefficient of van’t Hoff plot, ln k – 1/T curves

Figure 4.

Enantioselectivity free energy differences, Δ(ΔG°)_TAG – Δ(ΔG°)_T, between aglycone and native teicoplanin CSPs.

Chromatographic conditions: mobile phase, 0.1% TEAA(pH 4.1)/MeOH = 30/70 (v/v); flow rate, 0.5 ml min⁻¹; detection, 205 nm

In general, it seems that the steric hindrance effect of the sugar moieties predominates for the ß²-homoamino acids, while for the ß³-homoamino acids sugar units support the enantiorecognition. Better enantioseparation of the enantiomers was observed for α-amino acids and cyclic-ß-amino acids on the teicoplanin aglycone phase [14,47,48], whereas the ß³-homoamino acids were better separated on the native teicoplanin CSP [12,51].

The elution sequence was determined for ß²-1, ß²-2, ß²-3 and ß³-1, and in all cases was observed to be R < S.

3.2. Effects of temperature and thermodynamic parameters

In order to investigate the effects of temperature on the chromatographic parameters, a variable-temperature study was carried out between 280 and 318 K. For ß²-1, ß²-3, ß²-4, ß³-1, ß³-3 and ß³-4, the chromatographic data were measured and calculated in 10 K increments over the temperature range studied (Tables 2 and 3). A comparison of the retention factors in Tables 2 and 3 shows that all of the recorded values decrease with increasing temperature. It is evident that an increase in separation temperature lowers the separation factor, α, but it may also improve the peak symmetry. The dependence of the resolution, R_S, on temperature, however, is more complex. The data listed in Tables 2 and 3 show that R_S for ß²-3, ß³-1 and ß³-4 progressively decreased with increasing temperature on all three columns, while ß²-1, ß²-4 and ß³-3 exhibited a maximum R_S with increasing temperature. At higher temperatures, R_S decreased due to the smaller α, while the decrease in R_S at lower temperatures may be explained by the dominating effect of the decreased column efficiency. Since the effect of temperature on the separation was more complex, an extensive study dealing with the thermodynamics of enantiomer separation was carried out.

Table 2.

Retention factor of first-eluting enantiomer (k₁′), separation factor (α) and resolution (R_S) of enantiomers of ß²-homoamino acids as a function of temperature

Column	k1′, α, RS	Temperature (K)
		280	288	298	308	318
ß2-1
T	k1′	3.26	2.95	2.58	2.36	2.15
	α	1.22	1.19	1.15	1.12	1.09
	RS	1.15	1.80	1.40	1.30	1.05
T2	k1′	2.90	2.66	2.26	1.90	1.67
	α	1.21	1.16	1.11	1.07	1.03
	RS	0.40	0.60	0.65	0.40	<0.40
TAG	k1′	6.30	5.36	4.66	3.96	3.46
	α	1.10	1.09	1.07	1.05	1.05
	RS	0.80	0.85	0.70	0.60	0.55
ß2-3
T	k1′	2.43	2.23	2.10	1.95	1.80
	α	1.20	1.17	1.14	1.12	1.10
	RS	1.70	1.60	1.45	1.25	0.75
T2	k1′	2.05	1.79	1.63	1.43	1.31
	α	1.17	1.17	1.15	1.15	1.14
	RS	1.70	1.65	1.60	1.55	1.45
TAG	k1′	4.11	3.72	3.36	2.98	2.78
	α	1.28	1.26	1.22	1.19	1.17
	RS	3.20	3.10	3.00	1.75	1.50
ß2-4
T	k1′	2.94	2.64	2.35	2.05	1.85
	α	1.47	1.39	1.31	1.23	1.18
	RS	3.00	3.10	2.50	2.05	1.70
T2	k1′	2.75	2.35	1.95	1.64	1.43
	α	1.27	1.23	1.21	1.19	1.17
	RS	1.55	1.65	1.65	1.70	1.55
TAG	k1′	6.61	5.37	4.64	3.83	3.37
	α	1.57	1.47	1.38	1.31	1.24
	RS	2.50	2.80	3.10	3.05	2.95

Columns, T, Chirobiotic T, T2, Chirobiotic T2, TAG, Chirobiotic TAG; mobile phase, 0.1% TEAA (pH 4.1)/MeOH = 30/70 (v/v); flow rate, 0.5 ml min⁻¹; detection, 205 nm

Table 3.

Retention factor of first-eluting enantiomer (k₁’), separation factor (α) and resolution (R_S) of enantiomers of ß³-homoamino acids as a function of temperature

Column	k1′, α, RS	Temperature (K)
		280	288	298	308	318
ß3-1
T	k1′	5.10	4.53	4.02	3.59	3.26
	α	1.07	1.06	1.04	1.03	1.02
	RS	0.70	0.65	0.45	0.30	0.15
T2	k1′	4.32	3.72	3.11	2.70	2.38
	α	1.09	1.08	1.07	1.06	1.05
	RS	0.75	0.70	0.65	0.60	0.55
TAG	k1′	9.05	7.78	6.64	5.97	5.08
	α	1.07	1.06	1.00	1.00	1.00
	RS	0.55	0.45	0.00	0.00	0.00
ß3-3
T	k1′	4.07	3.72	3.28	2.92	2.64
	α	1.14	1.11	1.09	1.07	1.06
	RS	1.16	1.32	1.10	0.90	0.70
T2	k1′	2.79	2.39	2.05	1.82	1.63
	α	1.12	1.11	1.10	1.08	1.07
	RS	1.05	1.15	1.00	0.85	0.75
TAG	k1′	6.36	5.74	4.82	4.26	3.64
	α	1.00	1.00	1.00	1.00	1.00
	RS	0.00	0.00	0.00	0.00	0.00
ß3-4
T	k1′	3.93	3.47	3.10	2.68	2.30
	α	1.04	1.03	1.00	1.00	1.00
	RS	0.30	0.20	0.00	0.00	0.00
T2	k1′	3.33	2.94	2.49	2.18	1.92
	α	1.14	1.12	1.10	1.08	1.06
	RS	1.15	1.10	1.05	0.90	0.70
TAG	k1′	6.40	5.80	4.87	4.28	3.87
	α	1.00	1.00	1.00	1.00	1.00
	RS	0.00	0.00	0.00	0.00	0.00

Columns, T, Chirobiotic T, T2, Chirobiotic T2, TAG, Chirobiotic TAG; mobile phase, 0.1% TEAA (pH 4.1)/MeOH = 10/90 (v/v); flow rate, 0.5 ml min⁻¹; detection, 205 nm

In order to calculate the thermodynamic parameters and to acquire information of value for an understanding of the enantiomeric retention on these CSPs, van’t Hoff plots were constructed [Eq. (1)]. The ΔH° and ΔS°* values calculated from the slopes and intercepts of the plots of Eq. (1) for the enantiomers on all three columns were negative (Tables 4 and 5; since the thermodynamic parameters on Chirobiotic T and T2 for ß²-1, ß³-1, ß²-3 and ß³-3 were very similar, only the results obtained for ß²-4 and ß³-4 on Chirobiotic T2 are depicted in Table 4). Further, the ΔH°, ΔS°^* and ΔG°^* values for the first-eluting enantiomer were always less negative than those for the second-eluting enantiomer. It was also observed that the ΔG°* values for the Chirobiotic T and T2 columns were less negative than those for the Chirobiotic TAG column.

As concerns the six analytes, ß²-3 exhibited the smallest and ß²-4 the largest −ΔH° and −ΔS°^* values. The branched side-chain in ß²-3 is possibly responsible for steric repulsive effects when in the cavity. The largest −ΔH° and –ΔS°^* values for ß²-4 show that the aromatic ring may promote the steric/rigid or polar interactions with the CSP.

The data on the changes in Δ(ΔH°), Δ(ΔS°) and Δ(ΔG°) are depicted in Tables 4 and 5. For the ß²-homoamino acid analogs the −Δ(ΔH°) values were more negative than those for the ß³-homoamino acid analogs on the three different columns, (the exception was ß³-3 on Chirobiotic T2; data not shown). The largest values were obtained for ß²-4 on the Chirobiotic T and TAG columns, indicating the importance of steric/rigid effects or polar interactions. The −Δ(H°) value on Chirobiotic T2 was less for ß²-4 than for ß²-1 (data not shown). The linkage chain on Chirobiotic T2 is approximately twice as long as that on Chirobiotic T, and the coverage and spacing will be different for the two columns. The increased hydrophobicity of Chirobiotic T2 may improve the interaction between an amino acid with an alkyl side-chain and the CSP. The small negative Δ(ΔH°) values [−Δ(ΔH°) < 1.5 kJ mol⁻¹] for the ß³-homoamino acids on all three columns indicate that the enantiorecognition is relatively unfavorable enthalpically.

The trends in the change in −Δ(ΔS°) demonstrated that in most cases ß²-homoamino acids exhibited more negative Δ(ΔS°) values than did the ß³-homoamino acids; exceptions were ß³-3 and ß³-4 on the Chirobiotic T2 column. It was also observed that the Δ(ΔS°) values were more negative for the amino acids possessing an aromatic side chain than for those with an alkyl side-chain (Tables 4 and 5).

The data on the changes in −Δ(ΔG°) indicate that enantioseparation on macrocyclic glycopeptide CSPs is more favorable for ß²-homoamino acids (Tables 4 and 5). The largest negative Δ(ΔG°) values for ß²-4 suggest that the steric/rigid or polar interactions induce highly efficient binding to the selector. The big difference in −Δ(ΔG°) values on Chirobiotic TAG column for ß²-1 and ß²-3 may be explained by the difference in steric and hydrophobic effect of the alkyl side-chains. On the basis of the −Δ(G°) values, it can be stated that, of the three columns, Chirobiotic TAG (without sugar units) may promote the interactions of the enantiomers of ß²-homoamino acids, while for ß³-homoamino acids Chirobiotic T with sugar moieties seems to be more favorable. However, the −Δ(ΔG°) values also indicate that the enantioseparation of amino acids with alkyl side-chains is more favorable on native teicoplanin, while for amino acids possessing an aromatic side-chain the teicoplanin aglycone is favored.

In summary, complex formation that involved multiple intermolecular interactions was generally exothermic, and the corresponding entropic contribution was also negative. As is evident from the Δ(ΔG°) data listed in Tables 4 and 5, the separations for all the investigated analytes on this CSP were enthalpically favored.

4. Concluding remarks

The enantioseparation of ß²- and ß³-homoamino acid analogs was investigated by using macrocyclic glycopeptide-based CSPs, i.e. Chirobiotic T, T2 and TAG columns, at either subambient or elevated temperatures. The separations could be accomplished in reversed-phase mode with 0.1% TEAA(pH 4.1)/MeOH mobile phases of various compositions and at different temperatures. Linear van’t Hoff plots were observed in the studied temperature range, 280–318 K, and the apparent changes in enthalpy, ΔH°, entropy, ΔS°, and free energy, ΔG°, were calculated. The values of the thermodynamic parameters depended on the structures of the compounds and the CSPs applied. The elution sequence was determined in some cases. The macrocyclic glycopeptide-based CSPs for α- and β²-homoamino acids complement one another: if enantiomers exhibit partial resolution on one CSP, they would probably be baseline-resolved on the other macrocyclic glycopeptide-based CSP. However, this statement is not always valid for β³-homoamino acids.