Chemical Science
Edge Article
DV ¼ 1/Bt[A0] + Kd/Bt; Kd and Bt were obtained by plotting 1/DV ¼
f([A0]), DV being the difference between the retention volume of
the injected compound (V) minus the retention volume of the
void marker (V0) (i.e. breakthrough volume), and [A0] the total
concentration of the injected compound.
Validation of the system: injection of diclofenac (35) and
isoniazid (36). Solutions of diclofenac sodium salt (35) and
isoniazid (36) at 62.5 mM were prepared in PBS and sequentially
infused through the column containing HSA using the HPLC
method described in the previous section. Diclofenac was
monitored at 220 and 254 nm whereas isoniazid was followed at
220 and 262 nm. Retention volumes of 130 mL (s.d. ¼ 3.5 mL)
and 837.5 mL (s.d. ¼ 2.8 mL) were respectively obtained for these
two reference compounds.
Assessment of the HSA column with zolpidem (1): determi-
nation of Bt. The rst compound assessed using this assay was
zolpidem (1). In practice, from a 125 mM stock solution of zol-
pidem (1) in PBS were prepared concentrations of 62.5, 41.67,
31.25, 15.62 and 7.81 mM. Each of these mixtures were then
injected in triplicate following the same HPLC method, with
Bt and Kd values being calculated as explained above. In these
conditions, Kd and Bt were respectively found to be 61 mM and
20.7 nmoles. In order to show the reproducibility of this assay,
the same experiment was performed aer two months on the
same column; Bt did not change over the two months period,
validating the stability of the immobilised HSA on the column
as well as the reproducibility of the technique.
Conclusions
In summary we have demonstrated that by the use of enabling
ow chemistry methods we can assemble a collection of imi-
dazo[1,2-a]pyridines as potential GABAA agonists including
zolpidem (1) and alpidem (2), and evaluated these for binding
properties to HSA using frontal affinity chromatography. This
has been achieved via equipment in a footprint of one stan-
dard fume cupboard. We observed that affinities of zolpidem
(1) and alpidem (2) are in accordance with literature values14,15
and could obtain simple structure–activity relationship using a
series of related analogues. The results of this study correlate
with known drug-serum protein interactions.16 While we have
not yet fully integrated the synthesis with the biological assay,
the molecular components we have employed will form the
basis for future developments towards this goal. We anticipate
that the concentration-independent nature of FAC assay will
greatly facilitate the integration of biology and chemistry
which will ultimately lead to an automated drug discovery
platform.
Experimental section
Synthesis
Details for synthesis and full characterisation of all compounds
can be found in ESI.†
Frontal affinity chromatography assays
Determination of the affinity of the synthesised compounds
General considerations. Phosphate Saline Buffer (PBS), 1, 2, 15–34 for HSA. For each compound, stock solutions at
diclofenac sodium salt, isoniazid and immobilised HSA on 125 mM were prepared in PBS. 0.5% (vol) of either DMSO or 1 M
silica gel (Ø 5 mm, Suppelco) were bought from Sigma-Aldrich. sodium hydroxide were added when compounds were not
PBS was sterilised by ltration (Millipore, 0.22 mM) prior to use. completely soluble. Three concentrations (7.81, 31.25 and
DMSO (Alfa Aesar) was used without any further purication. 62.5 mM, all within the linear range of the detector) were
The 15 mL guard column (1 mm ꢃ 2 cm) was purchased from injected in triplicate and monitored simultaneously at two
Kinesis. FAC assays were run on an Agilent 1100 HPLC system wavelengths following the same HPLC method with Bt and Kd
using PBS as the eluent.
Retention volumes were calculated with an in-house Excel throughout the experiments; Kd values are reported in Table 1.
Macro le, which determines the time between the injection Ranking method based on a single injection from aliquots
point and the point when the concentration reaches 50% of the automatically prepared from the reaction output. In FAC
plateau. experiments, it is possible to rank compound affinities for the
values being calculated. An average Bt of 20 nmole was obtained
Methods. The HPLC system was set-up to inject 600 mL of target molecule based on their retention volumes obtained via
each concentration of analyte at 50 mL minꢁ1 for 40 min. Three single injections. Although this will not produce Bt and Kd
wavelengths were simultaneously monitored at 220 nm (analyte values, it is a quick way to compare relative affinities, especially
or DMSO), 254 nm (analyte) and 262 nm (analyte) using a Diode when dealing with a large collection of compounds or/and
Array UV Detector.
dealing with outputs straight from ow chemistry reactors. In
Serial dilutions of each compound were prepared from a practice, 20 mL of the aliquot taken by the fraction collector was
100 mL stock solution at 125 mM in PBS. When solubility issues diluted further in 2 mL of PBS and injected directly using the
were encountered, either 500 mL of DMSO or 500 mL of 1 M standard HPLC method.
sodium hydroxide were used to dissolve the compounds prior to
Compounds were simply ranked according to their retention
making the stock solution. A solution of DMSO (500 mL in volumes which correlated well with the trend seen in Table 1.
100 mL in PBS) was used as the void marker to determine the
minimum retention volume (V0 ¼ 127.4 mL, s.d. ¼ 2.3 mL) of
Acknowledgements
the column. Blank experiments (i.e. a column packed with
silica only) showed no non-specic interaction for any of the We would like to thank the EPSRC grant EP/F069685/1 (LG and
analytes.
NN), the Royal Society (IRB), the BP Endowment (SVL) for
Determination of binding constants (Kd) and the amount of nancial support and Dr Richard Turner for technical
HSA immobilised on the column (Bt) were performed using: 1/ assistance.
768 | Chem. Sci., 2013, 4, 764–769
This journal is ª The Royal Society of Chemistry 2013
Guetzoyan, Lucie
