SLAS Discovery - Special Issue on Advances in MALDI Mass Spectrometry for Drug Discovery - Volume 22 Issue 10 - December 2017 - 1199

1199

Heap et al.
Using the MALDI TOF assay, we determined that the
phosphorylation of CHKtide through SIK2 (Fig. 2B) was
linear up to 60 min between ~15% and 75% occupancy.
From these data, an optimal point of 30 min was chosen to
quench the reaction, thus allowing a conversion of approximately 50% of the substrate peptide to its phosphorylated
equivalent. When comparing ADP Hunter and MALDI
TOF assay enzyme concentrations, similar final enzyme
concentrations were used. However, the MALDI TOF assay
uses fourfold less enzyme due to the reduction in assay volume. Equally, peptide substrate consumption of the MALDI
assay is up to 300 times lower in concentration and 10 times
lower in volume, mostly due to the high sensitivity of the
mass spectrometer.
We further characterized enzymatic constants for CHKtide
(Fig. 2C) and ATP (Fig. 2D) using the MALDI-TOF assay.
Both the CHKtide and ATP Km constants were very similar
for SIK2 and comparable to the Km constants determined in
the ADP Hunter assay.
Furthermore, we determined for SIK2 a 15-point
dose-response curves for five known inhibitors (Fig. 2E).
Staurosporine is a well-known, broad, kinase inhibitor with
a high potency, whereas HG-9-91-01, MRT199665,
MRT67307, and KIN112 are all previously described as targeting SIKs at a nanomolar concentration, with HG-9-9101 being the most potent of the four.4 These inhibitors were
chosen as all five are ATP competitive, which reflects the
small molecular library of ATP mimetics that was selected
for the screen. Overall, the IC50 values were all in the low
nanomolar range and about 0.5 to 1 order of magnitude
higher potency than published biochemical assays,4 suggesting a specific trend in the MS-based assay (Suppl. Fig.
S2 and Suppl. Table S2). However, this is not unexpected
as previous IC50 values were determined by less direct
measurements, such as IL-10 messenger RNA (mRNA)
induction or cell-based methods. In cells, a higher ATP concentration and the presence of other potential targets such as
AMPKs and a number of protein tyrosine kinases, cellular
penetration of the compound, and inclusion of high concentrations of proteins that may bind to compound could contribute to differences seen for the inhibitor potencies.
Staurosporine, which serves as a control, appears significantly more potent than the four other control inhibitors
(Suppl. Fig. S2 and Fig. 2E), and we observe a poorer slope
fit, thus suggesting that other factors such as tight-binding
could be causing an overestimation in the potency of staurosporine. This could be a result of the IC50 value being near
or below the SIK2 concentration in the assay, which would
produce a tight-binding condition and therefore underrepresent the true potency of the inhibitor. Comparison of the
four other control inhibitors leads us to believe that the ADP
Hunter and MALDI-TOF assay have similar biochemical
pharmacology.

Table 1. Robustness of SIK2 Assays.
SIK2 Assay
MALDI TOF
ADP Hunter

Robust Z′

Robust s:b

0.5
0.9

3.0
2.8

MALDI TOF, matrix-assisted laser desorption/ionization time-of-flight;
s:b, signal to background; SIK2, salt-inducible kinase 2.

MALDI TOF SIK2 Assay Is Comparable to an
ADP Hunter Assay
Next, we screened the DDU small-molecule kinase library
against SIK2 using both the MALDI TOF assay and the
ADP Hunter assay. This library contains 2648 commercially available ATP mimetics describing a diverse region of
kinase pharmacophore space, consisting of nine 384-well
assay plates. Both assays were demonstrated to be robust
methods for screening, with calculated mean robust Z′ and
robust s:b values (Table 1).
The output of both screens for SIK2, expressed as percentage effect for MALDI TOF (Fig. 3A) and ADP Hunter (Suppl.
Fig. S3), exhibited a normal distribution, with calculated mean
percentage effects of 2.3 (MALDI TOF) and 7.6 (ADP Hunter)
for SIK2. Standard deviations of the whole screening data set
were comparable with 20.6% (MALDI TOF) and 14.6% (ADP
Hunter) effect. For both screens, compounds producing a statistically significant effect were defined as those producing a
percentage effect of greater than or equal to mean effect ± 3 SD
of the whole screening data set. In addition, the number of
compounds producing percentage effect greater than 40%
(Fig. 3B and Suppl. Fig. S4) was also determined for each
assay. A summary of the number of hit compounds is summarized in Table 2.
As expected for single-shot screens, the number of hits
was not identical for the two screens, but they did correlate
(R2 = 0.44) between the percentage effects measured for
compounds producing greater than 40% effect in the two
assays (Fig. 3D). With further replicates, we would expect a
tighter correlation between the hits of the two assays. Eighty
percent of the ADP Hunter assay hits (PE >40%; 59 compounds) were identified in both screens (28 with PE >3 SD).
There are also 16 hits that were only identified in the ADP
Hunter assay and a significant number of hits (45) that were
only present in the MALDI TOF assay (Fig. 3C), suggesting
that direct analysis of the peptide product might produce additional hits compared to the indirect ADP assay. Moreover, nine
compounds appeared to increase the kinase reaction rate
that were not identified in the ADP Hunter assay. These hits
appeared to be genuinely increasing the product formation
(Suppl. Fig. S5), and more future work will be needed to
characterize the differences between the ADP Hunter and
the MALDI TOF assays.

Table of Contents for the Digital Edition of SLAS Discovery - Special Issue on Advances in MALDI Mass Spectrometry for Drug Discovery - Volume 22 Issue 10 - December 2017