Soluble expression in Escherichia coli of active human cyclic nucleotide phosphodiesterase isoform 4B2 in fusion with maltose-binding protein

Article abstract of DOI:10.1271/bbb.90131

Recombinant expression in Escherichia coli of human cyclic nucleotide phosphodiesterase 4B2 (hPDE4B2) fused to maltose-binding-protein (MBP-hPDE4B2) was investigated. hPDE4B2 DNA amplified via nested RT-PCR with total RNAs from U937 cells was ligated with pMAL-p2x. After induction at 18 °C for 16 h, soluble MBP-hPDE4B2 was produced in E. coli. MBP-hPDE4B2 after amylose-resin chromatography showed 35% homogeneity, and its Michaelis-Menten constant was 10 ± 2 μM (n = 3). Rolipram had a dissociation constant of 9 ± 2 nM (n = 2), and zinc ion was a potent inhibitor. Hence, MBP-hPDE4B2 was expressed in E. coli as a soluble active protein.

Full text of DOI:10.1271/bbb.90131

Biosci. Biotechnol. Biochem., 73 (4), 968–970, 2009
Communication
Soluble Expression in Escherichia coli of Active Human Cyclic
Nucleotide Phosphodiesterase Isoform 4B2 in Fusion
with Maltose-Binding Protein
Sha ZHU,1 Genqing YANG,1;2 Xiaolan YANG,1 Yunsheng ZHAO,1 Xiang LI,1 Ping DENG,1
Yanling XIE,1 Zhiyong GAN,1 Yin LIU,1 Zhirong LI,1 Juan LIAO,1 Ming’an YU,1 and Fei LIAO
y
1;
1Chongqing Key Laboratory of Biochemical and Molecular Pharmacology, Chongqing Medical University,
Chongqing 400016, China
2Laboratory Medicine Department, Third Affiliated Hospital of Xinxiang Medical College,
Xinxiang, Henan 453003, China
Received February 17, 2009; Accepted March 12, 2009; Online Publication, April 7, 2009
[doi:10.1271/bbb.90131]
Recombinant expression in Escherichia coli of human
cyclic nucleotide phosphodiesterase 4B2 (hPDE4B2)
fused to maltose-binding-protein (MBP-hPDE4B2) was
investigated. hPDE4B2 DNA amplified via nested
RT-PCR with total RNAs from U937 cells was ligated
with pMAL-p2x. After induction at 18 ꢀC for 16 h,
soluble MBP-hPDE4B2 was produced in E. coli. MBP-
hPDE4B2 after amylose-resin chromatography showed
35% homogeneity, and its Michaelis-Menten constant
was 10 ꢁ 2 ꢀM (n ¼ 3). Rolipram had a dissociation
constant of 9 ꢁ 2 nM (n ¼ 2), and zinc ion was a potent
inhibitor. Hence, MBP-hPDE4B2 was expressed in
E. coli as a soluble active protein.
active human PDE4 isoform B2 in fusion with MBP
(MBP-hPDE4B2).
U937 cells were cultured in RPMI1640 medium
containing 10% fetal beef serum and 5% CO2 at 37 ꢀC
for 2 d to extract total RNAs with reagent kits from
Takara (all materials and reagent kits for operating DNA
and the vector were from Takara, Dalian, China).
According to the coding sequence of hPDE4B2
(Genebank accession no. M97515), three primers,
CTBF1 (50-GAATTCATGAAGGAGCACGGGGCAC-
30), CTBF2 (50-AAGAGACCTCCTAAAGACAT-30),
and CTBR (50-AAGCTTTTATGTATCCACGGGGG-
ACTTGT-30), were used to amplify human PDE4B2
(hPDE4B2) via nested RT-PCR with standard protocols.
Then the PCR product was digested with EcoR I and
Hind III, purified by agarose electrophoresis, and
inserted into pMD19-T for amplification in E. coli
JM109.
Key words: cyclic nucleotide phosphodiesterase 4;
recombinant expression; maltose-binding-
protein; kinetic properties; rolipram
Adenosine 30,50-cyclic monophosphate (cAMP) is a
key signal suppressing functions of immune cells
involved in inflammation reactions,1–3) and its action is
terminated via hydrolysis to adenosine 50-monophos-
phate (AMP) by cyclic nucleotide phosphodiesterase
(PDE). Among PDE isozymes, cAMP-specific PDE
(PDE4) predominates in inflammation reaction cells.
Therefore, PDE4 is a target in anti-inflammation
therapy,3–8) and some selective PDE4 inhibitors have
been proven to have promising efficiency to treat
inflammatory respiratory diseases.8–10)
For screening of human PDE4 inhibitors, active
PDE4 is required. Human PDE4 can be produced via
recombinant expression in eukaryotic cells, but it
shows high cost and low efficiency.11–14) Alternatively,
it can be produced via recombinant expression in
E. coli at low cost, but it suffers from insolubility and
laborious renaturation steps.14) The expression in
E. coli of eukaryotic genes in fusion with maltose-
binding-protein (MBP) is effective to produce soluble
active proteins after an easy purification step.15–18)
Herein, we report the soluble expression in E. coli of
After full-length sequencing, however, two point
mutations were identified in the inserted DNA fragment,
of which one coded a different amino acid. Hence, site-
directed mutation was used to correct the DNA sequence
to give the amino acid sequence for the native hPDE4B2
protein. Afterwards, hPDE4B2 DNA fragment was
released from pMD19-T by digestion with EcoR I and
Hind III, and ligated with pMAL-p2x after digestion
with the same two enzymes, to give pMAL-p2x-
hPDE4B2. Then competent E. coli BL21 cells were
transformed with pMAL-p2x-hPDE4B2, and positive
clones were verified by full sequencing. Finally, a
selected positive clone was cultured at 37 ꢀC for 3 h, and
induced by 1.0 mM isopropyl-ꢀ-D-thiogalactopyranoside
(IPTG, Sigma-Aldrich, St. Louis, MO, USA) for 16 h at
18 ꢀC to produce MBP-hPDE4B2. E. coli BL21 trans-
formed with the blank vector was also induced with
IPTG to produce endogenous PDEs.
After induction with IPTG, the E. coli cells were
harvested, diluted with a lysis buffer (50.0 mM sodium
citrate buffer at pH 6.4 containing 10.0 mM MgCl2,
5.0 mM EDTA, 2.0 mM p-aminobenzamidine, 1.0 mM
y
To whom correspondence should be addressed. Tel: +86-23-68485822; Fax: +86-23-68485161; E-mail: liaofeish@yahoo.com and
liaofeish@cqmu.edu.cn
Abbreviations: EDTA, ethylenediamine-tetraacetic acid; hPDE4B2, human PDE4 isoform B2; IEC, ion-exchange chromatography; Km, Michaelis-
Menten constant; MBP, maltose-binding-protein; MBP-hPDE4B2, human PDE4B2 fused to the C-terminal of MBP; PDE, cyclic nucleotide
phosphodiesterase

Soluble Expression in E. coli of PDE4B2 Fused to MBP
969
Fig. 2. Purification of MBP-hPDE4B2 from Cell Lysate.
The amounts of proteins in various lanes were approximately the
same. An arrow indicates the band for the desired MBP-hPDE4B2.
a, Cell lysate from the same positive clone of E. coli without
induction by IPTG; b, the supernatant after dialysis of the dissolved
precipitate from fractionation of the cell lysate with 60% saturation
of ammonium sulfate; c, MBP-hPDE4B2 eluted with maltose from
the amylase-resin column; d, Molecular weight marker as labeled;
e, proteins eluted with the equilibrating buffer alone from the
amylase-resin column.
the equilibrating buffer. To facilitate the folding of
proteins during syntheses, MBP-hPDE4B2 was induced
by IPTG at 18 ꢀC for 16 h. In this case, a few bound
proteins on DEAE-Cellulose could be eluted with 2.0 M
NaCl in the equilibrating buffer, which, however,
contained no MBP-hPDE4B2. Therefore, IEC was
ineffective to purify MBP-hPDE4B2.
Fig. 1. Induction of MBP-hPDE4B2 in E. coli with 1.0 mM IPTG.
The amounts of proteins in various lanes were approximately the
same. An arrow indicate the band for the desired MBP-hPDE4B2. a,
Molecular weight marker as labeled; b, pellets of cell lysate
dissolved in the sampling buffer for SDS–PAGE; c, the supernatant
of cell lysate after supersonic treatment; d, cell lysate from the
same positive clone of E. coli without induction by IPTG; e, cell
lysate from that transformed with the blank vector and induced
with IPTG.
In the cell lysate from E. coli after IPTG induction for
16 h, the proteins were fractionated with 60% ammo-
nium sulfate to yield about 16% proteins in the
precipitate, which was dissolved and dialyzed against
a buffer (20.0 mM Tris–HCl buffer at pH 7.4, 10.0 mM
MgCl2, 0.20 mM p-aminobenzamidine, and 0.10 mM
EDTA). After centrifugation at 4 ꢀC, an aliquot of the
supernatant containing 4.0 mg proteins was passed
through an amylose-resin column (5:0 cm ꢁ 1:0 cm)
pre-equilibrated with the equilibrating buffer (10.0 mM
Tris–HCl at pH 7.4 plus 10.0 mM MgCl2 and 0.1 mM
EDTA). After the unbound proteins were eluted with
this equilibrating buffer, 1.70 mg MBP-hPDE4B2 was
eluted with 10.0 mM maltose in the equilibrating buffer
(8% yield of proteins from cell lysate). Proteins from
the cells transformed with the blank vector were
fractionated with ammonium sulfate and eluted from
an amylose-resin column to give endogenous PDEs.
In these two PDE preparations, proteins were quanti-
fied by the Bradford method using bovine serum
albumin as the reference protein.19) Hydrolysis of cAMP
(Sigma-Aldrich, MO, USA) by PDEs was monitored by
coupled end-point assay of phosphate with the malachite
green (free base, Sigma-Aldrich), which was released
from AMP by the action of calf intestinal alkaline
phosphatase (Promega, Madison, WI, USA).20) PDE
reaction buffer contained 20.0 mM Tris–HCl buffer at
pH 7.4, 10.0 mM MgCl2 and 0.10 mM EDTA. One unit
of PDE4 was the amount hydrolyzing one micromole
cAMP per minute.
dithioerythritol, and 1.0 mM p-methylsulfonyl fluoride),
and treated with ultrasonic at 4 ꢀC for 15 min in an ice-
water bath. When the diluted cells had an absorbance
below 0.5 at 600 nm, there were no detectable proteins
in the pellet after ultrasonic treatment while MBP-
hPDE4B2 with an expected molecular weight of about
109 kDalton was easily detected in the soluble lysate by
sodium dodecyl polyacrylamide gel electrophoresis
(SDS–PAGE) (Fig. 1). When the diluted cells had an
absorbance above 0.9 at 600 nm, less than 3% proteins
was detected in the pellets, and their profiles were
similar to those of the soluble lysate. Therefore, MBP-
hPDE4B2 was soluble in E. coli, which was different
from that in E. coli with a FLAG tag,14) or in yeast with
a His tag.11)
The complete lysis of cells by ultrasonic treatment
gave a large amount of cell lysate for handling (the
culture medium in 700 ml gave about 320 mg protein in
120 ml lysate), which stimulated a trial purification of
MBP-hPDE4B2 by ion-exchange chromatography
(IEC). After induction with IPTG at 28 ꢀC, a few
proteins in the cell lysate passed out from a DEAE-
Cellulose column (DE32, Amersham Pharmacia, NJ,
USA), but no detectable proteins could be further eluted
out by 2.0 M NaCl in an equilibrating buffer (10.0 mM
Tris–HCl at pH 7.4, plus 10.0 mM MgCl2 and 0.1 mM
EDTA). When the proteins in the lysate were loaded on
a CM-Cellulose column, MBP-hPDE4B2 and other
proteins were found in the eluted solution without NaCl
in the equilibrating buffer, but few proteins could be
further eluted out from the column with 2.0 M NaCl in
MBP-hPDE4B2 eluted with 10 mM maltose from the
amylose-resin column showed about 35% homogeneity
by SDS–PAGE, and its molecular weight was about
100 kD, close to that deduced from its coding sequence

970
S. ZHU et al.
(Fig. 2). Its highest specific activity reached 0.12
Uꢂmgꢃ1 protein at 60.0 mM cAMP, and the total activity
was about 0.2 U from 4.0 mg proteins for affinity
chromatography. Thus MBP-hPDE4B2 was about
0.01 Uꢂmgꢃ1 protein in cell lysate and about 5 Uꢂlꢃ1 in
the culture medium. Based on the homogeneity, its
specific activity was about 50% of that expressed in
E. coli with a FLAG tag,14) and about 35% of that
expressed in yeast with a His tag.11) The MBP tag
accounted for 30% of the molecular weight of MBP-
hPDE4B2, which might partially account for its lower
specific activity than that with a FLAG tag in E. coli.14)
However, the productivity of MBP-hPDE4B2 was low,
which might have been due to an absence of phosphor-
ylation of hPDE4B2 synthesized in E. coli.13,21–23)
Zinc chloride at 50.0 mM had no effects on endogenous
PDEs, whereas it inhibited MBP-hPDE4B2 by 90%, an
effect consistent with that on human PDE4 expressed in
E. coli or in insect cells.14) Moreover, Michaelis-Menten
constant (Km) of endogenous PDEs from E. coli was
28 ꢄ 1 mM (n ¼ 3), while Km of MBP-hPDE4B2 was
10 ꢄ 2 mM (n ¼ 3), consistent with that of human PDE4
with a FLAG tag produced in E. coli.14) Furthermore,
the inhibition by rolipram (ICN 159810, stock solution
at 10.0 mM in dimethylsulfone) on these two prepara-
tions of PDEs was much different. The half-inhibition
concentration of rolipram was determined by regression
analysis of the linear part of a plot of inhibition
percentages to logarithmic concentrations of rolipram,
from which the dissociation constant was derived.20,24)
As to endogenous PDEs, rolipram showed an IC50 above
200 mM with a dissociation constant above 60 mM. As to
MBP-hPDE4B2, rolipram showed an IC50 of 52 ꢄ 2 nM
at 60.0 mM cAMP and a dissociation constant of 10 ꢄ
2 nM (n ¼ 3), consistent with that on human PDE4
having the high-affinity binding site.12,14,25) Besides, at
30.0 mM cAMP, rolipram had the dissociation constant
of 10 ꢄ 1 nM (n ¼ 2), confirming that rolipram was its
competitive inhibitor.24) The recombinant expression of
human PDE4 in yeast with a His tag gave both soluble
and insoluble hPDE4s of different kinetic properties.11)
Therefore, after one-step affinity chromatography with
amylose-resin, the MBP-hPDE4B2 preparation showed
kinetic homogeneity and was suitable for screening of
its inhibitors.
Acknowledgment
This study was supported by the National Natural
Science Foundation of China (no. 30472139, 30672009)
and Chongqing Medical University (2003CX03).
References
1) Conti M and Jin SL, Prog. Nucl. Acid Res. Mol. Biol., 63, 1–38
(1999).
2) Wong WSF and Koh DSK, Biochem. Pharmacol., 59, 1323–
1335 (2000).
3) Tasken K and Stokka AJ, Biochem. Soc. Trans., 34, 476–479
(2006).
4) Conti M, Richter W, Mehats C, Livera G, Park JY, and Jin C,
J. Biol. Chem., 278, 5493–5496 (2003).
5) Banner KH and Trevethick MA, Trends Pharmacol. Sci., 25,
430–436 (2004).
6) Houslay MD, Schafer P, and Zhang KY, Drug Discov. Today,
10, 1503–1519 (2005).
7) Lugnier C, Pharmacol. Ther., 109, 366–398 (2006).
8) Adcock IM, Caramori G, and Chung KF, Lancet, 372, 1073–
1087 (2008).
9) Lipworth BJ, Lancet, 365, 167–175 (2005).
10) Kroegel C and Foerster M, Exp. Opin. Investig. Drugs, 16, 109–
124 (2007).
11) Wilson M, Sullivan M, Brownn N, and Hously MD, Biochem.
J., 304, 407–415 (1994).
12) Wang P, Myers JG, Wu P, Cheewatrakoolpong B, Egan RW,
and Billah MM, Biochem. Biophys. Res. Commun., 234, 320–
324 (1997).
13) Rocque WJ, Holmes WD, Patel IR, Dougherty RW, Ittoop O,
Overton L, Hoffman CR, Wisely GB, Willard DH, and Luther
MA, Protein Expr. Purif., 9, 191–202 (1997).
14) Lario PI, Bobechko B, Bateman K, Kelly J, Vrielink A, and
Huang Z, Arch. Biochem. Biophys., 394, 54–60 (2001).
15) Terpe K, Appl. Microbiol. Biotechnol., 60, 523–533 (2003).
16) Feher A, Boross P, Sperka T, Oroszlan S, and Tozser J, Protein
Expr. Purif., 35, 62–68 (2004).
17) Kohl T, Schmidt C, Wiemann S, Poustka A, and Korf U,
Proteome Sci., 6, 4 (2008).
18) Li C, Tian M, Yuan Y, and Zhou Q, Assay Drug Dev. Technol.,
6, 803–810 (2008).
19) Bradford MM, Anal. Biochem., 72, 248–254 (1976).
20) Zhu S, Gan Z, Li Z, Liu Y, Yang X, Deng P, Xie Y, Yu M, Liao
H, Zhao Y, Zhao L, and Liao F, Anal. Chim. Acta, 636, 105–110
(2009).
21) Sette C and Conti M, J. Biol. Chem., 271, 16526–16534 (1996).
22) MacKenzie SJ, Baillie GS, McPhee I, MacKenzie C, Seamons
R, McSorley T, Millen J, Beard MB, van Heeke G, and Houslay
MD, Br. J. Pharmacol., 136, 421–433 (2002).
23) Houslay MD and Adams DR, Biochem. J., 370, 1–18 (2003).
24) Cheng Y and Prusoff WH, Biochem. Pharmacol., 22, 3099–
3108 (1973).
In conclusion, MBP-hPDEB2 was successfully ex-
pressed as a soluble active protein in E. coli, and the
MBP tag could also be used to produce fused proteins of
other PDE isozymes for screening of their selective
inhibitors, respectively.
25) Rocque WJ, Tian G, Wiseman JS, Holmes WD, Zajac-
Thompson I, Willard DH, Patel IR, Wisely GB, Clay WC,
Kadwell SH, Hoffman CR, and Luther MA, Biochem., 36,
14250–14261 (1997).