496
ZHANG ET AL.
the commercial teicoplanin CSPs (i.e., Chirobiotic T and under an argon atmosphere. After cooling, the dalbavancin
Chirobiotic T2) for comparison, and these two CSPs are bonded silica was washed with 50 ml portions in the
designated as T1 and T2.
sequence pyridine, water, methanol, acetonitrile, and
dichloromethane. It was dried under vacuum. Elemental
analysis showed 11.0% carbon loading (increased by 7.0%).
MATERIALS AND METHODS
Materials
Chromatographic condition. CSPs were slurry
packed into 250 mm 3 4.6 mm (i.d.) stainless steel col-
umns at 600 bar. Evaluation of the columns was conducted
on HP 1090 HPLC system with a DAD UV detector and
autosampler. Detection wavelengths were selected at 220,
230, and 254 nm. The injection volumes were 5 ll. All sam-
ple concentrations were 1 mg/ml. Separations were car-
ried out under isocratic conditions at a flow rate of 1 ml/
min at 218C. The mobile phases were premixed and
degassed under vacuum for 10 min. The column dead
times were tested by injection of solution of 1,3,5-tri-tert-
butylbenzene in 100% methanol.
All the racemic analytes tested in this study were pur-
chased from Sigma-Aldrich, and all HPLC-grade solvents
were obtained from VWR (Bridgeport, NJ). HPLC-grade
˚
Kromasil silica gel (particle size 5 lm, pore size 100 A,
and surface area 310 m2/g) was obtained from Akzo Nobel
(EKA Chemicals, Bohus, Sweden), and LiChrospher
˚
Si(100) silica gel (particle size 5 lm, pore size 100 A, and
surface area 400 m2/g) was purchased from Merck
(Darmstadt, Germany). All organosilane compounds were
obtained from Silar Laboratories (Wilmington, NC). These
include: (3-aminopropyl) dimethylethoxysilane, (3-amino-
propyl) triethoxysilane, [2-(carbomethoxy) ethyl] trichloro-
silane, [1-(carbomethoxy)ethyl] methyldichlorosilane, (3-
isocyanatopropyl) triethoxysilane, and (3-glycidoxypropyl)
triethoxysilane. Dalbavancin was the generous gift of
Pfizer (Washington, MO).
RESULTS AND DISCUSSION
The Structure of Dalbavancin
Dalbavancin is a second generation glycopeptide antibi-
otic molecule (see Fig. 2). The major difference between
dalbavancin and teicoplanin are as follows: (a) different
phenyl rings are chloro-substituted (see rings 2 and 3,
Figs. 1 and 2); (b) the b-D-N-acety-glucosamine unit of tei-
coplanin (see ring 5, Figs. 1 and 2) is replaced by a simple
hydroxyl group; (c) the primary hydroxyl group of N-acyl-
glucosamine unit of teicoplanin has been oxidized to a car-
boxylic acid, which can generate an anion; (d) the primary
amine group on the aglycone portion of teicoplanin is a
secondary amine substituted by methyl group; (e) the car-
boxylic group close to phenyl ring 7 is converted to an am-
ide group connected with three methylene groups and it
has a dimethylamino group at the end (in dalbavancin);
and (f) dalbavancin has 10 carbons in the carbon chain of
b-D-N-acyl-glucosamine, whereas teicoplanin has only nine
carbons. The last difference noted earlier is the least likely
Methods
Preparation of the D1 CSP. One gram of dried dal-
bavancin (0.53 mmol) was dissolved in 55 ml of anhydrous
DMF in a 250-ml three-neck round flask with mechanical
stirring. Then, triethylamine (0.72 ml, 5.16 mmol) and
3-(triethoxysilyl)propyl isocyanate (0.865 ml, 3.50 mmol)
were added into the solution at room temperature under
argon protection. The solution was heated to 958C for 5 h
and cooled to 608C. The dried Kromasil silica (3.50 g,
5 lm, 100 A8) was added into the solution. The mixture
was heated to 1058C over night and then cooled to room
temperature and filtered. The CSP was washed by metha-
nol, methanol/water (50/50, v/v), pure water, and metha-
nol (50 ml for each solvent), and dried in oven at 1008C
overnight. Elemental analysis showed 8.0% carbon loading.
Preparation of the D2 CSP. The D2 stationary phase to affect enantioseparation because one more methylene
was prepared as previously described for the teicoplanin group does not provide any additional interactions that are
CSP. Five grams of Lichrospher silica gel was first dried at beneficial to chiral recognition. Previous studies by our
1508C under vacuum, and then it is heated in toluene to group have shown that the teicoplanin carbohydrate units
reflux to remove azaeotropically all residual water. It is fol- play an important role in chiral recognition. It helps in the
lowed by adding 2.5 ml of 3-aminopropyl triethoxysilane, separation of nonamino acid compounds. However, they
and the reaction mixture was heated to reflux for 4 h. The also decrease the separation of many a-amino acid enan-
modified silica gel was filtered, washed with toluene, tiomers.14 Thus, the elimination of the b-D-N-acety-glucosa-
methanol, and dichloromethane, and dried at 908C over- mine unit in dalbavancin can substantially affect its enan-
night. Elemental analysis showed that the derivatized tioselectivity. The other changes made to carboxylic
silica gel has 4.0% carbon loading. A 2.5 ml portion of 1,6- groups, hydroxyl group, and amino groups can also con-
diisocyanatohexane (15 mmol) was added to an ice-bath- tribute to differences in the enantioselectivity of dalbavan-
cooled slurry of 2.5 g of 3-aminopropyl-Lichrospher in 50 cin relative to teicoplanin. Dalbavancin has one tertiary
ml of anhydrous toluene. Next, the mixture was heated at amine and secondary amine, respectively, and one carbox-
708C for 2 h. After cooling, the supernatant toluene phase ylic group on the N-acyl-glucosamine (Figs. 1 and 2).
was removed under an argon atmosphere. The excess Whereas, teicoplanin has only one carboxylic group con-
reactant was removed by dry toluene washing. A suspen- nected to the aglycone and one primary amino group. As
sion of 1 g of dalbavancin (0.53 mmol) in 100 ml of dry pyr- amine and carboxylic acids group are ionizable in aqueous
idine was added dropwise to the wet activated silica. Next, solution and can interact via electrostatic interactions with
the mixture was heated at 708C for 12 h with stirring charged analytes, these changes could lead to different
Chirality DOI 10.1002/chir