Intro / Stereochemistry / Principles / Columns / Mobile Phase Additives / Alternatives / Refs / Top


Online Guide to Chiral HPLC

By Mark Earll BSc(Hons) CChem MRSC (C) Copyright 1999, All rights reserved.

Return to Mark's Anlytical Chemistry Index Page


NB: You should have MDL's Chime installed to see these pages at their best!

Disclaimer: This article is for guidance and educational purposes only. The author can accept no responsibility for loss or damage however caused. The author recommends that manufacturers advice be consulted exclusively when using any HPLC products.

Contents:

  1. Introduction
  2. Stereochemistry
  3. Principles
  4. Columns
    1. Cellulose
    2. Cyclodextrin
    3. Brush type (Pirkle)
    4. Macrocyclic Antibiotics
    5. Protein
    6. Ligand exchange
    7. Crown ethers
  5. Mobile Phase Additives
  6. Alternative Methods


Intro / Stereochemistry / Principles / Columns / Mobile Phase Additives / Alternatives / Refs / Top

Introduction

This guide was written to help students and new workers in the field of Chiral HPLC. It is a fascinating area in which to work and requires a wide-ranging knowledge of chemical structure and physical interaction to become successful at it.

This review includes a large number of tips and anecdotes, which I have picked up during my twelve years experience of this field. These have come from practical experience and from discussions with my many colleagues friends and counterparts in the analytical community world-wide. I shall be continuing to update this site so please check back from time to time. If you have any suggestions, comments or corrections then please email me and I will be happy to consider them.

Specific analytes that may be separated are highlighted in red.
All trademarks of the respective manufacturer's are acknowledged.

Disclaimer: This article is for guidance and educational purposes only. The author can accept no responsibility for loss or damage however caused. The author recommends that manufacturers advice be consulted exclusively when using any HPLC products.


Intro / Stereochemistry / Principles / Columns / Mobile Phase Additives / Alternatives / Refs / Top

Stereochemistry

This guide is meant as a brief introduction to Chiral HPLC, so it is assumed that the reader will have a basic grasp of Chiral concepts. Since there is often great confusion with chirality this section contains some definitions and concepts which are essential to understand.

What causes chirality?
There are several ways that a molecule can display "handedness" (chirality).

Firstly if an atom such as carbon, silicon, nitrogen, phosphorus or sulphur forms a tetrahedral structure with four different groups attached then two non-superimposable mirror images will be formed. The most common and simplest example of this is with carbon. If you look at the diagram below you will see that there is no way that you can put the two molecules on top of one another and have all the groups lined up. One molecule is the mirror image of the other one. This is an example of a molecule with a
chiral centre.

R

S

The second way a molecule can show chirality is if there is a rigid feature in the molecule, which leads to two mirror images being formed. Examples are Allene, subtituted bi-aryls (o,o'-dinotrodiphenic acid) and metal alkene or metallocenes. These molecules posses a chiral axis.

Thirdly certain molecules can form helical structure due to steric effects which have high enough energy barriers to interconversion for two enantiomers to exist. Eamples are helicenes and binaphthylphosphoric acid. These are examples of molecules with a chiral plane.

Definitions:

Enantiomer

one mirror image form

Racemate

equimolar mixture of both enantiomers

Racemisation

the conversion of a single enantiomer to a racemate

Epimers

optically active molecules with greater than two optical centres differing at only one chiral centre

Epimerisation

the partial racemisation of one chiral centre in a molecule with two or more chiral centres

Diastereoisomer

stereoisomeric structures which are not enantiomers

Absolute stereochemistry

the absolute configuration of atoms in space

Meso

an optically inactive optical isomer of a compound which can exist in other optically active isomers


Number of chiral centres
If n = number of optical centres there will be
2n isomers. Molecules with 2 optical centres will form two diastereoisomers which can each be resolved into two enantiomers. In other words; two pairs of enantiomers. The exception to this is if two molecules have a plane of symmetry and therefore cancel out their net optical rotation. In such cases they are known as meso forms

Nomenclature:
Enantiomers possess a unique property in that they rotate plane-polarised light in equal and opposite directions. This is the basis of the (+) / (-) or d / l notation. It is not a preferred method of distinguishing between enantiomers as some molecules may change rotation on forming salts, hence d or l notation does not tell you anything about the absolute stereochemistry.

Direction of rotation of plane polarised light
d or (+) dextrorotatory (rotates to the right)
l or (-) levorotatory (rotates to the left)

Fischer Convention
The absolute stereochemistry in the Fisher notation gives the absolute spatial arrangement by reference to D-glyceraldehyde. The letters D or L are used (which sometimes caused confusion with lower case d or l ) The Fischer convention is used for carbohydrates but fails for more complex structures.

Cahn-Ingold-Prelog
The Cahn-Ingold-Prelog system is the definitive method of assigning stereochemistry. The letters R and S indicate spatial arrangements in the following manner.

  1. Assign values to the substituent groups by highest atomic number*
  2. Point the lowest value away
  3. If the remainder go from high to low clockwise then R (rectus)
  4. If the remainder go from high to low anti-clockwise then S (sinister)
  5. In case of a tie go to the next atoms along.

*The rules are in fact more detailed:
Highest atomic number > highest atomic mass > cis prior to trans > like pairs (RR) or (SS) prior to unlike > lone pairs are considered an atom of atomic no 0

R* indicates a single enantiomer obtained but with unknown stereochemistry


Intro / Stereochemistry / Principles / Columns / Mobile Phase Additives / Alternatives / Refs / Top

Principles

Chiral HPLC columns are made by immobilising single enantiomers onto the stationary phase. Resolution relies on the formation of transient diastereoisomers on the surface of the column packing. The compound which forms the most stable diastereoisomer will be most retained, whereas the opposite enantiomer will form a less stable diastereoisomer and will elute first. To achieve discrimination between enantiomers there needs to be a minimum of three points of interaction to achieve chiral recognition.

The forces that lead to this interaction are very weak and require careful optimisation by adjustment of the mobile phase and temperature to maximise selectivity. Chromatography is a multi-step method where the separation is a result of the sum of a large number of interactions. Typically a free energy of interaction difference of only 0.03 kJ/mol between the enantiomers and the stationary phase will lead to resolution.

The intermolecular forces involved with chiral recognition are polar/ionic interactions, pi-pi interactions, hydrophobic effects and hydrogen bonding. These can be augmented by the formation of inclusion complexes and binding to specific sites such as peptide or receptor sites in complex phases. The analyst may manipulate these intermolecular forces by choosing suitable mobile phases, for instance polar interactions may be controlled by the pH.

The effect of temperature is important in chiral HPLC. Lower temperature will increase chiral recognition, but as it alters the kinetics of mass transfer, it may actually make the chromatography worse by broadening peaks. There is often an optimum temperature for a separation and knowledge of this gives the analyst another factor to exploit in the method development process

The type of column used for separating a class of enantiomer is often very specific, this combined with the high cost of chiral columns, makes the choice of which column to use seem at first bewildering. Fortunately, taking the time to study the structure of chiral phases and visualising the potential interactions with the analyte can narrow down the choice significantly. There are still a few occasions where a brute force method of screening through many chiral columns is still necessary but these occasions are becoming rarer with time.


Intro / Stereochemistry / Principles / Columns / Mobile Phase Additives / Alternatives / Refs / Top

Columns - Chiral Stationary Phases

Chiral Stationary Phases (CSP's) may be classified according to their interaction mechanism with the solute. A scheme for classification was first proposed by Irving Wainer [3].

Type I CSP's are those which differentiate enantiomers by the formation of complexes based on attractive interactions. These may be hydrogen bonds, p-p interactions, dipole stacking.

Type II CSP's are those which involve a combination of attractive interactions and inclusion complexes to produce a separation. Most type II phases are based on cellulose derivatives.

Type III CSP's rely on the solute entering into chiral cavities to form inclusion complexes. The classic inclusion complex column is the cyclodextrin type of column developed by Prof.D.W. Armstrong [7]. Other CSP's in this class are crown ethers and helical polymers such as poly(triphenylmethyl methacrylate).

Type IV CSP's separate by means of diastereomeric metal complexes. This technique is also known as Chiral Ligand Exchange Chromatography (CLEC) and was developed by Davankov [2]

Type V CSP's are proteins where separations rely on a combination of hydrophobic and polar interactions.

Since this scheme was first proposed many more chiral columns have entered the marketplace and it is not clear in some cases to which of these categories thay belong. Instead I have classified them according to chemical type, reffering to the Wainer classification where helpful..

  1. Brush type (Pirkle)
  2. Cellulose
  3. Cyclodextrin
  4. Macrocyclic Antibiotics
  5. Protein
  6. Ligand exchange
  7. Crown ethers


Brush type (Pirkle)

Column: Brush type / Cellulose / Cyclodextrin / Macrocycles / Protein / Ligand exchange / Crown ethers

The columns within this group are mainly the result of the work of Bill Pirkle [4] and co-workers. Bill Pirkle applied work on chiral NMR to HPLC stationary phases in the 1960's. By a process of reciprocal design, that is immobilising analytes that had given large resolutions on earlier phases, he was able to refine phase design to give broad applicability and good chromatographic characteristics.

The columns are designed to give a strong three-point interaction with one of an enantiomer pair. This means they are classified as Type I in the Wainer classification. There are two main types of stationary phases, pi-acceptor or pi-donor phases. The most common pi-acceptor phase is N-(3,5-dinitrobenzoyl) -phenylglycine bonded to n-propylamino silica. These columns are capable of separating a large range of compounds which include a pi-donor aromatic group. This may be introduced by derivatization with naphthoyl chloride or other appropriate reagent.

pi-donor phases (typically naphthyl-amino-acid derivatives covalently bonded to silica) require the analyte to contain a pi-acceptor group such as the dinitrobenzoyl group. The dinitrobenzoyl group can easily be added to a wide range of compounds such as alcohols, amines, carboxylic acids etc. using dinitrobenzoyl chloride, isocyanate or dinitroanaline.

The advantages of the brush type phases are that they are easily synthesized, full details of their preparation are given by Bill Pirkle in many of his papers [5]. They are also readily available commercially at reasonable cost. Pirkle columns can give good selectivity factors resulting in relatively high loading factors. The disadvantages are that the columns will only work with aromatic compounds and that derivatization may need to be undertaken to aid the separation. It is important to note however that the derivatization is achiral and so does not present the problems involved in chiral derivatisation. Mobile phases are restricted to relatively non-polar organic solvents. This is not necessarily a disadvantage for preparative chromatography.

More recent developments in the field have been mixed pi-donor/acceptor phases such as the WHELK-O and BLAMO phases and the alpha-BURKE-II phase. The alpha-BURKE-II phase was developed specifically to separate beta blockers. Mobile phases consist of dichloromethane-ethanol-methanol mixtures typically in the ratio 85:10:5. Addition of 10mM ammonium acetate can modify the retention. [13]

The SS BLAMO II phase is a brush type phase that includes both a pi-donor and a pi-acceptor region in the same molecule. This forms a chiral cleft, which gives high selectivity for certain molecules. The criteria required for separation are an aromatic group and a hydrogen-bonding group on or near the chiral centre.
This phase will separate all the profens. The phase has a high coverage of 1.8 x 10.4 moles/gram and so it is very suited to preparative chromatography.

The SS WHELK-O1 phase is similar to the BLAMO in that both pi-acid and pi-basic groups are present and the phase forms a chiral cleft. The requirements for separation are an aromatic group, a hydrogen and a hydrogen bonding group attached to the chiral centre. This phase is also highly loadable and has been used in SFC where some exceptional loadings have been achieved (50mg trans-stilbene oxide on an analytical column!) . Normal phase solvents can be used such as Hexane-ethanol or IPA. The phase has broader applicability than the original Pirkle phases and structures it will separate include
Warfarin, Ibuprofen, aryl-amides, aryl-epoxides and aryl-sulphoxides.

New brush-type phases are appearing all the time. Quinine carbamates have recently been used which are suitable for the separation of acidic molecules through an ionic interaction with the basic quinine group. Large ranges of columns based on the Pirkle concept are available from Regis and Sumitomo.


Cellulose

Column: Brush type / Cellulose / Cyclodextrin / Macrocycles / Protein / Ligand exchange / Crown ethers

Cellulose columns involve a combination of attractive interactions and inclusion complexes to produce a separation. They are classified as Type II in the Wainer classification.

Columns available include microcrystalline triacetate- (MCTA), tribenzoate-, trisphenylcarbamate- and tris(3,5-dimethylphenylcarbamate)- cellulose. A wide range of compounds can be separated by type II phases and a wide range of column types are available. Low polarity mobile phases are used, typically alcohol-hexane mixtures though Chlorinated solvents must be avoided since they may strip the cellulose from the silica support.

One of the principle manufacturers of these columns is the Japanese company Daicell. Their range of cellulose ester and cellulose carbamate columns separate a large variety of alkaloids and pharmaceutical compounds. One column in particular, the OD, has given exceptional separations in some case with Rs values of over 25! [6] This means that the column may be loaded quite highly before the separation deteriorates.

Each unit of these cellulose phases displays a propellor-type shape and are believed to form helical polymeric structures which combine polar, pi-pi interactions with inclusion complexation. Complimentary columns have been made using amylose (starch). These columns show different selectivity to the cellulose columns but are less robust. Starch is water-soluble and so mobile phases must have a zero water content to ensure column longevity. By far the most useful columns to have in a column library are the OD and AD phases. These columns are widely acknowledged to separate about 80% of all chiral molecules you are likely to encounter.

These phases are used in the normal phase mode with hexane-ethanol, hexane-IPA mixtures. Good results have also been shown with pure ethanol (100%) mobile phases. A reverse phase version of the OD column is available which must be used with high concentration perchorate (0.5M) buffer solutions to prevent dissolution of the stationary phase. Even so this column will degrade with time, though it is very useful as many of the separations performed in normal phase still work in the reverse phase mode.

The types, structure and application of the Chiracel phases are shown in the table below:

Name

Structure

Applications

OA

-OCH3

small aliphatic compounds

OB

-OPh

small aliphatic and aromatic compounds

OC

-ONHPh

cyclopentenones

OK

-OC=CPh

aromatic compounds

OD*

-ONH3-5diMePh

alkaloids, tropines, amines, beta blockers

OF

-ONHpara-ClPh

beta lactams, dihydroxypryidines, alkaloids

OG

-ONHparaMePh

beta lactams, alkaloids

OJ

-OparaMePh

aryl methyl esters, aryl methoxy esters

AD*

-ONH3-5diMePh

alkaloids, tropines, amines, beta blockers

AS

-ONH(CH3)CPh

alkaloids, tropines, amines

*most used


Cyclodextrin

Column: Brush type / Cellulose / Cyclodextrin / Macrocycles / Protein / Ligand exchange / Crown ethers

Cyclodextrins are cyclic oligosaccharides containing from six to twelve D(+) glucopyranose units bonded through alpha-(1,4) linkages. They are produced by the action of Bacillus macerans amylase or cyclodextrin transglycosylase on starch. The latter enzyme can be made to produce cyclodextrins of specific sizes according to reaction conditions. Three sizes are commercially available alpha, beta & gamma corresponding to 6, 7 and 8 glucopyranose units respectively. The cyclodextrin molecule forms a truncated conical cavity the diameter of which depends on the number of glucopyranose units. The table below shows the size of the cavities and the type of molecules that may be accommodated.

CD

No.Of units

size A

Molecules included

Chiral centres

alpha

6

4.5-6.0

5-6 membered aromatic

30

beta

7

6.0-8.0

biphenyl or naphthalene

35

gamma

8

8.0-10.0

substituted pyrenes and Steroids

40

Alpha

Beta

Gamma

The cyclodextrin molecule has secondary 2- and 3- hydroxyl groups lining the mouth of the cavity and primary 6-hydroxyl groups at the rear of the molecule. This means that the cavity itself is a relatively hydrophobic region of the molecule and permits inclusion of hydrophobic portions of solute molecules. Interaction of any polar regions of a solute molecule with the surface hydroxyls combined with the hydrophobic interactions in the cavity provides the 3-point interaction required for chiral recognition. Cyclodextrin stationary phases are formed by bonding the CD molecules to silica gel. The coverage of cyclodextrin groups has been optimized by Ward and Armstrong [8] however there is still a limited coverage of the surface which results in limited preparative capacity. The kinetics of inclusion are relatively slow which results in poor peak shape which also hampers the use of cyclodextrins as preparative phases.

For chiral resolution to occur, a portion of the molecule must enter the hydrophobic cavity and a hydrogen bonding region of the molecule must interact with the mouth of the cavity. The beta form has been found to have the widest application. Recent structural information has shown that the beta form has a pronounced kink in its structure, whereas the alpha and gamma forms are more planar. This results in some of the primary OH groups being more acidic.

The selectivity of a Cyclodextrin Phase is dependent on the size of the analyte . Alpha-cyclodextrin will include single Phenyl groups or Napthyl groups end-on. Beta-cyclodextrin will accept Napthyl groups and heavily substituted phenyl groups. Gamma-Cyclodextrin is useful for bulky steriod-type molecules.

To achieve a separation on cyclodextrin the analyte must have an aromatic group for inclusion in the cavity. The exceptions to this rule are Steroids and Terpene alcohols. If there is a chiral centre between two aromatics or between a carbonyl and an aromatic group the compound will almost certainly be separable by a cyclodextrin phase. Aromatic groups will be preferentially adsorbed into the cavity if they have a strong including group as a substituent. The strength of inclusion into CD cavities has been found to follow the order I > Br > Cl > F > NO3 > SO3 > OH while the preferred order of H-bonding with surface hydroxyls is carboxyls > carbonyls > amines. Separations will be greater if the hydrogen bonding groups in the analyte are brought into close proximity to the surface hydroxyls. This is illustrated by the separation of D,L phenylalanine analogs on beta cyclodextrin. Meta-substituted analogs tilt the molecule thereby enhancing hydrogen bonding in the mouth of the cavity.

The influence of pH on hydrogen bonding is illustrated by the separation of Ibuprofen on beta-cyclodextrin. At pH 7 virtually no separation is observed, whereas at pH 4.1 a good separation may be achieved. In general amino-acids should be chromatographed at a low pH in order to suppress the ionisation of the acid groups and enhance the protonation of amine groups. Triethylamine Phosphate and Triethylamine Acetate have proved to be very good buffers for cyclodextrin columns. They are made by adjusting the pH of 0.1% TEA solution either with orthophosphoric or acetic acid.

Complexation with the CD cavity decreases with higher flow-rates. Low flow-rates produce better separations 0.5 to 1 ml/min being optimum. Increasing the concentration of the buffer can help to overcome the flow effect by enhancing the attractiveness of the CD cavity with respect to the mobile phase.

Typical buffer concentrations to use are as follows:

TEAA

0.01-2%

NH4NO3

10-500mM (used to reduce inclusion)

Citrate

10-200mM (esp for acidic structures)

NH4Acetate

10-200mM


Preferred
screening pH's are:

Alcohols and Amines

pH4 (enhance NH ionisation)

Acids

pH7


Clean up of columns and maintenance
Passing a buffer through the column at pH 4 should remove most impurities. Dichloromethane will stick to CD cavities and can only be removed by heating the column above 40 C. Columns should not be left in buffer since this can attack the cyclodextrin. Protein deposits can be cleared by backflushing with 25% IPA in water. Any triethylamine used should be of the highest quality since it often contains many contaminants.
Column Testing
Inject samples of ortho-, meta- and para- nitroaniline using 40% MeOH 60% Water. If they are separated it indicates that the CD ring is intact. To check chiral selectivity inject Ibuprofen using 0.01M TEAA pH4 18% MeCN
Optimisation of a separation

  1. Carry out a pH profile and plot resolution vs. pH and RT
  2. Carry out a Flow vs. resolution experiment
  3. Optimise column temperature, organic modifier and buffer concentration

Polar Organic Mode
It is possible to operate cyclodextrin columns in an alternative
Polar Organic Mode. This is akin to Non-Aqueous Reverse Phase (NARP) chromatography. The mobile phase consists of acetonitrile with up to 10% methanol plus up to 0.5% acetic acid and/or 0.5% triethylamine. The acetonitrile fills the hydrophobic cavity and the separation occurs at the surface of the cavity by means of hydrogen bonding to the chiral hydroxyls. A typical mobile phase might be 95:5:0.2:0.3 ACN/MeOH/HoAc/TEA. The separation is purely controlled by the relative amounts of triethylamine and acetic acid. By varying the relative concentration you may control the ionisation of the analyte until you achieve a favourable interaction. To achieve a "Polar Organic" separation you need an aromatic ring plus 2 hydrogen bonding groups, 1 close to the ring and another anywhere in the rest of the molecule. The polar organic mode will separate some molecules, which will not separate using aqueous mobile phases such as propranolol.

Modified Cyclodextrins
More recently a new range of modified cyclodextrins have been developed which expand the range of compounds which can be resolved and also extend their use into Gas Chromatography. The derivatives are formed by bonding various groups onto the surface hydroxyls of the Cyclodextrin cavity. This extends the area available for chiral interactions..They include acetylated, (analogous to acetylated cellulose), S-hydroxypropylated, S or R-Naphthylethylcarbamate (analogous to naphthyl pirkle type columns), 3,5 Dimethylphenylcarbamate (analogous to chiralcell OD type column) and Cyclobond PT para-Toluoylester.

These newer Cyclodextrin phases have several advantages. They are able to separate a much wider range of compounds than native cyclodextrins, they are competitively priced and the analogs of the more popular cellulose phases are considerably more stable even allowing reverse phase operation. The preparative capacity of these new phases are better than native cyclodextrins in cases of improved chiral recognition.

Name

Structure

Notes

Acetylated

 

(normal or reverse phase) Analogous to acetylated cellulose but more stable. Compounds separated include Scopalomine, Atropine, Homatropine and Cocaine, Phenylephrine, Epinephrine and Ephedrine. In the case of Epinephrine the aromatic ring is drawn into the cavity so strongly that the H-bonding at the mouth of the cavity is weakened. The addition of a sodium nitrate buffer weakens the inclusion and enhances the separation.

S or R,S-Hydroxypropylether

 

(reverse phase) relies on extended H-bonding interactions.

S or R-Napthylethylcarbamate

 

(reverse or normal phase) analogous to napthyl pirkle type columns, but stable enough to be able to withstand pure IPA or MeCN. Particularly useful for 3,5 DNB derivatives.

3,5 Dimethylphenylcarbamate

 

(normal phase) analogous to chiralcell OD type column. Several separations normally carried ou under normal phase conditions have been achieved under Reverse phase. This often results in a reversal of elution order.

Cyclobond PT para-Toluoylester

 

 

NB: phosphate buffers attack carbamate columns!

GC Phases
Three derivatives have been made for GC use. These derivatives in contrast to other cyclodextrins are liquid. This is believed to be due to the mixture of substituted species rather than a property of the pure cyclodextrin since all pure cyclodextrins are crystalline. The interaction in GC is believed to be more of a surface interaction than in LC. The columns have efficiencies of around 4,000 plates/metre and are stable over a temperature range of 0 to 260 degrees C, however an irreversible change occurs at 200 degrees which results in a 20% drop in efficiency!

There are three derivatives available each on alpha- , beta- or gamma- Cyclodextrin:

  1. S-hydroxypropyl (PH) hydrophilic
  2. Dialkyl (DA) hydrophobic
  3. Trifluoroacetyl (TA) hydrophobic(intermediate)

The most universal phase is beta-PH. The main advantages of the chiral GC phases are that they can separate non-aromatic compounds and enantiomers with little functionality. The efficiency of separation is often better than HPLC.

Phase

Comments

S-hydroxypropyl (PH) hydrophilic

This is the most widely applicable phase. It has a hydrophilic surface and a hydrophobic cavity. It is less size selective than HPLC or DA cyclodextrins. (Limonene oxide may be separated on both the alpha and gamma forms.) Other compounds separated are lactones and simple bi-cyclic structures.

Dialkyl (DA) hydrophobic

This phase has a hydrophobic surface and shows stronger size selectivity:
alpha: epoxides/cyclic ethers/linear substituted alkanes
beta : heterocyclic amines/sugars ( Also mfr by Chrompack)
gamma: napthyl analogs

Trifluoroacetyl (TA) hydrophobic(intermediate)

This phase is of intermediate hydrophobicity. It is particularly good for halocarbons. The gamma version is the most selective for aldehydes, carboxylic acids and epoxides; whereas the beta is better for alcohols alcohol amines and linear molecules.

The predominant mechanism of separation is believed to be dipole-dipole interaction. Halogens have strong cavity interactions. Separations may be improved by derivatisation. Acetic anhydride, chloroacetic anhydride, di- and tri- chloroacetic and trifluoroacetic anhydrides may be used for amines and alcohols. Methylesters and TMS esters for acids. Elution reversal has been observed both between cavity sizes and also across different phases. Chiral GC columns have been used with success in CZE.

Cyclodextrins columns have been commercialised by Astec Incorporated.


Macrocyclic Antibiotics

Column: Brush type / Cellulose / Cyclodextrin / Macrocycles / Protein / Ligand exchange / Crown ethers

Recently macrocyclic antibiotics have been immobilised on silica to form a new class of chiral column. Work in this area has again been pioneered by Dan Armstrong [12]. Three antibiotics have been used, Rifamycin, Vancomycin and Ticoplanin. Rifamycin has been more successful in the CE field where it has been used as a mobile phase additive. The glycopeptides Vancomycin and Ticoplanin have a cup like region and a sugar "flap". Both columns are robust enough to be used as multi-modal columns. These columns separate on the basis of pi-pi interactions, hydrogen bonding, inclusion complexation, ionic interactions and peptide binding.

Ticoplanin
has a molecular weight of 1885 has 20 chiral centres three sugar groups and four fused rings. An acid group at one end of the peptide cup/cleft and a basic group at the other may be involved with ionic interactions. The sugar groups are arranged in three flaps that can fold over to enclose a molecule in the peptide cup. The Ticoplanin column is very useful in that it will separate all amino acids in reverse phase mode using ethanol-water or methanol-water. This means the Ticoplanin column can replace Crown Ether or Ligand Exchange columns. It displays unusual behaviour by giving better resolution and longer retention with increasing organic modifier. Acids can be separated in reverse phase mode using a buffer to lower the pH and increase retention.

Amino acid derivatives such as methyl esters and peptides may be separated in a polar organic mode akin to the cyclodextrin technique except the typical mobile phase consists of 46:54:2:2 ACN/MeOH/HoAc/TEA. Normal phase separations are also possible with hexane/ethanol mobile phases.

Vancomycin has a molecular weight of 1449, 18 chiral centres and three fused rings. It has a basket like structure with a single flexible sugar flap that can enclose a molecule sitting in the basket. A carboxylic acid and a secondary amine group sit on the rim of the basket and can take part in ionic interactions. The Vancomycin column may be used in Reverse, Polar and Normal phase mode, though the polar organic mode is not as successful as it is with Ticoplanin

Vancomycin separates amines, amides neutrals and esters but is less selective for acidic molecules. In reverse phase mode THF, acetonitrile or methanol are used as modifiers with an aqueous buffer solution such as triethylamine-acetate. The pH range of the column is 4 to 7, with pH 7 being the preferred starting pH for basic compounds. THF and acetonitrile tend to show the greatest selectivity. Pure methanol or ethanol mobile phases have also yielded good separations in some cases. Normal phase operation is possible using hexane/ethanol mobile phases.

The Vancomycin columns have exceptional loading capacity and are very suited to preparative applications.


Protein

Column: Brush type / Cellulose / Cyclodextrin / Macrocycles / Protein / Ligand exchange / Crown ethers

Proteins CSP's are defined as Type V CSP's where separations rely on a combination of hydrophobic and polar interactions.

Several types of proteins have been used as chiral stationary phases. Currently available phases are Human a-Acid Glycoprotein (AGP), Human Serum Albumin (HSA), Bovine Serum Albumin (BSA) and Ovomucoid protein (OV).

Diagram of AGPHuman alpha-acid glycoprotein (shown on the left) is a polypeptide with 181 amino acid residues and 40 sialic acid residues. It has an acidic character, its iso-electricpoint being 2.7. It is very stable protein having two disulphide bridges. It may be covalently bonded to silica to produce a reverse-phase column. A very wide range of molecules have been separated using AGP columns.[9]

Typical mobile phases are phosphate buffer pH 4-7 with a low percentage of organic modifier. The first choice of modifier is IPA. If this doesn't produce a separation then Acetonitrile, Ethanol, Methanol or THF should be tried. Changing the modifier results in a temporary change in the protien structure. Column loading is of vital importance. Typical loadings are 20ul of 0.02mg/ml. Changing the pH has a critical effect on the selectivity, especially for amines. Lowering the pH will lower the negative charge on the protein causing amines to be less well retained, however this means that less organic modifier needs to be used and so enantioselectivity may increase along with improved peak shape. A good starting point for amines is 10mM sodium acetate buffer at pH 4.5, for acids 10mM phosphate at pH 7.0 would be a good choice.

Charged modifiers may be used when organic modifiers have failed. These result in a permanent change in the protein structure and should be used with caution. They include Butyric acid, Octanoic acid, Decanoic acid, Dimethyloctylamine. Ethylene Glycol, 1,2 butanol and Sodium Chloride, have also been used. Temperature affects the separation, as temperature increases the retention time and separation factors decrease.

For a compound to be resolved on an AGP column it requires the following properties:

  1. Ring close to chiral centre
  2. At least one Hydrogen bonding site
  3. The distance between the ring and the H-bonding site should not br greater than 3 atoms.(unless there are more than one).

AGP columns may be regenerated by washing with 25% aqueous IPA

Human serum albumin is a polypeptide with a molecular weight of 69,000 and an isoelectric point of 4.8. Two major drug binding sites are believed to exist in the protein, the warfarin-azapropazone site and the benzodiazapine-indole binding site.The column has proved particularly useful in separating benzodiazepine enantiomers using octanoic acid in the mobile phase. Warfarin and oxazepam may be separated using 84:10:6 100mM Sodium Phosphate pH7:MeCN:IPA.

Bovine serum albumin is a globular hydrophobic protein with a molecular weight of 66,000 and an isoelectric point of 4.7. The protein consists of a single amino acid chain with 17 disulphide bridges forming nine double loops. A wide range of compounds have been separated on BSA columns.[10] BSA is not as stable as AGP, several organic solvents denature the protein (MeOH MeCN) and more care has to be taken to prevent degradation.

Ovamucoid protein comes from chicken egg-whites and has a molecular weight of 55,000. It has resolved a number of amines and acids.

In all of the protein phases loading capacity is minute. The recommended loading for chiral resolutions is 3-5 nM (around 0.4 ug) on an analytical column. The maximum flow rate on a 4mm AGP column is 0.9ml/min. Many of the other protein phases are limited to 0.5ml/min. Both of these factors effectively rule out the use of Protein columns for preparative HPLC.

Protein columns have possibly the broadest application of any chiral columns but do not always offer the most efficient separations.

Picture of HSA

Structure of HSA


Ligand exchange

Column: Brush type / Cellulose / Cyclodextrin / Macrocycles / Protein / Ligand exchange / Crown ethers

Chiral Ligand Exchange Chromatography (CLEC), developed by Davankov [2] separates enantiomers by formation of diastereomeric metal complexes. They are classified as Type IV CSP's. The method is mainly used for separation of Amino Acids.

A chiral Amino-acid-copper complex is bound to silica or a polymeric stationary phase and Copper ions are included in the mobile phase to ensure there is no loss of copper. Amino Acids may then separated by forming diastereomeric copper complexes. Water stabilises the complex by coordinating in an axial position. Steric factors then determine which of the two complexes is most stable, one of the water molecules is usually sterically hindered from coordinating to the copper. Other transition metals have been used in Ligand Exchange chromatography, however copper has the most wide application. The complexation process is a comparatively slow one this is helped by running at elevated temperatures. The optimum temperature is around 50 C.

The use of CLEC is limited to a-amino acids and similar compounds. b-amino acids are very difficult to separate by CLEC
The preparative capacity of CLEC columns is reasonable, however the presence of copper in the mobile phase is not convenient. Copper can be removed using a iminodiacetic acid based resin column. Analytical Columns cost £400-1000 with a typical loading factor of 20 ug. Loose "Chiralsolve" polyacrimide polymer LEC material is available in kilogram amounts and has been used to separate multi gram amounts of simple amino acid derivatives on 2.5cm x 50cm columns.


Crown ethers

Column: Brush type / Cellulose / Cyclodextrin / Macrocycles / Protein / Ligand exchange / Crown ethers

18-crown-6

binaphthol 18-crown-6

Crown ethers have been immobilised to form HPLC phases which complex with primary amines. The amine must be protonated for the complex to form so acidic mobile phases are used such as perchloric acid. The most commonly used crown ether is "18 Crown 6" which has been made available commercially as the "Crownpak" column by Daicel. The columns give very efficient separations and are available in both (+) and (-) forms allowing the reversal of elution order in cases of trace analysis of the unwanted enantiomer. Crown ethers are also used as additives in CE and NMR but they are extremely toxic, being carcinogenic, so great care must be taken when using them.


Intro / Stereochemistry / Principles / Columns / Mobile Phase Additives / Alternatives / Refs / Top

Chiral Mobile Phase Additives.

Chiral mobile phase additives have the following advantages over Chiral Stationary phases:

  1. Can use standard (cheap) columns
  2. High loading capacities are possible
  3. Solute character may be modified (i.e. ion pairing)
  4. Wide Range of additives available.

The disadvantages are:

  1. The removal of the Chiral selector after chromatography
  2. Separations are difficult to develop
  3. May be expensive on a large scale without recycling of additives.

Chiral Mobile Phase Additives (CMPA's) are of several types: Ion Pair, Inclusion complexes, ligand exchange and protein interactions.

Ion Pair additives.
Chiral Ion pairing agents may be used with enantiomeric compounds on normal phase silica columns. The enantiomers are then separated by travelling down the column as diastereomeric ion pairs.
Propanolol has been separated in this way using Tartaric acid as the counter ion. Other amino alcohols have been separated using Quinine, N-benzoxycarbonyl-glycyl-L-Proline and (+)-10-Camphorsulphonic Acid.[11]

Unfortunately the loading capacity of Ion Pair separations is very low.

Inclusion complexes
Inclusion complexes may be formed by adding free cyclodextrin into aqueous mobile phases. Often resolution may be predicted by separations on Cyclodextrin stationary phases. A point of interest is that in most cases the elution order is reversed. This is due to the more strongly included enantiomer-CD pair travelling down the column faster than the less well included isomer. In a stationary phase the more strongly included species is retained.

Initial starting conditions can be found from the CD stationary phase. The amount of organic modifier and cyclodextrin then has to be optimised. The newer derivatives should extend the use of cyclodextrins as mobile phase additives.

There are two options for removal of cyclodextrins after chromatography. Firstly acid/base extractions can be used. For an acidic analyte the pH would be lowered to enable extraction of the acid from the soluble cyclodextrin. The second approach is to use solid phase extraction on a C18 column. The column is loaded with the fraction and then washed with 10:90 Methanol/Water to remove the cyclodextrin and buffer components. The compound is then washed from the column with Methanol.

Reasonable sized separations of a few milligrams have been achieved using Cyclodextrins. One advantage is that scaling up of a separation is not limited to the availability of a specialist column.

The toxicity of free crown ethers makes them unsuitable for HPLC use.

Ligand exchange
Ligand chromatography can be carried out by using a chiral selectors such as N-alkyl-L-Hydroxyproline with copper acetate in the mobile phase on a C18 column. The chiral selector is strongly adsorbed onto the C18 surface effectively forming a chiral stationary phase. If highly aqueous mobile phases are used there is virtually no column bleed so the selector may be left out of the mobile phase.

Large separations have been achieved on ligand exchange stationary phases, in principle similar separations should be possible using mobile phase additives.

Protein additives.
Both BSA and a-acid Glycoprotein have been used as mobile phase additives. BSA may be added to aqueous mobile phases at a concentration of around 3g/l on C18, CN or Diol columns. BSA has a UV cutoff at 340nm which limits its use. BSA is relatively cheap. a-acid Glycoprotein has the advantage of being UV transparent, however being a human blood product it is exceptionally expensive. As with protein stationary phases the preparative capacity of is very low.


Intro / Stereochemistry / Principles / Columns / Mobile Phase Additives / Alternatives / Refs / Top

Alternative Methods

Pure enantiomers may be obtained by one of three methods;

  1. Chiral synthesis.
  2. Achiral synthesis followed by indirect resolution.
  3. Achiral synthesis followed by resolution by chromatography.

Chiral Synthesis.
Chiral synthesis requires a chiral starting point, it is complex and requires care to avoid racemisation. The chiral purity must be monitored throughout the synthesis. The advantages are apparent in the long term due to the lack of wastage of the unwanted enantiomer and the ability to scale up the reaction to production size.

Achiral Synthesis followed by indirect resolution.
Achiral synthesis will tend to be quicker and cheaper than a chiral synthesis, however the resolution step can be time-consuming and difficult. The options for resolution are;

  1. Crystallization
  2. Enzymic reaction - destroy the unwanted isomer
  3. Diastereoisomer formation followed by crystallization

The major problem with these methods is that they are all single step processes. This means that a chiral selector is required with a very high degree of enantioselectivity. This also means that the chiral selector must be very pure itself. An impure selector will result in a loss of purity and yield of the enantiomers resolved. The difference in the free energy of interaction between the chiral selector and the two enantiomers to produce an a-value of 100 (98-99% purity) is 11 kJ/mol at 298 K.[1,2]. This is to be compared to the value in the next section for a chiral stationary phase.

Achiral Synthesis followed by resolution by chromatography.
There are three methods of separating enantiomers by chromatography;

  1. Form a diastereoisomeric derivative
  2. Use a chiral stationary phase (CSP)
  3. Use a chiral mobile phase additive (CMPA)

The first method suffers with the same disadvantages as the methods outlined above, in terms of the selectivity and purity of the derivatization agent. The chromatography can also be difficult and may take considerable development. The main advantage is that standard achiral stationary phases can be used.

Chiral stationary phases offer several advantages, the solutes are unmodified, the separations are rapid (useful for labile or racemizing compounds) and the separation relies on a multi-step process. The main disadvantages are those of high cost, low sample loading and the specificity of the stationary phase for the particular separation in hand.

Chiral mobile phase additives (CMPA) offer similar advantages to CSP's. CMPA's tend to be more versatile, since more chiral additives are available than chiral columns. The additives and columns are also cheaper than CSP's. The principle disadvantage is the separation of the CMPA from the compound of interest after chromatography. The discarding of the CMPA can also increase the cost.

Advantages of Multi-Step Methods.
Chromatography is a multi-step method where the separation is a result of the sum of a large number of interactions. This results in diminished requirements for the enantioselectivity of the stationary phase. Resolution can easily be achieved with a free energy of interaction difference of only 0.025 kJ/mol between the enantiomers and the stationary phase. This is three orders of magnitude lower than the free energy change required in single-step methods (11kJ/mol) [1,2]. The lowered enantioselectivity enables a larger range of chiral selectors to be used and makes possible the use of phases with a broad range of application. The multi-step nature of chromatography also means that the separations are not so dependant on the chiral purity of the selector. If a stationary phase contains a small quantity of the wrong isomer the effect will be countered by the combined action of the adsorbtions along the column as a whole. It is possible to achieve resolutions on chiral stationary phases that are somewhat less than 100% pure. Decreasing the purity of the stationary phase will simply decrease the enantioselectivity of the column.

Economics of Direct Preparative Chiral Resolutions.
The economics of chiral resolutions may be calculated by carefully considering all the costs involved. It is very easy to ignore factors that are not immediately obvious. The fact that a column may cost £5000 may be insignificant compared with the time taken to produce a resolution.

Resolution by Crystallisation
When considering an Achiral synthesis followed by crystallisation one must consider the following factors:

The advantage to this method is that it may be repeated on a large scale fairly easily.

Achiral synthesis followed by chromatography.
The factors to be considered with Achiral synthesis followed by chromatography are:

The disadvantage of this method is that the separation cannot be easily scaled up. The amount of material produced will increase linearly with time. The advantage is that small amounts of enantiomers can be produced fairly rapidly for initial screening.


Practical Points:

Conversion of basic salts to the free-base form for normal phase analysis.
It is often necessary to convert salts of bases into the free base form, to get the compound of interest to be soluble in normal phase solvents. This may be achieved simply by dissolving the compound in water, adding a few drops of base such as sodium bicarbonate or sodium hydroxide and then extracting into chloroform. This may be done in a small vial, shaken vigorously then left to stand. The unwanted aqueous layer (on top) can be decanted and the lower layer pipetted into a fresh vial. The chloroform can be evaporated in a stream of nitrogen to dryness in a fume hood. The residue may then be dissolved in the normal phase solvent ready for injection onto the column.


Intro / Stereochemistry / Principles / Columns / Mobile Phase Additives / Alternatives / Refs / Top

References:

1. Davankov V.A. Introduction to chromatographic resolution of enantiomers Chiral Separations by HPLC Ed. A.M.Krstalovic Ellis Horwood ISBN 0-74580331-8.
2. Davankov V.A. Resolution of racemates by ligand exchange Chromatography Advances in chromatography Vol 18 Marcel Dekker NY 1980 139.
3. Wainer I.W. Proposal for the classification of high performance liquid chromatographic chiral phases: how to chose the right column. Trends in analytical chemistry vol 6, no.5,1987.
4. Pirkle W.H. House D.W. Chiral HPLC stationary phases 1.J.Org.Chem. vol 44. no.12, p1957, 1979.
5. Pirkle W.H. Finn J.M. Chiral HPLC stationary phases 3.General resolution of Arylalkylcarbinols. J.Org.Chem. vol46 p2935,1981.
6. Okamoto Y. and Hatada K. et al. J. Chromatogr 448,454 1988 also Chem Lett 1125 1988
7. Daniel W. Armstrong Bonded phase material for chromatographic separations. 1985 U.S.patent 4539399.
8. Ward T.J. and Armstrong D.W. Improved cyclodextrin chiral phases: a comparison and review. J.Liq.Chromatography.,1986, 9, 407
9. Hermansson J. , a-acid Glycoprotein. J. Chromatogr 269,71 1983
10. Allenmark. S. Bomgren.B. Boren.H. J. Direct Liquid Chromatographic Separation of Enantiomers on Immobilized Protein Stationary Phases. Chromatogr 316 1984 617-624
11. Pettersson, C. Josefsson, M. Chiral Separation of amino-alcohols by Ion Pair Chromatography. Chromatographia 21 No.6 1986
12. Daniel W. Armstrong, Yubing Tang, Shushi Chen, Ylwen Zhou, Christina Bagwill and Jing-Ran Chen. Macrocyclic Antibiotics as a New Class of Chiral Selectors for Liquid Chromatography. Analytical Chemistry, Vol 66 No.9, May 1, 1994 p1473-1484
13. Pirkle W.H. Burke J.A. Chiral Stationary Phase for beta-blockers. J.Chromatogr, 557,173,1991