Forchlorfenuron-mimicking haptens: From immunogen design to antibody characterization by hierarchical clustering analysis

Article abstract of DOI:10.1039/c1ob05190c

To obtain highly-specific and selective forchlorfenuron binders, a collection of functionalized derivatives with different spacer arm locations and lengths was prepared. By immunization with target-mimicking haptens, a large battery of monoclonal and poly

Full text of DOI:10.1039/c1ob05190c

View Article Online / Journal Homepage / Table of Contents for this issue
Organic &
Biomolecular
Chemistry
Dynamic Article Links
Cite this: Org. Biomol. Chem., 2011, 9, 4863
www.rsc.org/obc
PAPER
Forchlorfenuron-mimicking haptens: from immunogen design to antibody
characterization by hierarchical clustering analysis†‡
Celia Sua´rez-Pantaleo´n,a Josep V. Mercader,a Consuelo Agullo´,b
Antonio Abad-Somovillab and Antonio Abad-Fuentes*a
Received 3rd February 2011, Accepted 24th March 2011
DOI: 10.1039/c1ob05190c
To obtain highly-specific and selective forchlorfenuron binders, a collection of functionalized
derivatives with different spacer arm locations and lengths was prepared. By immunization with
target-mimicking haptens, a large battery of monoclonal and polyclonal antibodies against this
synthetic cell regulator was produced and exhaustively characterized in two immunoassay formats
using homologous and heterologous conjugates. Antibodies with IC50 values lower than 0.3 nM were
successfully raised from the prepared immunogens, thus evidencing the efficacy of the explored
strategies. In order to identify significant epitopes in the antibody–antigen interaction, a series of new
chemical forchlorfenuron analogues, with slight modifications at both rings of the target molecule, were
synthesized and evaluated in competitive assays. As a novel approach in hapten recognition studies,
data processing was performed by computational classification methods based on hierarchical
clustering. This strategy was shown to be highly valuable for a straightforward profiling of antibodies
according to analogue recognition patterns. A relationship could be established between the antigen
binding properties of antibodies and the structure of the immunogen. Whereas antibodies with
equivalent affinities had been obtained from all of the derivatives, their specificity was found to be
largely influenced by the differential exposition of the molecule to the immune system.
composition, and length of the spacer arm are critical aspects
Introduction
in hapten design.2 Particularly, the derivatization site within the
Small organic molecules need to be covalently coupled to immuno-
immunizing hapten, and thus the differential presentation of the
genic carriers in order to obtain antibodies. Therefore, derivatives
immunogen to the immune system, is known to have a strong effect
on antibody properties, especially on their specificity.3 During the
last years, computer-assisted molecular modelling techniques have
been introduced and structure–activity relationship studies have
been carried out to provide useful information with regard to
immunogen design.4 However, prediction of the structure of the
perfect immunizing hapten still remains a challenge. Consequently,
for the generation of high-affinity and selective antibodies, it is
strongly recommended to prepare an assortment of compounds
mimicking the analyte of interest and containing a functional
linker at different positions, thus facilitating the presentation of
the antigen to the immune system from a variety of perspectives. In
addition, heterologous haptens in assay conjugates may contribute
to increase the sensitivity of analytical immunoassays.
Forchlorfenuron is a synthetic chloropyridyl phenylurea, also
named FCF or CPPU, which has been shown to exert varied
biological activities. Structure–activity relationship studies on
substituted ureas have revealed that compounds with an intact
NH–CO–NH bridge, a phenyl ring at the N-position, and
a substituted heterocyclic ring at the N¢-position significantly
enhance cell-promoting function in plants.5 In fact, CPPU displays
an outstanding phytohormonal activity,6 exceeding that of the
most active natural and synthetic adenine-substituted cytokinins,
incorporating a functional chemical group and a spacer arm ought
to be synthesized for most substances. A successful production
of antibodies mostly relies on the structure of the prepared
immunogen,1 and hence the immunizing hapten should resemble
the target compound almost perfectly. It is well established that a
certain immunizing derivative should preserve as much as possible
the physicochemical characteristics of the target molecule (steric,
electronic, and geometric properties), and therefore the position,
aDepartment of Biotechnology, IATA-CSIC, Agust´ı Escardino 7, 46980, Pa-
terna, Vale`ncia, Spain. E-mail: aabad@iata.csic.es; Fax: +34-963636301;
Tel: +34-963636301
bDepartment of Organic Chemistry, Universitat de Vale`ncia, Doctor Mo-
liner 50, 46100, Burjassot, Vale`ncia, Spain. E-mail: antonio.abad@uv.es;
Fax: +34-963544328; Tel: +34-963544509
† Limited amounts of the newly described chemicals and immunoreagents
are available upon request for evaluation.
‡ Electronic supplementary information (ESI) available: Full characteri-
zation data of haptens m2 and s3 and their intermediates, syntheses and
spectral data of haptens 4Fs5 and 4Ms5 and those of analogues thio-
CPPU and BzClPyA and their intermediates, details about the preparation
and characterization of protein conjugates, the generation of monoclonal
antibodies, tables for the standard curve parameters and antibody-distance
calculations, and copies of 1H NMR, 13C NMR, and mass spectra of
selected compounds. See DOI: 10.1039/c1ob05190c
This journal is
The Royal Society of Chemistry 2011
Org. Biomol. Chem., 2011, 9, 4863–4872 | 4863
©

View Article Online
Table 1 Chemical structure of CPPU and synthetic haptens
which has markedly determined its notable success as an agro-
chemical. Additionally, CPPU alters, in a specific and reversible
manner, the organization of septins, a family of highly conserved
GTPase proteins involved in diverse physiological events including
cytoskeleton organization, cell division, and morphogenesis in
fungi and animals.7 These findings have pointed out CPPU as
a very valuable candidate to help scientists elucidate the intricate
mode of action and function of septins, as well as to increase their
knowledge about disorders associated to defects in the regulation
of those proteins, like cancer, Parkinson, or Alzheimer diseases.
We have previously described the syntheses of functionalized
derivatives with different linker lengths or tethering sites and the
generation of monoclonal and polyclonal antibodies (mAbs and
pAbs, respectively) against CPPU.8 In the present study, new hap-
tens have been synthesized to screen novel derivatization positions
in the immunizing hapten that may influence the properties of
the produced antibodies. In the end, a large library of haptens,
conjugates, mAbs, and pAbs for CPPU has been constructed in
our lab and it is presented herein. In order to characterize the
affinity of the available antibodies and to rationally select the
best immunoreagents, competitive enzyme-linked immunosorbent
assays (cELISA) in the antibody-coated direct (d-cELISA) format
and in the conjugate-coated indirect (i-cELISA) format have
been employed. An extensive study was conducted using several
heterologous competitors, including conjugates with site, length,
and structure heterologies. Next, molecular epitope mapping was
carried out by scanning the antibody recognition towards a
collection of CPPU-analogous compounds that were specifically
designed and synthesized with the aim to establish a correlation
between antibody specificity and the structure of the employed
immunogen. Moreover, an innovative approach to the study
of the binding profiles of the antibodies was applied using an
agglomerative hierarchical clustering methodology, which was
revealed as a very suitable tool for the analysis of the diversity
of a miscellaneous collection of antibodies.
Hapten
R1
R2
R3
X
p2a
CH2CO2H
H
Cl
N
N
C
N
N
N
N
N
N
N
p6
(CH2)5CO2H
H
Cl
CldPhUp6
PhPyUp6
m2
m6
s3
s5
(CH2)5CO2H
(CH2)5CO2H
H
H
H
H
F
H
H
Cl
H
Cl
Cl
CH2CO2H
(CH2)5CO2H
H
H
H
H
S(CH2)2CO2H
S(CH2)4CO2H
S(CH2)4CO2H
S(CH2)4CO2H
4Fs5
4Ms5
Me
a Haptens used for immunogen preparation are highlighted in bold-type
letter.
pared with the spacer arm located at the phenyl ring, more
precisely, at the para and the meta positions (p-type and m-type
haptens, respectively), and at the C-2 position of the pyridyl ring
through substitution of the chlorine atom by a thioalkyl chain
(s-type haptens). The nucleophilic substitution of a chlorine atom
by a carboxyalkylthiol group in immunizing haptens has been
successfully employed before by other authors for the production
of antibodies against molecules containing this halogen.3a,10
For the preparation of new immunizing and assay competitor
haptens (m2, s3, 4Fs5, and 4Ms5), we followed the general
synthetic approach commonly used for this type of compound,
which is based on the formation of the ureido moiety via reaction
of a conveniently functionalized aryl- or pyridyl-amine with the
appropriate isocyanate. Thus, hapten m2 (4, Scheme 1A) was
efficiently obtained by reaction of commercially available 2-(3-
aminophenyl)acetic acid (1) with 2-chloro-4-isocyanatopyridine
(3), which was prepared in situ, due to its high instability, in a one-
pot reaction by Curtius rearrangement of 2-chloroisonicotinoyl
azide (2).8 On the other hand, the synthesis of hapten s3 (11,
Scheme 1B) required the previous preparation of the aminopy-
ridine 8, which was initiated by the reaction of 2-chloro-4-
nitropyridine N-oxide (5)12 with commercially available methyl
3-mercaptopropanoate (6) in the presence of Et3N as basic catalyst
at reflux of benzene.11 The nucleophilic aromatic substitution
reaction of the chlorine atom by the (3-methoxy-3-oxopropyl)thio
group to give compound 7 took place with moderate yield (54%)
due, at least in part, to the competitive substitution reaction of
the C-4 nitro group. Simultaneous reduction of the nitro and N-
oxide moieties of this product with metallic iron afforded the
required aminopyridine 8. Finally, the synthesis of hapten s3
was readily completed in high yield by reaction of 8 with phenyl
isocyanate followed by basic hydrolysis of the ester moiety of the
initially formed methyl ester 10. The synthesis of assay competitor
haptens (4Fs5 and 4Ms5) was effected in a similar way by reac-
tion of methyl 5-((4-aminopyridin-2-yl)thio)pentanoate8b with 4-
fluoroisocyanatobenzene and 1-isocyanato-4-methylbenzene, re-
spectively, as described in the ESI‡ (Scheme S1).
Results and discussion
Hapten synthesis and conjugate preparation
Immunogenic conjugates should allow a maximization of the
exposition of the main antigenic determinants that are present
in the hapten, as they presumably will rule the affinity of the
generated antibodies based on their contribution to the molecular
forces driving the antibody–antigen interaction. CPPU is a rather
symmetric molecule which is comprised of two aromatic systems
(phenyl and chloropyridyl rings) linked by a polar urea bridge
(Table 1). Concerning the chemical nature of the molecule, the
phenyl ring may contribute to hydrophobic interactions with
the active site of antibodies, whereas the more polarized 2-
chloropyridyl group may get involved in electrostatic interactions.
Additionally, the azomethynic nitrogen atom in the pyridyl ring
and the ureido group make possible the formation of hydrogen
bonds. Having all those considerations in mind, functionalized
derivatives of CPPU with different derivatization sites and linker
lengths were designed (Table 1). The ureido bridge was maintained
intact in all of the derivatives, given the relevant participation
generally attributed to this moiety in the molecular recognition
towards phenylurea compounds.9 Immunizing haptens were pre-
Haptens were activated and coupled to bovine serum albumin
(BSA) by the active ester method, and to ovalbumin (OVA) and
4864 | Org. Biomol. Chem., 2011, 9, 4863–4872
This journal is
The Royal Society of Chemistry 2011
©

View Article Online
Scheme 1 Preparation of immunizing haptens m2 and s3.
horseradish peroxidase (HRP) by the mixed anhydride method
to afford the immunogens and assay conjugates, respectively
(see the ESI‡). Coupling of hapten s3 to OVA and HRP by
the mixed anhydride method resulted in a poorly recognized
conjugate. Attempts to isolate the anhydride were unsuccessful,
and therefore s3 was conjugated to OVA and HRP by the active
ester method. The obtained hapten-to-protein molar ratios of all
available conjugates are listed in Table S1.‡ These values were
consistent with the number of free e-amines available in the
carriers.
absorbance (Amax) if the background signal approaches to zero.
As a foremost result, an IC50 value below 1.0 nM was displayed
by 70% of the mAbs in the i-cELISA format, and by 38%
of the mAbs in the d-cELISA format. High-affinity antibodies
could be produced from all immunogens, independently of the
derivatization site and linker length of the hapten. Based on this
study, 4 mAbs from each immunizing conjugate were selected.
With the aim of assessing the effect of hapten heterology on
the apparent affinity of the generated antibodies, heterologous
conjugates were also employed in both cELISA formats. For
each combination of immunoreagents different concentrations
were tested, so a collection of inhibition curves with different
Amax values was obtained. Table 2 summarizes this study for the
selected 24 mAbs and the 6 available rabbit pAbs. The diversity
of the prepared haptens allowed us to find some heterologous
combinations which could lower the IC50 values. Green cells in
Table 2 indicate the heterologous combinations that provided IC50
values lower than the homologous one. Around two thirds of the
mAbs clearly improved their binding to CPPU using heterologous
conjugates.
Although it was difficult to draw solid conclusions about the
type of heterology that exerted a higher improvement on the
IC50 values of antibodies, some general trends were observed.
First, linker-length heterology was not shown as a very effective
strategy. Second, structural heterology resulted in rare improve-
ments on assay sensitivity; conjugates based on haptens 4Fs5
and 4Ms5, with modifications at the phenyl ring, were better
recognized and performed better than those employing haptens
with changes at the chloropyridyl ring, i.e. haptens CldPhUp6
and PhPyUp6. Third, site-heterology was revealed as the most
successful approach, as the vast majority of the combinations that
afforded improved CPPU recognition were based on haptens with
the spacer arm at a position different to that of the immunizing
hapten. Fourth, enzyme tracers with short spacer arms were mostly
not recognized; this experimental observation has been frequently
documented for other antigens.13 Fifth, it was found that the
response of every particular antibody was not parallel in both
Animal immunization and monoclonal antibody production
In previous studies, mAbs against CPPU had been prepared
using haptens with the spacer arm at the para position of
the phenylurea ring and with different lengths (p2 and p6) as
immunizing haptens.8a To complete our collection of mAbs, mice
were immunized with conjugates of m-type and s-type haptens
(BSA-m2, BSA-m6, BSA-s3, and BSA-s5). A summary of the
mAb-generation process is reported in Table S2 of the ESI.‡ To a
set of 17 mAbs previously generated using haptens p2 and p6, 21
novel antibodies were added: 4 cell lines from mice immunized with
BSA-m2, 4 from BSA-m6, 6 from BSA-s3, and 7 from BSA-s5. All
of the produced mAbs were of the IgG1 isotype (only mAb m6#43
was of the IgG2b isotype) with a k-type light chain. Additionally, 6
pAbs had been generated from rabbits immunized with the three
site-heterologous haptens containing the longer linker (p6, m6,
and s5).8
Structure–activity relationships
Evaluation of antibody affinity. A preliminary characterization
of a set of 38 mAbs was performed in two cELISA formats using
coating conjugates or enzyme tracers carrying the homologous
hapten, that is, the same synthetic derivative used for immu-
nization. Antibody affinity was estimated from the concentration
of analyte at the inflection point of the fitted inhibition curve,
typically corresponding to a 50% reduction (IC50) of the maximum
This journal is
The Royal Society of Chemistry 2011
Org. Biomol. Chem., 2011, 9, 4863–4872 | 4865
©

View Article Online
Table 2 Influence of heterologous conjugates on antibody IC50 values (nM)a
assay formats. As it generally happens,14 the d-cELISA format
showed a more restrictive recognition pattern towards assay
conjugates than the i-cELISA format. Finally, the behaviour of the
pAbs was equivalent to that shown by the mAbs, although pAbs
generally displayed a broader recognition profile of heterologous
conjugates. This result may be attributed to the fact that the
pAbs represent the raw response of the animal immune system,
and therefore they comprise immunoglobulins with heterogeneous
properties.
Fig. 1 shows a representative inhibition curve for one mAb
from each functionalization site in both cELISA formats, using
as selection criteria the apparent affinity of the antibody and the
slope of the inhibition curve. Those curves had Amax values around
1.0 and showed IC50 values below 0.3 nM and slopes around -1.2.
A detailed list of the parameters of the depicted curves is provided
in Table S3 of the ESI.‡
Binding studies by hierarchical clustering analysis
The evaluation of antibody–antigen interactions has opened
the door to promising biomedical applications in the diagnosis,
prognosis, and therapeutic fields, or even in vaccine development
research.15 Assessment of antibody specificity is traditionally
conducted by cross-reactivity studies, where a set of molecules
structurally related to the target analyte are evaluated as com-
petitors. By studying the analogue recognition (AR) pattern of
antibodies, the major antigenic determinants within the target
molecule can be disclosed. Nevertheless, the heterogeneity some-
4866 | Org. Biomol. Chem., 2011, 9, 4863–4872
This journal is
The Royal Society of Chemistry 2011
©

View Article Online
Fig. 1 CPPU inhibition curves of the selected antibodies in two cELISA formats. Values are the average of three independent experiments.
Table 3 Chemical structures of the CPPU analogues used as competitors
in the AR studies
times displayed by antibodies can make it difficult to establish a
relationship between the structure of the immunizing hapten and
the binding properties of the antibodies, just by direct and non-
assisted evaluation of the competitive experiments. To overcome
this limitation, we have performed a wide analysis of the AR
pattern for a large panel of antibodies and we have applied
computer-assisted clustering methodologies in order to classify the
immunoglobulin repertoire. These informatics tools, commonly
employed in genomic and transcriptomic analysis, have been
proven as very helpful instruments to correlate and to assist the
comprehension of large sets of data from proteomic studies16
and they have also been applied to profiling antibody binding
properties to proteins or peptides.17 Recently, Pattathil et al.18
published the application of this agglomerative method in order to
group mAbs, based on plant cell wall polysaccharide recognition
patterns. In the present study, hierarchical clustering analysis has
been applied for the first time to the evaluation of the specificity
of antibodies generated against a small organic molecule.
To determine the binding properties of 38 mAbs and 6 rabbit
pAbs that were produced against CPPU, a study was carried out
using a group of synthetic compounds (see the Experimental
section) which are chemically related to CPPU (see the list
and their abbreviations in Table 3). These analogues could
be classified into three categories depending on the chemical
modification under consideration: first, chemicals with changes
at the phenyl ring (4FPhClPyU, 4MePhClPyU, F5PhClPyU, and
CldPyU); second, analogues with changes at the chloropyridyl ring
(PhBrPyU, PhdClPyU, PhPyU, dPhU, NdPhU, and CldPhU);
and third, those with modifications at the urea bridge (thio-
CPPU and BzClPyA). Additionally, the recognition of herbicide
thidiazuron (TDZ), which presents a thiadiazole ring instead of
the characteristic chloropyridyl ring of CPPU, was also assessed.
From the data matrix of AR values of those compounds, a
specificity heat map was generated as illustrated in Fig. 2A. Also
a dendrogram was created (see the left of the heat map) to easily
classify the antibodies according to their specificity profiles by
implementing a hierarchical clustering algorithm19 using the MeV
software (see the Experimental section). In that classification tree,
the antibodies represented in closer positions are mathematically
the most similar, and therefore the distance between clusters
(groups of antibodies) can be correlated to the similarity of the
binding profiles. As it can be observed, two well differentiated
groups of antibodies had been generated by immunization with
six CPPU derivatives. The first cluster, called A (light green bar),
included only antibodies obtained from immunizing haptens that
had been functionalized at the phenyl ring (p2, p6, m2, and m6),
whereas the second cluster, called B (pink bar), was composed
mainly of antibodies obtained from haptens derivatized at the
chloropyridyl ring (s3 and s5). Antibodies belonging to cluster
A seemed to be more permissive to changes in the structure
This journal is
The Royal Society of Chemistry 2011
Org. Biomol. Chem., 2011, 9, 4863–4872 | 4867
©

View Article Online
Fig. 2 Analysis of the binding properties of the antibodies to a set of CPPU analogous compounds. A) Heat map visualization of the logarithm of AR
data matrix and the generated dendrogram using the hierarchical Kendall’s Tau similarity metric and complete linkage as the aggregation method. Black
represents 100% binding. Intense green means that AR was much over 100%, whereas intense red means that binding was much lower than 100%. B)
AR profiles of representative antibodies of the main clusters of the dendrogram. C) Correlation between the original immunizing hapten and antibody
distribution within the clusters.
of the phenyl ring of the competing analogues (4FPhClPyU,
4MePhClPyU, F5PhClPyU, and CldPyU) than antibodies of
cluster B. A clear example of the differential response between both
groups was observed for compound CldPyU. Most antibodies of
cluster A were able to recognize this chemical to a similar extent
or even better than CPPU, whereas for antibodies of cluster B,
the distal substitution of the phenyl ring by a pyridyl ring (as
in CldPyU) caused a noticeable disturbance in the recognition.
On the contrary, lower AR values towards compounds with
modifications at the chloropyridyl ring (PhBrPyU, PhdClPyU,
PhPyU, dPhU, NdPhU, and CldPhU) were generally found for
antibodies in cluster A than for antibodies in cluster B. This trend
was especially evident for compounds NdPhU and CldPhU, which
were characterized by the substitution of the 2-chloropyridyl ring
by a nitrophenyl and chlorophenyl ring, respectively (see Table 3).
This observation highlighted the important role of the nitrogen
atom of the pyridine ring in the interaction of the antibodies with
CPPU.
the derivatization site were better tolerated. The classification of
the antibodies based on hierarchical clustering analysis clearly
evidenced that the position of the functionalized arm within the
immunizing hapten, and hence the variation in the displayed
orientation of the CPPU framework, had a significant influence
on antibody specificity.
Both clusters A and B could be further subdivided into two
branches, called A1 and A2, and B1 and B2, respectively. To
better visualize differences among clusters, the AR profile of a
representative mAb from each group has been depicted in Fig. 2B.
The main difference among antibodies of branches A1 (dark red
line) and A2 (orange line) was the recognition of compounds
F5PhClPyU and PhdClPyU. Whereas antibodies of branch A1
showed a certain tolerance towards the presence of five fluorine
atoms at the phenyl ring (F5PhClPyU), antibodies of branch
A2 were clearly more susceptible to this electronic modification,
following a more similar behavior to some antibodies in cluster
B. In contrast, the presence of two chlorine atoms at the pyridyl
ring (PhdClPyU) was quite well tolerated by most antibodies of
branch A2. Regarding cluster B, antibodies comprising branch
B1 (blue line) presented a narrower binding profile than those
included in branch B2 (dark green line), as evidenced by the lower
Altogether, these findings were in agreement with the Land-
steiner’s principle,20 as antibodies showed a higher interaction
with those moieties distally located from the attachment site of
the spacer arm. In contrast, modifications at positions closer to
4868 | Org. Biomol. Chem., 2011, 9, 4863–4872
This journal is
The Royal Society of Chemistry 2011
©

View Article Online
AR values of the former antibodies, thus being more sensitive
to modifications in the CPPU structure. Both Rp6-type and
both Rm6-type rabbit pAbs shared their binding profiles with
antibodies of cluster A (branch A1), whereas both Rs5-type rabbit
pAbs were included into cluster B (branch B1), as would be
expected attending to the hapten from which they were derived.
Some other significant findings with respect to relevant epitopes
in the target molecule were ascertained from the analysis of
the binding heat map (Fig. 2A). First, PhBrPyU was perfectly
recognized by all of the antibodies, independently of their origin.
On the contrary, a general loss of binding was observed for
compound PhPyU, that is, the analogue lacking the chlorine
atom (see Table 3). These results suggested that the presence
of an electron-withdrawing element at the C-2 position of the
pyridyl ring was highly important, chlorine and bromine atoms
being perfectly exchangeable, as also observed by other authors
with different targets.21 The significance of the presence of a
polarized area in the antigen was also supported by the high
recognition displayed by some antibodies towards analogue
PhdClPyU. Similar results were found by Schneider et al.22 for
antibodies produced against the phenylurea herbicide monuron.
Accordingly, the modification of the urea bridge in phenylurea
compounds has often been found to cause a disturbance in
the antibody recognition.23 The strong contribution of the urea
bridge to the antibody–antigen interaction was revealed by a
general decrease in the recognition of the thiourea analogue (thio-
CPPU), and especially by an almost undetectable binding to
the amide analogue (BzClPyA). The lower recognition of the
structurally similar thio-CPPU can be probably attributed to
the different hydrogen bond pattern showed by the ureido and
thioureido moieties due to the higher acidity of the thioureidic
hydrogens and, particularly, the much weaker hydrogen-bond
acceptor capability of the thiocarbonyl group.24 Concerning TDZ
recognition, a highly diverse binding was observed; some mAbs,
such as p6#23, showed low AR values, whereas other antibodies
such as s5#12 presented recognition over 100%. The medium-
to-high binding pattern displayed by most of the mAbs to TDZ
represents a clear indication of the bioisosterism25 that occurs
between the chloropyridine and the thiadiazole moieties. Finally,
chemicals with just one of the aromatic rings such as 4-amino-
2-chloropyridine and N-phenylurea, or with no ring at all such
as N,N¢-dimethylurea, were not recognized, thus indicating the
contribution of the whole molecule to a tight interaction with the
antibody binding site.
substrate interaction probably by the formation of hydrogen
bonding between the hydrogen atoms in the urea bridge and a
neighbouring aspartic acid residue within the active site.
As observed in Fig. 2C, no disparity on antibody specificity was
found regarding the derivatization positions, para or meta, at the
phenyl ring, as antibodies obtained from those immunogens were
similarly distributed within cluster A, with few antibodies included
in cluster B. On the other hand, s-type antibodies (coming from
haptens derivatized at the pyridyl ring) were all included in cluster
B and evenly distributed among both subgroups. With respect to
the spacer arm, it was observed that the assayed linker lengths
generated equivalent specificity responses.
As a final approach, the whole collection of antibodies was used
to create a heat map representing the antibody distance (Fig. 3).
The minimum and maximum Kendall’s tau (t) correlation coef-
ficients extracted from the comparison between the four clusters,
and the corresponding pairs of antibodies are listed in Table S4
in the ESI.‡ A higher similarity was found among antibodies
included in cluster A, especially in branch A2 (tmin = 0.60; tmax
=
0.98), than among antibodies in cluster B (tmin = 0.01; tmax = 0.80
for branch B1). Therefore, it could be concluded that immunogens
of the p-type and m-type generated a more homogeneous immune
response in terms of specificity than immunogens of the s-type.
Based on the results herein presented, it could be hypothesized
that the interaction between the generated antibodies and CPPU
could be in the first instance mainly driven by electrostatic
forces, being in this sense the polarization associated to the
2-chloropyridyl ring highly determinant. Then, the antibody–
antigen complex may be strengthened by formation of hydrogen
bonds involving the ureido group and/or the nitrogen atom of the
chloropyridyl ring. These results would fit with a cave-like binding
pocket characteristic of hapten–antibody interactions.26 Recently,
Kopecˇny´ et al.27 resolved the crystal structure of CPPU in a com-
plex with enzyme ZmCKO1, a cytokinin oxidase/dehydrogenase
of Zea mays which becomes inactivated upon binding with this
synthetic phytoregulator. In the most feasible conformation, the
2-chloropyridyl ring is internally oriented to interact with the
enzyme cofactor, thus allowing the stabilization of the enzyme–
Fig. 3 Heat map visualization of the antibody-distance matrix obtained
by implementation of Kendall’s tau similarity metric. The measures are
represented in a scale ranging from 0 (highest similarity, blue) to 1 (lowest
similarity, yellow), indicating perfect agreement and disagreement for every
pair of antibodies, respectively.
Conclusions
Monoclonal antibodies with peculiar or somehow deviated be-
haviours from the average response are occasionally generated.28
In this study, the high number of available antibodies allowed
us to establish a correlation between immunogen structure and
This journal is
The Royal Society of Chemistry 2011
Org. Biomol. Chem., 2011, 9, 4863–4872 | 4869
©

View Article Online
antibody specificity. Good molecular mimicking was achieved by
all immunizing derivatives as reflected by the high affinity of the
generated antibodies. For sensitivity improvement, modification
of the functionalization site of the assay competitor was demon-
strated as the most successful approach. Besides, the applied
strategy using chemical analogues was especially useful for gaining
some insights into the most significant epitopes involved in the
antibody–antigen interaction. Also, the potential of computer-
assisted classification techniques as complementary tools for the
comprehension of the diversity of the generated immune response
towards a small chemical molecule has been demonstrated. The
recently discovered influence of CPPU in septin organization could
make the synthesized chemicals and the developed antibodies
herein described very useful reagents in basic research about this
important family of proteins.
anhydrous toluene (1.2 mL) at reflux under an inert atmosphere.
The reaction mixture was stirred for 1.5 h at reflux and then
cooled to rt. The solid was collected by filtration and washed with
toluene and ethyl ether, and dried under vacuum to give pure
2-(3-(3-(2-chloropyridin-4-yl)ureido)phenyl)acetic acid, hapten
m2 (4, 200 mg, 85%) as a white solid: mp 197–202 ◦C (from
toluene).
Synthesis of hapten s3
2-(3-Methoxy-3-oxopropylthio)-4-nitropyridine N-oxide (7).
A
solution of the commercial methyl 3-mercaptopropionate (6,
904 mL, 8.35 mmol) and Et3N (1.16 mL, 8.35 mmol) in dry
benzene (10 mL) was dropwise added to a stirred suspension
of 2-chloro-4-nitropyridine N-oxide (5, 1.18 g, 6.76 mmol)29 in
benzene (23 mL) at reflux under argon. The reaction mixture was
refluxed overnight and then concentrated at reduced pressure. The
obtained residue was dissolved in CH2Cl2 and washed with water
and brine. The organic layer was then dried over Na2SO4 and
concentrated under vacuum. The residue obtained was purified
by silica gel chromatography, using hexane–EtOAc 1 : 1 as eluent,
affording the methyl ester 7 (938 mg, 54%) as a yellow solid: mp
112–114 ◦C (from benzene).
Experimental
Analytical-grade standard forchlorfenuron [1-(2-chloro-4-
pyridyl)-3-phenylurea, CAS Registry No. 68157-60-8, MW
247.7
g
mol-1] was purchased from Sigma–Aldrich–Fluka
(Madrid, Spain). Chemical reactions were monitored by thin-
layer chromatography with 0.25 mm precoated silica gel plates.
The synthesized chemicals were purified by flash column
chromatography on silica gel 60 (particle size = 40–63 mm).
Melting points were obtained in a Kofler hot-stage apparatus or
a Bu¨chi melting point apparatus and are uncorrected.
Carrier proteins, adjuvants, fetal bovine serum (FBS), cul-
ture media, and other biochemical reagents were from Sigma–
Aldrich (Madrid, Spain) or Roche Applied Science (Mannheim,
Germany). Purification of protein–hapten conjugates and an-
tibodies was carried out in Sephadex G-25 HiTrap Desalting
columns and HiTrap Protein G HP columns, respectively, ac-
quired from GE Healthcare (Uppsala, Sweden). P3-X63-Ag8.653
mouse plasmacytoma cell line was obtained from the European
Collection of Cell Cultures (Wiltshire, UK). Polyclonal goat anti-
rabbit immunoglobulin peroxidase conjugate (GAR-HRP) was
purchased from Bio-Rad (Madrid, Spain) and polyclonal rabbit
anti-mouse immunoglobulin peroxidase conjugate (RAM-HRP)
was from Dako (Glostrup, Denmark).
Methyl 3-((4-aminopyridin-2-yl)thio)propanoate (8). Iron pow-
der (1.00 g, 17.90 mmol) and NH4Cl (448 mg, 8.37 mmol) were
added to a refluxed solution of compound 7 (819 mg, 3.17 mmol)
in a 4 : 1 mixture of EtOH–H2O (25 mL). The reaction was stirred
vigorously at reflux for 30–40 min, then cooled down to rt and
filtered through a short pad of celite, using MeOH to wash.
The filtrate was concentrated under reduced pressure, diluted
with CHCl3, washed with saturated aqueous solution of Na2CO3
and brine, and dried over Na2SO4. The residue was purified
after concentration at vacuum by chromatography, using CHCl3–
MeOH 9 : 1 as eluent, to give the amino ester 8 (592 mg, 88%) as
a yellowish oil.
Methyl 3-(4-((3-phenylureido)pyridin-2-yl)thio)propanoate (10).
Phenyl isocyanate (9, 298 mL, 2.74 mmol) was added dropwise to
a solution of methyl 3-((4-aminopyridin-2-yl)thio) propanoate (8,
466 mg, 2.20 mmol) in anhydrous benzene (8 mL) under argon.
The reaction was stirred at reflux for 3 h and then concentrated
under reduced pressure. The residue was purified by column
chromatography, using CHCl3–MeOH 95 : 5 as eluent, to give
urea-methyl ester 10 (620 mg, 85%) as a white solid: mp 131–
133 ◦C (from benzene).
Hapten synthesis
Although most of the chemicals used in this work present minor or
theusualsafety concerns, itis recommendedto take special caution
for the handling of acyl azides and isocyanates. The syntheses of
haptens p2, p6, m6, s5, CldPhUp6, and PhPyUp6 were described
in previous articles.8 The procedures employed for the synthesis
of immunogens m2 and s3 (Scheme 1) were as explained below.
The details of the synthesis of assay derivatives 4Fs5 and 4Ms5
(Scheme S1), and the spectroscopic characterization data of the
new compounds are given in the ESI.‡
3-(4-((3-Phenylureido)pyridin-2-yl)thio)propanoic acid (11, hap-
ten s3). A solution of the urea-methyl ester 10 obtained above
(190 mg, 0.57 mmol) in a mixture of THF (4.5 mL) and H2O
(2.0 mL) was treated with LiOH·H2O (241 mg, 5.75 mmol). After
3 h under stirring at rt, the reaction mixture was diluted with
H2O (20 mL) and extracted with ethyl ether. Following removal of
the remaining organic solvent under vacuum, the aqueous layer
was cooled on ice and acidified with KHSO4 to nearly pH 3.
The obtained precipitate was filtered off, washed with water and
dried to give the crude product that was purified by column
chromatography, using CH2Cl2–MeOH 7 : 3 as eluent, to afford
hapten s3 (11, 128 mg, 70%) as a white solid: mp 178–183 ◦C
(from DMSO–H2O).
Synthesis of hapten m2
To prepare hapten m2 (4), the commercial 2-(3-
aminophenyl)acetic acid (1) was reacted with 2-chloro-4-
isocyianatopyridine (3). A solution of 2-chloroisonicotinoyl
azide (2, 149 mg, 0.82 mmol)12 in toluene (1.0 mL) was added
dropwise to a stirred solution of 1 (117 mg, 0.78 mmol) in
4870 | Org. Biomol. Chem., 2011, 9, 4863–4872
This journal is
The Royal Society of Chemistry 2011
©

View Article Online
CPPU analogues
mixing 50 mL per well of competitor solution and 50 mL per well
of the corresponding immunoreagent solution in PBS containing
0.05% (v/v) Tween 20 (PBST). In the case of i-cELISAs, the
retained immunoglobulins were detected with an additional step
using 100 mL per well of secondary antibody dilution in PBST
(1/10000 GAR-HRP and 1/2000 RAM-HRP for pAb and
mAb-based assays, respectively). After washing, the enzymatic
activity was revealed using 100 mL per well of 2.0 mg mL-1 o-
phenylenediamine and 0.012% (v/v) H2O2 in 25 mM sodium
citrate and 62 mM sodium phosphate buffer, pH 5.4. Colour
development was stopped by addition of 100 mL per well of 2.5 M
H2SO4. Assay times were 1 h for each immunological reaction and
10 min for signal development. All incubations were carried out at
rt. ELISA absorbance was read in dual wavelength mode (492 nm
with 650 nm as reference wavelength) and the signal intensity was
plotted against the analyte concentration in a logarithmic scale.
The resulting sigmoidal curve was mathematically fitted to a four-
parameter logistic equation using the SigmaPlot software package
from SPSS Inc. (Chicago, IL).
The chemical structures of all of the employed CPPU analogues
are depicted in Table 3. TDZ (1-phenyl-3-(1,2,3-thiadiazol-5-
yl)urea), dPhU (1,3-diphenylurea), CldPhU (1-(3-chlorophenyl)-
3-phenylurea), NdPhU (1-(3-nitrophenyl)-3-phenylurea), and
PhPyU (1-(4-pyridyl)-3-phenylurea) were obtained from Sigma-
Aldrich-Fluka (Madrid, Spain). The syntheses of 4FPhClPyU (1-
(2-chloro-4-pyridyl)-3-(4-fluorophenyl)urea), F5PhClPyU (1-(2-
chloro-4-pyridyl)-3-(2,3,4,5,6-pentafluorophenyl)urea), 4MePh-
ClPyU (1-(2-chloro-4-pyridyl)-3-p-tolylurea), PhdClPyU (1-
(2,6-dichloro-4-pyridyl)-3-phenylurea), CldPyU (1-(2-chloro-
4-pyridyl)-3-(4-pyridyl)urea), and PhBrPyU (1-(2-bromo-
4-pyridyl)-3-phenylurea) were previously described.8 The
preparation of compounds thio-CPPU (1-(2-chloro-4-pyridyl)-
3-phenylthiourea) and BzClPyA (N-(2-chloro-4-pyridyl)-2-
phenylacetamide) and their spectroscopic characterization data
are detailed in the ESI‡ (Scheme S2).
Protein–hapten conjugates
Binding studies and hierarchical clustering analysis
BSA was selected for the preparation of immunogens via activation
of the carboxylate group of haptens with the active ester method,
followed by formation of amide bonds with the free amino groups
of the carrier protein. OVA and HRP were used as carriers for the
i-cELISA and the d-cELISA formats, respectively. These assay
conjugates were prepared by the mixed anhydride method. For
details see the ESI.‡ The final hapten-to-protein molar ratios of
the conjugates were estimated by monitoring the UV–vis spectra,
assuming that the molar extinction coefficients of both species
were essentially the same before and after the conjugation.
The capability of the antibodies to recognize a battery of CPPU-
analogous compounds was evaluated by conducting competitive
experiments in the i-cELISA format using homologous coating
conjugates. Standard curves of all chemicals were prepared in
PBS from concentrated stock solutions. AR was expressed as
the percentage of the ratio between the IC50 value for CPPU
and the IC50 value for the analogue. The binding profiles of all
of the antibodies towards the evaluated chemicals were used to
determine the similarity between immunoglobulins with respect
to their specificity. With this purpose, a matrix of data was
generated from the base-2 logarithmic transformation of the
AR values normalized to the unit. The corresponding heat map
and the dendrogram were constructed with the MultiExperiment
Viewer (MeV) version 4.5.1 (http://www.tm4.org/mev), which is
a part of the TM4 Microarray Software Suite.31 The hierarchical
classification of antibodies was performed using non-parametric
Kendall’s Tau similarity metric and complete linkage as the ag-
gregation method. The antibody distance heat map was generated
in the MeV programme by representing the distance matrix, and
imposing the clustering result obtained previously.
Antibody production
Mice and rabbits were immunized with BSA conjugates. The
production of pAbs from haptens p6, m6, and s5 has been
described elsewhere.8 Also, a panel of mAbs directed against
CPPU was previously generated from haptens p2 and p6.8a For the
present study, our collection of anti-CPPU mAbs was increased
and diversified from mice immunized with BSA derivatives of
m2, m6, s3, and s5. The selection of hybridomas was conducted
by differential i-cELISA30 using microplates that were previously
coated with the homologous conjugate at 1.0 mg mL-1, and using
1 mM CPPU as competitor. Stable antibody-secreting clones were
finally cryopreserved in liquid nitrogen and cell cultures were
expanded for the production of mAbs. For further details see the
ESI.‡
Acknowledgements
This work was supported by Ministerio de Educacio´n y Ciencia
(AGL2009-12940-C02/01/02/ALI), and cofinanced by FEDER
Funds. C.S.P. and J.V.M. were hired by the CSIC, the former under
a predoctoral I3P contract and the latter under a Ramo´n y Cajal
postdoctoral contract.
Competitive immunoassays
Solutions of all competitors were prepared in anhydrous
N,N-dimethylformamide and stored in amber glass vials at
-20 ◦C. Ninety six-well polystyrene ELISA plates were coated
overnight with 100 mL per well of 50 mM sodium carbonate-
bicarbonate buffer, pH 9.6 containing the corresponding im-
munoreagent. After incubation steps, plates were washed
four times with 0.15 M NaCl containing 0.05% (v/v) Tween
20. Standard curves were prepared by serial dilutions in 10 mM
sodium phosphate-buffered saline (140 mM NaCl), pH 7.4 (PBS).
The competitive step was carried out in coated microplates by
We thank Laura Lo´pez-Sa´nchez and Ana Izquierdo-Gil for
excellent technical assistance.
Notes and references
1 (a) M. H. Goodrow and B. D. Hammock, Anal. Chim. Acta, 1998, 376,
83–91; (b) K. Shreder, Methods, 2000, 20, 372–379; (c) M. C. Este´vez-
Alberola and M.-P. Marco, Anal. Bioanal. Chem., 2004, 378, 563–575;
(d) H. Zhang and S. Wang, J. Immunol. Methods, 2009, 350, 1–13.
2 (a) P. Crabbe, C. van Peteghem, M. Salden and F. Kohen, J. Agric. Food
Chem., 2000, 48, 3633–3638; (b) J. V. Mercader, C. Sua´rez-Pantaleo´n,
This journal is
The Royal Society of Chemistry 2011
Org. Biomol. Chem., 2011, 9, 4863–4872 | 4871
©

View Article Online
C. Agullo´, A. Abad-Somovilla and A. Abad-Fuentes, J. Agric. Food
Chem., 2008, 56, 1545–1552; (c) J. Parra, J. V. Mercader, C. Agullo´,
A. Abad-Fuentes and A. Abad-Somovilla, Tetrahedron, 2011, 67, 624–
635.
3 (a) R. O. Harrison, M. H. Goodrow and B. D. Hammock, J. Agric. Food
Chem., 1991, 39, 122–128; (b) F. I. Carroll, P. Abraham, P. K. Gong,
R. R. Pidaparthi, B. E. Blough, Y. Che, A. Hampton, M. Gunnell, J. O.
Lay Jr., E. C. Peterson and S. M. Owens, J. Med. Chem., 2009, 52,
7301–7309; (c) C. Roucairol, S. Azoulay, M. C. Nevers, J. Golebiowski,
C. Cre´minon, J. Grassi, A. Burger and D. Duval, Bioorg. Med. Chem.
Lett., 2010, 20, 987–990.
4 (a) J. V. Mercader, C. Agullo´, A. Abad-Somovilla and A. Abad-Fuentes,
Org. Biomol. Chem., 2011, 9, 1443, DOI: 10.1039/C0OB00686F;
(b) Z. L. Xu, Y. D. Shen, R. C. Beier, J. Y. Yang, H. T. Lei, H. Wang
and Y. M. Sun, Anal. Chim. Acta, 2009, 647, 125–136; (c) L. Cao, D.
Kong, J. Sui, T. Jiang, Z. Li, L. Ma and H. Lin, Anal. Chem., 2009, 81,
3246–3251.
Zhou, L. Belov, P. Y. Huang, J. S. Shin, M. J. Solomon, P. H. Chapuis,
L. Bokey, C. Chan, C. Clarke, S. J. Clarke and R. I. Christopherson,
J. Immunol. Methods, 2010, 355, 40–51; (d) J. C. Gray, P. H. Corran, E.
Mangia, M. W. Gaunt, Q. Li, K. K. Tetteh, S. D. Polley, D. J. Conway,
A. A. Holder, T. Bacarese-Hamilton, E. M. Riley and A. Crisanti, Clin.
Chem., 2007, 53, 1244–1253; (e) A. Pitarch, C. Nombela and C. Gil,
Methods Mol. Biol., 2009, 470, 369–411.
16 B. Meunier, E. Dumas, I. Piec, D. Be´chet, M. He´braud and J. F.
Hocquette, J. Proteome Res., 2007, 6, 358–366.
17 (a) C. M. Story, E. Papa, C. C. Hu, J. L. Ronan, K. Herlihy, H. L. Ploegh
and J. C. Love, Proc. Natl. Acad. Sci. U. S. A., 2008, 105, 17902–17907;
(b) E. Shi, W. Fury, W. Li, W. Mikulka, T. Aldrich, A. Rafique, G. Chen,
S. Hoffenberg, T. J. Daly and C. Radziejewski, J. Immunol. Methods,
2006, 314, 9–20.
18 S. Pattathil, U. Avci, D. Baldwin, A. G. Swennes, J. A. McGill, Z.
Popper, T. Bootten, A. Albert, R. H. Davis, C. Chennareddy, R. Dong,
B. O’Shea, R. Rossi, C. Leoff, G. Freshour, R. Narra, M. O’Neil, W. S.
York and M. G. Hahn, Plant Physiol., 2010, 153, 514–525.
19 M. B. Eisen, P. T. Spellman, P. O. Brown and D. Botstein, Proc. Natl.
Acad. Sci. U. S. A., 1998, 95, 14863–14868.
20 K. Landsteiner, in The Specificity of Serological Reactions, Harvard
University Press, Cambridge, 1945.
21 (a) N. Sanvicens, B. Varela and M.-P. Marco, J. Agric. Food Chem.,
2003, 51, 3932–3939; (b) W. L. Shelver, L. M. Kamp, J. L. Church and
F. M. Rubio, J. Agric. Food Chem., 2007, 55, 3758–3763.
22 P. Schneider, M. H. Goodrow, S. J. Gee and B. D. Hammock, J. Agric.
Food Chem., 1994, 42, 413–422.
23 (a) P. Degelmann, J. Wenger, R. Niessner and D. Knopp, Environ. Sci.
Technol., 2004, 38, 6795–6802; (b) P. M. Kra¨mer, M. H. Goodrow and
E. Kremmer, J. Agric. Food Chem., 2004, 52, 2462–2471.
5 M. I. Bruce and J. A. Zwar, Proc. R. Soc. London, Ser. B, 1966, 165,
245–265.
6 T. Okamoto, K. Shudo, S. Takahashi, F. Kawachi and Y. Isogai, Chem.
Pharm. Bul., 1981, 29, 3748–3750.
7 (a) M. Iwase, S. Okada, T. Oguchi and A. Toh-e, Genes & Genetic
Systems, 2004, 79, 199–206; (b) Q. Hu, W. J. Nelson and E. T. Spiliotis,
J. Biol. Chem., 2008, 283, 29563–29571; (c) B. S. DeMay, R. A. Meseroll,
P. Occhipinti and A. S. Gladfelter, Cytoskeleton, 2010, 67, 383–399.
8 (a) C. Sua´rez-Pantaleo´n, J. V. Mercader, C. Agullo´, A. Abad-Somovilla
and A. Abad-Fuentes, J. Agric. Food Chem., 2008, 56, 11122–11131;
(b) C. Sua´rez-Pantaleo´n, J. V. Mercader, C. Agullo´, A. Abad-Somovilla
and A. Abad-Fuentes, J. Agric. Food Chem., 2010, 58, 8502–8511.
9 (a) J. M. A. Schlaeppi, W. Meyer and K. A. Ramsteiner, J. Agric. Food
Chem., 1992, 40, 1093–1098; (b) A. E. Karu, M. H. Goodrow, D. J.
Schmidt, B. D. Hammock and M. W. Bigelow, J. Agric. Food Chem.,
1994, 42, 301–309; (c) S. Wang, R. D. Allan, J. H. Skerrit and I. R.
Kennedy, J. Agric. Food Chem., 1998, 46, 3330–3338.
24 M. Kotke and P. Schreiner, in Hydrogen Bonding in Organic Synthesis,
ed. P. M. Pihko, Wiley-VCH, Weinheim, 2009, p. 145.
25 L. Moreira-Lima and E. J. Barreiro, Curr. Med. Chem., 2005, 12, 23–49.
26 M. Lee, P. Lloyd, X. Zhang, J. M. Schallhorn, K. Sugimoto, A. G.
Leach, G. Sapiro and K. N. Houk, J. Org. Chem., 2006, 71, 5082–5092.
10 J. J. Manclu´s and A. Montoya, J. Agric. Food Chem., 1996, 44, 3703–
3709.
ˇ
27 D. Kopecˇny´, P. Briozzo, H. Popelkova´, S. Marek, R. Koncˇit´ıkova´, L.
11 For examples of related nucleophilic aromatic substitution reactions,
see: C. Bouillon, G. Kalopissis and G. Lang, DE Patent 2165752, 1972.
12 A. Abad, C. Agullo´, A. C. Cun˜at, C. Vilanova and M. C. Ram´ırez de
Arellano, Cryst. Growth Des., 2006, 6, 46–57.
13 (a) J. V. Mercader and A. Montoya, J. Agric. Food Chem., 1999, 47,
1276–1284; (b) M. Nichkova, R. Galve and M.-P. Marco, Chem. Res.
Toxicol., 2002, 15, 1360–1370.
Sp´ıchal, J. Nisler, C. Madzak, I. Fre´bort, M. Laloue and N. Houba-
He´rin, Biochimie, 2010, 92, 1052–1062.
28 J. J. Manclu´s, M. J. Moreno, E. Plana and A. Montoya, J. Agric. Food
Chem., 2008, 56, 8793–8800.
29 G. C. Finger and L. D. Starr, J. Am. Chem. Soc., 1959, 81, 2674–2675.
30 J. V. Mercader, J. Primo and A. Montoya, J. Agric. Food Chem., 1995,
43, 2789–2793.
14 (a) A. Abad, M. J. Moreno and A. Montoya, J. Agric. Food Chem., 1998,
46, 2417–2426; (b) N. Pastor-Navarro, C. Garc´ıa-Bover, A. Maquieira
and R. Puchades, Anal. Bioanal. Chem., 2004, 379, 1088–1099.
15 (a) P. Ellmark, C. M. Ho¨gerkorp, S. Ek, L. Belov, M. Berglund, R.
Rosenquist, R. I. Christopherson and C. A. Borrebaeck, Cancer Lett.,
2008, 265, 98–106; (b) M. K. Han, Y. H. Oh, J. Kang, Y. P. Kim, S. Seo,
J. Kim, K. Park and H. S. Kim, Proteomics, 2009, 9, 5544–5552; (c) J.
31 (a) A. I. Saeed, N. K. Bhagabati, J. C. Braisted, W. Liang, V. Sharov,
E. A. Howe, J. Li, M. Thiagarajan, J. A. White and J. Quackenbush,
Methods Enzymol., 2006, 411, 134–93; (b) A. I. Saeed, V. Sharov, J.
White, J. Li, W. Liang, N. Bhagabati, J. Braisted, M. Klapa, T. Currier,
M. Thiagarajan, A. Sturn, M. Snuffin, A. Rezantsev, D. Popov, A.
Ryltsov, E. Kostukovich, I. Borisovsky, Z. Liu, A. Vinsavich, V. Trush
and J. Quackenbush, BioTechniques, 2003, 34, 374–378.
4872 | Org. Biomol. Chem., 2011, 9, 4863–4872
This journal is
The Royal Society of Chemistry 2011
©