Type II CSPS

Another example of modeling the structure of this type of CSP is presented by Francotte and Wolf [47]. They prepared benzoylcellulose beads, in a pure polymeric form as a sorbent, for the chromatographic resolution of racemic compounds like benzylic alcohols and acetates of aliphatic alcohols and diols. Their experimental results implicated multiple interaction sites to be involved in the complexation. Rationalizing the interaction mechanism required a more systematic investigation of the factors influencing separations and, to address the structural features of the cellulose tribenzoate, they carried out molecular modeling with molecular mechanics. The key question being addressed is to what extent is the polysaccharide backbone exposed to small molecules when sterically encumbered benzoates are attached  

Most of the computational studies of Type II CSPs do not consider the CSP directly. Instead, regression models are constructed to explain how a set of probe molecules interact with the CSP. We now present selected examples from the literature illustrating the diversity of such computational methodology used by chemists to address how these CSPs work. 

The first example is from Isaksson, Wennerstrom and Wennerstrom [51] who considered analyte binding to cellulose triacetate (CTA). They used statistics to assess the relationships between chiral recognition and analyte symmetry. Rather than attempt to compute the actual binding constants they addressed how symmetry 

Chromatographic results capacity factors fc i and k 2 and separation factor a 

Two criteria seemed important for chiral separations. First, as had been noted by many authors, the shape of the analyte seems critical and second, all compounds containing an oxygen adjacent to the stereogenic center give rise to large k 2 values 

---

Table 5.8 Representative Enantiomeric Drugs Resolved on Type II CSPs...

Cellulose is also a chiral polymer consisting of chiral subunits, D-p-glucose, a number of chiral environments created within the polymer including the areas (or cavities) formed between adjacent glucose units, and the spaces (or charmels) between the polysaccharide chains. Chiral recognition can take place through attractive interactions between the solute and the D-p-glucose units or by inclusion into the cavities or channels. Both these interactions take place on the type II CSPs. The commercially available type II CSPs are listed in Table 2,... 

Most of the molecular modeling studies involving Type II CSPs, as illustrated above, do not directly involve computations of the analjrfe with the CSP to discern where and how chiral recognition takes place. The reason for this is, clearly, the lack of structural information about these polymeric CSPs. This is in contrast to modeling studies of Type I CSPs described in an earlier section of this chapter and to those computational studies of Type III CSPs discussed below. 

The favorable effect of the introduction of a carbamate moiety into the cinchonan selectors was already proven by the prototype cinchonan carbamate CSPs (type I and type II) (Figure 1.9) [30], which showed enhanced enantioselectivities and a widened application range as compared to the CSPs with native cinchona alkaloid selectors and those reported earlier in the literature. 

FIGURE 1.9 Selection of cinchonan carbamate CSPs that have been prepared in the course of selector optimization studies (type I prototype type II, O-9-linked thiol-silica supported prototype type III, C-ll-linked thiol-silica supported CSPs type IV, dimeric selectors). (Adapted from M. Lammerhofer and W. Lindner, J. Chromatogr. A, 741 33 (1996) W. Lindner et al., PCT/EP97/02888, US 6,313,247 B1 (1997) P. Franco et ah, J. Chromatogr. A, 869 111 (2000) C. Czerwenka et ah. Anal. Chem., 74 5658 (2002).)... 

Different classifications for the chiral CSPs have been described. They are based on the chemical structure of the chiral selectors and on the chiral recognition mechanism involved. In this chapter we will use a classification based mainly on the chemical structure of the selectors. The selectors are classified in three groups (i) CSPs with low-molecular-weight selectors, such as Pirkle type CSPs, ionic and ligand exchange CSPs, (ii) CSPs with macrocyclic selectors, such as CDs, crown-ethers and macrocyclic antibiotics, and (iii) CSPs with macromolecular selectors, such as polysaccharides, synthetic polymers, molecular imprinted polymers and proteins. These different types of CSPs, frequently used for the analysis of chiral pharmaceuticals, are discussed in more detail later. 

Type II. When the primary mechanism for the formation of the solute/CSP complex is through attractive interactions, but inclusion complexes also play an important role... 

The Type IV CSPs are used with aqueous mobile phases that contain copper(II) sulfate (CS). The usual starting concentration of CS is 0.25 mM and retention times can be manipulated by increasing or decreasing the CS concentration an increase in CS concentration usually shortens retention, whereas a decrease has the opposite effect. Concentrations as low as 0.05 mM and as high as 20.0 mM have been used. 

The surface of the adsorbent in a chiral stationary phase (CSP) often contains two different types of adsorption sites [103, 110, 111], type-I and type-II sites. In most chiral columns only type-II site is chiral selective, but the saturation capacity for most columns is very low resulting in that when the sample size is increased, the selective retention mechanism is rapidly overloaded and the chiral separation disappears. The type-II saturation capacity is also called the true chiral saturation capacity. 

The type-I sites have, in protein based CSPs, identical behavior toward the two enantiomers, and cannot distinguish between them. Many columns contain mostly type-I sites. On type-I sites all possible molecular interactions, between the analyte molecules and atoms or groups of atoms belonging to the adsorbent surface, take place. These interactions can originate from the nonchiral parts of the protein and/or from the adsorbent (silica) matrix. The energies of each interaction on type-I sites are small. The other type of adsorption sites have, in protein based CSPs, much higher adsorption energy and are enantioselective (chiral). These sites, type-II sites, are responsible for the enantiomeric separations. On most CSPs the type-II sites are relatively few. 

Type II. Where the solute-CSP complexes are formed by attractive interactions and through the inclusion into a chiral cavity or ravine as represented by some cellulose based CSPs. 

Bi-Langmuir adsorption isotherms of enantiomeric pairs and CSPs were determined to gain information on chiral mechanisms. In the few cases fully studied, it was found that the two isomers interacted with type I nonselective sites as well as with type II enantioselective sites [18]. The bi-Langmuir equation is expressed as ... 

The PO mode is a specific elution condition in HPLC enantiomer separation, which has received remarkable popularity especially for macrocyclic antibiotics CSPs and cyclodextrin-based CSPs. It is also applicable and often preferred over RP and NP modes for the separation of chiral acids on the cinchonan carbamate-type CSPs. The beneficial characteristics of the PO mode may arise from (i) the offset of nonspecific hydrophobic interactions, (ii) the faster elution speed, (iii) sometimes enhanced enan-tioselectivities, (iv) favorable peak shapes due to improved diffusive mass transfer in the intraparticulate pores, and last but not least, (v) less stress to the column, which may extend the column lifetime. Hence, it is rational to start separation attempts with such elution conditions. Typical eluents are composed of methanol, acetonitrile (ACN), or methanol-acetonitrile mixtures and to account for the ion-exchange retention mechanism the addition of a competitor acid that acts also as counterion (e.g., 0.5-2% glacial acetic acid or 0.1% formic acid) is required. A good choice for initial tests turned out to be a mobile phase being composed of methanol-glacial acetic acid-ammonium acetate (98 2 0.5 v/v/w). 

Type IV includes chiral phases that usually interact with the enantiomeric analytes through the formation of metal complexes. There are usually used to separate amino acid enantiomers. These types of phases are also called ligand exchange phases. The transient diastereomeric complexes are ternary metal complexes between a transitional metal (usually Cu +), an amino acid enantiomeric analyte, and another compound immobilized on the CSP which is able to undergo complexation with the transitional metal (see also the ligand exchange section. Section 22.5). The two enantiomers are separated based on the difference in the stability constant of the two diastereomeric species. The mobile phases used to separate such enantiomeric analytes are usually aqueous solutions of copper (II) salts such as copper sulfate or copper acetate. To modulate the retention, several parameters—such as the pH of the mobile phase, the concentration of the copper ion, or the addition of an organic modifier such as acetonitrile or methanol in the mobile phase—can be varied. 

Although molecular imprinting is a fascinating tool for tailoring the enantioselectivity of a CSP, from a practical standpoint MlP-type CSPs are problematic for analytical applications. This is mainly due to (i) their poor efficiency, in particular for the high-affinity enantiomer and print molecule, and (ii) the limited range of applicability, i.e. only for the racemate of the print molecule and structurally closely related SAs for which cross-selectivity exists. These major limitations are the main reasons why there are no MlP-type CSPs currently available on the market. 

Since the introduction of CSPs based on macrocyclic antibiotics by Armstrong in 1994 [278], they have gained much interest owing to their (i) broad spectrum of applicability, (ii) complementary activity of the different types of macrocyclic antibiotics, (iii) multiple modes of operation (normal-phase, reversed-phase, polar-otganic phase modes) with complementary enantioselectivities in each mode, and (iv) the ability to separate the enantiomers of underivatized a- and P-amino acids. 

Drawbacks of the macrocyclic antibiotic type CSPs may be (i) the complexity of rationalizing and/or predicting enantiomer affinity, and accordingly the inability to predict the elution order so that chromatographic assignment of absolute configurations is not possible, and (ii) the total absence of the enantiomeric CSP which would facilitate the reversal of elution order of the SA enantiomers. 

Fig. 9.38. Loadability of different CSPs under bateh-ehromatography arnditions. (a) Triigcr base on Chi-ralpak AD methanol vs. aeetonitrilc (

HSA bears structural and functional resemblance to BSA, and HSA-type CSPs [164] also show similar enantioselective binding preferences for acidic and neutral drug molecules, such as 2-aryloxypropanoic acids [165], warfarin [166] and benzodiazepines [167]. The chiral recognition mechanism of HSA has been the subject of a number of investigations [168], which revealed that enantioselective binding occurs primarily at two well-defined hydrophobic sites. Acidic drugs have been shown to bind preferentially to the so-called warfarin-azapropazone (site I) and neutral drugs to the indol-benzodiazepine site (site II). 

---