Batch and in-flow kinetic resolution of racemic 1-(N-acylamino)alkylphosphonic and 1-(N-acylamino)alkylphosphinic acids and their esters using immobilized penicillin G acylase

Article abstract of DOI:10.1016/j.tetasy.2016.11.007

The kinetic resolution of racemic 1-(N-acylamino)alkylphosphonic acids 3 (R³= OH) and their dimethyl esters 1, as well as 1-(N-acylamino)alkylphosphinic acids 4 (R³= H or Ph) using penicillin G acylase (PGA) immobilized on three type

Full text of DOI:10.1016/j.tetasy.2016.11.007

Tetrahedron: Asymmetry xxx (2016) xxx–xxx
Contents lists available at ScienceDirect
Tetrahedron: Asymmetry
journal homepage: www.elsevier.com/locate/tetasy
Batch and in-flow kinetic resolution of racemic 1-(N-acylamino)
alkylphosphonic and 1-(N-acylamino)alkylphosphinic acids and their
esters using immobilized penicillin G acylase
a
b
a,
b,c
⇑
´
´
Katarzyna Zielinska , Katarzyna Szymanska , Roman Mazurkiewicz , Andrzej Jarze˛bski
a Department of Organic and Bioorganic Chemistry and Biotechnology, Faculty of Chemistry, Silesian University of Technology, Krzywoustego 4, 44-100 Gliwice, Poland
b Department of Chemical Engineering and Process Design, Faculty of Chemistry, Silesian University of Technology, Strzody 7, 44-100 Gliwice, Poland
c Institute of Chemical Engineering, Polish Academy of Sciences, Bałtycka 5, 44-100 Gliwice, Poland
a r t i c l e i n f o
a b s t r a c t
The kinetic resolution of racemic 1-(N-acylamino)alkylphosphonic acids 3 (R3 = OH) and their dimethyl
esters 1, as well as 1-(N-acylamino)alkylphosphinic acids 4 (R3 = H or Ph) using penicillin G acylase
(PGA) immobilized on three types of mesoporous silicas in both a batch slurry system and in a continu-
ous-flow reactor was studied. The initial hydrolytic deacylation rates in the presence of those catalysts
were measured and the relationships between the substrate structure and the enzyme efficiency are dis-
cussed. The stereospecific hydrolysis of the N-acyl group of both racemic N-acylated phosphorus ana-
logues of amino acids and their esters catalyzed by the immobilized PGA proved to be a highly
effective method for the kinetic resolution of all the investigated compounds, with the stereochemical
preference of PGA for (R)-substrates.
Article history:
Received 13 October 2016
Revised 8 November 2016
Accepted 11 November 2016
Available online xxxx
Ó 2016 Elsevier Ltd. All rights reserved.
1. Introduction
PGA has tremendous industrial application in the hydrolytic
deacylation of penicillin G to 6-aminopenicillanic acid (about
Chiral,
non-racemic
a-aminoalkylphosphonic
and
a-
20,000 t aꢀ1) and its reacylation to semi-synthetic b-lactam antibi-
otics (ampicillin, amoxicillin, cephalexin, cefaclor, cefadroxil and
others),9 with immobilized PGA accounts for 88% of worldwide
6-aminopenicillanic acid production.10–14
aminoalkylphosphinic acids have attracted significant attention
from chemists and biochemists for many years due to their diverse
biological activity and wide range of applications, mainly as phar-
maceuticals and agrochemicals.1–6
PGA has been immobilized on various matrices, ranging from
natural and synthetic organic polymers to inorganic carriers.10,14–
16 Inorganic mesoporous silicas such as SBA-15 or MCF offer some
special advantages, such as uniform, tunable pore size and struc-
ture, large surface area, opened pore structure, openness to a wide
variety of chemical modifications, great potential for high enzyme
loading and stability towards organic solvents.10–12,14 Diverse
methods for PGA immobilization were evaluated and compared in
a comprehensive review by Sheldon et al.10 Covalent immobiliza-
tion has generally been favored in the case of this enzyme.10,17,18
Herein, we report our studies on the immobilization of PGA
from Escherichia coli (EC 3.5.1.11) on powder mesoporous silica
of SBA-15 and MCF type and on the application of immobilized
PGA for the kinetic resolution of racemic 1-(N-acylamino)
alkylphosphonic and 1-(N-acylamino)alkylphosphinic acids and
their esters, including a comparative study on the effectiveness
of immobilized and native PGA in these processes (Scheme 1). A
Recently we reported the stereospecific hydrolysis of the N-acyl
group of both racemic 1-(N-acylamino)alkylphosphonic acids 3
(R3 = OH) and their dimethyl esters 1 catalyzed by the native peni-
cillin G acylase (PGA) as a highly effective method for the kinetic
resolution of these compounds.7 Looking at this problem from an
application perspective, we deemed it important to undertake a
study of the same process, but employing immobilized PGA. Immo-
bilization of the enzyme provides a number of important advan-
tages, including: (i) the ease of enzyme separation from the
reaction mixture, which in turn simplifies reaction mixture pro-
cessing; (ii) better handling and re-use of the enzyme; and (iii)
the possibility of employing an immobilized enzyme in continu-
ous-flow systems, with positive effects on the performance and
process economy. Other possible improvements include increased
operational and storage enzyme stability, and increased enzyme
resistance against extremes in pH, temperature and ionic strength.8
successful kinetic resolution of an
tive in a continuous-flow microfluidic system containing the PGA
a-aminophosphonic acid deriva-
⇑
Corresponding author.
E-mail address: roman.mazurkiewicz@polsl.pl (R. Mazurkiewicz).
http://dx.doi.org/10.1016/j.tetasy.2016.11.007
0957-4166/Ó 2016 Elsevier Ltd. All rights reserved.
´
Please cite this article in press as: Zielinska, K.; et al. Tetrahedron: Asymmetry (2016), http://dx.doi.org/10.1016/j.tetasy.2016.11.007

´
K. Zielinska et al. / Tetrahedron: Asymmetry xxx (2016) xxx–xxx
2
O
R2
R2
immobilized PGA
pH=7.0, 25oC
R2
O
R1
N
H
P(OMe)2
R1
N
H
P(OMe)2
+
H2N
P(OMe)2
O
O
O
(R)-2
1
(S)-1
R2
O
R2
immobilized PGA
pH=7.0, 25oC
R2
O
R3
OH
R3
OH
R3
OH
+
R1
N
H
P
R1
N
H
P
H2N
P
O
O
O
3, 4
(R)-5,6
(S)-3,4
Scheme 1. Kinetic resolution of 1-(N-acylamino)alkylphosphonic, 1-(N-acylamino)alkylphosphinic acids and their esters using immobilized forms of penicillin G acylase.
immobilized inside the multichannel siliceous monolith with
bi-modal pore structure is also reported.
bonded the enzyme via its amino groups with the free formyl
groups of the glutaraldehyde moiety.
The immobilized PGA was finally obtained in the form of a sus-
pension of silica powders in a phosphate buffer (pH 7.5), and it was
stored at 4 °C in this form. The measured specific activity of the
enzyme immobilized on SBA-15 and MCF was 273 and 2066 U/g,
respectively, whereas the amounts of the protein bonded were
equal to 20.9 mg/g for SBA-15 (immobilization yield 38%) and
49.4 mg/g for MCF (immobilization yield 95%). Thus the measured
activity was more than seven times larger for the MCF-based
enzyme than for SBA-15, despite there being only about two-fold
larger the amount of protein bonded on to MCF. This can be
explained by the specific structural features of the applied sup-
ports. Very open cellular pores (20–50 nm in diameter and ca.
2 cm3/g mesopore volume) connected by the wide windows (14–
18 nm) typical for MCFs allow for the unhindered diffusion of sub-
strates and products. The smaller hexagonally arranged cylindrical
pores of SBA-15 (6–8 nm, 0.8 cm3/g mesopore volume) may be big
enough to contain molecules of the attached enzyme, but may
restrict substrate or product diffusion.
2. Results and discussion
2.1. PGA immobilization
PGA was immobilized on two types of mesoporous silicas
(MPSs): SBA-15 (Santa Barbara Amorphous) and MCF (mesoporous
cellular foams). MPSs have a rigid, open-pore structure and a large
surface area that can be densely covered with various anchor
groups. They can be synthesized to obtain pore sizes ranging from
2 to 40 nm. Moreover they are environmentally acceptable, struc-
turally stable and resistant to microbial attacks. The mesoporous
silicas were prepared by the templating method reported in our
previous papers.20–22 The silicas were next activated by exposure
to water vapour at room temperature for 5 h, then dried at
200 °C and treated with (3-aminopropyl)triethoxysilane in toluene
to obtain mesoporous silicas functionalized with 3-aminopropyl
groups with amino groups density of ca. 1.5 mmol per 1 g of silica
(Scheme 2).24 Further activation of the functionalized silicas with a
solution of glutaraldehyde in phosphate buffer bonded the glu-
taraldehyde by the covalent azomethine bond. Treatment of these
silicas with a solution of PGA in phosphate buffer covalently
PGA was also immobilized in siliceous monolithic rods (MH)
with a hierarchical pore structure. The monoliths (45 mm in length
and 6 mm in diameter) were prepared by using the Nakanishi
method25 with minor modification, as recently described in our
papers (Fig. 1a).23,24 Their functionalization with aminopropyl
O
O
Si
Si
Si
O
O
O
Si OH
H
H
+
(EtO)3Si-(CH2)3-NH2
Si OH
Si OH
Si (CH2)3NH2
Si
Si
Si
O
Si
O
O
O
NH2-PGA
O
O
Si (CH2)3N=CH(CH2)3CH=N-PGA
Si (CH2)3N=CH(CH2)3CHO
Si
Si
Scheme 2. Immobilization of penicillin G acylase on mesoporous silicas.
(a)
(b)
monilithic rod
polymeric resin
cover
6 mm
21 mm
Figure 1. (a) PGA immobilized in a monolithic rod clad with polymeric resin. (b) A single-rod continuous-flow microreactor system.
´
Please cite this article in press as: Zielinska, K.; et al. Tetrahedron: Asymmetry (2016), http://dx.doi.org/10.1016/j.tetasy.2016.11.007

´
3
K. Zielinska et al. / Tetrahedron: Asymmetry xxx (2016) xxx–xxx
groups was achieved by immersing the rods in a solution of 3-
(aminopropyl)triethoxysilane in toluene for 72 h at 80 °C with
intensive stirring. Each monolith was then clad with an epoxide
polymer resin (L285MGS-H285MGS type) to obtain a single-rod
multichannel microreactor. Each microreactor was rinsed with a
series of solvents under flow, after which a solution of glutaralde-
hyde in phosphate buffer was pumped through the microreactor
and, after rinsing with water and phosphate buffer, a solution of
PGE in the phosphate buffer was pumped through the microreactor
in a circular system for 3 h. Finally, the microreactor was rinsed
again with water and a series of buffers.
The calculated amount of protein bounded inside the monolith
(14.5 mg, or 44.6 mg/g of monolith, for a monolith weight of
0.325 g) was very close to that obtained for MCF (49.4 mg/g). This
can also be explained by its very open pore structure (ca. 1 cm3/g of
mesopore volume, and over 2.5 cm3/g of macropore volume) as
determined by the nitrogen adsorption and mercury porosimetry
experiments.23,24
carried out in a solution of phosphate buffer (pH 7.0) at 25 °C with
intensive stirring using various reaction times and amounts of
immobilized enzyme depending on the specific reactivity of the
substrate to achieve substrate conversion close to 50% (Scheme 1,
Table 1). The reaction was stopped by filtering off the immobilizate
by using folded paper filters. The unconverted substrates and
hydrolysis products were separated by column chromatography
on silica gel (for the esters) or by ion-exchange chromatography
on DOWEX 50W X8 (H+ form) for the free acids. The enantiomeric
excesses in the unreacted substrates were determined as described
in our previous paper7 based on the integration of 31P NMR or 1H
NMR signals of both enantiomers and using quinine as the chiral
discriminating agent (Table 1). The enzyme enantioselectivities E,
as the most intrinsic expression of enzyme stereoselectivity, were
calculated as
a function of the enantiomeric excess in the
unchanged substrate and conversion of the substrate, as previously
described.7
As can be seen from Table 1, highly effective kinetic resolutions
with E-values exceeding 100 were obtained for all of the investi-
gated N-acylated substrates, i.e. for the phosphonic analogues of
valine, leucine, and asparagine 3a–c, the dimethyl esters of the
phosphonic analogues of alanine, phenylalanine leucine and aspar-
agine 1a–d, H-phosphinic analogues of alanine 4a–b and the
phenylphosphinic analogue of alanine 4c. It is noteworthy that for
two of the investigated substrates, i.e. the phosphonic analogue of
valine 3a and the dimethyl ester of leucine 1c, the kinetic resolution
catalyzed with native PGA gave only a moderate enzyme enantios-
electivity (22 and 66, respectively).7 The specific rotations of the
unchanged substrates and hydrolysis products were close to those
obtained for the same compounds in the kinetic resolution experi-
ment using native PGA7, which confirms the stereochemical prefer-
ence of both native and immobilized PGA for the (R)-substrate.
2.2. Initial hydrolysis rate measurements
The initial rates of enzymatic hydrolytic deacylation of N-acy-
lated 1-aminoalkylphosphonic and 1-aminoalkylphosphinic acids,
as well as their esters, were measured for each specific substrate
in a series of preliminary experiments in order to assure optimal
substrate conversion of approximately 50% in the further kinetic
resolution experiments (Scheme 1). The measurements were car-
ried out in a water solution of phosphate buffer (pH 7.0) at 25 °C,
with intensive stirring, using various amounts of the immobilized
enzyme depending on the specific reactivity of substrate (Table 1).
Approximately 2 mL of aliquots of the reaction mixture were
extracted at various reaction times through MinisartÒ SRP 4 syr-
inge filters to stop the reaction. Concentrations of the reaction pro-
duct and the unconverted substrate in the reaction mixtures were
determined by means of 1H NMR spectroscopy using pivalic acid as
the internal standard. The determined initial hydrolysis rates for
immobilized PGA were compared in Table 1 with those reported
in our previous paper for native PGA.7
2.4. Kinetic resolution of racemic 1-(N-phenylacetylamino)-2-
carbamoylethylphosphonic acid dimethyl ester 1d in a flow
system in a single-rod microfluidic microreactor
In
a preliminary experiment, a solution of substrate 1d
Most of the values for the initial hydrolysis rates for the reac-
tions catalyzed with PGA immobilized on MCF or SBA-15 were
similar or slightly lower in comparison with those obtained for
the reactions of the same compounds catalyzed by native PGA.7
However, the initial hydrolysis rates in the presence of MCF-immo-
bilized PGA were higher by a factor of 6.6 and 1.8, respectively, in
the case of the phosphonic analogue of asparagine 3c and the
H-phosphinic analogue of alanine 4b. Overall, the results obtained
with the immobilized PGA confirm the relationship between the
substrate structure and the enzyme efficiency observed for the
native PGA and reported in our previous paper.7 Thus the initial
hydrolysis rate of the N-phenylacetylamino group is higher in com-
parison with the N-benzyloxycarbonylamino group. cf. compounds
4a and 4b; the hydrolysis rates of the free acids are distinctly
higher when compared to the corresponding methyl esters, cf. 1c
and 3b, as well as 1d and 3c, while the hydrolysis rate rapidly
decreases when increasing the steric effect of the substituent at
(3 mmol/L) in phosphate buffer (pH 7) was pumped through the
microreactor (Fig. 1b) with a gradually diminishing flow rate, start-
ing with a value of 0.5 mL/min, to determine the dependency
between the flow rate and substrate conversion (Table 2). Assum-
ing that the free space inside the monolith is equal to 0.82 mL,
which corresponds to an estimated monolith porosity of 65% at
its geometrical volume of 1.27 mL, and that the activity of PGA in
the monolith is equal to 333.8 U, the average hydrolysis rate V of
the substrate was calculated at each specific flow rate (Table 2).
A good linear relationship between log(V) and substrate conver-
sion C was observed (Fig. 2). Extrapolation of this relationship to
C = 0 allowed us to determinate the initial hydrolysis rate in the
microreactor equal to 6.07 ꢁ 10ꢀ4
lmol/U min (which corresponds
to log(V) = ꢀ3.2169). The latter value is very close to the initial
hydrolysis rate of the same substrate catalyzed by the native
PGA, as reported in our previous paper (6.83 ꢁ 10ꢀ4
l
mol/U min).7
The experiment described above revealed that a substrate con-
version close to 50% can be achieved at a flow rate of 0.01 mL/min.
Therefore, in the final kinetic resolution experiment, a solution of
substrate (3 mmol/L, 100 mL) in phosphate buffer was pumped
through the microreactor at a flow rate of 0.01 mL/min. The uncon-
verted substrate (S)-1d and the hydrolysis product (R)-2d were iso-
lated from the outflow (87 mL) by column chromatography on
silica gel as described above in yields of 47% and 44%, respectively.
Their specific rotations, +5.4 and ꢀ2.4, respectively, were close to
those obtained for the same compounds obtained in the kinetic
resolution experiment using native PGA.7 The enantiomeric excess
in the unchanged substrate was determined to be 99% (1H NMR
the
a-position, cf. 1a versus 1b–d. The presence of a bulky sub-
stituent at the phosphorus atom in the phenylphosphinic acid
derivative 4c only slightly reduces its reactivity in comparison with
the corresponding H-phosphinic analogue 4b.
2.3. Kinetic resolution of racemic 1-(N-acylamino)alkylpho-
sphonic and 1-(N-acylamino)alkylphosphinic acids and their
esters in a batch system
The kinetic resolutions of N-acylated 1-aminoalkylphosphonic
and 1-aminoalkylphosphinic acids as well as their esters were
´
Please cite this article in press as: Zielinska, K.; et al. Tetrahedron: Asymmetry (2016), http://dx.doi.org/10.1016/j.tetasy.2016.11.007

Table 1
Enzymatic hydrolysis of N-acylated 1-aminoalkylphosphonic acids 3, their esters 1 and 1-aminoalkylphosphinic acids 4 catalyzed by immobilized PGA
Substrate
R2
Activity of PGA
[U]a/matrice
Initial hydrolysis [
Immobilized PGA
l
mol/U min]
Kinetic resolution results
Yield [%]b
½ ꢂ
a 2D0
No.
R1
R3
Native PGA7
Prod. No.
Time [h]
Conversion
Product config.
ees ꢁ 100 [%]
E
Substrate
Product
Substrate
Product
1a
1b
1c
1d
3c
4a
4b
4c
3a
3b
PhCH2
PhCH2
PhCH2
PhCH2
PhCH2
PhCH2O
PhCH2
PhCH2
PhCH2
PhCH2
Me
PhCH2
i-PrCH2
H2NCOCH2
H2NCOCH2
Me
Me
Me
i-Pr
i-PrCH2
—
—
—
—
OH
H
H
Ph
OH
OH
1.44/MCF
306/MCF
270/MCF
198/MCF
134/MCF
2.45/MCF
3.96/MCF
2.31/MCF
345/SBA
259/SBA
1.41 0.16 ꢁ 101
5.13 0.64 ꢁ 10ꢀ4
4.59 0.34 ꢁ 10ꢀ4
2.26 0.19 ꢁ 10ꢀ4
2.78 0.17 ꢁ 10ꢀ2
0.78 0.16 ꢁ 10ꢀ1
1.10 0.06 ꢁ 10ꢀ1
4.80 0.49
1.46 ꢁ 101
5.21 ꢁ 10ꢀ4
6.89 ꢁ 10ꢀ4
6.83 ꢁ 10ꢀ4
4.19 ꢁ 10ꢀ3
1.70 ꢁ 10ꢀ1
6.20
2a
2b
2c
2d
5c
6a
6a
6b
5a
5b
0.67
144
96
48
24
72
1
2
144
48
0.48
0.49
0.48
0.47
0.49
0.48
0.48
0.49
0.49
0.49
43
45
48
42
46
45
46
48
46
49
49
43
47
32
41
44
49
47
45
43
+43.9c
+28.7e
+31.7g
+5.4i
ꢀ6.4d
ꢀ25.2f
ꢀ17.2h
ꢀ2.4j
(R)
(R)
99
99
99
87
99
92
91
99
99
99
>100
>100
>100
300
>100
1990
450
>100
>100
>100
+7.6k
ꢀ32.8l
ꢀ6.7n
ꢀ6.8r
(R)
(R)
(R)
+47.6m
+67.8p
+81.1s
ꢀ0.7u
+31.6w
4.28
ꢀ38.4t
+1.0v
2.15 0.08 ꢁ 10ꢀ4
7.16 ꢁ 10ꢀ4
(R)
(R)
2.91 0.13 ꢁ 10ꢀ3
7.44 ꢁ 10ꢀ3
ꢀ24.1z
a
b
c
Activity of the immobilized PGA used.
Isolated yield after chromatography.
(c 2.4 MeOH), Lit.4
(c 1.5, MeOH).
(c 0.9, CHCl3).
[a]
D
20 = +44.5 (S) (c 1, MeOH).
d
e
f
(c 1.8, MeOH), Lit.26
(c 2, CHCl3).
[a]
20 = +25.9 (S) (c 0.9, CHCl3).
D
g
h
i
(c 2, CHCl3).
(c 0.1, CH3CN).
(c 1.6, CH3CN).
(c 1, 1 M NaOH).
j
k
l
(c 1, H2O), Lit.27
[
a
]
20 = ꢀ33.0 (R) (c 1, H2O).
D
m
n
p
r
(c 0.25, 1 M NaOH), Lit.28
[a
]
D
20 = +48.3 (R) (c 1, acetic acid).
(c 1, H2O), Lit.29
[
a
[
a
]
20 = ꢀ6.8 (R) (c 1, H2O).
D
(c 0.1, H2O), Lit.29
a
]
D
20 = +74.0 (S) (c 0.5, H2O).
(c 1, H2O), Lit.29
(c 1, 1 M NaOH).
[
]
D
20 = ꢀ6.8 (R) (c 0.5, H2O).
s
t
(c 0.4, 1 M NaOH).
(c 1, 1 M NaOH).
(c 2, 1 M NaOH).
(c 1, 1 M NaOH).
u
v
w
z
(c 0.5, 1 M NaOH), Lit.30
[a
]
D
20 = ꢀ25.0 (R) (c 1, 1 M NaOH).

´
5
K. Zielinska et al. / Tetrahedron: Asymmetry xxx (2016) xxx–xxx
Table 2
and 1d, H-phosphinic analogues of alanine 4a–b and the
phenylphosphinic analogue of alanine 4c, as well as for the phos-
phonic analogue of valine 3a and the dimethyl ester of leucine
1c. In the case of the last two compounds, the kinetic resolution
catalyzed with the native PGA only gave a moderate value of
enzyme enantioselectivity (22 and 66, respectively).7 As in the case
of the native PGA,7 the stereochemical preference of the immobi-
lized PGA for the (R)-substrates was found.
The dependence of the conversion of 1-(N-phenylacetylamino)- 2-car-
bamoylethylphosphonic acid dimethyl ester (R,S)-1d versus the flow rate
Flow rate
[mL/min]
Contact
time [min]
Conversion C
Initial hydrolysis
V [
mol/Uꢃmin]
l
0.5
0.4
0.3
0.2
0.1
0.05
0.01
1.65
2.07
2.75
4.13
8.25
16.5
82.6
0.085
0.095
0.118
0.162
0.242
0.312
0.490
2.31 ꢁ 10ꢀ4
1.65 ꢁ 10ꢀ4
1.16 ꢁ 10ꢀ4
7.05 ꢁ 10ꢀ5
2.63 ꢁ 10ꢀ5
8.50 ꢁ 10ꢀ6
5.33 ꢁ 10ꢀ7
The initial hydrolytic deacylation rate of racemic 1-(N-pheny-
lacetylamino)-2-carbamoylethylphosphonic acid dimethyl ester
1d in a flow system with PGA immobilized in the siliceous mono-
lith (6.07 ꢁ 10ꢀ4
lmol/U min) was proven to be very close to
those measured for the native PGA and the same substrate
(6.83 ꢁ 10ꢀ4
l
mol/U min).7 High yields of isolation of the hydroly-
-2
sis product (R)-2d and the unconverted (S)-1d with high enan-
tiomeric purity of the latter compound (99%) demonstrate that
the kinetic resolution of N-acylated phosphorus analogues of
amino acids in a flow system using siliceous monolithic rods with
a hierarchical pore structure as PGA carriers is potentially one of
the most convenient and effective methods for the kinetic resolu-
tion of this class of compounds.
0
0.05
0.1
0.15
0.2
0.25
0.3
-2.5
-3
y = -5,6994x - 3,2169
R² = 0,986
-3.5
-4
4. Experimental
4.1. General
-4.5
-5
Conversion C
Melting points were determined using capillary tubes in a Stir-
ling SMP 3 apparatus and are uncorrected. IR-spectra were mea-
sured on a Nicolett 6700 FT-IR spectrophotometer (ATR method).
1H and 13C NMR spectra were recorded on Varian instruments at
operating frequencies of 400 or 600 MHz and 100 or 150 MHz,
respectively, using TMS as a resonance shift standard. 31P NMR
spectra were recorded on a Varian instrument at an operating fre-
quency of 161.8 MHz with 80% orthophosphoric acid as an external
resonance shift standard. Column chromatography was conducted
using Merck 60 silica gel (70–230 mesh). Ion-exchange chromatog-
raphy was carried out using DOWEX 50W X8 in the H+ form. Rou-
tine monitoring of reactions was done with the use of thin layer
chromatography, using TLC plates coated with silica gel (Merck
60 F254). Optical rotations were measured on a JASCO V-650
polarimeter. Buffers used: phosphate buffer (Na2HPO4/KH2PO4)
pH 7.5, phosphate buffer (Na2HPO4/KH2PO4) pH 7.0, acetate buffer
(0.2 M CH3COOH/CH3COONa) pH 4.5, Tris–HCl buffer (tris(hydrox-
ymethyl)aminomethane/0.2 M HCl) pH 7.8.
Figure 2. Variation of the initial hydrolysis rate V of (R,S)-1d versus the substrate
conversion C in a semi-logarithmic system.
method). Thus the performed experiment proved that kinetic
resolution of racemic N-acylated phosphorus analogues of amino
acids in a flow system using siliceous monolithic rods as PGA car-
riers is potentially one of the most convenient and effective meth-
ods for kinetic resolution of this class of compounds.
3. Conclusions
Two kinds of mesoporous silicas, i.e. MCF and SBA-15 as well as
multichannel siliceous monoliths (MH) were successfully function-
alized with penicillin G acylase (PGA) to obtain biocatalysts with
the PGA specific activity of 2066, 273 and 1027 U/g, respectively,
at the immobilization yields of 95%, 38% and 75%, respectively.
The immobilization of PGA on MCF or SBA-15 has been proven
to give effective biocatalysts for the kinetic resolution of racemic 1-
(N-acylamino)alkylphosphonic and 1-(N-acylamino)alkylphos-
phinic acids and their esters carried out in a batch system. In most
of the investigated cases, the initial hydrolytic deacylation rates in
the presence of those catalysts were similar or slightly lower in
comparison with reactions of the same compounds catalyzed by
native PGA.7 However, the initial hydrolysis rate in the presence
of MCF-attached PGA was higher by factors of 6.6 and 1.8, respec-
tively, for the phosphonic analogue of asparagine 3c and the
H-phosphinic analogue of alanine 4b. The results obtained for the
immobilized PGA confirm the relationships between the substrate
structure and the enzyme efficiency as observed for native PGA and
reported in our previous paper.7
4.2. Chemicals and enzyme
The syntheses of racemic 1-(N-acylamino)alkylphosphonic
and 1-(N-acylamino)alkylphosphinic acids and their esters are
described in our previous paper.7 Penicillin G acylase from Escher-
ichia coli (20.34 U/mg) was purchased from CBC Biotech S.r.l.
(Italy). The specific activity of native penicillin G acylase and its
immobilized forms was determined with penicillin G as a substrate
according to the PDAB method.19
4.3. Immobilization of PGA
The stereospecific hydrolysis of the N-acyl group of both race-
mic N-acylated phosphorus analogues of amino acids and their
esters catalyzed by PGA immobilized on MCF or SBA-15 proved
to be a highly effective method for kinetic resolution of all the
investigated compounds. Highly effective kinetic resolutions with
E-values exceeding 100 were obtained for the phosphonic ana-
logues of leucine and asparagine 3b–c, dimethyl esters of the phos-
phonic analogues of alanine, phenylalanine and asparagine 1a–b
4.3.1. Immobilization on MCF and SBA-15
Mesoporous silica foams MCF and SBA-15 were prepared as
described in our previous papers.20–22 The supports (2 g) were then
treated with water vapour at room temperature for 5 h, after which
they were dried at 200 °C for 2 h and then treated with a solution
of (3-aminopropyl)triethoxysilane (0.54 g, 3 mmol) in toluene
(60 mL) at 80 °C for 24 h to obtain mesoporous silicas functional-
ized with 3-aminopropyl groups with a density of amino groups
´
Please cite this article in press as: Zielinska, K.; et al. Tetrahedron: Asymmetry (2016), http://dx.doi.org/10.1016/j.tetasy.2016.11.007

´
K. Zielinska et al. / Tetrahedron: Asymmetry xxx (2016) xxx–xxx
6
of ca. 1.5 mmol per 1 g of silica. The functionalized silica carriers
(1.5 g of each) were washed with ethanol (2 ꢁ 35 mL), water
(2 ꢁ 35 mL) and a phosphate buffer pH 7 (2 ꢁ 35 mL). Centrifuga-
tion (20 min, 9000 rpm) was applied for the separation of the silica
carriers. The carriers were then activated with a solution of glu-
taraldehyde in phosphate buffer pH 7 (2.5%, 30 mL) at room tem-
perature for 45 min. After activation, the carriers were washed
with water (3 ꢁ 35 mL) and phosphate buffer pH 7.5 (2 ꢁ 35 mL)
and, after separation, treated with a solution of PGA in phosphate
buffer pH 7.5 (25 mL, protein concentration 3.1 mg/mL) and gently
shaken at room temperature for 2.5 h, then left at 4 °C for 12 h.
After separation of the enzyme solution, the carriers were rinsed
with phosphate buffer pH 7.5 (2 ꢁ 35 mL), 0.5 M NaCl in phosphate
buffer pH 7.5 (35 mL), acetate buffer pH 4.5 (35 mL) and water
(2 ꢁ 35 mL), left in Tris–HCl buffer pH 7.8 (35 mL) at 4 °C for
12 h, rinsed once again with water (35 mL) and suspended in phos-
phate buffer pH 7.5 (35 mL).
The specific activity of the immobilizates was determined by
adopting the procedure of Shewale et al.,19 using the hydrolysis
of penicillin G as the standard reaction. The activity was measured
by adding 1 mL of suspension of immobilized enzyme in 0.1 M
phosphate buffer pH 7.5 to a solution of penicillin G in the same
buffer (9 mL, 10 mg/mL). After exactly 5, 10, 15 and 20 min of incu-
bation with stirring at 37 °C, a 0.5 mL sample was taken off and the
released product was determined by conventional tests.19
compound), the residue was dissolved in D2O, and the amount of
substrate and hydrolysis product was determined by 1H NMR.
4.4.2. Kinetic resolution experiments
4.4.2.1. Kinetic resolution of racemic 1-(N-acylamino)alkylphos-
phonic and 1-(N-acylamino)alkylphosphinic acids.
Racemic
free acids 3 or 4 (1 mmol) were dissolved in phosphate buffer
(0.2 M, pH 7.0, 75 mL) and the temperature of the mixture was
adjusted to 25 °C. The pH was adjusted to 7.0 by the addition of
1 M aqueous NaOH as necessary. A suspension of the appropriate
amount of immobilized penicillin G acylase (see Table 1) in phos-
phate buffer (5–15 mL) was then added and the resulting mixture
was stirred at 25 °C for the time given in Table 1. The reaction
was stopped by filtering off the immobilizate by using folded paper
filters. After evaporation of water under reduced pressure, the resi-
due was dissolved in water (5 mL) and passed through a DOWEX
50W X8 (H+ form) column (50 mL) which was rinsed with water.
The first fractions contained the remaining free acids 3a–c or 4a–
c, whereas the further ninhydrin-positive fractions contained deacy-
lation products 5a–c or 6a–b.
4.4.2.2. Kinetic resolution of racemic 1-(N-acylamino)alkylphos-
phonic acid esters.
Racemic esters 1 (1 mmol) were dissolved
in phosphate buffer (0.2 M, pH 7.0, 75 mL) and the temperature of
the mixture was adjusted to 25 °C. The reactions were carried out
as described above. After evaporation of the water from the reac-
tion mixture, the residue was extracted with CH2Cl2 (4 ꢁ 3 mL),
the extract was dried over MgSO4, the solvent was evaporated
under reduced pressure, and the residue was separated by column
chromatography (silica gel, CH2Cl2/MeOH 20:1).
4.3.2. Immobilization of PGA on silica monoliths
The preparation of silica monolith rods (45 mm in length and
6 mm in diameter) is described in our recent paper.23 The rods
(12 pieces, 3.5 g) were immersed in a solution of (3-amino-
propyl)triethoxysilane (3.14 g, 17.5 mmol) in toluene (140 mL)
for 72 h at 80 °C under intensive stirring. The monoliths were then
clad with epoxide polymer resin (L285MGS-H285MGS type) to
obtain a single-rod microfluidic microreactors.23 Each microreactor
was rinsed in a flow with ethanol (20 mL), water (20 mL), and
phosphate buffer pH 7 (20 mL) at a flow rate of 0.5 mL/min. A solu-
tion of glutaraldehyde (2.5%) in phosphate buffer (0.1 M, pH 7.0,
25 mL) was then pumped through the microreactor, and after rins-
ing with water (20 mL) and phosphate buffer pH 7.5 (20 mL), a
solution of PGE in phosphate buffer pH 7 (6 mL, protein concentra-
tion 3.26 mg/mL) was pumped through the microreactor in a circu-
lar system for 3 h. Finally, the microreactor was rinsed with
phosphate buffer pH 7.5 (20 mL), a solution of NaCl (0.5 M) in
phosphate buffer pH 7.5 (10 mL), acetate buffer pH 4.5 (10 mL),
water (15 mL), Tris–HCl buffer pH 7.8 (5 mL), water (15 mL) and
phosphate buffer pH 7.5 (5 mL), and stored in the last buffer.
4.5. Kinetic resolution of racemic 1-(N-phenylacetylamino)-2-
carbamoylethylphosphonic acid dimethyl ester 1d in a single-
rod microfluidic microreactor in a flow system
4.5.1. Dependency of the conversion of 1-(N-phenylacetyl-
amino)-2-carbamoylethylphosphonic acid dimethyl ester
versus the flow rate
A solution of substrate 1d (94.2 mg, 0.3 mmol, 3 mmol/L) in
phosphate buffer (100 mL, pH 7), was pumped through a microre-
actor (Fig. 1) at 25 °C with a gradually diminishing flow rate, start-
ing from the value of 0.5 mL/min. In order to determine the
conversion of the substrate 2 mL of the outflow was evaporated
under reduced pressure, a precisely measured amount of a solution
of pivalic acid in D2O was added as the internal standard (approx-
imately 0.5 mg of the pure compound), the residue was dissolved
in D2O and the amount of unconverted substrate and hydrolysis
product was determined by 1H NMR.
4.4. Kinetic resolution of racemic 1-(N-acylamino)alkylpho-
sphonic and 1-(N-acylamino)alkylphosphinic acids and their
esters in a batch system
4.5.2. Kinetic resolution experiment in a flow system
A solution of substrate 1d (94.2 mg, 0.3 mmol, 3 mmol/L) in
phosphate buffer (100 mL, pH 7), was pumped through the
microreactor (Fig. 1) at 25 °C with a flow rate of 0.01 mL/min.
The outflow was collected (87 mL) and worked up as described
in Section 4.4.2.1. The unreacted substrate 1d and the hydrolysis
product 2d were isolated by column chromatography on silica
gel (CH2Cl2/MeOH 10:1, v/v) in yields of 47% (18 mg) and 44%
(11 mg), respectively, with specific rotations of +5.4 and ꢀ2.4,
respectively. The determination of the enantiomeric excess in the
substrate gave a value of 99% (1H NMR method).
4.4.1. Initial hydrolysis rate measurement
Racemic substrates 1, 3, or 4 (0.2 mmol) were dissolved in phos-
phate buffer (0.2 M, pH 7.0, 5–15 mL) and the temperature of the
mixture was adjusted to 25 °C. The pH was then adjusted to 7.0
by the addition of 1 M aqueous NaOH as necessary. A suspension
of the appropriate amount of immobilized penicillin G acylase
(see Table 1) in phosphate buffer (0.45 mL) was then added, after
which the reaction mixture volume was adjusted to 20 mL and
the resulting mixture was stirred at 25 °C. The substrate conver-
sion was determined by extracting 2 mL of the aliquot of the reac-
tion mixture through MinisartÒ SRP 4 syringe filters to stop the
reaction. After evaporation of water under reduced pressure, a pre-
cisely measured amount of a solution of pivalic acid in D2O was
added as the internal standard (approximately 0.5 mg of the pure
4.6. The structure and the enantiomeric excesses determination
The structure of the hydrolysis products and the unchanged
substrates was determined by comparing their spectroscopic
´
Please cite this article in press as: Zielinska, K.; et al. Tetrahedron: Asymmetry (2016), http://dx.doi.org/10.1016/j.tetasy.2016.11.007

´
7
K. Zielinska et al. / Tetrahedron: Asymmetry xxx (2016) xxx–xxx
(1H-, 13C-, 31P-, IR) and HRMS data with the data reported in our
10. Kallenberg, A.; Rantwijk, F.; Sheldon, R. Adv. Synth. Catal. 2005, 347, 905–926.
11. Xue, P.; Xu, F.; Xu, L. Appl. Surf. Sci. 2008, 225, 1625–1630.
12. Zhao, J.; Wang, Y.; Luo, G. Bioresour. Technol. 2011, 102, 529–535.
13. Drtina, G.; Haddad, L.; Rasmussen, J.; Gaddam, B.; Williams, M.; Moeller, S.;
Fitzsimons, R.; Fansler, D.; Buhl, T.; Yang, Y.; Weller, V.; Lee, J.; Beauchamp, T.;
Heilmann, S. React. Funct. Polym. 2005, 64, 13–24.
7
previous paper for the same compounds. The determination of
the enantiomeric excesses was performed by 1H or 31P NMR meth-
ods using quinine as the chiral discrimination agent, as previously
described for the same compounds.7
14. Zhou, H.; Li, W.; Shou, Q.; Gao, H.; Xu, P.; Deng, F.; Liu, H. Chin. J. Chem. Eng.
2012, 20, 146–151.
15. Shah, P.; Srivedi, N.; Prabhune, A.; Ramaswamy, V. Microporous Mesoporous
Mater. 2008, 116, 157–165.
Acknowledgements
16. Srivedi, N.; Shah, P.; Prabhune, A. Res. J. Biotechnol. 2008, 3, 18–21.
17. Eldin, M.; Enshasy, H.; Hassan, M.; Haroun, B.; Hassan, E. J. Appl. Polym. Sci.
2012, 125, 3820–3828.
The financial support of the National Science Centre, Poland
(NCN) is gratefully acknowledged [Grant No. 2012/07/
18. Gao, B.; Wang, X.; Shen, Y. Biochem. Eng. J. 2006, 28, 140–147.
19. Shewale, J.; Kumar, K.; Ambekar, G. Biotechn. Technol. 1987, 1, 69–72.
20. Szyman´ ska, K.; Bryjak, J.; Jarze˛bski, A. Top. Catal. 2009, 52, 1030–1036.
´
N/ST5/00253, (K. Zielinska) and Grant No. 2013/09/B/ST8/02420
(K. Szyman´ ska and A. Jarze˛bski)].
_
_
21. Drozdz, A.; Chrobok, A.; Baj, S.; Szyman´ ska, K.; Mrowiec-Białon´ , J.; Jarze˛bski, A.
Appl. Catal. A 2013, 467, 163–170.
References
´
´
22. Rekuc, A.; Bryjak, J.; Szymanska, K.; Jarze˛bski, A. Process Biochem. 2009, 44,
191–198.
1. Kafarski, P.; Lejczak, B. Curr. Med. Chem. – Anti-Cancer Agents 2001, 1, 301–312.
2. Kukhar, V.; Hudson, H. Aminophosphonic and Aminophosphinic Acids: Chemistry
and Biological Activity; Wiley & Sons Ltd: Chichester, 2000.
3. Heisler, A.; Rabiller, C.; Douillard, R.; Goalou, N.; Hägele, G.; Levayer, F.
Tetrahedron: Asymmetry 1993, 4, 959–960.
´
23. Szymanska, K.; Pietrowska, M.; Kocurek, J.; Maresz, K.; Koreniuk, A.; Mrowiec-
Białon´ , J.; Widłak, P.; Magner, E.; Jarze˛bski, A. Chem. J. Eng. 2016, 287, 148–154.
24. Szyman´ ska, K.; Pudło, W.; Mrowiec-Białon´ , J.; Czardybon, A.; Kocurek, J.;
Jarze˛bski, A. Micropor. Mesopor. Mat. 2013, 170, 75–82.
25. Nakanishi, K.; Takahashi, R.; Nagakane, T.; Kitayama, K.; Koheiya, N.; Shikata,
H.; Soga, N. J. Sol-Gel. Sci. Technol. 2000, 17, 191–210.
4. Solodenko, V.; Kasheva, T.; Kukhar, V.; Kolzova, E.; Mironenko, D.; Svedas, V.
Tetrahedron 1991, 47, 3989–3998.
5. Brzezin´ ska-Rodak, M.; Klimek-Ochab, M.; Zyman´ czyk-Duda, E.; Kafarski, P.
26. Davis, F.; Wu, Y.; Yan, H.; McCoull, W.; Prasad, K. J. Org. Chem. 2003, 68, 2410–
2419.
_
Molecules 2011, 16, 5896–5904.
6. Kozyra, K.; Brzezin´ ska-Rodak, M.; Klimek-Ochab, M.; Zyman´ czyk-Duda, E. J.
27. Vasella, A.; Voeffray, R. Helv. Chim. Acta 1982, 65, 1953–1964.
28. Chen, H.; Noble, F.; Mothe, A.; Meudal, H.; Coric, P.; Danascimento, S.; Roques,
B.; George, P.; Fournie-Zaluski, M.-C. J. Med. Chem. 2000, 43, 1398–1408.
29. Solodenko, V.; Belik, M.; Galushko, S.; Kukhar, V.; kozlova, E.; Mironenko, D.;
Svedas, V. Tetrahedron: Asymmetry 1993, 4, 1965–1968.
_
Mol. Catal. B: Enzym. 2013, 91, 32–36.
7. Zielin´ ska, K.; Mazurkiewicz, R.; Szyman´ ska, K.; Jarze˛bski, A.; Magiera, S.; Erfurt,
K. J. Mol. Catal. B: Enzym. 2016, 132, 31–40.
8. Hanefeld, U.; Cao, L.; Magner, E. Chem. Soc. Rev. 2013, 42, 6211–6212.
9. Srirangan, K.; Orr, V.; Akawi, L.; Westbrook, A.; Moo-Young, M.; Chou, C.
Biotechnol. Adv. 2013, 31, 1319–1332.
30. Kafarski, P.; Lejczak, B.; Szewczyk, J. Can. J. Chem. 1983, 61, 2425–2430.
´
Please cite this article in press as: Zielinska, K.; et al. Tetrahedron: Asymmetry (2016), http://dx.doi.org/10.1016/j.tetasy.2016.11.007