.
Angewandte
Communications
To test this hypothesis, we synthesized the probe 6,
a biotinylated analogue of Cyplecksin-1 (Figure 3a). The non-
brominated variant 7 was used as a negative control. We then
incubated the cytohesin-1 PH domain, full-length cytohesin-1,
and the Sec7 domain of cytohesin-2, to which no binding of 6
should occur, with increasing concentrations of 6 and 7. After
denaturing polyacrylamide gel electrophoresis (PAGE) and
western blotting, covalently attached biotin groups in the
proteins were detected with neutravidin. Indicative of the
modification of the cytohesin-1 PH domain with the biotin-
ylated Cyplecksin-1 variant 6.
As a first step to identify the Cyplecksin-reactive nucle-
ophile in the protein, the cysteine residues of cytohesin-1 PH
were alkylated by iodoacetamide. Neither the binding of
TMR-PIP2 nor its inhibition by Cyplecksins 1–3 were affected
by the treatment, arguing against cysteine residues being the
sites of the covalent attachment of Cyplecksins (Figure S6
(SI)). To unambigously identify the binding mode and binding
site of Cyplecksins a combined mass spectroscopic/crystallo-
graphic approach would be required which, however, is
beyond the scope of this work.
To analyze whether Cyplecksins inhibit membrane target-
ing of cytohesin-2 PH, nanodisks[10] were used as a membrane
surrogate. Cytohesin-2 PH bound only to PIP2-doped nano-
disks, indicating that nanodisks are suitable to reliably detect
the PIP2-dependent membrane recruitment of cytohesins
(Figure 4a). Cyplecksins 1–3 but not the control compounds 4
and 5 inhibited the binding.
Having shown that Cyplecksins inhibit the binding of
cytohesin PH domains to PIP2/PIP3 phospholipids by a cova-
lent mechanism in vitro we next sought to test the activity of
these compounds in living cells. When HeLa cells are
stimulated with insulin, insulin receptor (IR)-dependent
signaling cascades lead to production of PIP3, which stim-
ulates the translocation of cytohesins to the inner leaflet of
the plasma membrane by means of their PH domains. To
analyze the effect of Cyplecksins on this process, we trans-
fected HeLa cells with a cytohesin-2 construct that was fused
to green fluorescent protein (GFP) (Figure 4b, top row,
Cyth2-GFP). The cells were then analyzed for membrane
recruitment of cytohesin-2 GFP by confocal fluorescence
microscopy. Without insulin stimulation almost no Cyth2-
GFP can be detected at the membrane (Figure 4b, column 1)
but after insulin stimulation, Cyt2-GFP translocates to the
membrane (white arrows, column 2) and colocalizes with
membrane proteins stained by wheat germ agglutinin (white
arrows, merge, column 2). In the presence of 50 mm Cypleck-
sins 1, 2, or 3, however, no insulin-dependent Cyth2-GFP
translocation to the plasma membrane can be detected
(columns 3–5). In contrast, the inactive Cyplecksin analogues
4 and 5 at similar concentrations have no influence on the
membrane recruitment of Cyth2-GFP (columns 6 and 7). The
analysis of Cyth2-GFP membrane recruitment in a large
number of cells revealed a statistically highly significant
inhibition by Cyplecksins 1–3 (Figure 4c). These results
clearly demonstrate that Cyplecksins 1–3 effectively inhibit
the binding of PIP3 to cytohesin PH domains also in the
context of living cells.
Figure 3. Cyplecksins bind covalently to cytohesins a) Chemical struc-
ture of biotinylated Cyplecksin-2 analogues 6 and 7. b) Increasing
concentrations of 6 and its inactive analogue 7 incubated with
cytohesin-1 PH domain (top panel), full-length cytohesin 1 (Cyth1-fl;
middle panel), and cytohesin-2 Sec7 domain (bottom panel), analyzed
by denaturing PAGE. Covalently bound biotin is detected by neutravi-
din and total protein by an anti-His5 antibody. Cytohesin-1 PH domain
and full-length cytohesin-1, but not the cytohesin-2 Sec7 domain, were
found to be biotinylated. In all experiments, compound 7 was inactive.
c) Competition of biotinylated adduct formation by 6 using Cyplecksins
1–5. Cyplecksins 1–3, but not 4 and 5, competed with 6 for binding to
the cytohesin-1 PH domain; analysis was the same as in (b).
formation of the covalent adduct, cytohesin-1 PH and full-
length protein both showed concentration-dependent biotin-
ylation with 6 but not with 7, whereas cytohesin-2 Sec7
remained unmodified (Figure 3b).
To further substantiate this result, we tested whether the
biotinylation of the cytohesin-1 PH domain with 6 was
competed by non-biotinylated Cyplecksins 1–3. At concen-
trations of 50 mm or higher, Cyplecksins 1–3 resulted in
a marked reduction of the biotinylation by 6, and at 5 mm
a reduction was already detectable. Neither 4 nor 5 were
able to compete with 6 for cytohesin-1 PH domain binding,
not even at 100 mm concentrations (Figure 3c). Altogether,
these results can only be explained by a specific covalent
In conclusion, our study demonstrates that the method-
ology of aptamer displacement screening is feasible for
identifying small-molecule inhibitors of the cythohesin PH
domain. Aptamer-directed screening assays have previously
led to useful druglike inhibitors.[3b,6,11] Our screening assay,
based on RiboGreen detection, revealed a series of substi-
tuted 5-bromopyrimidine-2,4,6-triones 1–3, termed Cypleck-
sins, which inhibit phospholipid binding to PH domains of
cytohesins. Within the subset of PH domains of other proteins
tested here, Cyplecksins exhibited high selectivity for cyto-
4
ꢁ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2013, 52, 1 – 6
These are not the final page numbers!