Journal of Biotechnology 219 (2016) 142–148

Contents lists available at ScienceDirect

Journal of Biotechnology
journal homepage: www.elsevier.com/locate/jbiotec

Efficient cascade synthesis of ampicillin from penicillin G potassium

salt using wild and mutant penicillin G acylase from Alcaligenes
faecalis
Senwen Deng a,c,1 , Xiaoqiang Ma c,1 , Erzheng Su b,∗ , Dongzhi Wei c,∗∗
a
School of life science, Hunan University of Science and Technology, Xiangtan 411201 Hunan, PR China
b
Enzyme and Fermentation Technology Laboratory, College of Light Industry Science and Engineering, Nanjing Forestry University, Nanjing 210037, PR
China
c
State Key Laboratory of Bioreactor Engineering, New World Institute of Biotechnology, East China University of Science and Technology, Shanghai 200237,
PR China

a r t i c l e i n f o a b s t r a c t

Article history: To avoid isolation and purification of the intermediate 6-aminopenicillanic acid (6-APA), a two-enzyme
Received 23 September 2015 two-step cascade synthesis of ampicillin from penicillin G was established. In purely aqueous medium,
Received in revised form penicillin G hydrolysis and ampicillin synthesis were catalyzed by immobilized wild-type and mutage-
24 November 2015
nized penicillin G acylases from Alcaligenes faecalis (Af PGA), respectively (Fig. 1). The ␤F24 G mutant Af
Accepted 21 December 2015
PGA (the 24th Phenylalanine of the ␤-subunit was replaced by Glycine) was employed for its superior per-
Available online 28 December 2015
formance in enzymatic synthesis of ampicillin. By optimizing the reaction conditions, including enzyme
loading, temperature, initial pH and d-PGME/6-APA ratio, the conversion of the second step of ampicillin
Keywords:
Penicillin G acylase
synthesis reached approximately 90% in 240 min and less than 1.7 mole d-PGME were required to pro-
Protein engineering duce 1 mole ampicillin. Overall, in a 285 min continuous two-step procedure, an ampicillin yield of 87%
Ampicillin was achieved, demonstrating the possibility of improving the cascade synthesis of ampicillin by mutage-
Enzyme catalysis nized PGA, providing an economically efficient and environmentally benign procedure for semi-synthetic
Cascade process penicillins antibiotics synthesis.
© 2015 Elsevier B.V. All rights reserved.

1. Introduction
Semi-synthetic penicillins antibiotics (SSPAs), one of the most
important families of anti-infection drugs in the world market
(Parmar et al., 2000; Chandel et al., 2008), are mainly produced
by a two-step fashion (Srirangan et al., 2013). First, bulk penicillin
G potassium (PGK) is hydrolyzed into 6-aminopenicillianic acid
(6-APA) and side-product phenylacetic acid (PAA) by immobilized
penicillin G acylase (PGA: EC 3.5.1.11) (Fig. 1 ). Next, purified 6-
APA is further processed into SSPAs by condensation with the acyl
side chain chemically (Wegman et al., 2001). As an environmental
Abbreviations: SSPA, Semi-synthetic penicillins antibiotic; PGK, penicillin G
potassium; PGA, Penicillin G acylase; Af PGA, Alcaligenes faecalis PGA; d-PGME,
d-phenylglycine methyl ester; 6-APA, 6-aminopenicillanic acid; AEH, ␣-amino
ester hydrolase; S/H, synthesis to hydrolysis; PAA, phenylacetic acid; Ec PGA,
Escherichia coli PGA; d-PG, d-phenylglycine; HPLC, High Performance Liquid Chro-
matography; PCR, Polymerase Chain Reaction.
∗ Corresponding author. Fax: +86 25 85428906.
∗∗ Corresponding author. Fax: +86 21 64250068.
E-mail addresses: ezhsu@njfu.edu.cn (E. Su), dzhwei@ecust.edu.cn (D. Wei).
1
These authors contributed equally to this work.
friendly alternative, the second step can also be catalyzed by PGA
(Fig. 1) or less studied ␣-amino ester hydrolase (AEH: EC 3.1.1.43)
(Bruggink et al., 1998), by coupling 6-APA with activated acyl donor
under kinetic control, but with lower yield (Bruggink, 2001).
Since both steps of SSPAs synthesis can be PGA-mediated, cas-
cade processes were proposed to avoid isolation and purification
of intermediates 6-APA in several reports (Schroën et al., 2002; Du
et al., 2009; Blum et al., 2010; Wu et al., 2010; Zhang et al., 2010),
rendering SSPAs synthesis more economically advantageous and
environmentally benign. Generally, isolation and purification of 6-
APA need several steps such as extraction and crystallization. Large
amounts of organic solvent are needed during these steps (Cao et al.,
2001; Sheldon, 2008). Moreover, the loss of 6-APA is about 10-15%
due to the purification procedures (Cao et al., 2001). In partially
organic media (ethylene glycol and water), Du et al. (2009) and Wu
et al. (2010) reported direct synthesis of ampicillin and amoxicillin
from PGK using immobilized wild PGA from Escherichia coli (E. coli).
After adding high concentration of ethylene glycol (60%, v/v), the
overall conversion reached 57.5% and 55.2% after 17 h and 11 h reac-
tion, respectively. In Zhang’s work (2010), zinc ions were added to
form complexes with amoxicillin in situ so that the equilibrium

http://dx.doi.org/10.1016/j.jbiotec.2015.12.034
0168-1656/© 2015 Elsevier B.V. All rights reserved.
 S. Deng et al. / Journal of Biotechnology 219 (2016) 142–148
Fig. 1. Scheme of two-enzyme two-step cascade synthesis of ampicillin.
was shifted to the amoxicillin synthesis, and yield of 71.5% was
achieved with immobilized wild penicillin G acylase from Kluyvera
citrophila. Blum et al. (2010) reported a two-enzyme system cat-
alyzed by immobilized wild PGA from E. coli and free AEH from
Xanthomonas campestris pv. campestris. The maximum conversion
was 47%. When catalyzed by PGA alone, the yield was less than 15%.
But poor thermostability of currently used AEHs is not favored for
industrial purpose (Blum and Bommarius 2010).
For processes mentioned above, relative low yield or long reac-
tion time were observed, which make them hard to fulfill the
practical industry requirements. Since penicillin G hydrolysis by
immobilized wild PGAs has already been commercialized for high
yield retained (>98%) (Bruggink et al., 1998; Cao et al., 2004), to
improve the productivity of these processes, the second step of
enzymatic SSPAs synthesis is of first concern (Fig. 1). Low yield of
enzymatic SSPAs synthesis can be mainly ascribed to poor kinetic
parameters of wild PGAs, as relative low S/H ratio is observed
and kcat /Km values of them for activated acyl donor are usually 10
times lower than those of antibiotic products (Alkema et al., 2002a;
Alkema et al., 2002b; Gabor and Janssen, 2004; Bečka et al., 2014).
Subsequently, enzyme-catalyzed initial hydrolysis of activated acyl
donor and secondary hydrolysis of antibiotic products were usu-
ally observed (Deaguero, 2011). Furthermore, wild PGAs are prone
to be inhibited by side product PAA (Švedas et al., 1997; Alkema
et al., 2002b), contributing to the prolonged reaction time, typi-
cally when high concentration of PAA was produced at the course
of direct synthesis. Protein engineering is commonly used to tai-
lor wild PGAs for industrial purposes like SSPAs synthesis (Alkema
et al., 2002a, 2002b, 2004; Gabor and Janssen, 2004; Jager et al.,
2007, 2008; Deaguero et al., 2012). In our previous work, ␤F24 G
mutant of thermo-stable Penicillin G acylase from Alcaligenes fae-
calis (Af PGA), in which the 24th Phenylalanine of the ␤-subunit was
replaced by Glycine, was isolated from a mutant library (Deng et al.,
2015). Kinetic analysis indicated higher kcat /Km value for activated
acyl donor d-phenylglycine methyl ester (358 vs 0.44 mM−1 S−1 of
wild-type), higher initial S/H ratio (4.2 vs 0.49 of the wild-type)
and less sensitivity to PAA inhibition (Ki = 0.27 vs 0.087 mM of the
wild-type) at the expense of reduced kcat /Km value for penicillin G
like phenylacetylated substrates (397 vs 11051 mM−1 S−1 of wild-
type). The space-time yield of ampicillin synthesis was improved
up to 138-fold by this mutant. The improvement in kinetic param-
eters make it promising for direct synthesis of ampicillin using the
penicillin G hydrolytic solution which contained 6-APA and equal
143
amount of PAA. Moreover, wild Af PGA showed higher catalytic
efficience for penicillin G than most of other reported wild PGAs
(Švedas et al., 1997).
Taking advantage of the substrate specificity of the wild and
mutant Af PGA, a two-enzyme two-step cascade synthesis of
ampicillin was established in this work. First, Penicillin G salt is
transformed into 6-APA by immobilized wild Af PGA. Then the
hydrolytic solution is filtered and further processed into ampicillin
by condensation with d-PGME. This step is catalyzed by immobi-
lized mutant Af PGA (␤F24G) (Fig. 1). As penicillin G hydrolysis
by immobilized wild Af PGA has been well studied in this lab, we
concentrated mainly on the second step to improve the cascade
synthesis of ampicillin.
2. Materials and methods
2.1. Chemicals and reagents
Molecular biology related reagents were bought from Toyobo
Biotechnology Co., Ltd. (Shanghai, China). Epoxy carrier (LH-EP)
was kindly supplied by Shanghai Bairui Biotech. Co., Ltd. (China).
Ampicillin was purchased from Sigma. PGK and 6-APA were pro-
vided by Tokyo chemical industry Co., Ltd. (Tokyo, Japan). The
6-APA using for reference standards was bought from Dr. Ehren-
storfer (German). d-Phenylglycine, d-PGME and PAA were bought
from Aoxin chemical industry Co., Ltd. (Yangzhou, China). All other
chemicals were of analytical reagent grade.
2.2. Gene cloning and vectors construction
pET-28a-Af PGA (wild-type) and pET-28a-Af PGA-24 G (␤F24
G mutant) were obtained by the method described below and
reserved in this lab (Deng et al., 2015). E. coli CGMCC 1.1023, har-
boring the plasmid psmlf-Af PGA, was obtained from China General
Microbiological Culture Collection Center (CGMCC). Using psmlf-Af
PGA as template, the Af PGA coding gene (GenBank: AF455356.1)
was amplified by two specific primers PGA-F and PGA-R (Table 1).
The PCR parameters were 94 ◦ C, 2 min for predenaturation; 94 ◦ C,
15 s for denaturation; 68 ◦ C, 1.5 min for extension, and continued
for 30 cycles. PCR product and plasmid pET-28a were all digested
with NcoI and HindIII, and fused together by T4 DNA Ligase, then
confirmed by sequencing and named pET-28a-Af PGA. Using pET-
28a-Af PGA as template, pET-28a-Af PGA-24 G was constructed by
144 S. Deng et al. / Journal of Biotechnology 219 (2016) 142–148
Table 1
Primers used in this study.
Primers Sequence (5 → 3 ) Restriction sites
PGA-F CATGCCATGGTGAAAGGGCTGGTTCGTACAG NcoI
PGA-R CCAAGCTCTAAGGCTGAGGCTGAATCAAAAGC HindIII
24G-F CTGATCAATGGCCCGCAGGGTGGCTGGTACAACCCGGCT –
24G-R AGCCGGGTTGTACCAGCCACCCTGCGGGCCATTGATCAG –
another PCR procedure with two specific primers (24G-F and 24G-
R, Table 1). PCR parameters were 94 ◦ C, 2 min for predenaturation;
94 ◦ C, 15 s for denaturation; 68 ◦ C, 4 min for extension, and contin-
ued for 15 cycles. The PCR products were then transformed into
E. coli DH5␣, and confirmed by sequencing.
2.3. Enzyme preparation and immobilization
pET-28a-Af PGA (wild-type) and pET-28a-Af PGA-24 G (␤F24 G
mutant) were tranformed into E. coli BL21(DE3) for protein expres-
sion. The cultivation of both recombinant strains following the
method reported previously (Wang et al., 2006). One milliliter of
fresh overnight culture of E. coli BL21 (DE3) was used to inoc-
ulate 50 ml LB broth containing 0.05 g/L kanamycin and 1 mM
calcium chloride. Bacteria were grown at 37 ◦ C and 200 rpm in
an orbital shaker till an optical density of 1.0 at 600 nm (OD600 ).
Expression was initiated by adding lactose to a final concentra-
tion of 0.6% (w/v). Then culture temperature was reduced to 20 ◦ C
and maintained for another 30 h at 200 rpm. The OD600 reached
about 8.2 for both strains. Cells were harvested by centrifugation at
8000 × g, 4 ◦ C for 10 min and resuspended in potassium phosphate
buffer (100 mM, pH 8.0), then disrupted by sonication (400 W, 100
cycles, working 3 s and intervals 5 s as one cycle). After centrifug-
ing at 8000 × g and 4 ◦ C for 10 min, the obtained supernatant was
collected and fractionated with ammonium sulfate. Protein that
precipitated between 20 and 30% (w/v) ammonium sulfate were
redissolved by potassium phosphate buffer (100 mM, pH 8.0) as
partly purified enzyme for immobilization. To 200 ml partly puri-
fied enzyme, 100.63 g Na2 HPO4 .12H2 O, 3.18 g NaH2 PO4 .2H2 O and
20 g epoxy carrier (LH-EP) were added, then stirred at 200 rpm and
25 ◦ C for 48 h (Gong et al., 2011). After immobilization, the solution
was filtered. The immobilized enzyme were resuspended in potas-
sium phosphate buffer (100 mM, pH 8.0) and stirred for half an hour,
then filtered. This process was repeated three times to make sure
that protein uncovalently linked to the epoxy carrier support were
washed away. Immobilized enzyme were stored at 4 ◦ C. The immo-
bilization supernatant and washing buffer were collected together
to determine the protein concentration and total volume. Then the
total amount of protein excluding those covalently linked to the
support could be calculated. The immobilization yield (%, w/w) was
calculated according to Eq. (1).
A−B
Immobilization yield=[( )] × 100%
A The partly purified acylases were then immobilized onto an
where:A: total protein of free enzyme solution before immobiliza-
tion.B: total amount of protein excluding those covalently linked
to the support after immobilization.
The activity recovery was calculated according to Eq. (2).
A toward PGK according to the different specific activities of the sol-
Activityrecovery=[ ] × 100%
B uble enzymes. The activties of the immobilized wild-type and ␤F24
where:A: total activity of immobilized enzyme after immobiliza-
tion.B: total activity of free enzyme solution before immobilization.
A unit of PGA (UPGK ) was defined as the amount of enzyme
required to release 1 ␮mol of 6-APA per minute in 4% (w/v) solu-
tion of PGK at pH 7.8 and 37 ◦ C. The enzyme activity was determined
by a previously reported method (Jaiprakash et al., 1987). Specific
activity is defined as UPGK per milligrams of protein. Protein con-
centration was determined using the Bradford method (Bradford,
1976).
2.4. Two-step enzymatic synthesis of ampicillin
The initial PGK concentration was 4% (w/v, approximately
107 mM). The first step was started by the addition of immo-
bilized wild Af PGA (0.03 g or 10.5 UPGK ml−1 ). The reaction was
conducted at 28 ◦ C, in 100 mM potassium phosphate buffer with
pH maintained at 8.0, and measured by High Performance Liquid
Chromatography (HPLC). The first step was terminated when the
yield of 6-APA reached approximately 97%. Immobilized wild Af
PGA was removed by filtration. Then medium pH was adjusted,
d-PGME and immobilized mutant Af PGA (␤F24G) were added
to initiate the second step without pH control, since the pH val-
ues were nearly unchanged over the reaction time. The volume of
the second step was 10 ml with a diluted 6-APA concentration of
100 mM. The conversion of 6-APA in the second step was calculated
by measuring the decrease of the 6-APA by HPLC. HPLC analysis
was performed on an Agilent 1100 system (Palo Alto, USA) with
a 4.6 × 150 mm Agilent ZORBAX Eclipse XDB C18 column in con-
nection with a quaternary pump and a diode-array detector (DAD)
measured at 214 nm and 30 ◦ C, at a flow rate of 1 ml/min, with
an eluting solvent system of water phase which contained 5 mM
potassium phosphate, 680 mg/L sodium dodecyl sulfate (SDS), pH
3.0 and acetonitrile (70:30, v/v).
3. Results
3.1. Protein expression
Expression of the proteins was conducted according to the
Novagen’s pET system manual (2002). Comparable results were
obtained for the wild and ␤F24 G mutant enzyme (Table 2). The
specific activity of the different crude extracts confirmed that
the hydrolytic activity of Af PGA was strongly affected by muta-
tion ␤F24G, which was consistent with the previous report about
mutant PGA from E. coli (Davide et al., 2012). The ␤F24 G mutant
acylase showed much lower hydrolytic activity (0.55 UPGK /mg)
than that of the wild-type (5.39 UPGK /mg). Since expression level
of enzymes was relatively low, the supernatant obtained from
sonication was fractionated with ammonium sulfate. PGA was
mainly precipitated between 20 and 30% (w/v) ammonium sulfate
(Table 2). Activity recovery of this step was approximately 44%, and
the specific activity of two enzymes increased by 4.11 or 4.31-fold.
A similar purity degree was checked by SDS-PAGE (data not shown).
3.2. Immobilization
(1)
epoxy carrier LH-EP under the optimized conditions as decribed
before (Gong et al., 2011). In all cases, approximately 70% of the total
protein were covalently linked to the support (Table 3). The activ-
ity recovery for both enzymes was identical (45%, Table 3). These
two immobilized acylases expressed different hydrolytic activities
(2)
G mutant acylase were 350 U/g and 37 U/g, respectively (Table 3).
The S/H ratio of immobilized wild-type and ␤F24 G mutant acy-
lases were 0.41 and 4.0, respectively, slightly lower than 0.47 and
4.26 of the free enzymes (Deng et al., 2015). Subsequently, these
two immobilized enzymes were used in the two-enzyme two-step
cascade synthesis of ampicillin (penicillin hydrolysis and ampicillin
synthesis).
 S. Deng et al. / Journal of Biotechnology 219 (2016) 142–148 145

Table 2
Ammonium sulfate fractionation of wild-type and ␤F24 G mutant Af PGA.

Enzyme Crude extract Partly purified enzyme

(mg/ml) Specific activity (UPGK /mg) (mg/ml) Specific activity (UPGK /mg) purification factor

WT 2.33 5.39 3.52 22.15 4.11

␤F24G 2.37 0.55 3.49 2.37 4.31

All data were the average of three independent experiments.

Table 3
Immobilization of wild-type and ␤F24 G mutant Af PGA.

Enzyme Offered enzyme(mg/g) Immobilization yield (%)a Activity recovery (%)b Activity (UPGK /g)c S/H ratioini d

WT 35.2 70 45 351 0.41

␤F24G 34.9 73 45 37 4.0

All data were the average of three independent experiments.

a
The immobilization yield (%, w/w) was calculated according to Eq. (1).
b
The activity recovery was calculated according to Eq. (2).
c
Activity toward PGK hydrolysis expressed per gram of support (UPGK /g).
d
S/H ratioini : initial rate of antibiotic formation to initial rate of d-PG formation.

Fig. 2. Time course of enzymatic synthesis of ampicillin at different concentration Fig. 3. Time course of enzymatic synthesis of ampicillin at different initial medium
of immobilized ␤F24 G mutant Af PGA. Conditions: initial pH 6.3, 28 ◦ C, 150 mM pH. Conditions: 28 ◦ C, 150 mM d-PGME, 0.19 g/mL immobilized ␤F24 G mutant Af
d-PGME. PGA.

3.3. Two-step enzymatic synthesis of ampicillin

Penicillin G hydrolysis by immobilized wild Af PGA has already
been optimized in this lab before (unpublished data). Thus, the
reaction conditions of the second step of enzymatic ampicillin syn-
thesis were investigated systematically. As shown in Fig. 2, the
concentration of immobilized ␤F24 G mutant Af PGA significantly
influenced the ampicillin synthesis. Increased enzyme loading
obviously led to faster reaction rate and higher 6-APA conversion
until a certain degree.When enzyme loading increased to more than
0.19 g/mL, 6-APA conversion can not be further improved. In con-
sideration of enzyme cost, 0.19 g/mL of enzyme loading was chosen
for further study.
Since the pH values were nearly unchanged over the reaction
time, we performed the reaction without pH control. The pattern of
ampicillin synthesis could be changed by initial reaction pH (Fig. 3).
Higher initial pH would increase the reaction rate, in accordance
with the slightly alkaline pH optimum of PGAs, but relative low S/H
ratio and more byproduct d-PG accumulation could be observed
at the same time (Schroën et al., 1999). At initial pH of 6.8, 6-APA
conversion decreased significantly after the time of its maximum
conversion, which could be explained by the hydrolysis of ampi-
cillin and residual d-PGME (Fig. 3). When the initial pH was reduced
to a certain degree, the 6-APA conversion also decreased with the
decline of reaction rate. The highest 6-APA conversion was achieved
at initial pH of 6.3. Meanwhile, it is noteworthy that acidic pH is
beneficial for substrates stability, especially 6-APA (Shewale and
Sudhakaran 1997; Kumar et al., 2008).
Unlike initial pH, the change of reaction temperature exhib-
ited smaller effect on ampicillin synthesis (Fig. 4). The reaction
rate increased slightly while temperature rising, as well as the 6-
APA conversion. The highest 6-APA conversion was achieved at
28 ◦ C. 6-APA conversion can not be improved when temperature
over 28 ◦ C due to the balance between hydrolysis and synthesis.
Higher temperature would speed up the chemical degradation of
substrates especially 6-APA (Shewale and Sudhakaran 1997; Kumar
et al., 2008). In consideration of 6-APA conversion, reaction time
and substrates stability, 28 ◦ C was chosen for further experiments.
As shown in Fig. 5, increase of the initial activated acyl donor
d-PGME concentration would accelerate the reaction rate and
improve the conversion of 6-APA at the same time. However, when
the d-PGME/6-APA ratio increased to a certain level (1.5:1), the
conversion of 6-APA can not be further improved. Considering the
process costs and ampicillin yield, d-PGME/6-APA ratio of 1.5:1 was
suitable.
146 S. Deng et al. / Journal of Biotechnology 219 (2016) 142–148

Table 4
Enzymatic cascade synthesis of ampicillin from PGK.

Methods Conditions Time (min) Yield (%)a Productivity Moles of d-PGME References
(mM h−1 )b per mole of ampicillin
−1 c
(mol mol )

One-step Ethylene glycol/H2 O 600 55.2 8.3 5.4 Wu et al., 2010

(60:40, v/v) medium
150 mM PGK + 450 mM
d-PGME
Immobilized wild-type
PGA from E.coli
Two-step Purely aqueous 160 47.0 3.5 6.0 Blum et al. 2010
medium
Step 1: 40 mM PGK
Immobilized wild-type
PGA from E. coli
Step 2: 20 mM
6-APA + 120 mM
d-PGME
Free AEH from X.
campestris
pv.campestris
Two-step Purely aqueous 285 87.3 19.0 1.7 This study
medium
Step 1: 107 mM PGK
Immobilized wild-type
Af PGA
Step 2: 100 mM
6-APA + 150 mM
d-PGME
Immobilized mutant Af
PGA
a
Calculated by the moles of ampicillin produced from per moles of PGK.
b
Calculated by mM ampicillin produced per hour.
c
Calculated by the amount of d-PGME required to produce one mole of ampicillin under the maximum conversion.

Fig. 4. Time course of enzymatic synthesis of ampicillin at different temperature. Fig. 5. Time course of enzymatic synthesis of ampicillin at different d-PGME/6-APA
Conditions: initial pH 6.3, 150 mM d-PGME, 0.19 g/mL immobilized ␤F24 G mutant molar ratio. Conditions: initial pH 6.3, 28 ◦ C, 0.19 g/mL immobilized ␤F24 G mutant
Af PGA. Af PGA.

According to the above results, the optimal reaction conditions
of the second step were determined: 1.9 g immobilized mutant Af
PGA in 10 ml reaction mixture (7.03 UPGK /mL), initial pH of 6.3,
28 ◦ C and d-PGME/6-APA ratio of 1.5:1. By controlling reaction
conditions in two separated steps, the conversion of the first and
second step reached 97% and 90%, respectively (Fig. 6). Overall, a
final yield of 87% was obtained by a 285 min continuous proce-
dure in aqueous medium. The productivity of this procedure was
obviously higher than previously reported two-enzyme two-step
or one-enzyme one-step methods (Table 4). It is noted that this
yield was achieved at low d-PGME/6-APA ratio (1.5:1). Less than
1.7 mole d-PGME was required to produce 1 mole ampicillin, much
lower than the reported 6 and 5.4 in previous work (Blum et al.,
2010; Du et al., 2009) (Table 4).
4. Discussion
As described in other studies and this report, direct synthesis
of SSPAs can be conducted in either one-step or two-step way.
For the one-step process, hydrolysis of PGK and synthesis of SSPAs
are performed simultaneously, equipment and operation cost were
deemed to be the least. However, it is noted that these two reac-
tions share the same mechanism, but required different conditions
(Srirangan et al., 2013). Hydrolysis of penicillin G is usually per-
 S. Deng et al. / Journal of Biotechnology 219 (2016) 142–148
Fig. 6. Time course of two-enzyme two-step cascade synthesis of ampicillin. Dotted
line is the demarcation line of the first step and the second step. Conditions of the first
step: pH 8.0, 28 ◦ C, 4% (w/v) PGK, 0.03 g/mL immobilized wild Af PGA; Conditions
of the second step: initial pH 6.3, 28 ◦ C, 150 mM d-PGME, 0.19 g/mL immobilized
␤F24 G mutant Af PGA.
formed at a slightly alkaline pH (8.0-8.5) while the synthesis of
ampicillin prefers an acidic one (Srirangan et al., 1999). On the other
hand, it is difficult to screen an enzyme efficient for both hydroly-
sis and synthesis, as structure difference between penicillin G like
phenylacetylated substrates and activated acyl donor (the amide
or ester of d-4-hydroxyphenylglycine and d-phenylglycine deriva-
tives), and higher S/H ratio of mutant enzymes were usually gained
at the expense of hydrolytic activity (Alkema et al., 2002a; Alkema
et al., 2002b; Gabor and Janssen, 2004). From above viewpoints,
two-step process seems to be more feasible as each step can be
optimized with different enzymes under optimal reaction condi-
tions separately. But, some questions still remain before they can
fulfill the industry demand.
For direct synthesis of SSPAs, especially the two-step process,
inhibition of high concentration of PAA and subsequently long pro-
cess time can be always noted (Du et al., 2009; Blum et al., 2010; Wu
et al., 2010; Zhang et al., 2010). To overcome this, various methods
have been developed. In Zhang’s work (2010), zinc ions were added
in the second step to form complexes with product in situ to shift
the equilibrium to amoxillin production. AEH, used in Blum’s work
(2010), is promising due to negligible PAA inhibition and low activ-
ity towards antibiotics product (Blinkovsky and Markaryan 1993;
Barends et al., 2003), but its poor thermostability should be fur-
ther improved (Blum and Bommarius 2010). 800-fold increase in
kcat /Km value for d-PGME and less sensitivity to PAA, make ␤F24 G
mutant Af PGA used in this work a promising choice (Deng et al.,
2015).
However, inhibition of high concentration of PAA (approxi-
mately 100 mM) was still observed. We have noticed that 6-APA
conversion was associated with reaction rate at the very begin-
ning of this process. It suggested further evolution needed for this
mutant enzyme to reduce its sensitivity to PAA inhibition. Also,
to some extent, improvements for reaction speed were needed
in consideration of process productivity. As further results in this
study described, higher reaction rate could be realized by increased
enzyme loading, relative high reaction pH, high reaction tempera-
ture and high concentration of d-PGME. But these measures have
to be taken carefully in consideration of downstream purification
process and total operation costs. For example, in addition to higher
enzyme preparation costs, increased enzyme loading could also
lead to mass transfer limitations (Kallenberg et al., 2005). Besides,
acidic medium pH and low temperature were usually favorable for
147
higher S/H ratio, as well as better substrate stablity (Shewale and
Sudhakaran 1997; Kumar et al., 2008). To a certain degree, that’s
why slight acidic initial pH (6.3) and relative mild temperature
(28 ◦ C) were chosen in this work.
In addition to long process time required, the low product yield
is also noted in previous reports (Du et al., 2009; Blum et al.,
2010; Wu et al., 2010; Zhang et al., 2010). To increase the con-
version of ␤-lactam nucleus, water-miscible organic solvents and
excess acyl donor were usually employed (Illanes et al., 2007;
Kallenberg et al., 2005). Du et al. (2009) and Wu et al. (2010)
reported direct synthesis of ampicillin and amoxicillin from PGK
with high concentration of ethylene glycol (60%, v/v). However,
PGAs are usually unstable in partly organic solvents (Illanes et al.,
2007). From the economic viewpoint, it is important to keep the
molar ratio acyl donor/nucleophile as low as possible to reduce
the cost of raw materials and downstream processing. And in pre-
vious reports, even after adding of excess acyl donor, the 6-APA
conversion were still relative low due to poor S/H ratio of wild-
type PGAs (Du et al., 2009; Blum et al., 2010; Wu et al., 2010).
S/H ratio of wild-type PGAs could be directly changed by protein
engineering (e.g., Alkema et al., 2002a,b, 2004; Gabor and Janssen,
2004; Jager et al., 2007; Jager et al., 2008; Deng et al., 2015). But
in most cases, protein engineering also result in decreased activ-
ity for acyl donor, make most mutant enzymes unsuitable for the
directly sythesis. In our previous work, ␤F24 G mutant of Af PGA
was isolated for 8-fold increased initial S/H ratio and kinetic anal-
ysis indicated much higher kcat /Km value for activated acyl donor
d-PGME at the same time (Deng et al., 2015). So, in the second
step of this study, mutant Af PGA is tested in cascade synthesis of
ampicillin and different results were observed. Much higher con-
version of expensive nucleus 6-APA (90%) was obtained at lower
d-PGME/6-APA ratio (1.5:1). More importantly, only 1.7 mole d-
PGME were required to produce 1 mole ampicillin, much lower than
previous reports (Table 4), resulted in reduced acyl donor cost and
simplified product purification process. Finally, the overall conver-
sion of this two-step procedure reached approximately 87%, higher
than reported values, and so as the process productivity (Table 4).
All those make it an economically efficient and environmentally
benign new process. The catalytic performance of ␤F24 G mutant
was in accordance with its kinetic parameters, further suggesting
that enzyme property plays the most important role in SSPAs syn-
thesis. Protein engineering could not only improve the productivity
of traditional enzymatic process, but also meet the critical demand
of new process.
In conclusion, we demonstrated the feasibility of improving cas-
cade synthesis of ampicillin by ␤F24 G mutant Af PGA. This strategy
can be easily implemented, economize the effort of 6-APA isola-
tion and purification. Due to the superior catalytic property of the
mutant enzyme, higher 6-APA conversion and shorter reaction time
were achieved at the same time. These results indicated an econom-
ically efficient and environmentally benign process, favorable for
future industrial production of ampicillin and other SSPAs. Further
studies on product isolation and purification are being conducted
in this lab.
Conflict of interest
The authors declare that they have no conflict of interest.
Acknowledgements
This work was supported by the National High Technology
Research and Development Program of China (863 program, No.
2012AA022201C), Six talent peaks project in Jiangsu Province
148 S. Deng et al. / Journal of Biotechnology 219 (2016) 142–148
(2015-JY-016) and the Open Funding Project of the State Key Lab-
oratory of Bioreactor Engineering.
References
Alkema, W.B.L., de Vries, E., Janssen, D.B., 2002a. Role of ␣ Arg (145) and ␤ Arg
(263) in the active site of penicillin acylase of E. coli. Biochem. J. 365, 303–309.
Alkema, W.B.L., Dijkhuis, A.J., de Vries, E., Janssen, D.B., 2002b. The role of
hydrophobic active-site residues in substrate specificity and acyltransfer
activity of penicillin acylase. Eur. J. Biochem. 269, 2093–2100.
Barends, T.R.M., Polderman-Tijmes, J.J., Jekel, P.A., Hensgens, C.M.H., de Vries, E.J.,
Janssen, D.B., Dijkstra, B.W., 2003. The sequence and crystal structure of the
␣-amino acid ester hydrolase from Xanthomonas citri define a new family of
␤-lactam antibiotic acylases. J. Biol. Chem. 278, 23076–23084.
Bečka, S., et al., 2014. Penicillin G acylase from Achromobacter sp. CCM 4824: an
efficient biocatalyst for syntheses of beta-lactam antibiotics under conditions
employed in large-scale processes. Appl. Microbiol. Biotechnol. 98, 1195.
Blinkovsky, A.M., Markaryan, A.N., 1993. Synthesis of ␤-lactam antibiotics
containing ␣-aminophenylacetyl group in the acyl moiety catalyzed by
d-(−)-phenylglycyl-␤-lactamide amidohydrolase. Enzyme Microb. Technol. 15,
965–973.
Blum, J.K., Bommarius, A.S., 2010. Amino ester hydrolase from Xanthomonas
campestris pv. campestris, ATTC 33913 for enzymatic synthesis of ampicillin. J.
Mol. Catal. B Enzym. 67, 21–28.
Blum, J.K., Deaguero, A.L., Perez, C.V., Bommarius, A.S., 2010. Ampicillin synthesis
using a two-enzyme cascade with both ␣-amino ester hydrolase and penicillin
G acylase. Chem. Catal. Chem. 2, 987–991.
Bradford, M.M., 1976. A rapid and sensitive method for the quantitation of
microgram quantities of protein utilizing the principle of protein-dye binding.
Anal. Biochem. 72, 248–254.
Bruggink, A., Roos, E.C., de Vroom, E., 1998. Penicillin acylase in the industrial
production of ␤-lactam antibiotics. Org. Process Res. Dev. 2, 128–133.
Bruggink, A., 2001. Bruggink Synthesis of Beta-Lactam Antibiotics, 335. Kluwer
Academic Publishers, Geleen.
Cao, X.J., Wu, X.Y., Wu, T., Jin, K., Hur, B.K., 2001. Concentration of
6-aminopenicillanic acid from penicillin bioconversion solution and its mother
liquor by nanofiltration membrane. Biotechnol. Bioprocess Eng. 6, 200–204.
Cao, X.J., Wu, X.Y., Fonseca, L.J.P., Cabral, J.M.S., Marcos, J.C., 2004. Production of
6-aminopenicillanic acid in aqueous two-phase systems by recombinant E. coli
with intracellular penicillin acylase. Biotechnol. Lett. 26, 97–101.
Chandel, A.K., Rao, L.V., Narasu, M.L., Singh, O.V., 2008. The realm of penicillin G
acylase in ␤-lactam antibiotics. Enzyme Microb. Technol. 42, 199–207.
Davide, A.C., Roberto, P., Sar, S., Simona, D., Roberto, X., Massimo, P., Marco, T.,
2012. Efficient biocatalyst for large-scale synthesis of cephalosporins, obtained
by combining immobilization and site-directed mutagenesis of penicillin
acylase. Appl. Microbiol. Biotechnol. 95, 1491–1500.
Deaguero, L.A., 2011. Ph.D. Thesis. Georgia Institute of Technology, Georgia.
Deaguero, A.L., Blum, J.K., Bommarius, A.S., 2012. Improving the
diastereoselectivity of penicillin G acylase for ampicillin synthesis from
racemic substrates. Protein Eng. Des. Sel. 25, 135.
Deng, S.W., Su, E.Z., Ma, X.Q., Yang, S.L., Wei, D.Z., 2015. Efficient enzymatic
synthesis of ampicillin by mutant Alcaligenes faecalis penicillin G acylase. J.
Biotechnol. 199, 62–68.
Du, L.L., Wu, Q., Chen, C.X., Liu, B.K., Lin, X.F., 2009. A two-step, one-pot enzymatic
synthesis of ampicillin from penicillin G potassium salt. J. Mol. Catal. B Enzym.
58, 208–211.
Gabor, E.M., Janssen, D.B., 2004. Increasing the synthetic performance of penicillin
acylase PAS2 by structure-inspired semi-random mutagenesis. Protein Eng.
Des. Sel. 17, 571–579.
Gong, X., Su, E., Wang, P., Wei, D., 2011. Alcaligenes faecalis penicillin G
acylase-catalyzed enantioselective acylation of dl-phenylalanine and
derivatives in aqueous medium. Tetrahedron Lett. 52, 5398–5402.
Illanes, A., Wilson, L., Altamirano, C., Cabrera, Z., Alvarez, L., Aguirre, C., 2007.
Production of cephalexin in organic medium at high substrate concentrations
with CLEA of penicillin acylase and PGA-450. Enzyme Microb. Technol. 40,
195–203.
Jaiprakash, G.S., Kamalesh, K.K., Gangadhar, R.A., 1987. Evaluation of
determination of 6-aminopenicillanic acid by p-dimethyl aminobenzaldedyde.
Biotechnol. Tech. 1 (1), 69–72.
Jager, S.A.W., Jekel, P.A., Janssen, D.B., 2007. Hybrid penicillin acylases with
improved properties for synthesis of beta-lactam antibiotics. Enzyme Microb.
Technol. 40, 1335–1344.
Jager, S.A.W., Shapovalova, I.V., Jekel, P.A., Alkema, W.B.L., Svedas, V.K., Janssen,
D.B., 2008. Saturation mutagenesis reveals the importance of residues ␣R145
and ␣F146 of penicillin acylase in the synthesis of ␤-lactam antibiotics. J.
Biotechnol. 133, 18–26.
Kallenberg, A.I., van Rantwijk, F., Sheldon, R.A., 2005. Immobilization of penicillin G
acylase: The key to optimum performance. Adv. Synth. Catal. 347, 905–926.
Kumar, A., Prabhune, A., Suresh, G.C., Pundle, A., 2008. Characterization of smallest
active monomeric penicillin V acylase from new source a yeast Rhodotorula
aurantiaca (NCIM 3425). Process Biochem. 43, 961–967.
Parmar, A., Kumar, H., Marwaha, S., Kennedy, J.F., 2000. Advances in enzymatic
transformation of penicillins to 6-aminopenicillanic acid (6-APA). Biotechnol.
Adv. 18, 289–301.
Schroën, C.G.P.H., Nierstrasz, V.A., Kroon, P.J., Bosma, R., Janssen, A.E.M., Beeftink,
H.H., Tramper, J., 1999. Thermodynamically controlled synthesis of ␤-lactam
antibiotics: Equilibrium concentrations and side-chain properties. Enzyme
Microb. Technol. 24 (8–9), 498–506.
Schroën, C.G.P.H., Nierstrasz, V.A., Kroon, P.J., Bosma, R., Tjeerdsma, P.S., DeVroom,
E., VanderLaan, J.M., Moody, H.M., Beeftink, H.H., Janssen, A.E.M., Tramper, J.,
2002. Integrated reactor concepts for the enzymatic kinetic synthesis of
cephalexin. Biotechnol. Bioeng. 80, 144–155.
Sheldon, R.A., 2008. E factors, green chemistry and catalysis: an odyssey. Chem.
Commun. 29, 3352–3365.
Shewale, J.G., Sudhakaran, V.K., 1997. Penicillin V acylase: its potential in the
production of 6-amino penicillanic acid. Enzyme Microb. Technol. 20, 402–410.
Srirangan, K., Orr, V., Akawi, L., Westbrook, A., Moo-Young, M., Chou, C.P., 2013.
Biotechnological advances on penicillin G acylase: pharmaceutical
implications, unique expression mechanism and production strategies.
Biotechnol. Adv. 31 (8), 1319–1332.
Švedas, V., Guranda, D., van Langen, L., van Rantwijk, F., Sheldon, R., 1997. Kinetic
study of penicillin acylase from Alcaligenes faecalis. FEBS Lett. 417, 414–418.
Wang, T., Zhu, H., Ma, X., Fei, Z., Ma, Y., Wei, D., 2006. Enhancing enzymatic activity
of penicillin G acylase by coexpressing pcm gene. Appl. Microbiol. Biotechnol.
72, 953–958.
Wegman, M.A., Janssen, M.H.A., van Rantwijk, F., Sheldon, R.A., 2001. Towards
biocatalytic synthesis of beta-lactam antibiotics. Adv. Synth. Catal. 343,
559–576.
Wu, Q., Chen, C.X., Du, L.L., Lin, X.F., 2010. Enzymatic synthesis of amoxicillin via a
one-pot enzymatic hydrolysis and condensation cascade process in the
presence of organic co-solvents. Appl. Biochem. Biotechnol. 160, 2026–2035.
Zhang, Y.W., Liu, R.J., Xu, X.M., 2010. One-pot: two-step enzymatic synthesis of
amoxicillin by complexing with Zn2+ . Appl. Microbiol. Biotechnol. 88, 49–55.