Preparation and evaluation of regioselectively substituted amylose derivatives for chiral separations

Article abstract of DOI:10.1002/chir.22720

Six novel regioselectively substituted amylose derivatives with a benzoate at 2-position and two different phenylcarbamates at 3- and 6-positions were synthesized and their structures were characterized by ¹H nuclear magnetic resonance (NMR) spectroscopy. Their enantioseparation abilities were then examined as chiral stationary phases (CSPs) for high-performance liquid chromatography (HPLC) after they were coated on 3-aminopropyl silica gels. Investigations indicated that the substituents at the 3- and 6-positions played an important role in chiral recognition of these amylose 2-benzoate serial derivatives. The derivatives demonstrated characteristic enantioseparation and some racemates were better resolved on these derivatives than on Chiralpak AD, which is one of the most efficient CSPs, utilizing coated amylose tris(3,5-dimethylphenylcarbamate) as the chiral selector. Among the derivatives prepared, amylose 2-benzoate-3-(phenylcarbamate/4-methylphenylcarbamate)-6-(3,5-dimethylphenylcarbamate) exhibited chiral recognition abilities comparable to that of Chiralpak AD and may be useful CSPs in the future. The effect of mobile phase on chiral recognition was also studied. In general, with the decreased concentration of 2-propanol, better resolutions were obtained with longer retention times. Moreover, when ethanol was used instead of 2-propanol, poorer resolutions were often achieved. However, in some cases, improved enantioselectivity was achieved with ethanol rather than 2-propanol as the mobile phase modifier.

Full text of DOI:10.1002/chir.22720

Received: 30 March 2017
DOI: 10.1002/chir.22720
Revised: 12 May 2017
Accepted: 15 May 2017
R E G U L A R A R T I C L E
Preparation and evaluation of regioselectively substituted amylose
derivatives for chiral separations
Shouwan Tang1
| Zhaolei Jin2 | Baishen Sun1 | Fang Wang1 | Wenyuan Tang1
1 Department of Chemistry, College of
Pharmaceutical and Chemical Engineering,
Taizhou University, Taizhou, People's
Republic of China
2 College of Pharmaceutical Science,
Zhejiang University, Hangzhou, People's
Republic of China
Abstract
Six novel regioselectively substituted amylose derivatives with a benzoate at 2‐posi-
tion and two different phenylcarbamates at 3‐ and 6‐positions were synthesized and
their structures were characterized by 1H nuclear magnetic resonance (NMR) spec-
troscopy. Their enantioseparation abilities were then examined as chiral stationary
phases (CSPs) for high‐performance liquid chromatography (HPLC) after they were
coated on 3‐aminopropyl silica gels. Investigations indicated that the substituents at
the 3‐ and 6‐positions played an important role in chiral recognition of these amy-
lose 2‐benzoate serial derivatives. The derivatives demonstrated characteristic
enantioseparation and some racemates were better resolved on these derivatives than
on Chiralpak AD, which is one of the most efficient CSPs, utilizing coated amylose
tris(3,5‐dimethylphenylcarbamate) as the chiral selector. Among the derivatives pre-
pared, amylose 2‐benzoate‐3‐(phenylcarbamate/4‐methylphenylcarbamate)‐6‐(3,5‐
dimethylphenylcarbamate) exhibited chiral recognition abilities comparable to that
of Chiralpak AD and may be useful CSPs in the future. The effect of mobile phase
on chiral recognition was also studied. In general, with the decreased concentration
of 2‐propanol, better resolutions were obtained with longer retention times. More-
over, when ethanol was used instead of 2‐propanol, poorer resolutions were often
achieved. However, in some cases, improved enantioselectivity was achieved with
ethanol rather than 2‐propanol as the mobile phase modifier.
Correspondence
S. Tang, Department of Chemistry, College
of Pharmaceutical and Chemical
Engineering, Taizhou University, No. 1139
Shifu Avenue, Taizhou 318000, People's
Republic of China.
Email: tangswtz@163.com
Funding information
Department of Science and Technology of
Zhejiang Province, Grant/Award Number:
No. 2015C37034
KEYWORDS
amylose derivative, chiral separation, chiral stationary phase, HPLC, polysaccharide
1 | INTRODUCTION
However, the regioselective introduction of different substitu-
ents at 2‐ and 3‐positions was not achieved until 2008.18
For their distinguished chiral recognition abilities, polysac-
charide derivatives, especially the benzoates and
phenylcarbamates of cellulose and amylose, are recognized
as the most powerful chiral stationary phases (CSPs) in
high‐performance liquid chromatography (HPLC).1-10 Gen-
erally, these derivatives are homogeneously substituted,
which means that they bear the same substituent at 2‐, 3‐,
and 6‐positions of a glucose ring. The heterosubstituted poly-
saccharide derivatives having different substituents at 2, 3‐
positions and 6‐position has also been obtained.11-17
Based on the regioselective esterification at 2‐position of
amylose reported by Dicke,19 two novel amylose derivatives
with three different substituents at 2‐, 3‐, and 6‐positions
have been successfully synthesized by the Okamoto group
and they exhibited enantioseparation abilities comparable to
or better than the commercial amylose‐based column,
Chiralpak AD.18 In 2010, the Okamoto group extended the
above work by synthesizing a series of amylose 2‐(substituted
benzoate) derivatives and found that amylose 2‐(4‐tert‐
butylbenzoate) and amylose 2‐(4‐chlorobenzoate) series
Chirality. 2017;1–10.
wileyonlinelibrary.com/journal/chir
© 2017 Wiley Periodicals, Inc.
1

2
TANG ET AL.
showed high enantioseparation abilities.20 In 2012, Tang and
coworkers prepared four novel regioselectively amylose
was obtained from Acros (Somerville, NJ). 3,5‐
Dichlorophenyl isocyanate was bought from Zhejiang
Tongfen Chemical and Pharmaceutical Enterprise (Zhejiang,
China). Triphenylmethyl chloride, anhydrous lithium chlo-
ride, and triphosgene were purchased from Aladdin (Shang-
hai, China). 3,5‐Dimethylphenyl, 4‐methylphenyl, and
phenyl isocyanates were prepared by reaction of the corre-
sponding anilines with triphosgene in dry toluene utilizing
the traditional method. Wide pore silica gel (Daiso gel, SP‐
7‐1000) with a mean particle size of 7 μm and a mean pore
diameter of 100 nm, purchased from Daiso Chemical (Osaka,
Japan), was silanized using (3‐aminopropyl)triethoxysilane in
dry toluene at 80 °C, together with a small amount of pyri-
dine as catalyst. All solvents used in preparation of amylose
derivatives were of analytical reagent grade, carefully dried,
and distilled before use.
derivatives bearing
a
chiral substituent, (S)‐1‐
phenylethylcarbamate at 3‐ or 6‐position, and found that
increased enantioselectivity could be achieved by introducing
a chiral substituent at the 6‐achiral rather than the 3‐chiral
position.21 Among the derivatives, amylose 2‐benzoate‐3‐
(3,5‐dimethylphenylcarbamate/3,5‐dichlorophenylcarbamate)‐
6‐((S)‐1‐phenylethylcarbamate) exhibited chiral recognition
abilities comparable to that of Chiralpak AD.21 Recently, the
regioselective introduction of substituents at 2‐, 6‐positions
and 3‐postion have been carried out through the regioselective
protection at 2‐ and 6‐positions using a bulky trialkylsilyl
chloride.22
In order to investigate the regioselective substituents
introduced between 3‐ and 6‐positions on the
enantioseparation abilities of 2‐benzoyl amylose derivatives
as well as to develop new polysaccharide derivatives‐based
CSPs with high chiral recognition, six novel amylose deriva-
tives (Figure 1, 1a–f) were synthesized. Their structures were
characterized by 1H nuclear magnetic resonance (NMR)
spectroscopy. After being coated on 3‐aminopropyl silica
gels, they were then utilized as HPLC CSPs. Their chiral rec-
ognition abilities were evaluated and compared with com-
mercially available Chiralpak AD, which is regarded as one
of the most promising CSPs. The effects of the nature and
the position of the substituents and mobile phase on
enantioseparation abilities of the derivatives were examined.
2.2 | Synthesis of regioselectively substituted
amylose derivatives
As represented in Figure 2, six amylose 2‐benzoate deriva-
tives, 1a–f, bearing two different phenylcarbamate substitu-
ents at 3‐ and 6‐positions, were prepared in a sequential
procedure. To regioselectively esterify only 2‐position,19
amylose (9.0 g) was first dissolved in dry DMSO (180 ml)
at 80 °C and then vinyl benzoate (2.3 equiv. to 2‐position)
and Na2HPO4 (2 wt%) as catalyst were added at 40 °C. The
above mixture was magnetically stirred slowly at 40 °C for
216 h under N2 atmosphere. 2‐Benzoyl amylose was isolated
as a 2‐propanol‐insoluble fraction. To selectively protect 6‐
position as trityl ether, the obtained monoester was allowed
to react with triphenylmethyl chloride (2 equiv. to 6‐position)
in pyridine at 80 °C for 24 h. 2‐Benzoyl‐6‐O‐trityl amylose
was obtained as a methanol‐insoluble fraction. In order to
convert 3‐hydroxyl to the corresponding phenylcarbamate
group, 2‐benzoyl‐6‐O‐trityl amylose was allowed to react
2 | MATERIALS AND METHODS
2.1 | Chemicals and reagents
Vinyl benzoate was purchased from Alfa Aesar (Haverhill,
MA, Germany). Amylose (DP, 300) was kindly supplied by
Prof. Y. Okamoto. Anhydrous dimethyl sulfoxide (DMSO)
FIGURE 1 Structures of amylose derivatives

TANG ET AL.
3
FIGURE 2 Schematic synthesis of amylose derivatives, 1a–f
with an excess of R1NCO in pyridine at 80 °C for 24 h. The
product was isolated as an insoluble fraction in methanol.
Subsequently, the obtained 2‐benzoyl‐3‐carbamoyl‐6‐O‐trityl
amylose was suspended in methanol containing a small
amount of hydrochloric acid (2%, v/v) at 50 °C for 24 h to
cleave the triphenylmethyl group. Finally, the hydroxyl group
at 6‐position of the recovered 2‐benzoyl‐3‐carbamoyl amy-
lose was treated with an excess of R2NCO in pyridine for
24 h at 80 °C. 1a–f having different substituents at 2‐, 3‐,
and 6‐positions were isolated as methanol‐insoluble
fractions.
obtained above were then packed into stainless‐steel tubes
(25 × 0.20 cm i.d.) by a slurry method at 30–40 MPa.5
Solvents used in chromatographic experiments were of
HPLC grade and mobile phases were filtered through a
0.22‐μm pore diameter membrane and degassed by
ultrasonication prior to use. 1,3,5‐Tri‐tert‐butylbenzene was
utilized as the nonretained compound to determine dead time
(t0).23 HPLC enantioseparations were carried out at a flow
rate of 0.1 ml min−1 and solutes were detected at 254 nm.
3 | RESULTS AND DISCUSSION
2.3 | Apparatus, preparation of packed
columns, and chromatography
3.1 | Synthesis of regioselectively amylose
derivatives
Chromatographic experiments were performed on an Agilent
1200 series liquid chromatography (Agilent Technologies,
Palo Alto, CA) consisting of a degasser, a quaternary pump,
and a UV detector at ambient temperature. In addition, a cir-
cular dichroism (CD) detector (CD 2095, Jasco, Tokyo,
Japan) was equipped to monitor the signals of the eluted
enantiomers. Fourier transform infrared (FTIR) spectra were
measured using a Shimadzu (Kyoto, Japan) FTIR‐8400 spec-
The amylose derivatives 1a–f with three different substitu-
ents at 2‐, 3‐, and 6‐positions were synthesized as previously
described.21 First, in order to esterify only 2‐hydroxyl group
of amylose, it was dissolved in DMSO and reacted with vinyl
benzoate at 40 °C in the presence of a catalyst, Na2HPO4.
Then, by reacting with triphenylmethyl chloride, the 6‐
hydroxyl group was selectively protected as trityl ether.
Subsequently, the 3‐hydroxyl group was converted to the
corresponding carbamate group by reacting with an excess
of R1NCO. Finally, the 6‐hydroxyl group was regenerated
by the removal of the trityl group in a methanol/HCl mixture
and reacted with R2NCO to form the corresponding
carbamate. It should be pointed out that in this study
triphenylmethyl chloride was used instead of 4‐
methoxytriphenylmethyl chloride to selectively protect the
6‐hydroxyl group and both of them were effective to protect
the 6‐hydroxyl group as trityl ether.11-13,15,18,20,21
1
trophotometer as KBr pellets. H and 13CNMR spectra were
carried out on a Bruker (Billerica, MA) Avance III spectrom-
eter (500 MHz). Thermo gravimetric analysis (TGA) was per-
formed on an SDT Q600 thermo analyzer (TA Instruments,
New Castle, DE).
The amylose derivatives (0.15 g each) were dissolved in
tetrahydrofuran (THF) (5 ml) and then coated on 3‐
aminopropyl silica gels (0.6 g) according to the method
described elsewhere.5 The chiral packing materials (CPMs)

4
TANG ET AL.
Figure 3 demonstrated the 1H and 13C NMR spectra of 2‐
benzoyl amylose. In the 13C NMR spectrum, only one car-
bonyl peak at 165 ppm occurred, which was assigned as the
carbonyl peak at 2‐position of amylose.19 Thus, the 2‐position
of amylose was successfully regioselectively esterified and the
1
degree of substitution was calculated to be 0.95 from the H
1
NMR spectrum. The H NMR spectra of 1a and 1b with
reversed substituents at 3‐ and 6‐positions are shown in
Figure 4. Based on the spectrum of amylose tris(3,5‐
dimethylyphenylcarbamate) reported by Kaida and
Okamoto,11 the peaks at about 2.0 and 2.4 ppm in Figure 4A,
B, individually, were assigned to the methyl groups on the phe-
nyl moieties at 3‐ and 6‐positions, respectively. Thus, it con-
firmed that 1a and 1b had a 3,5‐dimethylphenylcarbamate at
3‐ and 6‐positions, individually. Minor methyl signals are also
observed in Figure 4, which indicated that the regioselective
substitution was not 100%. The structures of the other amylose
1
derivatives were similarly confirmed by H NMR. Moreover,
the derivatives were characterized by elemental analysis and
the data are listed in Table 1. The obtained CPMs were charac-
terized by FTIR spectroscopy and TGA. Compared with the
FTIR spectrum of 3‐aminopropyl silica gel, when the amy-
lose derivatives were coated onto 3‐aminopropyl silica gel,
new peaks at about 1730 (C = O) and 1630 (−C6H5) cm−1
occurred, demonstrating that the coating was successful.
FIGURE 4 1H NMR spectra of amylose derivatives, (A) 1a and (B)
1b in pyridine‐d5 at 80 °C
TABLE 1 Element analysis results of amylose derivatives, 1a–f
Calculated (%)a
H
Found (%)
Derivatives
C
N
C
H
N
1a
1b
1c
1d
1e
1f
65.41
65.41
66.18
66.18
56.54
56.54
5.26
5.26
5.51
5.51
3.84
3.84
5.26
5.26
5.05
5.05
4.89
4.89
58.67
63.48
63.69
64.68
54.28
55.09
4.68
5.28
5.43
5.52
3.76
3.94
4.77
5.20
5.10
5.09
4.76
4.73
aEstimated based on a repeated glucose unit.
Figure 5 shows the FTIR spectra of 3‐aminopropyl silica
gel before (Figure 5A) and after (Figure 5B) being coated
with the amylose derivative, 1a. The final polymer contents
for CPMs 1a–f estimated by TG were 18.8, 18.3, 18.6,
18.1, 19.0, and 18.6%, respectively.
3.2 | Chiral recognition of novel amylose
derivatives
1
FIGURE 3 (A) H and (B) 13C NMR spectra of 2‐benzoyl amylose
The resolving abilities of the obtained CSPs were evaluated
in DMSO‐d6 at 80 °C
using the racemates shown in Figure 6: Tröger's base (2),

TANG ET AL.
5
FIGURE 7 Chromatographic resolution of racemate 7 on 1a.
Column size, 25 × 0.20 cm i.D.; mobile phase, hexane/2‐propanol
(90:10, v/v); flow rate, 0.1 ml min−1
FIGURE 5 FTIR spectra of 3‐aminopropyl silica gel before (A) and
after (B) coated with the amylose derivative 1a
opposite for racemates 3 and 6. As for racemates 4 and 8,
very close α values were achieved on 1b, 1d, and Chiralpak
AD. Thus, 1b and 1d demonstrated equivalent
enantioseparation abilities in contrast with Chiralpak AD,
which appeared to be useful CSPs with wide versatility.
The other derivatives exhibited somewhat decreased but
characteristic chiral recognition. For example, exhibiting a
relatively lower enantioseparation, the highest α value, 1.33,
for racemate 3 was obtained on 1e. Another example is the
separation of racemates 2 and 6; the highest α values, 2.91
and 1.18, respectively, were achieved on 1f. Especially race-
mate 6 was only discriminated on 1f. The different substitu-
ents introduced at 3‐ and 6‐positions are responsible for
their different and unique chiral recognition since they are
expected to change the 3D structure and local polarity of
the polysaccharide derivatives, which are critical factors to
keep a regular higher‐order structure as well as the
enantioselective interaction between enantiomeric pairs and
derivatives for efficient chiral recognition.5,24
Investigations indicated that for regioselectively
substituted polysaccharide derivatives, their chiral recogni-
tion abilities were affected by the nature of the substituent
as well as its position on a glucose unit.18,20,21 The 2‐benzoyl
amylose derivatives, 1a and 1b, 1c and 1d, and 1e and 1f,
which had reversed substituents at 3‐, and 6‐positions, dem-
onstrated somewhat different chiral recognition. For 1a and
1b, better resolutions were often achieved on 1b. Especially
trans‐2,3‐diphenyloxirane (3), benzoin (4), 2‐phenylcyclo-
hexanone (5), 1‐(9‐anthryl)‐2,2,2‐trifluoroethanol (6), cobalt
(III) tris(acetylacetonate) (7), flavanone (8), and trans‐
cyclopropanedicarboxylic acid dianilide (9). The chromato-
gram for resolution of racemate 7 on 1a is shown in
Figure 7 with hexane/2‐propanol (90:10, v/v) as mobile
phase. The enantiomers were eluted at retention times of t1
and t2 with a baseline separation. The retention factors, k1'
[= (t1 ‐ t0)/t0] and k2' [= (t2 ‐ t0)/t0], were obtained as
1.51 and 2.20, respectively, which led the separation factor
α (= k2'/k1') to be 1.46.
In Table
2 are summarized the chromatographic
resolution results of racemates 2–9 on 1a–f using a
conventional eluent, hexane/2‐propanol (90:10, v/v).
Moreover, the data on the commercially available chiral
column, Chiralpak AD,18 utilizing coated amylose tris(3,5‐
dimethylphenylcarbamate) as chiral selector, which is
regarded as one of the most useful CSPs, are included. Com-
pared with Chiralpak AD, 1a–f showed higher chiral recogni-
tion abilities to racemates 2 and 7. Especially, racemate 7,
which is usually rather difficult to resolve on commercially
available columns, was completely resolved on these phases
with α values larger than 1.20. Racemates 2, 5, 7, and 9 were
better resolved on 1b and 1d than on Chiralpak AD, while the
FIGURE 6 Molecular structures of racemates 2–9

6
TANG ET AL.
TABLE 2 Resolution results of racemates 2–9 on 1a–h and Chiralpak AD
a, b
a, b
a, b
a, b
1a
k1'
1b
k1'
1c
k1'
1d
k1'
1e a, b
k1'
1f a, b
k1'
1g a, c
1h a, c Chiralpak ADc, d
α
α
α
α
α
α
α
α
α
2
3
4
5
6
7
8
9
0.89 (+) 2.03 1.22 (+) 2.44 0.84 (+) 2.02 1.15 (+) 2.13 1.02 (+) 1.76 1.87 (+) 2.91 1.85 (+) 2.31 (+)
0.69 (−) 1.16 0.64 1.00 0.64 1.00 0.78 (+) 1.14 0.59 (−) 1.33 1.13 1.00 1.30 (−) ca. 1 (−)
2.54 (−) 1.13 2.83 (−) 1.35 2.41 (−) 1.15 3.04 (−) 1.23 2.92 (−) 1.20 8.43 (−) 1.11 1.73 (−) 1.97 (−)
1.70 (+)
2.81 (+)
1.31 (−)
1.02 (−)
1.39 (+)
ca. 1 (−)
1.04 (+)
1.59 (+)
1.24
1.40
1.00 1.38 (+) 1.19 1.09
1.00 0.77 1.00 1.11
1.00 1.30 (+) 1.08 1.78 (−) 1.08 2.51
1.00 1.15 (−) 1.24 (+)
1.00 1.14 1.00 0.51 1.00 0.60 (+) 1.18 1.18 (−) 1.23 (+)
1.51 (+) 1.46 1.22 (+) 1.30 1.72 (+) 1.29 1.47 (+) 3.14 2.17 (+) 1.49 1.39 (+) 1.55 1.75 (−) 2.46 (−)
1.97 1.00 1.80 (−) 1.11 1.76 1.00 2.34 (+) 1.04 2.47 1.00 2.64 1.00 1.17 (−) 1.13 (+)
1.13 (+) 1.48 1.19 (+) 1.80 1.24 (+) 1.69 1.32 (+) 2.05 1.08 (+) 1.63 3.26 (+) 1.57 3.21 (+) 3.71 (+)
Mobile phase, hexane/2‐propanol (90:10, v/v). The structures of 1a–h are shown in Figure 1.
aColumn size, 25 × 0.2 cm i.d.; flow rate, 0.1 ml/min.
bThe signs in parentheses represent the CD detection of the first eluted enantiomer.
cData taken from Ref. 18. The signs in parentheses represent the optical rotation of the first eluted enantiome.
dColumn size, 25 × 0.46 cm i.d.; flow rate, 0.5 ml min−1
.
racemates 5 and 8, which were not resolved on 1a, were sep-
arated on 1b, with α values 1.19 and 1.11, respectively. How-
ever, larger α values for racemates 3 (α: 1a, 1.16; 1b, 1.00)
and 7 (α: 1a, 1.46; 1b, 1.30) were obtained on 1a rather than
1b. The chromatographic resolutions of racemates 3 and 8 on
1a and 1b are shown in Figure 8. While for 1c and 1d, 1d
always showed a higher enantioselective ability except race-
mate 6, which were not resolved on both derivatives, indicat-
ing that for 2‐benzoyl amylose derivatives the combination of
a
4‐methylphenylcarbamate at 3‐position and a 3,5‐
dimethylphenylcarbamate at 6‐position is superior for chiral
recognition of the tested racemates than that of reversed
FIGURE 8 Chromatographic resolution of racemates 3 and 8 on 1a and 1b. Conditions are the same as in Figure 7

TANG ET AL.
7
substituents at 3‐ and 6‐positions. For example, racemates 3,
5, and 8, which were not resolved on 1c, were resolved on 1d
and their α values were 1.14, 1.08, and 1.04, individually.
Another example is the resolution of racemate 7: its α value
was significantly larger on 1d (3.14) than that on 1c (1.29).
As for 1e and 1f, 1e exhibited a higher chiral recognition
for most racemates, except racemates 2 (α: 1e, 1.76; 1f,
2.91) and 6 (α: 1e, 1.00; 1f, 1.18). Especially racemates 3
and 5 were realized on 1e (α: 3, 1.33; 5, 1.08) rather than
on 1f. These results implied that, as well as the substituent
at 2‐position,20 the substituents at 3‐ and 6‐positions have a
large influence on enantioseparation.20,21
racemates. As for 1b and 1e, their chiral recognition abilities
differed greatly. For racemate 2, a much larger α value, 2.44,
was obtained on 1b than that on 1e, 1.76. In addition, race-
mate 8, which was not resolved on 1e, was partially resolved
on 1b (α, 1.11). On the contrary, with α value, 1.33, 1e rather
than 1b (α, 1.00) demonstrated a fairly good separation for
racemate 3, indicating that the introduction of an electron‐
withdrawing 3,5‐dichlorophenylcarbamate instead of an elec-
tron‐donating 3,5‐dimethylphenylcarbamate at 6‐position of
2‐benzoyl‐3‐phenylcarbamoyl amylose is preferable for the
chiral recognition of this racemate. Moreover, in the resolu-
tion of racemate 5, reversed elution orders of the enantiomers
were achieved on 1b and 1e, suggesting that the separation
mechanisms of racemate 5 were different. It seems that the
electronic effect of methyl and chloro groups are responsible
for the significantly changed chiral recognition between 1b
and 1e, which was the same for 3,5‐dimethylphenylcarbamtes
and 3,5‐dichlorophenylcarbamates of cellulose5 and amy-
lose.25 In the case of 1f and 1h, 1h always exhibited a
higher chiral recognition, except racemates 2 (α: 1f, 2.91;
1h, 2.31) and 3 (α: 1f, 1.00; 1h, ca. 1), demonstrating that
for 2‐benzoyl‐3‐(3,5‐dichlorophenylcarbamoyl) amylose,
the introduction of a more powerful electron‐donating
3,5‐dimethylphenylcarbamate rather than a less powerful
phenylcarbamte at 6‐position is favorable for enantioseparation.
In addition, racemates 5 and 8, which were realized on 1h (α:
5, 1.24; 8, 1.13), but were not on 1f. As indicated above, the
substituent at 6‐position influenced the chiral recognition of
amylose derivatives with three different substituents at 2‐,
3‐, and 6‐positions, and the extent depending on the nature
of substituent.
The data18 on 1g and 1h are also included in Table 2 for
comparison. Having the same substituents at 2‐ and 3‐posi-
tions, 1a, 1c, and 1g, 1b and 1e, and 1f and 1h exhibited var-
ied chiral recognition abilities depending on the substituent at
6‐position. As indicated in Table 2, the k1' values for all eight
racemates were close on 1a and 1c, suggesting that the inter-
actions between the racemates and the substituents at 6‐posi-
tion, being phenylcarbamate and 4‐methylphenylcarbamate
moieties, respectively, were similar, while their α values on
1a and 1c were somewhat different. For racemates 2, 4, 5,
6, and 8, close or identical α values were achieved on these
two derivatives. For racemates 3 and 7, higher α values were
realized on 1a (α: 3, 1.16; 7, 1.46) than on 1c (α: 3, 1.00; 7,
1.29). Especially racemate 3, which was partially resolved on
1a, was not resolved on 1c. For racemate 9, its α value on 1a
(α, 1.48) was smaller than on 1c (α, 1.69). Thus, 1a generally
exhibited a higher chiral recognition ability than that of 1c.
Their different chiral recognition abilities may be
ascribed to the different electronic effects of the substituents
at 6‐position, that is, phenylcarbamate is
a
less
Investigations indicated that the substituent at 3‐position
has a large influence on chiral recognition of regioselectively
substituted amylose derivatives,20,21 which is also the
same in this study. Bearing the same substituents at 2‐ and
6‐positions, 1b, 1d and 1h, 1a and 1f, and 1e and 1g, demon-
strated changed chiral recognition relying on the substituent
at 3‐position. Having different substituent at 3‐position, being
phenylcarbamte, 4‐methylphenylcarbamate, and 3,5‐
dichlorophenylcarbamate, respectively, 1b, 1d, and 1h
showed different chiral recognition abilities. In general, 1b
and 1d exhibited somewhat similar chiral recognition, while
1h showed the highest chiral recognition ability to most
racemates. Among these three derivatives, racemate 3 was
separated only on 1d and racemate 6 only on 1h,
demonstrating that the introduction of an electron‐donating
4‐methylphenylcarbamate and an electron‐withdrawing 3,5‐
dichlorophenylcarbamate at 3‐position was favorable for the
chiral recognition of the corresponding racemates. The elu-
tion orders of racemate 8 on 1b and 1d were reversed, mean-
ing that the separation mechanisms were different, which
might be ascribed to the different higher‐order structures of
the derivatives due to the different substituent at 3‐position.
powerful electron‐donating group compared with 4‐
methylphenylcarbamate. However, with the introduction of
an electron‐withdrawing 3,5‐dichlorophenylcarbamate at 6‐
position instead of an electron‐donating substituent being
phenylcarbamate or 4‐methylphenylcarbamate, the chiral rec-
ognition of 1g significantly differed from those of 1a and 1c,
indicating that the electronic effect of the chloro group
changed the chiral recognition of 1g dramatically, which
might be due to the changed local polarity and even the
higher‐order structure of the polymer. All eight racemates
were resolved on 1g, while only 5 and 4 racemates were sep-
arated on 1a and 1c, respectively. Among the three deriva-
tives, 1g always exhibited the highest enantioseparations for
the tested racemates, except racemates 2 and 9, whose α
values were slightly higher on 1a and 1c. Especially race-
mates 5, 6, and 8, which could not be resolved on 1a and
1c, were resolved on 1g, with α values, 1.15, 1.18, and
1.17, respectively, implying that the combination of an elec-
tron‐donating 3,5‐dimethylcarbamate at 3‐position and an
electron‐withdrawing 3,5‐dichlorophenylcarbamate at 6‐posi-
tion is favorable for chiral recognition of these three

8
TANG ET AL.
The highest α values obtained on 1b, 1d, and 1h were 1
(racemate 2), 2 (racemates 3 and 7), and 5 (racemates 4, 5,
6, 8, and 9) racemates, respectively. Racemates that were not
separated on 1b, 1d, and 1h were 2 (racemates 3 and 6), 1
(racemate 6), and 1 (racemate 3), individually. These results
imply that for 2‐benzoyl‐6‐(3,5‐dimethylphenylcarbamoyl)
amylose, the introduction of an electron‐withdrawing 3,5‐
dichlorophenylcarbamate rather than an electron‐donating
phenylcarbamate or 4‐methylphenylcarbamate at 3‐position is
more favorable for chiral recognition. For 1a and 1f, close α
values for most racemates were realized on these two deriva-
tives, except racemates 2 (α: 1a, 2.03; 1f, 2.91), 3 (α: 1a,
1.16; 1f, 1.00), and 6 (α: 1a, 1.00; 1f, 1.18). Thus, the intro-
duction of an electron‐donating 3,5‐dimethyphenylcarbamate
and an electron‐withdrawing 3,5‐dichlorophenylcarbamate at
3‐position of 2‐benzoyl‐6‐phenylcarbamoyl amylose is prefer-
able for the chiral recognition of racemates 3 and 6, respec-
tively. As for 1e and 1g, higher α values were always
achieved on 1g, except a slightly decreased α value for race-
mate 3. Especially racemates 6 and 8, which were not resolved
on 1e, were partially resolved on 1g (α: 6, 1.18; 8, 1.17). Also,
1g demonstrated a significantly larger α value, 3.21, for race-
mate 9 than 1e (α, 1.63). It seems that for 1e and 1g, a more
powerful electron‐donating 3,5‐dimethylphenylcarbamate
at 3‐position is more suitable than a less powerful
phenylcarbamate for chiral recognition.
The effect of mobile phase on chiral recognition was also
investigated and the results are listed in Table 3. Generally,
the k1' and α values were increased with the decreased con-
centration of 2‐propanol in mobile phase. A typical example
is the resolution of racemate 7 on 1d; its k1' values were
increased from 0.73 to 1.47 and 3.23, and α values from
2.74 to 3.14 and 3.25, when the concentration of 2‐propanol
decreased from 20 to 10 and 5%. The increased k1' values are
due to the increased H‐bonding interactions between the
racemates and the derivatives ascribed to the decreased H‐
bonding solvent, 2‐propanol. When ethanol was used instead
of 2‐propanol, poorer separations were often obtained. A
typical example is the resolution of racemate 3 on 1d. As
shown in Figure 9, when the eluent was switched from
TABLE 3 Influence of mobile phase on optical resolution of racemates 2–9 on 1a–f
1a 1b 1c
1d
1e
1f
Eluent
k1'
α
k1'
α
k1'
α
k1'
α
k1'
α
k1'
α
2
3
4
5
6
7
8
9
A
B
C
D
0.64 (+)
0.89 (+)
1.17 (+)
0.83 (+)
1.97
2.03
2.10
1.98
0.67 (+)
1.22 (+)
1.73 (+)
1.13 (+)
1.84
2.44
2.07
2.02
0.49 (+)
0.84 (+)
1.12 (+)
0.79 (+)
1.72
2.02
2.08
1.93
0.76 (+)
1.15 (+)
1.27 (+)
0.87 (+)
1.99
2.13
2.04
1.88
0.76 (+)
1.02 (+)
1.64 (+)
0.91 (+)
1.53
1.76
1.54
1.50
1.43 (+)
1.87 (+)
2.69 (+)
1.79 (+)
2.69
2.91
3.05
2.85
A
B
C
D
0.56 (−)
0.69 (−)
0.80 (−)
0.64 (−)
1.17
1.16
1.17
1.16
0.59
0.64
0.88 (−)
0.65
1.00
1.00
1.09
1.00
0.41
0.64
0.76
0.60
1.00
1.00
1.00
1.00
0.63
1.00
1.14
1.21
1.00
0.56 (−)
0.59 (−)
0.60 (−)
0.54 (−)
1.33
1.33
1.33
1.39
1.07
1.13
1.32
0.94 (−)
1.00
1.00
1.00
1.13
0.78 (+)
0.81 (+)
0.65
A
B
C
D
1.44 (−)
2.54 (−)
3.57 (−)
2.33 (−)
1.12
1.13
1.15
1.13
1.27 (−)
2.83 (−)
5.28 (−)
2.22 (−)
1.14
1.35
1.24
1.15
1.31 (−)
2.41(−)
3.65 (−)
2.09 (−)
1.12
1.15
1.16
1.16
1.65 (−)
3.04 (−)
6.19 (−)
2.14 (−)
1.21
1.23
1.28
1.15
1.75 (−)
2.92 (−)
4.57 (−)
2.25 (−)
1.20
1.20
1.20
1.18
5.69
1.00
1.11
1.17
1.34
8.43 (−)
13.86 (−)
6.15 (−)
A
B
C
D
0.84
1.24
1.62
1.18
1.00
1.00
1.00
1.00
0.78 (+)
1.38 (+)
2.08 (+)
1.11 (+)
1.14
1.19
1.21
1.14
0.61
1.09
1.49
1.00
1.00
1.00
1.00
1.00
0.82 (+)
1.30 (+)
1.88 (+)
1.00 (+)
1.09
1.08
1.10
1.13
1.60 (−)
1.78 (−)
3.38 (−)
1.74
1.07
1.08
1.12
1.00
1.96
2.51
3.33
2.49
1.00
1.00
1.00
1.00
A
B
C
D
0.60
1.40
2.62 (+)
1.31
1.00
1.00
1.07
1.00
0.38
0.77
1.73 (+)
0.79
1.00
1.00
1.11
1.00
0.48
1.11
2.30
1.05
1.00
1.00
1.00
1.00
0.39
1.14
2.45 (+)
0.80
1.00
1.00
1.10
1.00
0.33
0.51
1.56 (+)
0.48 (+)
1.00
1.00
1.17
1.54
0.29
1.00
1.18
1.28
1.27
0.60 (+)
1.01 (+)
0.53 (+)
A
B
C
D
0.80 (+)
1.51 (+)
2.41 (+)
1.45 (+)
1.35
1.46
1.34
1.39
0.64 (+)
1.22 (+)
5.84 (+)
1.13 (+)
1.25
1.30
1.21
1.30
0.92 (+)
1.72 (+)
3.21 (+)
1.54 (+)
1.21
1.29
1.28
1.26
0.73 (+)
1.47 (+)
3.23 (+)
0.97 (+)
2.74
3.14
3.25
1.95
1.19 (+)
2.17 (+)
5.42 (+)
1.74 (+)
1.39
1.49
1.56
1.11
0.77 (+)
1.39 (+)
2.73 (+)
1.28 (+)
1.68
1.55
1.60
1.52
A
B
C
D
1.42
1.97
2.41
1.93
1.00
1.00
1.00
1.00
1.32 (−)
1.80 (−)
3.09 (−)
1.79 (+)
1.13
1.11
1.15
1.13
1.08
1.76
2.26
1.55
1.00
1.00
1.00
1.00
1.73
1.00
1.04
1.07
1.11
1.77
2.47
4.47
2.01 (+)
1.00
1.00
1.00
1.10
2.20
2.64
3.74
2.58
1.00
1.00
1.00
1.00
2.34 (+)
3.23 (+)
1.72 (+)
A
B
C
D
0.61 (+)
1.13 (+)
3.18 (+)
1.02 (+)
1.58
1.48
1.93
1.37
0.58 (+)
1.19 (+)
3.39 (+)
1.15 (+)
1.51
1.80
1.46
1.57
0.55 (+)
1.24 (+)
3.29 (+)
1.21 (+)
1.46
1.69
1.80
1.62
0.55 (+)
1.32 (+)
3.68 (+)
1.08 (+)
2.00
2.05
1.98
1.64
0.45 (+)
1.08 (+)
2.67 (+)
0.95 (+)
1.50
1.63
1.87
1.27
1.38 (+)
3.26 (+)
8.86 (+)
2.18 (+)
1.77
1.57
1.44
1.60
Mobile phase: A hexane/2‐propanol (80:20, v/v), B hexane/2‐propanol (90:10, v/v), C hexane/2‐propanol (95:5, v/v), D hexane/ethanol (90:10, v/v). The signs in paren-
theses represent the CD detection of the first eluted enantiomer. Column size, 25 × 0.2 cm i.d.; flow rate, 0.1 ml min−1
.

TANG ET AL.
9
FIGURE 9 Chromatographic resolution of racemate 3 on 1d and 6 on 1e. Column size, 25 × 0.20 cm i.D.; flow rate, 0.1 ml min−1. H, hexane; I, 2‐
propanol; E, ethanol
hexane/2‐propanol (90:10, v/v) (α, 1.14) to hexane/ethanol
(90:10, v/v), it could not be resolved. However, in the
resolution of racemate 6 on 1e (Figure 9) and 3 on 1f, better
resolutions were obtained when ethanol was used in place of
2‐proppanol, which may be due to the fact that the changes of
the stereo environment of the chiral cavities in the
polymer chains caused by ethanol is favorable for the
enantioseparations of the corresponding racemates.26
phase on chiral recognition was also investigated. In general,
with the decreased concentration of 2‐propanol in eluent,
better resolutions were obtained. Moreover, when ethanol
was used instead of 2‐propanol, poorer resolutions were
often achieved. However, in some cases, improved
enantioselectivity was achieved with ethanol rather than
2‐propanol as the mobile phase modifier.
ACKNOWLEDGMENTS
4 | CONCLUSION
Financial support from the Science and Technology Plan
Projects of Zhejiang Province (Grant No. 2015C37034) to
W. Tang and valuable materials support from Professor Y.
Okamoto are gratefully acknowledged.
In this study, six amylose derivatives having a benzoate at
2‐postition and two different phenylcarbamates at 3‐ and
6‐positions were synthesized and their structures were char-
acterized by 1H NMR spectroscopy. After being coated onto
3‐aminopropyl silica gels, they were evaluated as CSPs for
HPLC enantioseparations of eight racemates. The results
indicated that the derivatives demonstrated characteristic
enantioseparation abilities and some racemates were better
resolved on these derivatives than on Chiralpak AD. The sub-
stituents at 3‐ and 6‐positions have a large influence on their
chiral recognition. Among the derivatives prepared, amylose
2‐benzoate‐3‐(phenylcarbamate/4‐methylphenylcarbamate)‐
6‐(3,5‐dimethylphenylcarbamate) exhibited chiral recogni-
tion abilities comparable to that of Chiralpak AD and may
be potential useful CSPs in the future. The effect of mobile
REFERENCES
1. Tang S. Chiral stationary phases. In: Zou H, Ou J, eds. Liquid
Chromatographic Separation Material: The Preparation and Appli-
cation. Beijing: Chemical Industry Press; 2016:228‐241.
2. Salvatore F. Nano‐liquid chromatography applied to enantiomers
separation. J Chromatogr A. 2017;1486:20‐34.
3. Daniele S, Rosella F, Leo Z, Adriano C, Roberto C. Analytical and
semipreparative high performance liquid chromatography
enantioseparation of bicalutamide and its chiral impurities on an
immobilized polysaccharide‐based chiral stationary phase.
Chromatogr A. 2016;1445:166‐171.
J

10
TANG ET AL.
4. Francotte E, Zhang T. Preparation and evaluation of immobilized
16. Shen J, Ikai T, Shen X, Okamoto Y. Synthesis and immobilization of
4‐methylbenzoylcellulose stationary phases for enantioselective
separations. J Chromatogr A. 2016;1467:214‐220.
amylose derivatives bearing a 4‐tert‐butylbenzoate group at the
2‐position
and
3,5‐dichlorophenylcarbamate/3‐(triethoxysilyl)
propylcarbamate groups at 3‐ and 6‐positions as chiral packing
material for HPLC. Chem Lett. 2010;39:442‐444.
5. Okamoto Y, Kawashima M, Hatada K. Chromatographic resolution.
XI. Controlled chiral recognition of cellulose triphenylcarbamate
17. Shen J, Li P, Liu S, Shen X, Okamoto Y. Immobilization and
chromatographic evaluation of novel regioselectively substituted
amylose‐based chiral packing materials for HPLC. Chirality.
2011;23:878‐886.
derivatives supported on silica gel.
1986;363:173‐186.
J
Chromatogr.
6. Peluso P, Mamane V, Cossu S. Liquid chromatography
enantioseparations of halogenated compounds on polysaccharide‐
based chiral stationary phases: Role of halogen substituents in
molecular recognition. Chirality. 2015;27:667‐684.
18. Kondo S, Yamamoto C, Kamigaito M, Okamoto Y. Synthesis and
chiral recognition of novel regioselectively substituted amylose
derivatives. Chem Lett. 2008;37:558‐559.
7. Kurka O, Kučera L, Bednř P. Analytical and semipreparative
chiral separation of cis‐itraconazole on cellulose stationary phases
by high‐performance liquid chromatography.
2016;39:2736‐2745.
19. Dicke R. A straight way to regioselectively functionalized polysac-
charide esters. Cellul. 2004;11:255‐263.
J
Sep Sci.
20. Shen J, Ikai T, Okamoto Y. Synthesis and chiral recognition of novel
amylose derivatives containing regioselectively benzoate and
phenylcarbamate groups. J Chromatogr A. 2010;1217:1041‐1047.
8. Shen J, Ikai T, Okamoto Y. Synthesis and application of
immobilized polysaccharide‐based chiral stationary phases for
enantioseparation by high‐performance liquid chromatography.
J Chromatogr A. 2014;1363:51‐61.
21. Sun B, Li X, Jin Z, et al. Preparation of regioselectively modified
amylose derivatives and their applications in chiral HPLC.
Chromatographia. 2012;75:1347‐1354.
9. Bezhitashvili L, Bardavelidze A, Ordjonikidze T, et al. Effect of
pore‐size optimization on the performance of polysaccharide‐based
superficially porous chiral stationary phases for the separation of
enantiomers in high‐performance liquid chromatography.
J Chromatogr A. 2017;1482:32‐38.
22. Shen J, Li G, Yang Z, Okamoto Y. Synthesis and chiral recognition
of amylose derivatives bearing regioselective phenycarbamate
substituents at 2,6‐ and 3‐positions for high‐performance liquid
chromatography. J Chromatogr A. 2016;1467:199‐205.
23. Koller H, Rimböck KH, Mannschreck A. High‐pressure liquid chro-
matography on triacetylcellulose. J Chromatogr. 1983;282:89‐94.
10. Shen J, Okamoto Y. Efficient separation of enantiomers using
stereoregular chiral polymers. Chem Rev. 2016;11:1094‐1138.
24. Mobian P, Nicolas C, Francotte E, Burgi T, Lacour J. Synthesis, res-
olution, and VCD analysis of an enantiopure diazaoxatricornan
derivative. J Am Chem Soc. 2008;130:6507‐6514.
11. Kaida Y, Okamoto Y. Optical resolution on regioselectively
carbamoylated cellulose and amylose with 3,5‐dimethylphenyl and
3,5‐dichlorophenyl
1993;66:2225‐2232.
isocyanates.
Bull
Chem
Soc
Jpn.
25. Okamoto Y, Aburatani R, Fukumoto T, Hatada K. Useful chiral
stationary
phases
for
HPLC.
Amylose
tris(3,5‐
12. Chassaing C, Thienpont A, Félix G. Regioselective carbamoylated
and benzoylated cellulose for the separation of enantiomers in
dimethylphenylcarbamate) and tris(3,5‐dichlorophenylcarbamate)
supported on silica gel. Chem Lett. 1987;16:1857‐1860.
high‐performance liquid chromatography.
1996;738:157‐167.
J Chromatogr A.
26. Sardella R, Ianni F, Lisanti A, Marinozzi M, Scorzoni S, Natalini B.
The effect of mobile phase composition in the enantioseparation of
pharmaceutically relevant compounds with polysaccharide‐based
stationary phases. Biomed Chromatogr. 2014;28:159‐167.
13. Chassaing C, Thienpont A, Soulard M‐H, Félix G. Regioselectively
carbamoylated polysaccharides for the separation of enantiomers in
high‐performance liquid chromatography.
1997;786:13‐21.
J Chromatogr A.
How to cite this article: Tang S, Jin Z, Sun B, Wang
F, Tang W. Preparation and evaluation of
regioselectively substituted amylose derivatives for
chiral separations. Chirality. 2017;0:1‐10. https://doi.
org/10.1002/chir.22720
14. Felix G. Regioselectively modified polysaccharide derivatives as
chiral stationary phases in high‐performance liquid chromatography.
J Chromatogr A. 2001;906:171‐184.
15. Tang S, Li F, Wang F, Liu G, Li Y, Pan F. Synthesis and HPLC
chiral recognition of regioselectively carbamoylated cellulose deriv-
atives. Chirality. 2012;24:167‐173.