Structure activity studies with xenobiotic substrates using carboxylesterases isolated from Arabidopsis thaliana

Article abstract of DOI:10.1016/j.phytochem.2006.12.014

Carboxylesterases (CXEs) catalyse the hydrolysis of xenobiotics and natural products radically altering their biological activities. Whereas the substrate selectivity of animal CXEs, such as porcine liver esterase (PLE) have been well studied, the respective enzymes in plants have yet to be defined and their activities determined. Using Arabidopsis thaliana (At) as a source, five representative members of the α/β hydrolase AtCXE family of proteins have been cloned, expressed and the purified recombinant proteins assayed for esterase activity with xenobiotic substrates. Two members, AtCXE5 and AtCXE18 were found to be active carboxylesterases, though AtCXE5 proved to be highly unstable as a soluble protein. AtCXE18 and the previously characterised S-formylglutathione hydrolase from Arabidopsis (AtSFGH) were assayed against a series of esters based on methylumbelliferone in which the acyl moiety was varied with respect to size and conformation. The same series was used to assay crude esterase preparation from Arabidopsis plants and the results compared with those obtained with the commonly used PLE. With straight chain esters, AtCXE18 behaved like PLE, but the Arabidopsis hydrolases proved less tolerant of branched chain acyl components than the mammalian enzyme. While none of the enzyme preparations accurately reflected all the activities determined with crude Arabidopsis protein extracts, the plant enzymes proved more useful than PLE in predicting the hydrolysis of the more sterically constrained esters.

Full text of DOI:10.1016/j.phytochem.2006.12.014

PHYTOCHEMISTRY
Phytochemistry 68 (2007) 811–818
www.elsevier.com/locate/phytochem
Structure activity studies with xenobiotic substrates using
carboxylesterases isolated from Arabidopsis thaliana
*
Ian Cummins, Marie Landrum, Patrick G. Steel, Robert Edwards
Centre for Bioactive Chemistry, Department of Chemistry, Durham University, Durham DH1 3LE, United Kingdom
Received 20 October 2006; received in revised form 13 December 2006
Available online 31 January 2007
Abstract
Carboxylesterases (CXEs) catalyse the hydrolysis of xenobiotics and natural products radically altering their biological activities.
Whereas the substrate selectivity of animal CXEs, such as porcine liver esterase (PLE) have been well studied, the respective enzymes
in plants have yet to be defined and their activities determined. Using Arabidopsis thaliana (At) as a source, five representative members
of the a/b hydrolase AtCXE family of proteins have been cloned, expressed and the purified recombinant proteins assayed for esterase
activity with xenobiotic substrates. Two members, AtCXE5 and AtCXE18 were found to be active carboxylesterases, though AtCXE5
proved to be highly unstable as a soluble protein. AtCXE18 and the previously characterised S-formylglutathione hydrolase from Ara-
bidopsis (AtSFGH) were assayed against a series of esters based on methylumbelliferone in which the acyl moiety was varied with respect
to size and conformation. The same series was used to assay crude esterase preparation from Arabidopsis plants and the results com-
pared with those obtained with the commonly used PLE. With straight chain esters, AtCXE18 behaved like PLE, but the Arabidopsis
hydrolases proved less tolerant of branched chain acyl components than the mammalian enzyme. While none of the enzyme preparations
accurately reflected all the activities determined with crude Arabidopsis protein extracts, the plant enzymes proved more useful than PLE
in predicting the hydrolysis of the more sterically constrained esters.
ꢀ 2006 Elsevier Ltd. All rights reserved.
Keywords: Arabidopsis thaliana; Carboxylesterase; Ester series; a/b Hydrolases; Methylumbelliferone
1. Introduction
in mammals and insects these enzymes have attracted
attention due to their roles in drug and insecticide metabo-
lism, respectively (Satoh and Hosokawa, 1998; Oakeshott
et al., 1999). In contrast, the corresponding enzymes in
plants have received far less attention. This is surprising,
given their importance in determining the uptake and bio-
logical activity of important classes of herbicides, fungi-
cides and insecticides (Cummins et al., 2001; Haslam
et al., 2001). For example, insecticides such as the pyre-
throids are detoxified by plant esterases, limiting their bio-
availability in crop protection (Preiss et al., 1988).
Esterases also bioactivate pro-herbicides to their phyto-
toxic alcohols or acids, with the differential rates of hydro-
lysis in different plants contributing to selective weed
control (Cummins and Edwards, 2004; Haslam et al.,
2001). The relationship between chemical structure and
the rates of pesticide hydrolysis in plants would therefore
Serine hydrolases are a diverse group of enzymes, which
catalyse hydrolytic and transacylating reactions with both
natural and synthetic substrates (Heikinheimo et al.,
1999). In the case of drugs and pesticides, hydrolysis radi-
cally alters biological activity and trans-membrane trans-
port and so the respective hydrolases are key proteins in
determining bioavailability and bioactivity (Satoh and
Hosokawa, 1998). In microorganisms, the hydrolases
active toward synthetic amides, carboxyesters and thioes-
ters have been studied in some detail due to their impor-
tance in the biodeterioration of polymers and the
bioremediation of pollutants (Ro et al., 2004). Similarly,
*
Corresponding author. Tel.: +44 191 334 1318; fax: +44 191 334 1201.
E-mail address: Robert.Edwards@durham.ac.uk (R. Edwards).
0031-9422/$ - see front matter ꢀ 2006 Elsevier Ltd. All rights reserved.
doi:10.1016/j.phytochem.2006.12.014

812
I. Cummins et al. / Phytochemistry 68 (2007) 811–818
be a useful tool in agrochemistry for predicting the relative
rates of pesticide detoxification/bioactivation in crops and
weeds. A greater knowledge of these enzymes would also
be useful in determining their largely uncharacterized roles
in plant secondary metabolism.
in a wide variety of plants, suggesting they are of wide-
spread importance (Marshall et al., 2003). The CXE class
of esterases therefore represent a useful example of a novel
plant-specific carboxylesterases with which to compare
activities with PLE.
Currently, porcine liver esterase (PLE) is used as a
model hydrolase to predict the likely rates of cleavage of
synthetic esters in both the pharmaceutical and agrochem-
ical industries (Redinbo and Potter, 2005). While this
model enzyme may usefully reflect the activities of drug-
metabolising hydrolases in animals, its utility in accurately
predicting the activity of esterases in plants has not been
reported. It would therefore be useful to identify ‘model’
esterases in plants and compare their structure–activity
profiles with those of PLE. PLE is a classic serine hydrolase
with a characteristic a/b structural fold and the catalytic
motif G · S · G, associated with this class of protein (Hei-
kinheimo et al., 1999). Informatic analysis based on the
presence of the catalytic motif and predicted a/b structural
modules shows that plants contain multiple classes of ser-
ine hydrolases. After eliminating those classic serine hydro-
lases with known functions in hydrolysing proteins and
carbohydrates, two classes of enzymes with potential roles
in biotransforming small bioactive metabolites have been
identified. These are the hydroxynitrile lyase (hnl)-like
(Wa¨spi et al., 1998) and the hsr203J-like proteins, the latter
originally being identified in tobacco as being linked to the
hypersensitive response to microbial pathogens (Baudouin
et al., 1997). Members of both classes of enzymes have been
shown to hydrolyse xenobiotic esters (Baudouin et al.,
1997; Wa¨spi et al., 1998). The model plant Arabidopsis tha-
liana (At) contains both types of putative hydrolases, with
the hsr203J-like proteins being encoded by a group of 20
genes termed the carboxylesterase (CXE) family (Marshall
et al., 2003). Based on sequence relatedness, the Arabidop-
sis AtCXE family can be arrayed into five clades (Fig. 1).
CXE-like sequences have subsequently been determined
In addition to looking at plant-specific esterases, we
have also been interested in the possibility that the hydro-
lysis of synthetic esters can be conserved in plants and ani-
mals. Thus, S-formylglutathione hydrolase (SFGH), is an
example of a serine hydrolase which is highly conserved
in both prokaryotes and eukaryotes (Kordic et al., 2002;
Haslam et al., 2002), where it performs a key role in C-1
metabolism (Fig. 1). Interestingly, irrespective of their
source SFGHs are also known to be highly active in hydro-
lysing xenobiotic carboxylic esters including methylumbel-
liferyl acetate (Kordic et al., 2002; Cummins et al., 2006).
We have therefore used SFGH from Arabidopsis
(AtSFGH) as an example of a serine hydrolase which will
be found in plants as well as animals. To test their diversity
in substrate usage, we have prepared a custom series of
esters of methylumbelliferone (MU) and used these to test
for carboxylesterase activities in crude plant preparations
from Arabidopsis. The profile of activities has then been
compared with those determined with commercially avail-
able PLE, a representative member of the AtCXE super-
family and AtSFGH.
2. Results
2.1. Cloning and expression of AtCXEs from Arabidopsis
Representatives from five of the six clades of the
AtCXEs were cloned from Arabidopsis foliage by PCR,
namely AtCXE5, AtCXE13, AtCXE16, AtCXE17,
AtCXE18 and AtCXE20 (Fig. 2). The sequences were then
sub-cloned into the pET24 plasmid for the expression of
Formaldehyde + GSH
S -hydroxymethylglutathione
NAD(P)+
S -hydroxymethylglutathione dehydrogenase
NAD(P)H +H+
S-formylglutathione
S -formylglutathione hydrolase
Formate + GSH
Fig. 1. Role of AtSFGH in C-1 metabolism. After the spontaneous reaction of formaldehyde with glutathione to form S-hydroxymethylglutathione, the
conjugate is selectively oxidised to S-formylglutathione, which is then hydrolysed by AtSFGH to release formic acid and glutathione.

I. Cummins et al. / Phytochemistry 68 (2007) 811–818
813
Fig. 2. Dendrogram of the AtCXE family in Arabidopsis as determined by sequence identity/similarity with the values showing the number of bootstrap
replicates (of 100) that concurred at each branch-point. Clade numbers are shown as assigned by Marshall et al. (2003). Family members selected for
cloning and expression are shown in bold.
the respective C-terminal His-tagged proteins in Esche-
richia coli. As determined by SDS–PAGE, recombinant
protein expression was not seen with AtCXE13, AtCXE16
and AtCXE17. With the other three clones, the expressed
recombinant proteins largely accumulated in the insoluble
fraction (Fig. 3). Attempts to enhance the expression of
these proteins in the soluble fraction by growing the
induced bacteria at low temperature proved unsuccessful.
When the respective soluble fractions of these cultures were
analysed for hydrolytic activity toward p-nitrophenyl ace-
tate (pNA), AtCXE5 and AtCXE18 were found to show
high levels of carboxyesterase activity after correcting for
the endogenous activity present in E. coli. No activity
was found in extracts from AtCXE20, even though soluble
protein was detected following SDS–PAGE (Fig. 3). Solu-
ble proteins were purified by Ni-chelate affinity chromatog-
raphy and all found to migrate as 40 kDa polypeptides
when analysed by SDS–PAGE, with a trace of contaminat-
ing higher molecular weight polypeptides (Fig. 3). No carb-
oxylesterase activity could be determined with purified
AtCXE20. When assayed with pNA, activity was deter-
mined with freshly prepared pure AtCXE5 (6.5 lkat mgꢀ1
protein) and AtCXE18 (3.9 lkat mgꢀ1). Both enzymes
were unstable, AtCXE5 being particularly labile with the
respective enzyme activity having a half-life of only
30 min when stored on ice. As determined by gel perme-
ation chromatography, instability was due to aggregation
to high molecular weight inactive complexes (data not
shown). Attempts to stabilize the activity by adding reduc-
ing agents, salts or glycerol all proved ineffective and fur-
ther characterization of AtCXE5 was not attempted.
Instead, studies focussed on AtCXE18 which was analysed
by time-of-flight mass spectrometry after electrospray ioni-
sation and shown to have a parent mass of 38,649 Da.
Based on the predicted mass of 38,754 Da this suggested
that the N-terminal methionine of AtCXE18 had been
cleaved during expression in E. coli.
2.2. Structure activity studies with an MU ester series
AtCXE18, AtSFGH and PLE were subjected to struc-
ture activity studies with a range of MU carboxylesters pre-
pared from the respective acid chlorides (Fig. 4). For
reference, these activities were compared to those deter-
mined with crude protein extracts obtained from Arabid-
opsis plants. With the straight chain acyl ester series,
extension from C2 to C5 (compounds 1–4) was poorly tol-
erated by the carboxylesterases present in the plant
extracts. Interestingly, the hexanoate ester (5) was an excel-
lent substrate for the Arabidopsis esterase preparation,
while further extension of the acyl moiety resulted in a total
loss in activity. The substrate preferences observed with the

814
I. Cummins et al. / Phytochemistry 68 (2007) 811–818
1
2
3
6
7
8
9
11
4
10
5
kDa
45
31
Fig. 3. SDS–PAGE of recombinant AtCXEs showing expression in E. coli with molecular mass reference markers shown in lanes 1, 4 and 8 (lane 2 = pre-
stained Mr standards mixture). AtCXE18, the family member used in the structure activity studies is shown as the purified recombinant enzyme in lane 3.
The other two AtCXEs which were not analysed further are also shown for reference. For AtCXE5, the analysis of the respective total E. coli pellet is
shown in lane 5, the soluble crude protein in lane 6 and the purified enzyme in lane 7. Similarly for AtCXE20 the protein samples are total extract (lane 9),
soluble fraction (lane 10) and purified protein (lane 11).
straight chain ester series in the plant extracts were quite
different from those determined with the individual Arabid-
opsis esterases, which were far more tolerant of the longer
acyl chains. Of the two Arabidopsis enzymes, AtCXE18
behaved most like PLE showing high activities with acyl
components up to C8 (6), albeit with 10-fold lower specific
activities than the mammalian carboxylesterase. With all
enzyme preparation the hexanoate ester (5) proved to be
an excellent substrate.
observed in the effect of the position of the bulky acyl
group along the series such that 12 > 7 < 13 > 14. Finally,
whilst the aryl containing ester substituted 14 was hydroly-
sed by all the preparations, it was a particularly good sub-
strate for AtCXE18.
To examine the reasons for the differences in activity
observed with the MU ester series, kinetic analyses were
performed with compounds 1–14 using AtSFGH as a
test-bed (Table 1). Based on the kcat and Km determina-
tions, it could be seen that binding affinity for the MU
esters was little affected by varying the acyl group. Instead,
varying the structure had a more profound effect on turn-
over. For example, when comparing the simple C5 ester
(4) with the branched C5 (7) and the unsaturated C5
(11), Km values of 50 lM, 80 lM and 60 lM were very sim-
ilar, while kcat/Km values (sꢀ1 mMꢀ1) were 291, 103 and 9,
respectively. Similar detailed kinetic analysis was not con-
ducted with AtCXE18. However, preliminary studies with
outlying members of the MU ester series showed that dif-
ferences in activity were due to altered kcat rather than
Km values, as had been determined with AtSFGH.
The enzymes were then assayed with branched chain
and unsaturated acyl esters (Fig. 4b, compounds 7–11).
As with the longer chain derivatives, with the crude protein
extracts from Arabidopsis, turnover was low or below the
limit of detection. Generally, the introduction of an alkene
(10, 11) resulted in lower turnover, consistent with the gen-
eral reactivity of a,b-unsaturated carbonyl compounds.
The effect of branching identified clear differences in the
three carboxylesterases. Compound 9 was not tolerated
by any of the enzymes, whereas compounds 7 and 8 probed
subtle differences in their respective active sites. Thus,
whereas AtCXE18 was intolerant of both esters, they were
turned over with similar activities by PLE, whereas
AtSFGH only acted on 7. To extend the series of sterically
hindered acyl esters, a set of cyclic containing substrates
moieties were then tested (Fig. 4b, compounds 12–14).
These gave more varied results but some basic trends were
observed. The two cyclopentane esters (12, 13) were hydro-
lysed by the plant extracts at similar rates. This would not
have been predicted from the studies conducted with any of
the enzyme preparations, with compound 12 being a much
better substrate than 13 with PLE, AtCXE18 and
AtSFGH, respectively. Interestingly, given the similar ste-
3. Discussion
Representative members of the AtCXE superfamily
from Arabidopsis were cloned and expressed in E. coli,
with AtCXE5 and AtCXE18, being shown to be active
carboxylesterases when assayed with xenobiotic substrates.
All 20 members of the AtCXE superfamily are classic ser-
ine hydrolases (Marshall et al., 2003), with the potential
of being catalytically active. Therefore, it is quite possible
that the other AtCXEs expressed were also active hydro-
lases, but their activity was not apparent due to their insta-
i
ric volumes (A-values) of an Pr (2.15), cyclopentyl (ꢁ2.1)
and phenyl group (2.7) an alternate odd-even trend is

I. Cummins et al. / Phytochemistry 68 (2007) 811–818
815
CXE
O
RCOOH
H2O
O
O
O
R
O
O
OH
Specific Activity (nkat mg-1 protein)
Compound
#
R
Theoretical
mass
Observed
mass
At Protein
Extract
PLE
AtCXE18 AtSFGH
238 588 12
218 14 399
CH3
1
2
2120 33
4
15
2
2
218.2
C2H5
m/z+1 219.2
m/z+1 233.2
1796 78
9
0.1
232.2
3
4
5
3566 124 225 13 437 11
7
2
0.25
2
246.3
260.3
274.3
m/z+1 247.2
m/z+1 261.2
m/z+1 275.2
4135 228 338 32
330
8
1
8645 241
1374 62
474
105
9
2
604 15
21
6
7
162
230
2
3
0
302.4
260.3
260.3
m/z+1 303.4
m/z+1 261.2
m/z+1 261.2
1989 42
1948 67
8
4
0.2
0.3
2
0.5
8
9
1
0
9
12
2
0
12
34
11
3
0
274.3
244.3
m/z+1 275.2
m/z+1 245.1
10
11
794 13
220 27
45
4
4
1
1
0.35
7
0.5
0.2 0.1
258.3
272.3
m/z+1 259.2
m/z+1 273.2
m/z+1 301.2
12
13
1971 54
158 18
344 28
226
76
1
6
6
1
3
63
2
0
300.4
O
14
988 112
272 34
184
6
3
0.2
324.3
m/z+1 325.1
Fig. 4. (a) The enzymic hydrolysis of the MU ester series. (b) The acyl (R) series used with compound numbers referred to in the text and predicted and
obtained (m/z + 1) parent molecular masses of the pure ester substrates shown. Rates of hydrolysis for each of the enzymes are shown as means of
duplicate determinations with the error bars showing the variation in the replicates.

816
I. Cummins et al. / Phytochemistry 68 (2007) 811–818
Table 1
non-enzymic hydrolysis (data not shown). In the case of
AtSFGH, this result was somewhat surprising, as the
endogenous substrate of this enzyme is the thioester S-
formylglutathione, with the enzyme performing an impor-
tant role in formaldehyde detoxification (Fig. 1). When
the related S-acetylglutathione was assayed as a substrate,
no thiohydrolase activity was determined with AtCXE18 or
PLE (data not shown), whereas this thioester was actively
hydrolysed by AtSFGH. From this it was concluded that
the thioesterase activity of AtSFGH was strictly dependent
on the substrate being a thioester of glutathione alone,
reflecting its conserved essential role in C-1 metabolism
in both prokaryotes and eukaryotes (Fig. 1). In fact, thio-
esters of MU such as the hexanoate derivative actually
acted as competitive inhibitors of carboxylesterase activity
with substrate 1, with a Ki of 0.31 mM being determined.
The results of this study suggest that the acyl binding
pocket of the Arabidopsis esterases studied here is more
constrained than that determined for the large hydropho-
bic pocket of PLE (Provencher and Jones, 1994), In the
case of AtSFGH, we have recently solved the structure of
Kinetic constants determined with SFGH using the MU ester series as
numbered in Fig. 4b
AtSFGH Km (mM)
AtSFGH Kcat/Km (sꢀ1 mMꢀ1
)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
0.08
0.08
0.07
0.05
0.09
0.04
0.08
0.1
0.09
0.08
0.06
0.04
0.05
0.08
283
202
264
291
243
217
103
4
5
19
9
224
68
97
bility as soluble recombinant proteins or their inability to
hydrolyse the limited range of MU ester substrates tested.
AtCXE18, AtSFGH and PLE were incubated with the
MU ester series with the intention of comparing their activ-
ities to one another as well as identifying which enzymes, if
any, could predict the spectrum of carboxylesterases pres-
ent in crude Arabidopsis protein extracts. With the straight
chain acyl esters (compounds 1–6), the pure proteins
proved of limited value in predicting the likely rates of
hydrolysis in the crude plant extracts, with the exception
of highlighting the rapid hydrolysis of the hexanoate ester
(5). Interestingly, the results of the current study with carb-
oxylesterases are quite different from those obtained in ear-
lier studies with MU esters incubated with lipase
preparations from wheat germ, castor bean and peanut,
where the hydrolysis of the acyl moiety was sustained to
C9 and found to be identical to that determined with a
mammalian lipase (Jacks and Kircher, 1967). This confirms
that the Arabidopsis hydrolases selected were indeed CXEs
likely to act on small esters rather than classic lipases which
due to their active site architecture show a preference for
long chain acylated substrates (Grabulkeda et al., 1997).
With the more sterically constrained branched chain and
cyclised esters, the Arabidopsis enzymes did prove to be
more useful in reflecting the restricted activities seen in
the plant extracts. This was particularly apparent with
compounds 7 and 8, which were good substrates of PLE
but poorly tolerated by the plant enzymes. Since PLE
would tend to over-estimate the rates of cleavage of esters
bearing sterically constrained acyl functions, this result
suggests that the incorporation of plant esterases into
screens of carboxylesterase-mediated metabolism of pesti-
cide esters would be potentially useful predicting rates of
hydrolysis in crops and weeds. Importantly, when assayed
with MU-derived substrates, both Arabidopsis enzymes
and PLE were true carboxylesterases. Thus, when incu-
bated with a series of thioesters and amides based on the
same substitutions shown in Fig. 3, negligible activities
with any substrate were determined after correcting for
˚
the protein to a resolution of 1.7 A (Cummins et al.,
2006), but were unable to map the active site of the enzyme
due to difficulties in co-crystallising the protein with xeno-
biotic ester substrates. We are therefore unable to rational-
ise why AtSFGH showed differing activities toward the
ester series due to variations in kcat rather than Km. The
kinetic studies clearly demonstrate the recognition of the
MU-esters is dominated by the selective binding of the
alcohol rather than the acyl moiety. By extending the struc-
ture activity series with the MU esters to determine which
acyl components are tolerated and therefore accommo-
dated, it would be possible to functionally map the dimen-
sions of the acyl binding site of the plant esterases.
Currently, we would predict that the acyl binding pocket
for AtSFGH to be somewhat smaller than the
3
˚
6.2 · 4.7 · 3.1 A determined for PLE (Provencher and
Jones, 1994). A large acyl-binding pocket was functionally
determined with a lipase from Pseudomonas cepacia (Gra-
bulkeda et al., 1997). Our results suggest that such an exer-
cise with the plant enzymes isolated here would be of
limited value, since the structure activity studies would sug-
gest that they do not accurately reflect the hydrolyses of
esters determined in planta. This is in contrast to mamma-
lian tissues, where a single class of esterase can dominate
the hydrolysis of xenobiotics (Rojas-Garbanzo and Mata-
Segreda, 1998). In plants, this appears to be a consequence
of the wide diversity of serine hydrolases recruited to cata-
lyse a much wider diversity of hydrolytic and acyl-transfer
reactions in endogenous secondary metabolism. It is there-
fore more likely that the development of esterases as targets
for selective activation or detoxification of pesticides in
crops or weeds will have to be developed in a species-spe-
cific manner, to reflect the differing ranges of hydrolases
present. To this end, we are currently identifying dominant
esterases in major crops and weeds and characterising their
ability to hydrolyse pesticide and xenobiotic esters. Inter-

I. Cummins et al. / Phytochemistry 68 (2007) 811–818
817
estingly, while we have recently determined major differ-
ences between plant species in their ability to hydrolyse
pesticide esters, our results demonstrate that Arabidopsis
possesses an unusually representative set of hydrolytic
activities toward diverse chemistries (Gershater et al.,
2006).
extracting in 0.1 M Tris–HCl buffer, pH 7.4, containing
5 mM EDTA, 2 mM DTT and polyvinylpolypyrrolidone
(5% w/v). After centrifuging (12,000g, 20 min) the superna-
tant was treated with (NH4)2SO4 to 80% saturation and the
protein pellet collected by re-centrifuging. The protein was
re-dissolved in 0.1 M Tris–HCl pH 7.4 and desalted in the
same buffer by gel filtration prior to assay, at a final protein
concentration of 2.4 mg mlꢀ1. PLE (Sigma, Dorset, UK)
was diluted to 0.2 mg mlꢀ1 protein in assay buffer prior
to use. The purified recombinant Arabidopsis hydrolases
(0.2 mg mlꢀ1) were desalted in 100 mM K.Pi buffer, pH
7.2, prior to use. Hydrolysis of MU esters was measured
continuously using fluorescence-based assays (Kordic
et al., 2002). In each case the reaction rate obtained for
heat-denatured enzyme was subtracted from the enzymatic
rate. Kinetic data were obtained by varying substrate con-
centrations from 20 lM to 500 lM with Km and Vmax val-
ues obtained from double reciprocal plots.
4. Materials and methods
4.1. Synthesis of MU esters
Ester substrates (approximately 50 mg) were prepared
using standard synthetic chemistry procedures. Carboxylic
acids were treated with oxalyl chloride (1.2 equivalents) in
dichloromethane (ꢁ0.2 M solution) containing a catalytic
amount of dimethylformamide. Following concentration
under vacuum, acid chlorides (0.4 mmol) were dissolved
in CH2Cl2 (1 ml) containing 0.8 mmol N-ethyldiisopropyl-
amine (125 ll of stock) and reacted with 4-meth-
ylumbelliferolne (0.32 mmol) added in CH2Cl2 (2 ml).
Similar ratios of 7-mercapto-4-methyl coumarin and 7-
amino-4-methyl coumarin were used for the preparation
of the respective thioesters and amides, respectively. After
16 h at ambient temperature, the reaction mixture was par-
titioned two times with saturated aqueous NaHCO3 (1 v/v)
and the organic phase dried under reduced pressure and
applied to silica extraction columns (3 ml) in Et2O (1 ml).
Columns were eluted with Et2O in 1 ml aliquots which were
assessed for purity by LC–MS. Fractions judged to be
>95% pure were dried under reduced pressure, weighed
and then prepared as stock solutions in acetone for enzyme
assays.
4.4. Cloning and expression of AtCXEs
Members of the AtCXE family of Arabidopsis (Marshall
et al., 2003), were clustered into clades based on sequence
similarity (Fig. 1). Representatives of each of the five major
clades were identified and PCR forward (F) and reverse (R)
primers designed for the amplification of the respective full
coding sequences as follows. AtCXE5 (At1g49660); F = 50
TCATATGGAATCTGAAATCGCCTCC 30 and R = 50
TCTCGAGACCAATAATAAACTCGAC 30. AtCXE13
(At3g48700); F = 50 TCATATGGATTCCGAGATCGCCG
30 and R = 50 TCTCGAGCTTGTCTCCCTTGATA 30.
AtCXE16 (At5g14310); F = 50 TCATATGCCAGGGG-
TAGCTGTGCTC 30 and R = 50 TCTCGAGGTAAGA-
GAACTCATGGCC 30. For AtCXE17 (At5g16080); F = 50
4.2. LC–MS
TCATATGGCAACCATTTCCTTCTCTCAC
30
and
R = 50 TCTCGAGTGATGGATGTATGAAGTTGTG
30. AtCXE18 (At5g23530); F = 50 TCATATGGCGACA-
GATTCTCAACCA 30 and R = 50 TCTCGAGAGCGGA-
GAGCGAAGC 30. AtCXE20 (At5g62180); F = 50
TCATATGTCCGAACCAAGTCCAATC 30 and R = 50
TCTCGAGCAGAACAGAGAATATGAA 30.
Samples of esters were injected onto a C18 column (Phe-
nonemex, Ultrosphere 3 lM, 30 · 2 mm) and eluted at
0.2 ml minꢀ1 with a linear solvent gradient starting with
MeCN:H2O (5:95 v/v) both containing 0.5% (v/v)
HCOOH (5:95 v/v) and ending with 100% MeCN contain-
ing 0.5% (v/v) HCOOH over 10 min. The column eluate
was sequentially analysed using a diode array detector
(k = 200–400 nm) prior to injection into a Micromass
LCT time-of-flight mass spectrometer operating with elec-
trospray ionization (positive ion mode) with a cone voltage
of 25 V and nitrogen used as cone and desolvation gas.
Positive mass ions were detected in the range m/z 100–
1000, with the instrument calibrated using sodium iodide.
The mass ions (m/z + 1) of the MU esters used in the study
are shown in Fig. 3.
Total RNA was isolated from aerial tissues of flowering
Arabidopsis plants using Tri-Reagent (Sigma) and cDNA
prepared using oligo(dT) to prime reverse transcription as
the template for PCR. Amplification and cloning of prod-
ucts into pGEMT (Promega) prior to sub-cloning of the
authenticated sequence into pET 24 (Novagen) was as
described previously (Kordic et al., 2002). Following trans-
formation of E. coli strain ROSETTA DE3 (pLysS)(Nova-
gen), cultures were grown to an OD of 0.6 at 37 ꢁC, then
cooled to 10 ꢁC prior to induction with IPTG (1 mM). Fol-
lowing growth at 10 ꢁC for 16 h, bacteria were pelleted and
then lysed by resuspending in 1/100 original volume B-
PER (Pierce). Recombinant His-tagged proteins were then
purified by nickel-chelate affinity chromatography (Kordic
et al., 2002), with Triton X-100 (0.1% v/v) added to all
chromatographic buffers. Polypeptide composition was
4.3. Esterase assays
For the assay of the crude Arabidopsis proteins, 30-day-
old plants were frozen in liquid nitrogen and then homog-
enized to a powder using a pestle and mortar, prior to

818
I. Cummins et al. / Phytochemistry 68 (2007) 811–818
from wheat that catalyse deesterification of herbicides. Pestic. Bio-
chem. Physiol. 71, 178–189.
assessed by SDS–PAGE (12% gels) with coomassie
staining.
Haslam, R., Rust, S., Pallet, K., Cole, D., Coleman, J., 2002. Cloning and
characterisation of S-formylglutathione hydrolase from Arabidopsis
thaliana: a pathway for formaldehyde detoxification. Plant Physiol.
Biochem. 40, 281–288.
Acknowledgements
Heikinheimo, P., Golman, A., Jeffries, C., Ollis, D.L., 1999. Of barn owls
and bankers: a lush variety of a/b hydrolases. Structure 7, R141–R146.
Jacks, T.J., Kircher, H.W., 1967. Fluorometric assay for the hydrolytic
activity of lipase using fatty acyl esters of 4-methylumbelliferone. Anal.
Biochem. 21, 279–285.
The authors acknowledge the support from One North-
East through the County Durham Sub-Regional Partner-
ship, with Marie Landrum supported by a co-operative
award in science and engineering by Dako Cytomation.
The authors thank Markus Gershater for performing the
phylogenetic analysis of the AtCXEs.
Kordic, S., Cummins, I., Edwards, R., 2002. Cloning and characterization
of a S-formylglutathione hydrolase from Arabidopsis thaliana. Arch.
Biochem. Biophys. 399, 232–238.
Marshall, S.D.G., Putterill, J.J., Plummer, K.M., Newcomb, R.D., 2003.
The carboxylesterase gene family from Arabidopsis thaliana. J. Mol.
Evol. 57, 487–500.
References
Oakeshott, J.G., Claudianos, C., Russell, R.J., Robin, G.C., 1999.
Carboxyl/cholinesterases: a case study of the evolution of a successful
multigene family. BioEssays 21, 1031–1042.
Baudouin, E., Charpenteau, M., Roby, D., Marco, Y., Ranjeva, R.,
Ranty, B., 1997. Functional expression of a tobacco gene related to the
reine hydrolase family: esterase activity towards short-chain dinitro-
phenyl acylesters. Eur. J. Biochem. 248, 700–706.
Preiss, U., Wallno¨fer, P.R., Engelhardt, G., 1988. Partial purification and
properties of an esterase from tomato cell suspension cultures
hydrolysing the pyrethroid insecticide cyfluthrin. Pestic. Sci. 23, 13–24.
Provencher, L., Jones, J.B., 1994. A concluding specification of the
dimensions of the active site model of pig liver esterase. J. Org. Chem.
59, 2729–2732.
Cummins, I., Edwards, R., 2004. Purification and cloning of an esterase
from the weed black-grass (Alopecurus myosuroides), which bioacti-
vates aryloxyphenoxypropionate herbicides. Plant J. 39, 894–904.
Cummins, I., Burnet, M., Edwards, R., 2001. Biochemical characteriza-
tion of esterases active in hydrolysing xenobiotics in wheat and
competing weeds. Physiol. Plant. 113, 477–485.
Redinbo, M.R., Potter, P.M., 2005. Mammalian carboxylesterases: from
drug targets to protein therapeutics. Drug Discovery Today 10, 313–
325.
Ro, H.-S., Hong, H.P., Kho, B.H., Kim, S., Chung, B.H., 2004. Genome-
wide cloning and characterisation of microbial esterases. FEMS
Microbiol. Lett. 233, 97–105.
Cummins, I., McAuley, K.M., Fordham-Skelton, A., Schwoerer, R., Steel,
P.G., Davis, B.G., Edwards, R., 2006. Unique regulation of the active
site of the serine esterase S-formylglutathione hydrolase. J. Mol. Biol.
359, 422–432.
Rojas-Garbanzo, R.E., Mata-Segreda, J.F., 1998. Assessment of the steric
tolerance of the P sector in the catalytic site of porcine liver esterase.
Biorg. Med. Chem. Lett. 8, 7–10.
Gershater, M., Sharples, K., Edwards, R., 2006. Carboxylesterase
activities toward pesticide esters in crops and weeds. Phytochemistry
67, 2561–2567.
Satoh, T., Hosokawa, M., 1998. The mammalian carboxylesterases: from
molecules to function. Annu. Rev. Pharmacol. Toxicol. 38, 257–288.
Wa¨spi, U., Misteli, B., Hasslacher, M., Jandrositz, A., Kohlwein, S.D.,
Schwab, H., Dudler, R., 1998. The defense-related rice gene Pir7b
encodes an a/b hydrolase fold protein exhibiting esterase activity
towards naphthol AS-esters. Eur. J. Biochem. 254, 32–37.
Grabulkeda, X., Jaime, C., Guerrero, A., 1997. Estimation of the lipase PS
(Pseudomonas cepacia) active site dimensions based on molecular
mechanics calculations. Tetrahedron: Assymetry 8, 3675–3683.
Haslam, R., Raveton, M., Cole, D.J., Pallett, K.E., Coleman, J.O.D.,
2001. The identification and properties of apoplastic carboxylesterases