Online fluorescence enhancement assay for the acetylcholine binding protein with parallel mass spectrometric identification

Article abstract of DOI:10.1021/jm100230k

The acetylcholine binding protein (AChBP) is considered an analogue for the ligand-binding domain of neuronal nicotinic acetylcholine receptors (nAChRs). Its stability and solubility in aqueous buffer allowed the development of an online bioaffinity analysis system. For this, a tracer ligand which displays enhanced fluorescence in the binding pocket of AChBP was identified from a concise series of synthetic benzylidene anabaseines. Evaluation and optimization of the bioaffinity assay was performed in a convenient microplate reader format and subsequently transferred to the online format. The high reproducibility has the prospect of estimating the affinities of ligands from an in-house drug discovery library injected in one known concentration. Furthermore, the online bioaffinity analysis system could also be applied to mixture analysis by using gradient HPLC. This led to the possibility of affinity ranking of ligands in mixtures with parallel high-resolution mass spectrometry for compound identification.

Full text of DOI:10.1021/jm100230k

4720 J. Med. Chem. 2010, 53, 4720–4730
DOI: 10.1021/jm100230k
Online Fluorescence Enhancement Assay for the Acetylcholine Binding Protein
with Parallel Mass Spectrometric Identification
Jeroen Kool,*,†, Gerdien E. de Kloe,‡, Ben Bruyneel,† Jon S. de Vlieger,† Kim Retra,† Maikel Wijtmans,‡
Rene van Elk,§ August B. Smit,§ Rob Leurs,‡ Henk Lingeman,† Iwan J.P. de Esch,‡ and Hubertus Irth†
†BioMolecular Analysis, Department of Chemistry and Pharmaceutical Sciences, Faculty of Sciences, VU University Amsterdam,
De Boelelaan 1083, 1081HV Amsterdam, The Netherlands, ‡Medicinal Chemistry, Department of Chemistry and
Pharmaceutical Sciences, Faculty of Sciences, VU University Amsterdam, The Netherlands, and Molecular and Cellular Neurobiology,
§
Center for Neurogenomics and Cognitive Research, Neuroscience Campus Amsterdam, VU University Amsterdam, The Netherlands.
Both authors contributed equally to this work.
Received February 22, 2010
The acetylcholine binding protein (AChBP) is considered an analogue for the ligand-binding domain of
neuronal nicotinic acetylcholine receptors (nAChRs). Its stability and solubility in aqueous buffer allowed
the development of an online bioaffinity analysis system. For this, a tracer ligand which displays enhanced
fluorescence in the binding pocket of AChBP was identified from a concise series of synthetic benzy-
lidene anabaseines. Evaluation and optimization of the bioaffinity assay was performed in a convenient
microplate reader format and subsequently transferred to the online format. The high reproducibility
has the prospect of estimating the affinities of ligands from an in-house drug discovery library injected in
one known concentration. Furthermore, the online bioaffinity analysis system could also be applied to
mixture analysis by using gradient HPLC. This led to the possibility of affinity ranking of ligands in
mixtures with parallel high-resolution mass spectrometry for compound identification.
Introduction
which can be produced in large quantities, opened new oppor-
tunities for biochemical assays. Methods like SPR and NMR10
can be used for screening and also, together with tryptophan
fluorescence,11 for gaining insights into AChBP function-
ing. All these technologies, however, are hampered by back-
ground interferences, low sensitivities and/or long analysis
times. Furthermore, they are most suitable for analyzing pure
compounds.
We now present an online bioaffinity analysis methodology
for AChBP based on fluorescence-enhancement with parallel
mass spectrometric detection. Online bioaffinity analysis is
based on continuous flow assays12,13 and could be perfectly
used for screening compound libraries when based on fluore-
scence enhancement of a tracer ligand and performed in flow
injection mode. These assay formats, including the current
one, are usually easily automated, robust, and sensitive.
Furthermore, online bioaffinity analysis is ideally suited for
mixture analysis when performed in postcolumn format,
coupled in parallel to mass spectrometry (MS) to directly link
affinities to the corresponding mass spectra.14
For development of the online bioaffinity analysis system,
a small set of ligands with anticipated fluorescence enhance-
ment properties in the AChBP’s binding pocket was synthe-
sized and analyzed. A disubstituted benzylidene anabaseine
ligand was chosen as the most efficient tracer. Subsequent
focus was on obtaining a very high reproducibility of the final
system in order to allow affinity estimation of ligands from
our in-house compound library when injected at only one
concentration. Finally, the system was used in gradient HPLC
mode to rank affinities of ligands in a mixture with a single
concentration analysis.
The nicotinic acetylcholine receptor (nAChRa) family is
considered one of the most intensively studied ligand-gated
ion channels and constitutes valid pharmaceutical targets
for pain relief, Alzheimer’s disease, Parkinson’s disease,
epilepsy, anxiety, and several cognitive and attention defi-
cits.1-5 Analysis of ligand binding to nAChRs is usually
conducted by low-throughput electrophysiological patch-
clamping methodologies or radioligand binding analysis.
Although patch-clamping is automated today, this is still a
cumbersome and time-consuming endeavor. Radioligand
binding assays (RBAs), on the other hand, suffer from pro-
blems associated with the use of radioactive materials and
heterogeneous assay formats. For screening purposes, assays
based on frontal affinity chromatography have been deve-
loped.6-8 These methodologies, however, suffer from rela-
tively long analysis times.
The acetylcholine binding protein (AChBP) is a structural
analogue of the extracellular ligand binding domain of nAChRs.
It displays comparable ligand pharmacology to the R7 nAChR
in particular.1,5 AChBPs originate from mollusks, and since
the first crystal structure was published in 2001,9 it has become
a model to identify novel ligands for nAChRs by design-
intensive methods. The stable and water-soluble protein,
*To whom correspondence should be addressed. Phone: 0031-
205987542. Fax: 0031-205987543. E-mail: j.kool@few.vu.nl.
aAbbreviations: AChBP, acetylcholine binding protein; Ls, Lymnaea
stagnalis; nAChR, nicotinic acetylcholine receptor; SPR, surface plasmon
resonance; NMR, nuclear magnetic resonance; HPLC, high performance
liquid chromatography; HR-MS, high resolution mass spectrometry; RBA,
radioligand binding assay.
r
pubs.acs.org/jmc
Published on Web 06/02/2010
2010 American Chemical Society

Article
Journal of Medicinal Chemistry, 2010, Vol. 53, No. 12 4721
Scheme 1. Synthetic Route to Benzylidene Anabaseines (6-17) and of Precursors 19-20a
a (i) NaH, THF, 0°C-rt; (ii) conc aq HCl, reflux; (iii) potassium phthalimide, K2CO3, NMP, 90°C; (iv) conc aq HBr, reflux; (v) substituted benzalde-
hyde, acetic acid, sodium acetate, methanol, reflux; (vi) Me2CO3, K2CO3, DMF, reflux; (vii) DMF, POCl3, 1,2-DCE, reflux.
Results and Discussion
fluorescence enhancement in the active site of AChBP. Two
strategies included in the SAR had the aim of improving
the push-pull character of 7. Incorporation of the aniline
moiety in a ring should improve its electron-donating pro-
perties by increasing the overlap of the lone pair of the nitro-
gen with the π-system (16), while N,N-diethylation combined
with introduction of an extra electron-donating substituent
at the ortho-position yielded 17. Indeed, compounds 16 and
17 both have high λmax values and, especially 17, display
significant fluorescence enhancement upon binding to
AChBP (Table 1). On the basis of a balanced evaluation of
spectral properties, 4-diethylamino-2-hydroxybenzylidene
anabaseine (17, DAHBA) emerged as the most suitable dis-
placement ligand for the online bioaffinity analysis system.
Fluorescence Enhancement Assay Development in Micro-
plate Reader Format. The bioaffinity assay was first optimi-
zed in microplate reader format; the results obtained were
used to determine the assay conditions for the online bio-
affinity analysis system. The principle of the assay is based on
the competition of ligands with tracer ligand DAHBA that
shows fluorescence enhancement in the binding pocket of
AChBP. To this end, the high affinity compound epibatidine
was used as competing ligand in all evaluation and optimiza-
tion experiments. For the first evaluation experiment, the
concentration of DAHBA was varied (see Figure 1A), while
for every concentration of DAHBA, a complete epibatidine
IC50 curve was constructed. From Figure 1A, it is seen that
the assay window increased with higher concentrations of
DAHBA until the Bmax value was reached. At that point,
only the background increased as a result of higher concen-
trations of nonbound DAHBA, but not the assay window.
At much higher DAHBA concentrations than its pKi, also
higher concentrations of epibatidine are needed for displa-
cement as seen in Figure 1A by shifting of the IC50 curves.
Development of a Suitable Tracer Ligand. For the design of
our fluorescence enhancement ligand, we turned to benzyl-
idene anabaseine as a core structure based on previously
demonstrated fluorescence enhancement of 4-dimethylamino-
benzylidene anabaseine (7; see Scheme 1) in the active site
of AChBP.15 Fluorescence enhancement in the bound state
is thought to be a result of reduced rotational freedom of
the ligand due to its binding in the active site.15 A concise
structure-activity relationship (SAR) study around the
anabaseine scaffold was initiated. Starting from ethylnicoti-
nate (1) and δ-valerolactone (2), chloro-precursor 3 was syn-
thesized and substituted by an amine via introduction and
deprotection of a phthalimide to give the open (salt) form of
anabaseine (5). This anabaseine salt could be reacted with
several commercially available substituted (benz)aldehydes.
Only 1-methylindoline-5-carbaldehyde (20) was not avail-
able, and therefore synthesized from indoline (18) by methy-
lation and Vilsmeier-Haack introduction of the aldehyde.
Various substituents at the benzylidene moiety were ex-
plored, and an overview of the spectral properties of all ben-
zylidene anabaseines is given in the Supporting Information,
Table S-1. A large distribution of λmax values is obtained,
representing the differences in electronic properties of the
substituents on the benzylidene moiety (8-15). For com-
pounds with λmax values below 400 nm, no fluorescence enhan-
cement was measured because higher excitation wavelengths
are preferred to diminish protein fluorescence background
and avoid protein destruction. For compounds with λmax
values above 400 nm, fluorescence enhancement in the
presence of AChBP was measured (see Table 1). The only
two compounds of the initial series with λmax values higher
than 400 nm, the 3-methoxy-4-hydroxybenzylidene anaba-
seine 10 and indole-containing anabaseine 13, showed minor

4722 Journal of Medicinal Chemistry, 2010, Vol. 53, No. 12
Table 1. Spectral Properties of Benzylidene Anabaseines
Kool et al.
fluorescence
without AChBPa
fluorescence
with AChBPb
no.
λmax (nm)
ε (L mol-1 cm-1
)
enhancementc
λem. (nm)
7
474
476
425
493
489
42617
17418
142
1.15
0.58
4.85
0.65
3.60
6.15
0.58
5.35
1
1.06d
2.46
5.78
537
524
493
543
533
10
13
16
17
5.15
1.60
21745
65261
20.80
a Measured for 5 μM conc in buffer, pH=7.4. b Measured with a final concentration Ls-AChBP of 42 nM. c Defined as fluorescence in presence of
Ls-AChBP divided by fluorescence in absence of Ls-AChBP. d Measured at 5 ꢀ 10-4 M conc.
Figure 1. Optimization results in microplate reader format. (A) Variation of tracer ligand ([AChBP] = 2.65 nM). (B) Variation of AChBP
([tracer ligand]=5.21 ꢀ 10-8 M). (C) Tracer saturation curve ([AChBP]=15.76 nM; nonspecific binding was measured in presence of 10-5
M
epibatidine). (D) IC50 curves of two ligands. [AChBP]=15.76 nM; [tracer ligand]=5.21 ꢀ 10-8 M.
Analysis of the corresponding IC50 values gave 6.98 ( 0.16,
7.65 ( 0.06, 8.07 ( 0.08, 8.32 ( 0.05, 8.37 ( 0.05, 8.26 ( 0.05,
and ∼8.39, respectively.
indicating a titration instead of real binding kinetics). The
Kd of DAHBA was therefore determined indirectly with the
radioligand binding assay, resulting in a pKi of 7.59 ( 0.01.
Next, attention was paid to additives and organic modi-
fiers that can either positively influence the assay in general
(BSA, ELISA blocking buffer, and Tween 20) or are neces-
sarily introduced in the online bioaffinity analysis format
(MeOH, ACN, and DMSO). ELISA blocking reagent and
BSA did not influence the assay performance up to 5 mg/mL.
MeOH, ACN, DMSO, and Tween 20 decreased assay win-
dow to lower than 50% at concentrations of 12%, 13%, 4%,
and 5 mg/mL, respectively. It was shown that the optimized
assay was stable in a microplate for at least 5 days. The Z0-factor
was 0.80 at day one and 0.85 after storage at 4 °C for 5 days.
The use of 384-well microplates was also demonstrated and
showed the feasibility of this microplate format (Z0-factor of
0.83; data not shown).
In the next step, the effect of the AChBP concentration
was studied. These results are depicted in Figure 1B, where
no displacement and full displacement are shown. Decreas-
ing the AChBP concentration 3 and 9 times, respectively, also
gave approximately the same changes in the assay window.
The highest AChBP concentration was used further in order
to get a good assay window in the microplate reader format
for straightforward evaluation of the assay prior to transfer
to the online bioaffinity analysis system (where only 0.158 nM
AChBP was used in the assay).
Then, a Bmax (tracer saturation) curve was constructed as
shown in Figure 1C. A concentration of 10-5 M epibatidine
was used for analysis of nonspecific binding. The three curves in
Figure 1C represent the nonspecific binding curve, the total
binding curve, and the resulting Bmax curve. The use of the
high concentration of AChBP (15.76 nM) prevented deter-
mination of the Kd through the fluorescence enhancement
based assay (the apparent Kd found, 47 ( 13 nM, was appro-
ximately half of the AChBP binding sites concentration,
Validation of the assay was done with the ligands nicotine
and acetylcholine. IC50 curves obtained are shown in Figure 1D.
Table 2 shows the resulting pKis for the ligands (calculated
from the IC50 values by using the Cheng-Prusoff equation.16
For the ligands measured, small assay discrepancies between

Article
Journal of Medicinal Chemistry, 2010, Vol. 53, No. 12 4723
the fluorescence enhancement microplate reader assay and
the radioligand binding assay were observed.
Evaluation of the Online Bioaffinity Analysis System. The
optimal assay conditions found in microplate reader format
were then used in the online bioaffinity analysis system, shown
in Figure 2. In the system, injected ligands elute directly from
the column in flow injection analysis mode or during gra-
dient HPLC. The eluate is split to the MS and to the bio-
affinity analysis system. Here, eluting ligands interact with
continuously infused AChBP via the superloop operated by
pump P1a in the first reaction coil (4). In the second reaction
coil (6), tracer ligand is continuously infused (via the super-
loop operated by pump P2a) and interacts with the remain-
ing free binding sites of AChBP. The resulting fluorescence
enhancement is monitored by a fluorescence detector. MeOH
(70%) with 0.05% formic acid was used through the HPLC
column in order to allow direct elution of injected ligands at
t0 (flow injection mode). This setup was preferred instead of a
real flow injection analysis setup17 as injection peaks are
minimized. All measurements in this setup were conducted in
triplicate. The maximum signal-to-noise ratio obtained was
Table 2. pKi Values Calculated through the Cheng-Prusoff Equation
from Measured IC50 Curvesa
compd
FE plate reader (pKi)
online (pKi)
RBA (pKi)
acetylcholine
nicotine
4.93 ( 0.16
5.91 ( 0.09
4.97 ( 0.12
5.93 ( 0.02
5.09 ( 0.03
6.40 ( 0.01
a Results from three assay formats are compared to evaluate and
validate the on-line bioanalysis system. FE, fluorescence enhancement;
RBA, radioligand binding assay.
Figure 2. General schematic setup of the online bioaffinity analysis system. Separation (1) with split (2) to the online assay, the online auto-
fluorescence control assay and the MS (8). Online addition of AChBP and tracer ligand (3 and 5). Reaction coils (4 and 6). Fluorescence
detection (7). P1a=HPLC pump operating AChBP superloop; P1b is HPLC pump operating a superloop with only buffer. P2a and P2b=HPLC
pumps operating superloops with tracer ligand.
Figure 3. Post column online bioaffinity analysis, minutes vs relative fluorescence. (A) Typical displacement peaks (overlays) from epibatidine
injected at different concentrations (highest [C] = 5.1ꢀ10-6 M; then 1/3 serially diluted). (B) Overlays of triplicate injections from three
different concentrations of epibatidine injected (6.3 ꢀ 10-8 M, 2.1 ꢀ 10-8 M, and 7.0 ꢀ 10-9 M).

4724 Journal of Medicinal Chemistry, 2010, Vol. 53, No. 12
Kool et al.
214 and resulted from injecting epibatidine at a concentra-
tion of 10-4 M with an AChBP concentration of ∼0.158 nM
in the online bioaffinity analysis system. Peaks obtained with
the system from all concentrations of epibatidine injected are
shown in Figure 3A. Injecting different concentrations of
epibatidine showed negative peaks of which the peak heights
represented the percentages of displacement in the assay. At
full displacement, injecting even higher concentrations of
epibatidine evidently did not give higher signals but only
broader signals due to tailing effects. Figure 3B shows the
peaks of three different concentrations of epibatidine injec-
ted in triplicate. The figure shows that the triplicate injections
for every concentration injected can be perfectly overlaid
which demonstrates the reproducibility of the system. The
reproducibility was further studied and described analyti-
cally and pharmacologically in the next sections.
Pharmacological Validation. As injected compounds are
diluted in online bioaffinity analysis systems during the
HPLC separation and the online assay, a dilution factor to
determine the actual assay concentrations is required. The
procedure for determining the dilution factor is described in
the Experimental Section. It was found that, upon injection,
compounds were diluted 19.7 ( 0.5 times in the online bio-
affinity analysis system.
IC50 curves were measured for several reference ligands
and shown in Figure 4. As ligands are diluted in the online
bioaffinity analysis system, it was not possible to measure a
complete IC50 curve for the low affinity ligands due to
solubility problems. For comparison reasons, the IC50 values
were used to calculate pKi values (Table 2). The calculated
pKi values correspond well with the results obtained from the
fluorescence enhancement based microplate reader assay
and the RBA.
Intraday Repeatability. Intraday repeatability was demon-
strated by measuring three sequential IC50 curves for nicotine
during a measuring day (24 h; see Supporting Information
Figure S-1A). The online assay did not show a significant
decrease in the assay window over time, thus allowing repea-
table results over a measuring day. This was reflected by
the calculated IC50 values, which are depicted in Table 3.
Furthermore, Z0-factors were determined with data obtained
from injecting nicotine at two different concentrations for
multiple times (see Supporting Information Figure S-1B). For
the higher and lower concentrations of nicotine, Z0-factors
of 0.95 and 0.94 were obtained, respectively, indicating an
excellent assay performance.18
Interday Reproducibility. For interday reproducibility, an
IC50 curve of nicotine was measured every day for three mea-
suring days with all reagents freshly prepared every day. Here,
significant differences in assay window were seen (See Sup-
porting Information Figure S-1C), which were mainly caused
by the freeze-thawing process of AChBP causing nonrepro-
ducible inactivation. Therefore, the assay window was nor-
malized every day with nicotine injected at a concentration,
giving approximately 100% displacement. Subsequently, all
other data points during that measuring day were recalculated
with the same factor. Supporting Information Figure S-1D
gives the normalized IC50 curves. Calculation of the IC50
values from three measuring days is shown in Table 3, clearly
showing very reproducible results.
Ruggedness. As an important factor influencing the online
bioaffinity analysis system is the stability of AChBP, rug-
gedness experiments were focused on this parameter. For
this, a batch of AChBP was thawed and used for the first
measuring day. Some of the thawed AChBP was kept at 4 °C
overnight and used during the second measuring day in a
10 times lower concentration. The rest of the thawed AChBP
was frozen again and thawed at the last measuring day of the
ruggedness experiments. For every experiment, an IC50 curve
was measured and normalized the same way as done with the
interday experiments. Supporting Information Figure S-1E
and S-1F show the original and normalized IC50 curves, respec-
tively. It was shown (see Supporting Information Figure S-1F)
that widely differing active AChBP concentrations still result
in reproducible end results after normalization. This was also
supported by the calculated IC50 values as seen in Table 3.
Analytical Validation. Sigmodial-Concentration Response
Analysis. The online bioaffinity analysis system was inspected
by curve fitting of the normalized sigmoidal concentration-
response curves obtained during the interday, intraday, and
ruggedness experiments. Curve fitting by Graphpad Prism gave
curves perfectly crossing all data points measured for every
curve analyzed. The quality of fit was determined as the co-
efficient of determination (r2) and was found to be higher than
0.979 in all cases (Table 3).
Sensitivity. It should be mentioned that two kinds of assay
sensitivity can be distinguished, namely pharmacological
and analytical sensitivity. For too high AChBP concentra-
tions, the pharmacological sensitivity decreases for high
affinity ligands as displacement binding becomes a titration
thus lowering the IC50 values found. Second, higher tracer
ligand concentrations also decrease the pharmacological
sensitivity as more ligand is needed for displacement of
the tracer ligand. Lower concentrations of AChBP, therefore,
will give a pharmacologically more sensitive assay format
in terms of IC50 values found. However, for lower AChBP
Figure 4. Concentration-response curves of ligands obtained with
the online bioaffinity analysis system in flow injection mode.
Table 3. Pharmacological Validationa
parameter
1 (IC50
)
r2
2 (IC50
)
r2
3 (IC50
)
r2
intraday
interday
ruggedness
5.39 ( 0.03
5.48 ( 0.02
5.62 ( 0.01
0.9983
0.9990
0.9999
5.31 ( 0.01
5.52 ( 0.03
5.55 ( 0.02
0.9999
0.9973
0.9990
5.31 ( 0.01
5.46 ( 0.07
5.38 ( 0.03
0.9999
0.9791
0.9983
a The IC50 value of nicotine was determined three times to inspect the intraday repeatability, interday reproducibility, and the ruggedness. The
r2 represents goodness of the curve fitting. The table correlates with Figure S-1 and Table S-2 (Supporting Information).

Article
Journal of Medicinal Chemistry, 2010, Vol. 53, No. 12 4725
Table 4. Estimated pKi Values Analyzed with the Online Bioanalysis System after One Concentration (n=3) Injecteda
pKi at Log [C]=
-5.30 injected
pKi at Log [C]=
-5.77 injected
pKi at Log [C]=
-6.25 injected
pKi at Log [C]=
-6.73 injected
pKi from RBA
5
6.09 ( 0.06
7.13 ( 0.05
6.29 ( 0.03
8.55 ( 0.02
7.09 ( 0.03
6.87 ( 0.03
7.00 ( 0.05
8.71 ( 0.03
8.46 ( 0.02
7.38 ( 0.03
5.77 ( 0.00
6.81 ( 0.02
5.97 ( 0.01
7.52 ( 0.03*b
6.70 ( 0.01
6.55 ( 0.02
6.80 ( 0.02
7.85 ( 0.04*b
7.46 ( 0.05*b
6.77 ( 0.02
5.81 ( 0.00
6.76 ( 0.01
5.98 ( 0.01
8.07 ( 0.02
6.68 ( 0.01
6.54 ( 0.00
6.89 ( 0.02
8.41 ( 0.20
8.08 ( 0.07
6.89 ( 0.00
5.84 ( 0.01
6.75 ( 0.01
6.00 ( 0.01
8.27 ( 0.06
6.69 ( 0.01
6.60 ( 0.00
nd
5.90 ( 0.02
6.74 ( 0.00
6.02 ( 0.01
8.23 ( 0.01
6.70 ( 0.00
6.62 ( 0.00
nd
6
8
9
10
11
12
13
14
15
nd
8.06 ( 0.02
6.93 ( 0.02
nd
7.94 ( 0.02
6.95 ( 0.01
a The pKi of every compound was estimated independently after injection at four different concentrations (Log [C] injected of -5.30, -5.77, -6.25,
and -6.73). nd, not determined; RBA, radioligand binding assay. b *, injected in a too high concentration for good pKi estimation.
Figure 5. Correlation of estimated affinities from the online bioanalysis system and the RBA. For the online bioanalysis system, the average
estimated pKi of each ligand was calculated with the estimated pKi values per ligand injected at four different concentrations. (A) Nine
individual compounds were analyzed separately at four different concentrations (r2=0.986). (B) A mixture of compounds was analyzed at four
different concentrations (r2 =0.813). For the RBA, pKi values were determined by constructing complete IC50 curves (n=3).
and/or tracer ligand concentrations, assay windows will dec-
rease and thus the associated analytical sensitivity. Here, the
analytical sensitivity is determined as the limit of detection
(LOD), which is the ligand concentration that gives a signal
of at least three times the noise. The decrease in analytical
sensitivity is nicely observed with the ruggedness experiments
for the IC50 curve measured at a 10 times lower AChBP con-
centration. All LODs for nicotine during the validation experi-
ments are depicted in Supporting Information Table S-2 as
the lowest depicted concentration for every IC50 curve.
Intraday Repeatability, Interday Reproducibility, and Rug-
gedness. Supporting Information Table S-2 shows the displa-
cement percentages for all concentrations of nicotine tested
(n=3) during intraday, interday, and ruggedness experiments.
The ruggedness data demonstrated that alterations in the
most crucial factor determining assay performance, the AChBP
concentration, and/or handling, did not influence the final
data markedly. The data shows only large RSDs near the detec-
tion limits. Other parameters influencing the assay were
evaluated in microplate reader format (e.g., stability toward
organic modifiers) and subsequently used in the online bio-
affinity analysis system in such a way as to minimize varia-
tions in the online bioaffinity analysis readout.
For screening purposes, a reliable and reproducible ranking
of hits after single concentration injections would be highly
valuable.
Pure Compounds. This section demonstrates the affinity
estimation of 10 ligands after injection in the online bioaffi-
nity analysis system in one known concentration. Each ligand
was subsequently injected in three different concentrations
(5, 25, and 125 times diluted), which should result in the esti-
mation of a similar pKi for every independent ligand. For this,
ligand concentrations in the online assay were calculated with
use of the dilution factor (see Experimental Section). Then
IC50 values were calculated from the percentage of displacement
measured at the concentration in the assay with the formula
used for construction of sigmoidal concentration-response
curves (%displacement = 100%/(1 þ 10log IC -log [ligand])).
50
For these affinity estimations, the Hill slope was assumed to
be 1. From there, estimated pKi values were calculated (Table 4).
The correlation of these estimated pKi values with pKi values
from the RBA are represented in Figure 5A, from which an
excellent correlation (r2 = 0.986) was found. Any pKi estima-
tions from very high or very low displacement percentages,
however, are less accurate as they represent the shallow curved
areas of the sigmoidal concentration-response curves. This was
indeed seen for the three ligands with the highest affinity at
the highest concentration injected (Table 4; indicated with an
asterisk and omitted from the correlation). By a subsequent
injection at a lower concentration, as can be deduced from the
results, a more accurate result can be obtained.
Application of the Online Bioaffinity Analysis System for
Single Concentration Based Affinity Ranking. As proof-of-
principle demonstration, 10 pure compounds (nonfluorescent
anabaseine 5, benzylidene anabaseines 6 and 8-15) as well as
a mixture of nine compounds were analyzed with the online
bioaffinity analysis system. This was done in an effort to esti-
mate affinities of the ligands by injection of one concentration.
Mixture of Compounds. For proof-of-concept, a small focused
library was constructed from our running drug discovery

4726 Journal of Medicinal Chemistry, 2010, Vol. 53, No. 12
Kool et al.
Figure 6. Mixture analysis with the online bioaffinity analysis system. (A) Four bioaffinity chromatograms of a test mixture injected in diffe-
rent concentrations (10-4 M of every compound in the mixture for the lowest bioaffinity chromatogram, then 1/5 serially diluted). The com-
pounds were identified with parallel MS and UV detection and correlated with their bioaffinity. All compounds identified are depicted in the
figure, where the dashed arrows indicate the same compound in a splitted peak. (B) UV trace (220 nm extracted from DAD) and MS traces of
the parent ions from all compounds in the mixture (at 10-4 M injected).
programs (nine compounds, 21-29, see Supporting Informa-
tion Table S-4). Mixture analysis was performed in the same way
as discussed for the pure compound library (see previous section)
with the difference that the analyses were performed with gra-
dient HPLC runs. In Figure 6A, the bioaffinity traces of a mix-
ture injected in four concentrations are superimposed with the
highest concentration represented as the lowest trace. The bio-
affinity peaks were assigned to the compounds in the mixture
with help of the MS trace, which is shown for the highest
concentration of mixture injected in Figure 6B. High resolution
IT-TOF MS-MS spectra provided accurate MS-MS data for
further structure elucidation. Additionally, Figure 6B shows the
UV absorbance at 220 nm extracted from diode array detection
(200-300 nm). The strength of the system was highlighted by the
identification of an impurity in the mixture (tr = 37.0 min.),
which was determined to be a desmethyl-analogue (m/z 334.21)
of 23 (m/z348.23). Furthermore, the system showed that lobeline
(27) was present as two isomers, which was confirmed by an inde-
pendent LC-MS analysis of lobeline (data not shown). Epimer-
ization of lobeline has been described in literature, both in (polar)
organic solvents and in buffer solutions.19,20 Table 5 shows the
estimated pKi values for the individual compounds in the mixture
at every mixture concentration injected. As done for the analyses
of pure compounds, the estimated pKi values were correlated
with pKi values from the RBA and presented in Figure 5B, from
which a reasonable correlation (r2 = 0.813) was found.
One critical note is the peak splitting observed, of which the
severity decreased upon increasing retention time (Figure 6;
indicated per compound by the dashed arrows). As a high
injection volume (10 μL) was used to introduce sufficient
amounts of compounds in the online bioaffinity assay and
the test mixtures were dissolved in MeOH-H2O (1:1) to
prevent precipitation, this was caused by breakthrough of
especially the more polar compounds. This thus resulted
in division of the compounds over two peaks, which is not a
major issue as the bioaffinity trace can be correlated with
individual extracted ion chromatograms of the ligands. The
actual concentrations of compound in the bioaffinity signals
that are analyzed are only slightly lower for most compounds.
Only for the early eluting 24 and 28, severe peak splitting gave

Article
Journal of Medicinal Chemistry, 2010, Vol. 53, No. 12 4727
Table 5. Estimated pKi Values of Compounds in a Mixture Analyzed with the Online Bioanalysis System in Gradient HPLC Mode after One Con-
centration of Mixture Injected (n=1 for [Mix C1]; n=2 for [Mix C2,3,4])a
pKi from HPLC
[mix C1]
pKi from HPLC
[mix C2]
pKi from HPLC
[mix C3]
pKi from HPLC
[mix C4]
mixture contains
pKi from RBA
21
22
7.61 ( 0.12
7.59 ( 0.05
7.15 ( 0.02
6.66 ( 0.07
6.65 ( 0.09
6.59 ( 0.03
6.29 ( 0.01
nd
7.32
7.30
7.14 ( 0.15
7.07 ( 0.02
6.60 ( 0.04
5.83 ( 0.11
6.17 ( 0.05
6.34 ( 0.05
6.03 ( 0.08
coeluting with 26
7.01 ( 0.06
7.07 ( 0.05
6.76 ( 0.08
6.09 ( 0.19
6.34 ( 0.20
6.51 ( 0.13
6.26 ( 0.19
coeluting with 26
5.86
7.14 ( 0.15
7.18 ( 0.14
6.98 ( 0.24
nd
23
24
6.44
5.67
25
26
6.08
6.12
6.62 ( 0.43
nd
nd
27 (lobeline)
isomer of lobeline
5.97
coeluting with 26
5.58
coeluting with 25
coeluting with 26
nd
coeluting with 25
28
29
6.39 ( 0.16
5.42 ( 0.07
5.72 ( 0.12
coeluting with 25
coeluting with 25
a The pKi of every compound in the mixture was estimatedindependently after mixture injection at four different concentrations (see Figure 6). nd, not
determined; RBA, radioligand binding assay.
significantly lower concentrations of compound in the bio-
affinity signals analyzed and consequently lower pKi values
estimated (see Figure 5B and Table 5). However, this could
be anticipated for, which would result in a better correlation
in Figure 5B.
For 6-17, a Shimadzu LC-20AD liquid chromatograph pump
system with a Shimadzu SPD-M20A diode array detector was
used, with the MS detection performed with a Shimadzu LCMS-
2010 liquid chromatograph mass spectrometer. Conditions: an
Xbridge (C18) 5 μm column (4.6 mmꢀ100 mm) with solvent A
(90% MeCN-10% buffer) and solvent B (90% water-10%
buffer), flow rate of 1.0 mL/min, start 5% A, linear gradient to 90%
A in 8 min, then linear gradient to 5% A in 0.5 min, then 6.5 min
at 5% A, total run time of 15 min. For 21-29, a Shimadzu
LC-8A preparative liquid chromatograph pump system with a
Shimadzu SPD-10AV UV-vis detector with the MS detection
performed with a Shimadzu LCMS-2010 liquid chromatograph
mass spectrometer. Conditions: an Xbridge (C18) 5 μm column
(4.6 mm ꢀ 100 mm) with solvent A (90% MeCN-10% buffer)
and solvent B (90% water-10% buffer), flow rate of 2.0 mL/min,
start 5% A, linear gradient to 90% A in 10 min, then 10 min at
90% A, then 10 min at 5% A, total run time of 30 min. The buffer
was a 0.4% (w/v) NH4HCO3 solution in water, adjusted to
pH 8.0 with NH4OH. Compound purities were calculated as the
percentage peak area of the analyzed compound by UV detec-
tion at 254 nm. The compounds were g95% pure, unless stated
otherwise. A list of all purities is given in the Supporting Infor-
mation, Table S-3 and S-4.
Synthetic Methods. 5-Chloro-1-(pyridin-3-yl)pentan-1-one (3).
In a three-neck flask under N2, δ-valerolactone 2 (5.0 g, 33 mmol)
was dissolved in 200 mL dry THF. This colorless solution was
cooled on ice. NaH (60% in mineral oil, 2.0 g, 50 mmol) was
added. The gray suspension was stirred at 0 °C for 30 min and at
room temperature for another 30 min. The mixture was cooled to
0 °C. To this, a solution of ethylnicotinate 1 (3.0 mL, 22 mmol) in
10 mL dry THF was added dropwise. The reaction mixture was
allowed to warm to room temperature overnight while stirring.
The solvent was removed under reduced pressure. The resulting
solid was suspended in diethylether (200 mL) and collected by
vacuum filtration. This solid was mixed with 30 g of crushed ice
and concentrated HCl (100 mL) was added. The resulting solu-
tion was refluxed during 1 h, cooled to room temperature and
poured on 100 g ice. The mixture was basified to pH=8 and ex-
tracted with EtOAc (2ꢀ 150 mL). The organic layers were com-
bined, dried with Na2SO4, filtered, and concentrated. The dark-
yellow oil was purified with flash column chromatography (eluent:
30% EtOAc/cyclohexane). The product was obtained as a yellow
oil (1.47 g, 7.44 mmol, 34%). 1H NMR (250 MHz, CDCl3)
δ (ppm) 9.16 (d, J = 1.60 Hz, 1H), 8.77 (dd, J = 4.82, 1.66 Hz,
1H), 8.26-8.17 (m, 1H), 7.42 (ddd, J=7.98, 4.83, 0.84 Hz, 1H),
3.58 (t, J = 6.18 Hz, 2H), 3.03 (t, J = 6.82 Hz, 2H), 1.99-1.80
(m, 4H).
Conclusion
This work describes the development of an online bio-
affinity analysis system for the acetylcholine binding protein
(AChBP). For this, DAHBA, which showed fluorescence
enhancement upon binding to AChBP, was first developed.
The system underwent a thorough analytical as well as phar-
macological validation, revealing a very robust and suitable
system delivering data comparable with a traditional radi-
oligand binding assay. The key strength of the system was the
capability of estimating and thus ranking affinities of pure
compounds at single concentration injections. Furthermore,
the system was also capable of estimating affinities of com-
pounds in a focused mixture in combination with their iden-
tification after a single gradient HPLC-MS analysis. Thus, we
believe that the major surplus value of our methodology is the
ability to rapidly and conveniently screen compound libraries
and even mixtures of compounds for AChBP affinity. After
ranking of the hits with the online bioaffinity analysis system,
validation of interesting compounds by measuring complete
IC50 curves should be performed by injections of multiple
concentrations, for further, e.g. SAR, studies. We are cur-
rently investigating the screening of a fragment library on
AChBP using this methodology.
Experimental Section
General Remarks. The ELISA blocking reagent (BR) was obta-
ined from Hoffmann-La Roche (Mannheim, Germany) and the
bovine serum albumin (BSA) came from Gibco BRL (Breda, NL).
For the AChBP assays (in microplate reader and in online format),
a TRIS/PBS buffer was used with the following composition:
1 mM KH2PO4, 3 mM Na2HPO4, 0.16 mM NaCl, and 20 mM
Trizma-base at pH 7.5 with 400 mg/L ELISA BR.
General Synthetic. The compounds for mixture analyses were
synthesized in-house. Chemicals and reagents were obtained
from commercial suppliers and were used without further
purification. Dry THF was obtained by distillation from
CaH2. Flash column chromatography was typically carried
out on a Biotage flash chromatography system, using prepacked
Biotage Si 12þM columns, with the UV detector operating
2-(5-Oxo-5-(pyridin-3-yl)pentyl)isoindoline-1,3-dione (4). Chlo-
ride 3 (0.5 g, 2.5 mmol) was dissolved in N-methylpyrrolidinone
(8 mL). Phthalimide potassium salt (0.47 g, 2.5 mmol) and K2CO3
(0.38 g, 2.8 mmol) were added. The dark-yellow suspension
was warmed to 90 °C and stirred at this temperature for 4 h.
1
at 254 nm. All H and 13C NMR spectra were measured on a
€
Bruker 250 or Bruker 400. Analytical HPLC-MS analyses for
organic compounds were conducted using two different systems.
€

4728 Journal of Medicinal Chemistry, 2010, Vol. 53, No. 12
Kool et al.
The reaction mixture was diluted with water (100 mL). This
aqueous solution was extracted with DCM (2 ꢀ 50 mL). The
organic layers were combined, washed thoroughly with water
(6 ꢀ 50 mL), dried with Na2SO4, filtered, and concentrated.
A dark-brown oil was obtained (450 mg, 1.46 mmol, 58%). 1H
NMR (250 MHz, CDCl3) δ (ppm) 9.16 (d, J=1.60 Hz, 1H), 8.77
(dd, J = 4.82, 1.66 Hz, 1H), 8.23-8.17 (m, 1H), 7.82 (m, 2H),
7.70 (m, 2H), 7.39 (ddd, J=7.98, 4.83, 0.84 Hz, 1H), 3.73 (t, J=
6.18 Hz, 2H), 3.04 (t, J =6.82 Hz, 2H), 1.82-1.75 (m, 4H).
13C NMR (126 MHz, DMSO) δ 166.83, 157.44, 149.27, 149.11,
148.85, 136.68, 136.17, 130.66, 130.54, 125.59, 122.93, 110.11, 102.92,
97.55, 49.39, 43.77, 25.87, 22.24, 12.58. HR-MS m/z [M þ H]
336.2086, error[ppm] =-4.6. LC-purity: >99% (220 and 254 nm).
1-Methylindoline (19). Indoline 18 (2.8 mL, 25.2 mmol) was
added to K2CO3 (1.5 g, 10.9 mmol) in 20 mL of DMF. Dimethyl-
carbonate (6.4 mL, 76 mmol) was added. The yellow mixture
was refluxed overnight. It was cooled to room temperature and
50 mL of water was added. The product was extracted with
60 mL of tert-butylmethylether. The organic layer was washed
with water (3 ꢀ 50 mL), dried with Na2SO4, filtered, and
concentrated. The product was purified with flash column chro-
matography, eluent: gradient of hexane/ethyl acetate. Yield:
500 mg (3.75 mmol, 15%) of a pale-yellow oil. 1HNMR (250MHz,
CDCl3) δ (ppm) 7.12 (m, 2H), 6.70 (t, J =7.36 Hz, 1H), 6.58-
6.49 (d, J=8.18 Hz, 1H), 3.35 (t, J=8.08 Hz, 2H), 3.00 (t, J=
8.16 Hz, 2H), 2.82 (s, 3H).
5-Amino-1-(pyridin-3-yl)pentan-1-one 2HBr (5). Phthalimide
3
4 (450 mg, 1.5 mmol) was mixed with concentrated aq HBr
(48%, 7 mL), and the mixture was refluxed overnight. The next
day, the reaction mixture was concentrated and recrystallized
from hot isopropyl alcohol. The product was obtained as a brown
solid (245 mg, 0.72 mmol, 36%). 1H NMR (250 MHz, DMSO)
δ (ppm) 9.24 (s, 1H), 8.91 (dd, J=5.00, 1.51 Hz, 1H), 8.60-8.47
(m, 1H), 7.88-7.56 (m, 4H), 3.16 (d, J = 6.71 Hz, 2H), 2.83
(s, 2H), 1.75-1.56 (m, 4H). 13C NMR (126 MHz, DMSO) δ
198.19, 149.46, 146.03, 140.59, 133.67, 126.23, 39.01, 38.27, 26.78,
20.42.
1-Methylindoline-5-carbaldehyde (20). DMF (0.70 mL, 13.7
mmol) and POCl3 (0.27 mL, 3.0 mmol) were dissolved in 1,2-
dichloroethane (2.5 mL). Then 1-methylindoline (19) in 1,2-
dichloroethane (2.5 mL) was added. The dark-yellow solution
was refluxed during 3 h. Saturated NaHCO3 solution (20 mL)
and dichloromethane (20 mL) were added. The layers were
separated, and the water layer was extracted with dichloro-
methane (20 mL). The combined organic layers were dried with
General Method for Benzylidene Anabaseines. A general pro-
cedure reported by Sultana et al. was followed.21 Amine salt 5
(93 mg, 0.29 mmol) was dissolved in methanol (2 mL). Sodium
acetate (33 mg, 0.4 mmol), acetic acid (72 μL, 1.2 mmol), and a
substituted benzaldehyde (0.59 mmol) were added to this solu-
tion. The dark-yellow solution was refluxed overnight. The reac-
tion mixture was concentrated and dissolved in 1 M aq HCl
solution (25 mL). This acidified solution was washed twice with
dichloromethane (25 mL). The water layer was basified with
Na2CO3 and extracted with dichloromethane (3 ꢀ 25 mL). The
organic layers were combined, dried with Na2SO4, filtered, and
concentrated. For a few compounds, still present aldehyde was
removed by making an HCl salt of the imine and filtering it from
ether. If necessary, the compounds were purified with flash
column chromatography (eluent: EtOAc/2%TEA), or by for-
mation of the HCl-salt (2 M HCl solution in Et2O).
1
Na2SO4, filtered, and concentrated. Yield: 244 mg. H NMR
analysis revealed that the crude product contained 60% product
and 40% doubly formylated product as byproduct. The product
was used without further purification. 1H NMR (250 MHz,
CDCl3) δ (ppm) 9.72 (s, 1H), 7.60 (s, 2H), 6.44 (d, J=8.26 Hz,
1H), 3.60 (t, J=8.46 Hz, 2H), 3.06 (dd, J=18.05, 9.48 Hz, 2H),
2.93 (s, 3H).
Protein Production and Purification. Ls-AChBP (from snail
species Lymnaea stagnalis) was expressed from baculovirus
using the pFastbac I vector in Sf9 insect cells and purified from
medium as described by Celie et al.1
(E)-3-(3-(4-Dimethylaminobenzylidene)-3,4,5,6-tetrahydropyridin-
2-yl)pyridine (7). Synthesized from 5-amino-1-(pyridin-3-yl)pentan-
1-one 2HBr (5) and 4-dimethylaminobenzaldehyde by the general
UV and Fluorescence Enhancement Properties of Benzylidene
Anabaseines. Dilution series, consisting of five concentrations
between 5ꢀ10-5 M and 5ꢀ10-6 M, were made from a 1ꢀ10-2
M
3
method described. Yield: 93 mg (0.32 mmol, 36%) of a brown oil.
1H NMR (250 MHz, CDCl3) δ (ppm) 8.78 (d, J = 1.69 Hz, 1H),
8.74-8.64 (m, 1H), 7.89 (dt, J= 7.70, 1.87 Hz, 1H), 7.38 (dd, J=
7.79, 4.83 Hz, 1H), 7.30 (dd, J=6.80, 1.83 Hz, 3H), 6.72 (d, J=8.93
Hz, 2H), 6.63(s, 1H), 3.88(t, J=5.53Hz, 2H), 3.04(s, 6H), 2.93(td,
J = 6.80, 1.87 Hz, 2H), 1.96-1.82 (m, 2H). 13C NMR (63 MHz,
CDCl3) δ 167.97, 150.29, 149.96, 149.68, 137.42, 136.48, 131.56,
127.56, 123.61, 123.02, 111.74, 49.73, 40.19, 26.23, 22.26.
solution in DMSO, using the TRIS/PBS buffer described. UV
spectra were recorded using a Cary 50 Conc UV/vis spectro-
meter. The λmax values and ε values were determined. This λmax
was used to measure the fluorescence of the compounds at 5 ꢀ
10-6 M concentrations using a Perkin-Elmer LS50B fluores-
cence meter in the absence and presence of ∼42 nM Ls-AChBP
(625 ng/μL, 5 μL added to 3 mL compound solution in buffer).
Fluorescence of Ls-AChBP is neglectable when λexcitation>400 nm.
For compounds with lower λmax values, no fluorescence enhan-
cement was found, either because there was no measurable enhan-
cement or due to background fluorescence of the protein.
Fluorescence Enhancement Microplate Reader Assay. A Novostar
(with software version 1.20-0) from BMG Labtechnologies
(Offenburg, Germany) was used in fluorescence mode for the plate
reader assays. The excitation and emission wavelengths were 485
and 520 nm, respectively. Black-bottomed PP-96-well (and PP-384-
well) microtiter plates from Greiner Bio-one (Alphen a/d Rijn, NL)
were used. For all microplate reader measurements, 30 μL of ligand
or different concentrations organic modifier, blocking reagents or
detergent, 30 μL of Ls-AChBP, and 30 μL of DAHBA were added
subsequently. DAHBA was added 5 min after addition of the
AChBP, followed by incubation at room temperature for 10 min
before readout. In the manuscript, only the final concentrations are
depicted.
(E)-1-methyl-5-((2-(pyridin-3-yl)-5,6-dihydropyridin-3(4H)-yli-
dene)methyl)indoline (16). Synthesized from 5-amino-1-(pyridin-
3-yl)pentan-1-one 2HBr (5) and 1-methylindoline-5-carbaldehyde
3
(20) by the general method described. Yield: 63 mg (0.21 mmol,
24%) of a red oil. 1H NMR (250 MHz, CDCl3) δ (ppm) 8.77 (dd,
J=2.18, 0.82 Hz, 1H), 8.65 (dd, J=4.86, 1.71 Hz, 1H), 7.87-7.79
(m, 1H),7.35(ddd, J=7.82, 4.87, 0.86 Hz, 1H), 7.12 (d, J=7.35Hz,
2H), 6.58 (s, 1H), 6.45 (d, J=8.83 Hz, 1H), 3.86 (dd, J=9.80, 4.19
Hz, 2H), 3.42 (t, J=8.31 Hz, 2H), 3.00 (t, J= 8.30 Hz, 2H), 2.90
(ddd, J=6.76, 4.69, 2.12 Hz, 2H), 2.83 (s, 3H), 1.87 (dt, J=12.04,
6.21 Hz, 2H). 13C NMR (63 MHz, CDCl3) δ 167.82, 153.43, 149.91,
149.42, 136.98, 136.72, 136.38, 130.58, 130.31, 127.70, 125.94,
125.18, 122.93, 106.14, 55.64, 50.09, 35.35, 28.32, 26.34, 22.37.
(E)-3-(3-(4-Diethylamino-2-hydroxybenzylidene)-3,4,5,6-tetra-
hydropyridin-2-yl)pyridine (17, DAHBA).Synthesized from 5-amino-
1-(pyridin-3-yl)pentan-1-one 2HBr (5) and 4-diethylamino-2-hy-
3
droxybenzaldehyde by the general method described. Yield: 50 mg
(0.15 mmol, 51%) of a dark-red oil. 1H NMR (500 MHz, DMSO) δ
(ppm) 9.28 (s, 1H), 8.64-8.52 (m, J= 4.89 Hz, 2H), 7.78 (dt, J=
7.80, 1.93 Hz, 1H), 7.41 (dd, J=7.80, 4.84 Hz, 1H), 7.23 (d, J=8.85
Hz, 1H), 6.73 (s, 1H), 6.19 (dd, J=8.84, 2.50 Hz, 1H), 6.12 (d, J=
2.51 Hz, 1H), 3.70 (t, J=5.48Hz,2H),3.29(q,J=7.00 Hz, 4H), 2.73
(t, J=5.65 Hz, 2H), 1.81-1.63 (m, 2H), 1.08 (t, J=7.00 Hz, 6H).
Radioligand Binding Assay. A radioligand binding assay (96-
well format) was conducted to compare the results obtained
with the fluorescence enhancement assay. Ls-AChBP was dilu-
ted in PBS-Tris binding buffer (final concentration of 1.4 mM
KH2PO4, 4.3 mM Na2HPO4, 137.0 mM NaCl, 2.7 mM KCl,
20 mM Trizma-base, 4% DMSO, 0.05% Tween 20, pH 7.4) to
obtain a quantity of 1.3 ng per well. Ls-AChBP was incubated

Article
Journal of Medicinal Chemistry, 2010, Vol. 53, No. 12 4729
with 10-4-10-11 M of ligands (from stock solutions of 10 mM
in DMSO) in the presence of approximately 1.5 nM [3H]epiba-
tidine (KD = 0.875 nM, Perkin-Elmer Life Science, Inc., USA)
and 0.2 mg PVT Copper His-Tag SPA beads (GE healthcare).
Final well volume was 100 μL and the incubation time was 60 min
followed by 3 h in the absence of light. Thereafter the label-
bead complexes were counted in a micro beta counter.
Software. LCMS Solution from Shimadzu was used for analysis
of all data obtained with the online bioaffinity analysis system (LC,
MS and fluorescence enhancement data). Graphpad Prism 4 was
used for curve fitting and determination all biochemical parame-
ters (e.g., IC50 values, pKi values, Kds and Bmax values).
Acknowledgment. We thank Shimadzu for their instrumen-
tal support. We thank Ewald Edink for providing compounds
for the focused compound library. ABS, GdK, and IdE were
supported by Top Institute Pharma grant D2-103.
Online Bioaffinity Analysis. A similar setup described by de
Vlieger et al. was used.22 Figure 2 shows a schematic drawing of
the system used. A Shimadzu (‘s Hertogenbosch, The Netherlands)
SIL20 autoinjector introduced the samples into the system (10 μL
injections) followed by separation with a Waters (Milford, MA)
Xterra C18MS column (2.1 mmꢀ100 mm, 3.5 μm particles) and
guard column. Mobile phase A consisted of H2O-MeOH-formic
acid (98.9%-1%-0.1%) and mobile phase B of MeOH-H2O-
formic acid (98.9%-1%-0.1%). In flow-injection analysis
mode for analysis of pure compounds, 30/70% mobile phase
A/B was used at a flow rate of 125 μL/min. For gradient elution,
the following gradient was used: 0-2 min at 20% mobile phase
B followed by a linear increase to 50% mobile phase B in 33 min,
then a linear increase to 100% mobile phase B in 2 min followed
by 3 min at 100% mobile phase B. Equilibration occurred by
performing a linear decrease to 20% mobile phase B in 1 min
followed by 4 min at 20% mobile phase B. After the LC column,
a split of 125:125:1000 directed the flow to the parallel assays and
MS, respectively. The LC-MS system consisted of a Shimadzu
LCMS-IT-TOF with a SPD20 M photodiode array detector.
The ion-trap time-of-flight (IT-TOF) MS was operated in data
dependent acquisition mode switching between full-spectrum
MS and MS-MS modes. The electrospray (ESI) source was ope-
rated in positive ion mode. The temperature of the heating block
and curved desolvation line was set to 200 °C. Interface voltage
was set at 4.5 kV, the nebulizing gas flow (nitrogen, 99.999%
purity, Praxair, Oevel, Belgium) was 1.5 L/min, and drying gas
was applied at a pressure of 65 kPa. Full MS spectra were
acquired from m/z 50-700. MS-MS data was obtained auto-
matically with 25 ms accumulation time from m/z 50-700 and a
precursor isolation width of 3.0 using argon (99.9995% purity,
Praxair) as collision gas. The bioaffinity analysis part employed
four Shimadzu LC10ADvp pumps and two Shimadzu RF10-
XL fluorescence detectors (λex=485 nm, λem=526 nm). One of
the online assays was used as autofluorescence control assay in
which AChBP was omitted from its superloop (via P1b). This
allowed the detection of false positive ligands that display
intrinsic fluorescence at the wavelengths used for bioaffinity
analysis. For the actual online assay, the AChBP concentration
was 0.345 nM in superloop 1 (representing 0.158 nM in the assay).
The concentration of tracer ligand DAHBA was 150 nM in
superloop 2 (representing 68 nM in the assay). The superloops
were operated at a flow rate of 70 μL/min and were kept on ice.
The assays were performed in a thermostatted oven at 37 °C.
The volumes of the first and second reaction coils (of PEEK
material) were 25 and 20 μL, respectively. For analysis of IC50
values, 10-2 M solutions of ligands or mixtures of ligands in
DSMO were diluted in MeOH-H2O (1:1) to 10-4 M solutions.
These solutions, or dilutions thereof, were injected into the
system and the resulting percentages of displacement were used
for IC50 value calculations.
Supporting Information Available: Pharmacological and ana-
lytical validation of the on-line system. Purities and HR-MS data
for all in-house synthesized compounds. Spectral properties and
synthetic routes of some of the compounds used. This material is
available free of charge via the Internet at http://pubs.acs.org.
References
(1) Celie, P. H.; Kasheverov, I. E.; Mordvintsev, D. Y.; Hogg, R. C.;
van Nierop, P.; van Elk, R.; van Rossum-Fikkert, S. E.; Zhmak,
M. N.; Bertrand, D.; Tsetlin, V.; Sixma, T. K.; Smit, A. B. Crystal
structure of nicotinic acetylcholine receptor homolog AChBP in
complex with an alpha-conotoxin PnIA variant. Nat. Struct. Mol.
Biol. 2005, 12, 582–588.
(2) Dutertre, S.; Ulens, C.; Buttner, R.; Fish, A.; van Elk, R.; Kendel,
Y.; Hopping, G.; Alewood, P. F.; Schroeder, C.; Nicke, A.; Smit,
A. B.; Sixma, T. K.; Lewis, R. J. AChBP-targeted alpha-conotoxin
correlates distinct binding orientations with nAChR subtype selec-
tivity. EMBO J. 2007, 26, 3858–3867.
(3) Hogg, R. C.; Bertrand, D. Nicotinic acetylcholine receptors as drug
targets. Curr. Drug Targets CNS Neurol. Disord. 2004, 3, 123–130.
(4) Hogg, R. C.; Raggenbass, M.; Bertrand, D. Nicotinic acetylcholine
receptors: from structure to brain function. Rev. Physiol. Biochem.
Pharmacol. 2003, 147, 1–46.
(5) Smit, A. B.; Syed, N. I.; Schaap, D.; van Minnen, J.; Klumperman,
J.; Kits, K. S.; Lodder, H.; van der Schors, R. C.; van Elk, R.;
Sorgedrager, B.; Brejc, K.; Sixma, T. K.; Geraerts, W. P. A glia-
derived acetylcholine-binding protein that modulates synaptic
transmission. Nature 2001, 411, 261–268.
(6) Sharma, J.; Besanger, T. R.; Brennan, J. D. Assaying small-
molecule-receptor interactions by continuous flow competitive
displacement chromatography/mass spectrometry. Anal. Chem.
2008, 80, 3213–3220.
(7) Jozwiak, K.; Haginaka, J.; Moaddel, R.; Wainer, I. W. Displace-
ment and nonlinear chromatographic techniques in the investigation
of interaction of noncompetitive inhibitors with an immobilized
alpha3beta4 nicotinic acetylcholine receptor liquid chromatogra-
phic stationary phase. Anal. Chem. 2002, 74, 4618–4624.
(8) Jozwiak, K.; Moaddel, R.; Yamaguchi, R.; Ravichandran, S.;
Collins, J. R.; Wainer, I. W. Qualitative assessment of IC50 values
of inhibitors of the neuronal nicotinic acetylcholine receptor using
a single chromatographic experiment and multivariate cluster
analysis. J. Chromatogr., B: Anal. Technol. Biomed. Life Sci. 2005,
819, 169–174.
(9) Brejc, K.; van Dijk, W. J.; Klaassen, R. V.; Schuurmans, M.; van
Der Oost, J.; Smit, A. B.; Sixma, T. K. Crystal structure of an ACh-
binding protein reveals the ligand-binding domain of nicotinic
receptors. Nature 2001, 411, 269–276.
(10) Gao, F.; Mer, G.; Tonelli, M.; Hansen, S. B.; Burghardt, T. P.;
Taylor, P.; Sine, S. M. Solution NMR of acetylcholine binding
protein reveals agonist-mediated conformational change of the
C-loop. Mol. Pharmacol. 2006, 70, 1230–1235.
(11) Hansen, S. B.; Radic, Z.; Talley, T. T.; Molles, B. E.; Deerinck, T.;
Tsigelny, I.; Taylor, P. Tryptophan fluorescence reveals conforma-
tional changes in the acetylcholine binding protein. J. Biol. Chem.
2002, 277, 41299–41302.
Dilution Factor. The dilution factor after injection into the
online bioaffinity analysis system was determined by a calibra-
tion process. For this, the fluorescent compound resorufin (100 nM
in buffer) was pumped into the online bioaffinity analysis system
via one superloop, while the other superloop only contained
buffer. The resulting increased fluorescent signal was recorded.
Then, the resorufin solution was replaced by buffer and the
resorufin solution was subsequently injected (10 μL) into the
online bioaffinity analysis system. The peak heights of eluting
resorufin were compared with the elevated fluorescent signal
from which the dilution factor of resorufin in the online bioaffi-
nity analysis system was be determined.
(12) van Elswijk, D. A.; Diefenbach, O.; van der Berg, S.; Irth, H.;
Tjaden, U. R.; van der Greef, J. Rapid detection and identification
of angiotensin-converting enzyme inhibitors by on-line liquid
chromatography-biochemical detection, coupled to electrospray
mass spectrometry. J. Chromatogr., A 2003, 1020, 45–58.
(13) Schenk, T.; Breel, G. J.; Koevoets, P.; van den Berg, S.; Hogenboom,
A. C.; Irth, H.; Tjaden, U. R.; van der Greef, J. Screening of natural
products extracts for the presence of phosphodiesterase inhibitors
using liquid chromatography coupled online to parallel biochemi-
cal detection and chemical characterization. J. Biomol. Screening
2003, 8, 421–429.

4730 Journal of Medicinal Chemistry, 2010, Vol. 53, No. 12
Kool et al.
(14) Schobel, U.; Frenay, M.; van Elswijk, D. A.; McAndrews, J. M.;
Long, K. R.; Olson, L. M.; Bobzin, S. C.; Irth, H. High resolution
screening of plant natural product extracts for estrogen receptor
alpha and beta binding activity using an online HPLC-MS bio-
chemical detection system. J. Biomol. Screening 2001, 6, 291–303.
(15) Talley, T. T.; Yalda, S.; Ho, K. Y.; Tor, Y.; Soti, F. S.; Kem, W. R.;
Taylor, P. Spectroscopic analysis of benzylidene anabaseine com-
plexes with acetylcholine binding proteins as models for ligand-
nicotinic receptor interactions. Biochemistry 2006, 45, 8894–8902.
(16) Cheng, Y.; Prusoff, W. H. Relationship between the inhibition
constant (K1) and the concentration of inhibitor which causes 50%
inhibition (I50) of an enzymatic reaction. Biochem. Pharmacol.
1973, 22, 3099–3108.
(18) Zhang, J. H.; Chung, T. D.; Oldenburg, K. R. A Simple Statistical
Parameter for Use in Evaluation and Validation of High Through-
put Screening Assays. J. Biomol. Screening 1999, 4, 67–73.
(19) Compere, D.; Marazano, C.; Das, B. C. Enantioselective access to
Lobelia alkaloids. J. Org. Chem. 1999, 64, 4528–4532.
(20) Crooks, P. A.; Dwoskin, L. P. Use of lobeline epimers in the treat-
ment of central nervous system diseases, pathologies, and drug abuse.
WO Patent Application 2009/059260, 2009.
(21) Sultana, I.; Ikeda, I.; Ozoe, Y. Structure-activity relationships of
benzylidene anabaseines in nicotinic acetylcholine receptors of
cockroach nerve cords. Bioorg. Med. Chem. 2002, 10, 2963–2971.
(22) de Vlieger, J. S.; Kolkman, A. J.; Ampt, K. A.; Commandeur, J. N.;
Vermeulen, N. P.; Kool, J.; Wijmenga, S. S.; Niessen, W. M.; Irth,
H.; Honing, M. Determination and identification of estrogenic
compounds generated with biosynthetic enzymes using hyphena-
ted screening assays, high resolution mass spectrometry and off-
line NMR. J. Chromatogr., B: Anal. Technol. Biomed. Life Sci.
2010, 878, 667–674.
(17) Kool, J.; van Liempd, S. M.; Ramautar, R.; Schenk, T.; Meerman,
J. H.; Irth, H.; Commandeur, J. N.; Vermeulen, N. P. Development
of a novel cytochrome p450 bioaffinity detection system coupled
online to gradient reversed-phase high-performance liquid chromato-
graphy. J. Biomol. Screening 2005, 10, 427–436.