based on the reversible binding of constructs to affinity resins
avoiding protein purification and washing steps. In this assay, soluble lysates of recombinant proteins overexpressed in bacteria (e.g.,
PDZ domain) are incubated with avidin beads, previously fully
saturated by either a biotinylated peptide (e.g., PBM) or a reference
(biotin). Once equilibrium is achieved, the resin–liquid mixture is
subjected to a fast filtration. The resulting flow-through contains
the remaining nonbound recombinant protein, which is further
quantified using a capillary electrophoresis instrument. The electropherogram obtained for the PBM–PDZ pair is then superimposed and compared with the one for the reference with biotin–
PDZ. A depletion of the recombinant protein of interest observed
in the flow-through of the resin containing the peptide, as compared to the reference flow-through, indicates that a protein–peptide binding event has occurred. The stronger the depletion of
the recombinant protein of interest, as compared to the reference,
the stronger the protein–peptide binding interaction. Ultimately,
the binding intensity (BI), based on PDZ peak intensities
and related to the interaction strength, is determined, from which
in turn the equilibrium dissociation affinity constant (K D ) can be
deduced [8].
A proper estimate of the BI relies on the precise and accurate
pairwise comparison of electropherograms. When dealing with the
same protein extract, the pattern of two electropherograms is
highly conserved in most parts of the graphs, except for the PDZ
peak which may have partly disappeared from the extract by being
specifically retained on the PBM-coated resin. This makes possible
to compare them in an automated way, using computational methods. However, capillary electrophoresis migration may be subject
to variations linked to slight differences of input volume, migration
in the capillary, buffer, well position, or measurement temperature,
possibly altering the peak intensity or the migration of molecules
and therefore their apparent sizes [9, 10]. As a consequence, slight
fluctuations of the peaks on the X- (migration) and/or Y- (sensitivity) axis and baseline perturbation may alter the reproducibility
of the electropherograms and subsequently the accuracy of BI
values (Fig. 1a). For this reason, a strict comparison of two different
electropherograms is not always easy to perform, even when data
are recorded in the same experimental conditions. Therefore, the
use of semiautomated methods to correct and compare the electropherograms, can dramatically improve their superimposition
(Fig. 1b).
Here we present the main lines of a computational protocol to
deal with these issues using commands available in the free Python
Spike package. The accuracy of the results dramatically increases
when one uses this automated approach to improve superimposition of the data. This can be achieved by applying up to five
consecutive processing steps to the electropherograms of both,
the reference and the sample (Fig. 2):
62
Pau Jane ´ et al.
avoiding protein purification and washing steps. In this assay, soluble lysates of recombinant proteins overexpressed in bacteria (e.g.,
PDZ domain) are incubated with avidin beads, previously fully
saturated by either a biotinylated peptide (e.g., PBM) or a reference
(biotin). Once equilibrium is achieved, the resin–liquid mixture is
subjected to a fast filtration. The resulting flow-through contains
the remaining nonbound recombinant protein, which is further
quantified using a capillary electrophoresis instrument. The electropherogram obtained for the PBM–PDZ pair is then superimposed and compared with the one for the reference with biotin–
PDZ. A depletion of the recombinant protein of interest observed
in the flow-through of the resin containing the peptide, as compared to the reference flow-through, indicates that a protein–peptide binding event has occurred. The stronger the depletion of
the recombinant protein of interest, as compared to the reference,
the stronger the protein–peptide binding interaction. Ultimately,
the binding intensity (BI), based on PDZ peak intensities
and related to the interaction strength, is determined, from which
in turn the equilibrium dissociation affinity constant (K D ) can be
deduced [8].
A proper estimate of the BI relies on the precise and accurate
pairwise comparison of electropherograms. When dealing with the
same protein extract, the pattern of two electropherograms is
highly conserved in most parts of the graphs, except for the PDZ
peak which may have partly disappeared from the extract by being
specifically retained on the PBM-coated resin. This makes possible
to compare them in an automated way, using computational methods. However, capillary electrophoresis migration may be subject
to variations linked to slight differences of input volume, migration
in the capillary, buffer, well position, or measurement temperature,
possibly altering the peak intensity or the migration of molecules
and therefore their apparent sizes [9, 10]. As a consequence, slight
fluctuations of the peaks on the X- (migration) and/or Y- (sensitivity) axis and baseline perturbation may alter the reproducibility
of the electropherograms and subsequently the accuracy of BI
values (Fig. 1a). For this reason, a strict comparison of two different
electropherograms is not always easy to perform, even when data
are recorded in the same experimental conditions. Therefore, the
use of semiautomated methods to correct and compare the electropherograms, can dramatically improve their superimposition
(Fig. 1b).
Here we present the main lines of a computational protocol to
deal with these issues using commands available in the free Python
Spike package. The accuracy of the results dramatically increases
when one uses this automated approach to improve superimposition of the data. This can be achieved by applying up to five
consecutive processing steps to the electropherograms of both,
the reference and the sample (Fig. 2):
62
Pau Jane ´ et al.