A practical high-throughput screening system for feruloyl esterases: Substrate design and evaluation

Article abstract of DOI:10.1016/j.molcatb.2011.08.011

Feruloyl esterases (FAEs) are a group of important industrial enzymes, which could hydrolyze the ester bonds between hydroxycinnamic acids and arabinose or galactose in hemicellulose present in plant cell walls. To establish a practical high-throughput screening system of feruoyl esterases, two new substrates with modified chromo- or fluoro-phore, 2-chloro-4-nitrophenyl ferulate (CNPF; 1b) and umbelliferyl 5-O-feruolyl-α-L-arabinofuranoside (UFA; 2a), were designed and synthesized. Both substrates provided significantly improved signal-to-noise ratio compared with previously known structural analogues. The chromogenic substrate CNPF could be readily adapted to the high-throughput screening of feruloyl esterase A from Aspergillus niger, one of the most studied FAEs, with the coefficient of variance of as low as 10.3% as determined for a single sequence library. .

Full text of DOI:10.1016/j.molcatb.2011.08.011

Journal of Molecular Catalysis B: Enzymatic 74 (2012) 36–40
Contents lists available at SciVerse ScienceDirect
Journal of Molecular Catalysis B: Enzymatic
journal homepage: www.elsevier.com/locate/molcatb
A practical high-throughput screening system for feruloyl esterases: Substrate
design and evaluation
Shuai-Bing Zhanga,b,1, Xiao-Feng Maa,b,1, Xiao-Qiong Peia, Jing-Yuan Liua,
Hua-Wu Shaoa,∗∗, Zhong-Liu Wua,∗
a Chengdu Institute of Biology, Chinese Academy of Sciences, P.O. Box 416, Chengdu 610041, China
b Graduate University of the Chinese Academy of Sciences, Beijing 100049, China
a r t i c l e i n f o
a b s t r a c t
Article history:
Received 12 April 2011
Received in revised form 11 August 2011
Accepted 27 August 2011
Available online 3 September 2011
Feruloyl esterases (FAEs) are a group of important industrial enzymes, which could hydrolyze the ester
bonds between hydroxycinnamic acids and arabinose or galactose in hemicellulose present in plant cell
walls. To establish a practical high-throughput screening system of feruoyl esterases, two new substrates
with modified chromo- or fluoro-phore, 2-chloro-4-nitrophenyl ferulate (CNPF; 1b) and umbelliferyl 5-
O-feruolyl-␣-L-arabinofuranoside (UFA; 2a), were designed and synthesized. Both substrates provided
significantly improved signal-to-noise ratio compared with previously known structural analogues. The
chromogenic substrate CNPF could be readily adapted to the high-throughput screening of feruloyl
esterase A from Aspergillus niger, one of the most studied FAEs, with the coefficient of variance of as
low as 10.3% as determined for a single sequence library.
Keywords:
Feruloyl esterases
High-throughput screening
Pichia pastoris
© 2011 Elsevier B.V. All rights reserved.
2-Chloro-4-nitrophenol
Umbelliferyl
1. Introduction
such as hydroxycinnamic esterified polysaccharide, methyl feu-
late and their analogues [6–9]. Recently, several chlorogenic
Feruloyl esterases (FAEs, EC 3.1.1.73) consist of a diverse group
of hydrolases catalyzing the cleavage of ester bonds between arabi-
nose or galactose and hydroxycinnamic acid, and act as the integral
part of the enzyme system involved in the complete hydrolysis of
hemicellulose. The last decade, in particular, has seen considerable
efforts to the application of FAEs in biomass degradation, food and
pharmaceutical industries [1]. One of the most extensively stud-
ied FAEs, feruloyl esterase A (AnFaeA) from Aspergillus niger [2], has
widely been applied to enhance the yield of sugar products from
lignocellulose [3,4].
FAEs generally display broad substrate specificities with the
requirement of phenylalkanoate substrates but have different
specificity on a specific group at a specific location [5]. How-
ever, they always have specificity on monomeric ferulate moiety
and are active on its synthetic substrate analogues. The assay
of FAEs’ activity commonly depends on HPLC, using enzymatic
hydrolysis of a variety of natural and synthetic compounds,
substrates of FAEs have been reported for quantitative assays,
such as 4-nitrophenyl ferulate [10,11] (Fig. 1), 2-O- and 5-O-
feruloylated 4-nitrophenyl-l-arabinofuranosides [12], feruloylated
p-nitrophenyl ␣-l-arabinofuranosides and ␤-d-xylopyranosides
[13], and 2-chloro-4-nitrophenyl 5-O-hydroxycinnamoyl-␣-l-
arabinofuranoside (Fig. 1) [14] as well as substrates based
on (2-chloro-4-nitrophenyl)-1,2,4-butanetriol that require addi-
tional non-enzymatic reactions of periodate oxidation and
beta-elimination to trigger the liberation of the chromophore [15].
A microplate screening based on the pH indicator, 4-nitrophenol,
has also been reported [16]. However, the establishing and evalu-
ation of a reliable and practical high-throughput screening system
is still required, which would otherwise facilitates the directed
evolution of FAEs, as well as the novel enzyme mining from envi-
ronmental microorganisms [17,18].
In the current work, we designed and synthesized two
new mimic substrate 2-chloro-4-nitrophenyl ferulate (CNPF)
and umbelliferyl-5-O-hydroxycinnamoyl-␣-l-arabinofuranoside
(UFA) (Fig. 1), and developed a high-throughput screening system
for FAEs. Both of them and their individual structural analogue
reported were evaluated with feruloyl esterase A from A. niger as a
model enzyme. The result shows that the substrate CNPF could be
readily obtained and applied to the high-throughput screening of
FAEs with desirable sensitivity, while the substrate UFA is difficult
to synthesize but has a higher sensitivity, which may be used
∗
Corresponding author at: Chengdu Institute of Biology, Chinese Academy of Sci-
ences, 9 South Renmin Road, 4th Section, Chengdu, Sichuan 610041, PR China.
Tel.: +86 28 85238385; fax: +86 28 85238385.
∗∗
Corresponding author.
E-mail addresses: shaohw@cib.ac.cn (H.-W. Shao), wuzhl@cib.ac.cn (Z.-L. Wu).
Contributed equally to the present work.
1
1381-1177/$ – see front matter © 2011 Elsevier B.V. All rights reserved.
doi:10.1016/j.molcatb.2011.08.011

S.-B. Zhang et al. / Journal of Molecular Catalysis B: Enzymatic 74 (2012) 36–40
37
(20 mL) and washed with a 5% solution of aqueous Na2CO3 until
the solution was completely colourless. The aqueous layers were
extracted with CH2Cl2, and the product was purified by flash chro-
matography to yield 4a as a white foam (0.71 g, 52%); [˛]2D0 = −96
(c 20.6, CHCl3); 1H NMR (600 MHz, CDCl3): ı 2.11, 2.16, 2.17 (3× s,
9H, 3× CH3C O), 4.27–4.30 (m, 1H, H-5a), 4.40–4.45 (m, 2H, H-
4, H-5b), 5.14 (d, 1H, J3,2 4.8 Hz, H-3), 5.40 (s, 1H, H-2), 5.78 (s,
1H, H-1), 6.29 (d, 1H, J 9.2 Hz, HC CHCO), 6.98 (dd, 1H, J 8.5 Hz, J
2.2 Hz, H-Ar), 7.02 (d, 1H, J 2.2 Hz, H-Ar), 7.42 (d, 1H, J 8.5 Hz, H-Ar),
7.67 (d, 1H, J 9.5 Hz, HC CHCO); 13C NMR (150 MHz, CDCl3): ı 20.6
(CH3C O), 20.7 (CH3C O), 20.9 (CH3C O), 62.9 (C-5), 76.8, 81.4,
81.9 (C-2, C-3, C-4), 103.9 (C-1), 105.6 (HC CHCO), 113.7 (C-Ar),
113.9 (C-Ar), 114.1 (C-Ar), 128.8 (C-Ar), 143.1 (C-Ar), 155.4 (C-Ar),
158.8(C-Ar), 160.7 (HC CHCO), 169.5 (CH3C O), 170.0 (CH3C O),
170.4 (CH3C O); ESI-HRMS exact mass: m/z calcd. for C20H20NaO10
[M+Na] 443.0954; found 443.0950.
Umbelliferyl ␣-l-arabinofuranoside (5a):
A mixture of 4a
(0.15 g, 0.347 mmol) and Et3N (3.5 ml, 0.5 M in MeOH) was refluxed
for 2 h. The reaction mixture was left overnight at room temper-
ature, and was neutralized with silica. The crude product was
purified by flash column chromatography to give compound 5a
as a white foam (0.082 g, 80%); [˛]2D0 = −195 (c 0.87, CH3OH); 1H
NMR (600 MHz, CD3OD): ı 3.68 (dd, 1H, J5a,4 4.8 Hz, J5a,5b 12.1 Hz,
H-5a), 3.78 (dd, 1H, J5b,4 4.3 Hz, J5b,5a 6.8 Hz, H-5b), 4.01 (dd, 1H,
J3,4 4.1 Hz, J3,2 6.2 Hz, H-3), 4.05–4.08 (m, 1H, H-4), 4.28 (dd, 1H,
J2,1 1. Hz, J2,3 4.0 Hz, H-2), 5.63(d, 1H, J1,2 1.9 Hz, H-1), 6.26(d, 1H,
J 9.5 Hz, HC CHCO), 7.02–7.04 (m, 2H, H-Ar), 7.53(d, 1H, J 8.4 Hz,
H-Ar), 7.87(d, 1H, J 9.5 Hz, HC CHCO); 13C NMR (150 MHz, CD3OD):
ı 61.3 (C-5), 76.8, 82.3, 85.5 (C-2,C-3,C-4), 103.4 (C-1), 106.5
(HC CHCO), 112.6 (C-Ar), 113.5 (C-Ar), 114.0 (C-Ar), 129.0 (C-Ar),
144.2 (C-Ar), 155.3 (C-Ar), 160.3 (C-Ar), 161.8 (HC CHCO); ESI-
HRMS exact mass: m/z calcd. for C14H14NaO7 [M+Na] 317.0638;
found 317.0619.
Fig. 1. The substrates of FAEs: structures of 1a/1b (A), and synthesis of 2a/2b
(B). Compounds 1b and 2a were newly designed substrates. 1a: 4-nitrophenol
ferulate (pNPF); 1b: 2-chloro-4-nitrophenol ferulate (CNPF); 2a: umbelliferyl-5-
O-feruloyl-L-arabinofuranoside (UFA); 2b: 2-chloro-4-nitrophenyl 5-O-feruloyl-L-
arabinofuranoside (CNPFA). Reagents and conditions: (i) BF3·Et2O/TEA, 7-hydroxy-
2H-chromen-2-one (for 2a), or BF3·Et2O, 2-chloro-4-nitrophenol (for 2b), −10 ◦C;
(ii) TEA/CH3OH, reflux (for 2a), or NaOCH3, CH3OH, rt (for 2b); (iii) Pyridine, 4-
acetyl feruloyl chloride, toluene, 0 ◦C – rt, 4 h; (iv) pyrroline/CH2Cl2 (1:12), rt. The
enzymatic cleavage sites of the substrates are marked with dotted lines.
for the detection of trace amount of FAEs existing in biological
samples.
Umbelliferyl 5-O-feruloyl-␣-l-arabinofuranoside (2a): The acid
chloride (0.2 g, 0.8 mmol) was dissolved in toluene (30 mL) and
added dropwise to a continuously stirring ice-cold mixture of
5a (0.29 g, 1.0 mmol) in pyridine (10 mL). Once the addition was
complete (approximately 35 min), the mixture was left unstirred
overnight. The resulting mixture was concentrated to a syrup and
twice diluted with toluene and evaporated. The syrup reaction
product was dissolved in CH2Cl2 (10 mL) and pyrrolidine (0.84 mL,
10 mmol), stirred at room temperature for 1 h, and neutralized
with silica. The crude product was purified by flash column chro-
matography (CH2Cl2–MeOH, 99:1 v/v) to give compound 2a as a
brown solid (0.21 g, 58%); [˛]2D0 = −48 (c 0.22, CH3OH); 1H NMR
(600 MHz,CD3OD): ı 3.90 (s, 3H, OMe), 4.08 (dd, 1H, J3,2 4.2 Hz, J3,4
4.2 Hz, H-3), 4.28–4.34 (m, 1H, H-4), 4.35–4.38 (m, 2H, H-5a, H-2),
4.48 (dd, 1H, JH5b,4 3.2 Hz, JH5b,H5a 11.9 Hz, H-5b), 5.70 (s, 1H, H-
1), 6.29 (d, 1H, J 9.4 Hz, HC CHCO), 6.41 (d, 1H, J 15.7 Hz, HC CH),
6.82 (d, 1H, J 8.0 Hz, H-Ar), 7.06–7.08 (m, 3H,), 7.20 (s, 1H, H-Ar),
7.57 (d, 1H, J 8.2 Hz, H-Ar), 7.67 (d, 1H, J 16.1 Hz, HC CH), and 7.89
(d, 1H, J 9.6 Hz, HC CHCO); 13C NMR (150 MHz, CD3OD): ı 55.1
(OMe), 63.4 (C-5), 77.5, 82.2, 82.6 (C-2,C-3, C-4), 103.4 (C-1), 106.5
(HC CHCO), 110.3 (HC CHCO), 112.7 (C-Ar), 113.7 (2 C,C-Ar), 113.9
(C-Ar), 115.1 (C-Ar), 122.9 (C-Ar), 126.3 (C-Ar), 129.1 (C-Ar), 144.2
(HC CHCO), 145.9 (HC CHCO), 148.0 (C-Ar), 149.4 (C-Ar), 155.3
(C-Ar), 160.1 (C-Ar), 161.8 (HC CHCO), and 167.5 (HC CHCO); ESI-
HRMS exact mass calcd:. m/z for C24H20NaO10 [M+Na] 493.1111;
found 493.1092.
2. Materials and methods
2.1. Chemicals
Ferulic acid, 2-chloro-4-nitrophenol and umbelliferone were
purchased from Alfa Aesar (Tianjin, China). 4-Nitrophenyl feru-
late (pNPF) (1a) was synthesized with established method [11].
1,2,3,5-Tetra-O-acetyl-␣-l-arabinofuranose (3) was synthesized
from l-arabinose according to the literature [19]. Spectra of 1b, 2a,
2b, 4a, 4b, 5a, 5b are provided in the Supplementary information.
2.2. Synthesis of 2-chloro-4-nitrophenyl ferulate (1b)
CNPF was synthesized with established method [11] except
that 2-chloro-4-nitrophenol was used instead of 4-nitrophenol. The
product was purified by column chromatography to give 1b as a
light yellow solid (2.79 g, 80%). 1H NMR (CDCl3, 600 MHz) ı 3.97
(s, 3H), 6.50 (d, 1H, J 15.8 Hz), 6.97 (d, 1H, J 8.1 Hz), 7.11 (d, 1H, J
1.7 Hz), 7.18 (dd, 1H, J 1.7 Hz and J 8.1 Hz), 7.44 (d, 1H, J 9.1 Hz),
7.88 (d, 1H, J 15.8 Hz), 8.21 (dd, 1H, J 2.7 Hz and J 9.1 Hz), 8.39 (d,
1H, J 2.7 Hz); ESI-HRMS exact mass: m/z calcd. for C16H12ClNNaO6
[M+Na] 372.0245; found 372.0243.
2.3. Synthesis of umbelliferyl-5-O-feruloyl-a-l-arabinofuranoside
(2a)
2.4. Synthesis of 2-chloro-4-nitrophenyl
Umbelliferyl 2,3,5-tri-O-acetyl-␣-l-arabinofuranoside (4a): To
a solution of 3 (1.0 g, 3.14 mmol) in CH2Cl2 (5 mL), 7-hydroxy-2H-1-
benzopyran-2-one (0.25 g, 1.57 mmol), Et3N (0.24 mL, 1.57 mmol)
and BF3·OEt2 (1 mL, 7.85 mmol), were successively added. The reac-
tion mixture was stirred for 19 h at 0 ◦C, then diluted with CH2Cl2
5-O-feruloyl-˛-l-arabinofuranoside (2b)
2-Chloro-4-nitrophenyl 2,3,5-tri-O-acetyl-␣-l-arabinofurano-
side (4b): To a solution of 3 (1.3 g, 4.1 mmol) in dry CH2Cl2 (35 mL)
˚
containing fresh activated 4 A molecular sieves (1.3 g) was added

38
S.-B. Zhang et al. / Journal of Molecular Catalysis B: Enzymatic 74 (2012) 36–40
2-chloro-4-nitrophenol (0.72 g, 4.1 mmol). The reaction was then
cooled to −20 ◦C and BF3·Et2O (3.3 mL, 25.9 mmol) was added
dropwise. Then the mixture was slowly warmed up to room tem-
perature and stirred for additional 4 h. The mixture was filtered
through Celite, diluted with CH2Cl2, washed successively with 5%
aq. Na2CO3 solution and H2O, and dried with anhydrous Na2SO4.
The crude product was purified by flash column chromatography
to give compound 4b as a white foam (1.11 g, 63%); [˛]2D0 = −129 (c
2.0, CHCl3); 1H NMR (600 MHz, CDCl3):ı 2.11, 2.15, 2.17 (3× s, 9H,
3× CH3C O), 4.29 (dd, 1H, J4,5 6.8 Hz, J5a,5b 13.2 Hz, H-5a), 4.44–4.45
(m, 2H, H-4, H-5b), 5.12 (d, 1H, J3,4 3.5 Hz, H-3), 5.46 (s, 1H, H-2),
5.86 (1H, s, H-1), 7.28 (d, 1H, J 9.1 Hz, H-Ar), 8.14 (dd, 1H, J 9.1,
2.7 Hz, H-Ar), 8.31 (d, 1H, J 2.7 Hz, H-Ar).
2-Chloro-4-nitrophenyl-␣-l-arabinofuranoside (5b): Com-
pound 4b (1.16 g, 2.7 mmol) was suspended in dry MeOH (8 mL),
and NaOMe (17 mg, 0.31 mmol, 0.12 eqiv.) was added at 0 ◦C. After
2 h, the reaction was neutralized with silica. The crude product was
purified by flash column chromatography to give compound 5b as
a yellowish solid (0.612 g, 77%); [˛]2D0 = −208 (c 0.2, CH3OH); 1H
NMR (600 MHz, CD3OD): ı 3.67 (dd, 1H, J5,4 5.2 Hz, J5a,5b 12.1 Hz,
H-5a), 3.78 (dd, 1H, J5,4 3.3 Hz, J5a,5b 12.1 Hz, H-5b), 4.02 (dd, 1H,
J3,4 4.1 Hz, J3,2 6.6 Hz, H-3), 4.07–4.10 (m, 1H, H-4), 4.39 (dd, 1H,
J2,1 1.9 Hz, J2,3 4.4 Hz, H-2), 5.71 (d, 1H, J1,2 1.5 Hz, H-1), 7.44 (d, 1H,
J 9.1 Hz, H-Ar), 8.16 (dd, 1H, J 2.9 Hz, 9.2 Hz, H-Ar), 8.29 (d, 1H, J
2.5 Hz, H-Ar).
AnFaeA-D93G/S187F was purified as described previously using
a Bio-scale Mini UNOsphere Q column [21]. Protein estimations
were done with the BCA Protein Assay Kit (Beyotime, Shanghai,
China) with bovine serum albumin as a standard.
2.6. Measurement of enzymatic activity
To evaluate the feasibility of the four mimic substrates for the
high-throughput screening system of FAEs in a 96-well microplate
format, the reaction mixture was prepared by mixing 160 ␮L of
100 mM sodium phosphate buffer (pH 6.4) containing 2.5% Tri-
ton X-100 with 10 ␮L of 20 mM substrate in DMSO [10], followed
by the addition of 20 ␮L supernatant and sufficient amount of ␣-
l-arabinofuranosidase (␣-AF) (Megazyme, Wicklow, Ireland) only
in case of CNPFA and UFA. The reactions were performed in trip-
licate at 40 ◦C for 10 min with shaking at 100 rpm. The minimal
amount of ␣-l-arabinofuranosidase needed per reaction was con-
firmed under the reaction condition above with different amount of
␣-l-arabinofuranosidase. The formation of 2-chloro-4-nitrophenol
or 4-nitrophenol was measured at 425 nm, and the formation of
umbelliferone was determined fluorometrically with ꢀex of 360 nm
and ꢀem of 460 nm on a Varioskan Flash microplate reader (Thermo
scientific, USA). The signal-to-noise ratio (S/N) of each reaction sys-
tem was defined as the ratio of signal absorbance to the background
noise absorbance. The background noise absorbance was acquired
for the same reaction system without enzyme. The specific activity
of AnFaeA-D93G/S187F toward each substrate was measured with
0.16 ␮M purified enzyme and 1 mM substrate.
2-Chloro-4-nitrophenyl
5-O-feruloyl-␣-l-arabinofuranoside
(2b): The procedure was essentially the same as that for com-
pound 2a. Under N2 atmosphere, acid chloride (0.3 g, 1.28 mmol)
in toluene (35 mL) was added dropwise to a stirring solution
of 5b (0.45 g, 1.47 mmol) in pyridine (10 mL) at 0 ◦C. Once the
addition was complete (approximately 25 min), the mixture was
left unstirred overnight. The resulting mixture was concentrated
to a syrup and twice diluted with toluene and evaporated. The
solid reaction product was dissolved in CH2Cl2 (9 mL), pyrrolidine
(1.1 mL, 14.7 mmol, 10 eqiv.) was added, and the reaction mixture
was stirred at room. After 2 h, the reaction was complete and was
neutralized with silica. The crude product was evaporated and
purified by flash column chromatography (CH2Cl2–MeOH, 99:1,
v/v) to give the title compound 2b as a yellowish solid (0.32 g, 50%).
[˛]2D0 = −51 (c 0.14, CH3OH); 1H NMR (600 MHz, CD3OD): ı 3.88 (s,
3H, OMe), 4.07 (dd, 1H, J3,4 4.3 Hz, J3,2 6.8 Hz,H-3), 4.28–4.34 (m,
2H, H-5a, H-5b), 4.43–4.47 (m, 2H, H-4 H-2), 5.75 (d, 1H, J1,2 1.6 Hz,
H-1), 6.38 (d, 1H, J 15.8 Hz, HC CH), 6.80 (d, 1H, J 7.7 Hz, H-Ar),
7.06 (dd, 1H, J 1.7 Hz, 8.3 Hz, H-Ar), 7.18 (d, 1H, J 1.5 Hz, H-Ar), 7.42
(d, 1H, J 9.1 Hz, H-Ar), 7.64 (d, 1H, J 15.8 Hz, HC CH), 8.18 (dd, 1H, J
2.7 Hz, 9.3 Hz, H-Ar), 8.30 (d, 1H, J 2.7 Hz, H-Ar); ESI-MS: m/z calcd
for [C21H20ClNO10Na]+ 504.07; found 504.01.
The kinetic parameters of purified AnFaeA-D93G/S187F toward
new substrates 1b and 2a were determined under standard
conditions with various substrate concentrations. All measure-
ments were performed in triplicates. Data were fitted to the
Michaelis–Menten equation using GraphPad Prism v5.0 (GraphPad
Software, San Diego, CA, USA) to generate estimates of apparent
Michaelis constant (Km) and catalytic turnover frequency (kcat) val-
ues.
3. Results and discussion
3.1. Design and synthesis of mimic substrates
In order to develop the high-throughput screening system of
FAEs, we firstly synthesized the previously reported substrate pNPF
(1a) (Fig. 1A). And then, we also designed a novel substrate, CNPF
(1b) (Fig. 1A) with an almost pH-independent chromophore, 2-
chloro-4-nitrophenol, with a view to obtain higher molar extinction
coefficiency (ε) at neutral pH. The substrate CNPF was easily syn-
thesized through the same synthetic route of substrate pNPF.
According to the previous report [14], AnFaeA displayed over
10-fold catalytic efficiency toward 2-chloro-4-nitrophenyl 5-O-
feruloyl-l-arabinofuranoside (CNPFA) 2b (Fig. 1B) than methyl
ferulate due to the resemblance of CNPFA to natural sub-
strates. Thus, we synthesized this substrate, as well as further
designed a novel fluorogenic substrate umbelliferyl-5-O-feruloyl-
a-l-arabinofuranoside (UFA) 2a (Fig. 1B) with the expectation of
achieving better sensitivity.
The synthesis of CNPFA was previously described by
Marmuse et al., starting from 1-O-acetyl-2,3,5-tri-O-benzoyl-
l-arabinoarabinofuranose [14]. The procedure requires the
application of various protecting/deprotecting strategies that
was quite time-consuming and suffered from poor yield. In
order to shorten the lengthy protection/deprotection sequences,
1,2,3,5-tetra-O-acetyl-␣-l-arabinofuranose 3 were used instead of
benzoylated counterpart as starting material, which was synthe-
sized from l-arabinose according to the method for the synthesis
2.5. Expression of AnFaeA in Pichia pastoris KM71
The enzyme was heterologously expressed in Pichia pastoris
KM71 [20] as an extracellular protein following established method
[21] in a miniature format. Briefly, the recombinant plasmid
pGAPZ␣A/faeA-D93G/S187F [21] encoding a variant of FaeA from
A. niger CIB 423.1 was linearized by BspHI digestion and electro-
transformed into P. pastoris KM71 (Invitrogen, Carlsbad, CA, USA).
The transformants were grown on YPD medium containing 1%
(w/v) yeast extract, 2% (w/v) peptone, 2% (w/v) dextrose supple-
mented with 1 M sorbitol and 100 ␮g mL−1 zeocin. After incubation
at 30 ◦C for 72 h, single colonies were picked with sterile tooth-
picks and inoculated into 150 ␮L YPD medium on sterile 96-well
microplates, and incubated for 48 h at 30 ◦C with gyratory shaking
at 280 rpm on an INFORS Multitron incubator shaker (INFORS HT,
Swissland). The plate was subjected to centrifugation at 1800 g for
10 min. The supernatant of each well was transferred into another
96-well plate to assay the enzymatic activity.

S.-B. Zhang et al. / Journal of Molecular Catalysis B: Enzymatic 74 (2012) 36–40
39
of d-arabinofuranose [19]. Then glycosyl donor 3 was coupled
with 2-chloro-4-nitrophenol smoothly promoted with BF3·Et2O.
The product 4b was obtained in 63% yield. Deacetylation of 4b
with 1 M NaOMe in MeOH afforded the expected compound 5b
in 77% yield. Selectively acylation of 5b with 4-O-acetylferuloyl
chloride [22] under mild basic conditions [8] and subsequent
De-O-acetylation with pyrrolidine in dichloromethane [12] gave
2-chloro-4-nitrophenyl 5-O-feruloyl-␣-l-arabinofuranoside 2b
(Fig. 1B) in 50% yield.
Compound UFA 2a (Fig. 1B) is an analogue of reported
CNPFA, in which the 2-chloro-4-nitrophenyl aglycon was replaced
by coumarin. Unexpectedly, we were unable to get the gly-
cosylated compound 4a according to the above coupling
method. Interestingly, under the BF3·Et2O/TEA condition, the
glycosylation of umbelliferone reacted with 1,2,3,5-tetra-O-
acetyl-l-arabinofuranose 3 smoothly, producing the expected
␣-l-arabinofuranoside 4a in 52% yield [23]. Similarly, deacetyla-
tion of 4a with 0.5 M TEA in MeOH gave compound 5a in 80%
yield [24]. Then, the coupling reaction of 5a and 4-O-acetylferuloyl
chloride was carried out with pyridine as base and followed by
De-O-acetylation to produce the desired umbelliferyl 5-O-feruloyl-
␣-l-arabinofuranoside 2a (Fig. 1B) in 55% yield.
Fig. 2. Release of chlorophore (closed circle) or fluorophore (open circle) with the
addition of arabinosidase.
However, a considerable amount of ␣-AF was required for the
release of the fluorophore, which would add approximately $0.65
for each well according to the Megazyme catalog (Bray, Ireland),
and thus made it impractical for the assay of large amount of sam-
ples. In addition, the synthesis of both CNPFA and UFA involved a
multi-step process, and required expertise in carbohydrate chem-
istry, which limited their availability. In contrast, the synthesis of
CNPF was easily achieved in one step, and no ␣-AF was required
for the assay. A high-throughput screening system based on sub-
strate CNPF would be more sensitive than that based on NPF, and at
the same time, more convenient than that based on UFA or CNPFA.
The application of CNPF afforded an S/N of 7.1, which was sensi-
tive enough for an industrially relevant enzyme. Hence, CNPF was
chosen as the substrate for the high-throughput screening system.
3.2. Evaluation of the four mimic substrates
Both NPF and CNPF are unstable under alkaline conditions. Cat-
alyzed by the FAEs, they could directly release the chlorophore after
the cleavage of the ester bond. After 10-min reactions under pH 6.4,
the chlorogenic substrate, pNPF, resulted in a low signal-to-noise
ratio (S/N) of 2.4 (Table 1), while the newly designed substrate,
CNPF, with Km and kcat of 5.49 0.66 mM and 116 9.31 min−1
,
3.3. Application of CNPF in a high-throughput manner
showed an significantly improved S/N value of 7.1 under the same
conditions, around 3-fold of that of the substrate pNPF (Table 1).
The release of chlorogenic or fluorogenic group of CNPFA
and UFA needs a two-step enzymatic process including an FAE-
catalyzed release of an intermediate chromogenic arabinoside and
the liberation of the chlorophore/umbelliferone catalyzed by ␣-AF
[14]. Under the reaction condition, at least 37.5 or 25 ng of ␣-
AF was required to fully release the chlorophore or fluorophore,
respectively, for a 10-min reaction with 20-␮L crude AnFaeA (about
13.6 U) (Fig. 2). The apparent Michaelis constant (Km) and cat-
alytic turnover frequency (kcat) of the novel substrate UFA was
1.84 0.19 mM and 63.6 3.8 min−1. The hydrolysis rates of CNPFA
were higher than that of pNPF or CNPF, or UFA in 10-min reac-
tions (Table 1). However, UFA displayed the best S/N value of 70
(Table 1). It was also observed that substrate CNPFA was unstable
under alkaline conditions, while substrate UFA could well tolerate
alkaline conditions. In addition, umbelliferone is a pH-dependent
fluorophore. Therefore, for substrate UFA, glycine buffer (400 mM,
pH 10.8) was added before fluorescence detection to enhance the
signal intensity, which led to a S/N value of as high as 540 (Table 1).
The results clearly indicated that UFA afforded the highest sen-
sitivity for the assay of AnFaeA expressed in P. pastoris on 96-well
plate, which could be beneficial for enzymes with low activity.
The reproducibility and uncertainty of the high-throughput
screening system using CNFA was evaluated with AnFaeA variant
expressed in P. pastoris. A total of 96 transformants of P. pastoris
expressing AnFaeA were incubated in the 96-well plate format
for a single sequence library and the supernatant of induced cul-
ture was subjected to enzymatic assay (Fig. 3). The reaction time
was extended to 15 min, which raised the S/N value to 9.8. The
mean and standard deviation of the absorbance values from 96
samples were 1.74 and 0.18, respectively (Fig. 3), making the coef-
ficient of variance (CV) of the screening to be 10.3% (CV = [standard
deviation/(mean) × 100%]), which indicates lower variability of the
high-throughput screening systems [25,26]. The expected false
positives exhibiting higher than 1.5 times the mean value would
be close to zero in a typical library of 104–105 colonies according
to statistical calculations [26].
Table 1
Hydrolysis of four mimic substrates catalyzed with AnFaeA.
Substrate
Specific activity
S/N
␣-AF (ng well−1
)
(␮mol min−1 mg−1
)
pNPF (1a)
CNPF (1b)
UFA (2a)
0.52 0.01
0.62 0.01
0.67 0.01
1.0 0.01
2.4
7.1
No
No
>25.0
>37.5
70 (540)a
20
CNPFA (2b)
Fig. 3. Evaluation of the feasibility of using 2-chloro-4-nitrophenol ferulate in
a high-throughput manner. Enzyme activities are plotted versus plate position
(columns) and in descending order (squares).
a
Data in bracket was measured after the addition of equivalent volume of 400 mM
glycine buffer (pH 10.8) to quench the reaction.

40
S.-B. Zhang et al. / Journal of Molecular Catalysis B: Enzymatic 74 (2012) 36–40
4. Conclusions
References
[1] D.W.S. Wong, Appl. Biochem. Biotechnol. 133 (2006) 87–112.
[2] A.E. Fazary, Y.-H. Ju, Biotechnol. Mol. Biol. Rev. 3 (2008) 095–110.
[3] M.G. Tabka, I. Herpoël-Gimbert, F. Monod, M. Asther, J.C. Sigoillot, Enzyme
Microb. Technol. 39 (2006) 897–902.
[4] M.J. Selig, E.P. Knoshaug, W.S. Adney, M.E. Himmel, S.R. Decker, Bioresour.
Technol. 99 (2008) 4997–5005.
[5] A.E. Fazary, Y.-H. Ju, Acta Biochim. Biophys. Sinica 39 (2007) 811–828.
[6] W.S. Borneman, R.D. Hartley, D.S. Himmelsbach, L.G. Ljungdahl, Anal. Biochem.
190 (1990) 129–133.
[7] N. Juge, G. Williamson, A. Puigserver, N.J. Cummings, I.F. Connerton, C.B. Faulds,
FEMS Yeast Res. 1 (2001) 127–132.
During the last decade, many researchers have been devoted to
the design and synthesis of mimic substrates for the rapid assay of
FAEs. In this study, we designed and synthesized two new chromo-
or fluoro-genic substrates, CNPF and UFA, to improve the spec-
trophotometric and fluorometric assay of FAEs. Compared with
their structural analogues, pNPF and CNPFA, respectively, CNPF and
UFA displayed significantly enhanced sensitivity. The fluorogenic
substrate UFA has better stability at alkaline pH and higher signal-
to-noise ratio, and may be suitable for detecting small changes of
activity. The chromogenic substrate CNPF could be easily acquired,
and well adapted to high-throughput screening. A practical high-
throughput screening system with a CV of 10.3% was established for
AnFaeA expressed in P. pastoris in the 96-well-plate format using
the chromogenic substrate CNPF. This system has been success-
fully applied to our ongoing work on the identification of AnFaeA
variants with improved thermostability from random mutagenesis
library.
[8] R.D. Hatfield, R.F. Helm, J. Ralph, Anal. Biochem. 194 (1991) 25–33.
[9] T. Koseki, A. Hori, S. Seki, T. Murayama, Y. Shiono, Appl. Microbiol. Biotechnol.
83 (2009) 689–696.
[10] V. Mastihuba, L. Kremnicky´, M. Mastihubová, J.L. Willett, G.L. Côté, Anal.
Biochem. 309 (2002) 96–101.
[11] S. Hegde, P. Srinivas, G. Muralikrishna, Anal. Biochem. 387 (2009) 128–129.
[12] M. Mastihubová, J. Szemesová, P. Biely, Tetrahedron Lett. 44 (2003)
1671–1673.
[13] M. Mastihubová, P. Biely, Carbohydr. Res. 345 (2010) 1094–1098.
[14] L. Marmuse, M. Asther, E. Fabre, D. Navarro, L. Lesage-Meessen, M. Asther, et al.,
Org. Biomol. Chem. 6 (2008) 1208–1214.
[15] L. Marmuse, M. Asther, D. Navarro, L. Lesage-Meessen, M. Asther, S. Fort, et al.,
Carbohydr. Res. 342 (2007) 2316–2321.
Acknowledgements
[16] L. Ramírez, J. Arrizon, G. Sandoval, A. Cardador, R. Bello-Mendoza, P. Lappe,
et al., Appl. Biochem. Biotechnol. 151 (2008) 711–723.
[17] J.-L. Reymond, P. Babiak, Adv. Biochem. Eng. Biotechnol. 105 (2007) 31–58.
[18] P. Lorenz, C. Schleper, J. Mol. Catal. B: Enzym. 19–20 (2002) 13–19.
[19] D. Crich, C.M. Pedersen, A.A. Bowers, D.J. Wink, J. Org. Chem. 72 (2007)
1553–1565.
[20] R. Daly, M.T.W. Hearn, J. Mol. Recognit. 18 (2005) 119–138.
[21] S.-B. Zhang, Z.-L. Wu, Bioresour. Technol. 102 (2011) 2093–2096.
[22] A. Hosoda, E. Nomura, K. Mizuno, H. Taniguchi, J. Org. Chem. 66 (2001)
7199–7201.
This work was supported by the Program of 100 Distinguished
Young Scientists and the Knowledge Innovation Program of the
Chinese Academy of Sciences (No. KSCX1-YW-11B2), the National
Natural Science Foundation of China (20802073 and 21072183), the
Province Science Foundation of Sichuan, China (Nos. 08ZQ026-023
and 2010SZ0128).
[23] Y.S. Lee, E.S. Rhoa, Y.K. Min, B.T. Kim, K.H. Kim, J. Carbohydr. Chem. 20 (2001)
503–506.
Appendix A. Supplementary data
[24] M. Jesˇelnik, J. Plavec, S. Polanc, M. Kocˇevar, Carbohydr. Res. 328 (2000)
591–597.
[25] Z.-G. Zhang, Z.-L. Yi, X.-Q. Pei, Bioresour. Technol. 101 (2010) 9272–9278.
[26] O. Salazar, L. Sun, Methods Mol. Biol. 230 (2003) 85–97.
Supplementary data associated with this article can be found, in
the online version, at doi:10.1016/j.molcatb.2011.08.011.