Psoromic acid is a selective and covalent rab-prenylation inhibitor targeting autoinhibited rabggtase

Article abstract of DOI:10.1021/ja211305j

Post-translational attachment of geranylgeranyl isoprenoids to Rab GTPases, the key organizers of intracellular vesicular transport, is essential for their function. Rab geranylgeranyl transferase (RabGGTase) is responsible for prenylation of Rab proteins

Full text of DOI:10.1021/ja211305j

Article
pubs.acs.org/JACS
Psoromic Acid is a Selective and Covalent Rab-Prenylation Inhibitor
Targeting Autoinhibited RabGGTase
Celine Deraeve,†,¶,⊥ Zhong Guo,‡,#,⊥ Robin S. Bon,†,◆ Wulf Blankenfeldt,‡,○ Raffaella DiLucrezia,∥
́
Alexander Wolf,∥ Sascha Menninger,∥ E. Anouk Stigter,† Stefan Wetzel,† Axel Choidas,∥
Kirill Alexandrov,‡,# Herbert Waldmann,†,§ Roger S. Goody,‡ and Yao-Wen Wu*,‡
‡
†Department of Chemical Biology and Department of Physical Biochemistry, Max Planck Institute of Molecular Physiology,
Otto-Hahn-Strasse 11, 44227 Dortmund, Germany
§Technische Universitat Dortmund, Fachbereich Chemie, 44227 Dortmund, Germany
̈
∥Lead Discovery Center GmbH, Otto-Hahn-Strasse 15, 44227 Dortmund, Germany
S
* Supporting Information
ABSTRACT: Post-translational attachment of geranylgeranyl
isoprenoids to Rab GTPases, the key organizers of intracellular
vesicular transport, is essential for their function. Rab
geranylgeranyl transferase (RabGGTase) is responsible for
prenylation of Rab proteins. Recently, RabGGTase inhibitors
have been proposed to be potential therapeutics for treatment
of cancer and osteoporosis. However, the development of
RabGGTase selective inhibitors is complicated by its structural
and functional similarity to other protein prenyltransferases.
Herein we report identification of the natural product
psoromic acid (PA) that potently and selectively inhibits
RabGGTase with an IC50 of 1.3 μM. Structure−activity relationship analysis suggested a minimal structure involving the
depsidone core with a 3-hydroxyl and 4-aldehyde motif for binding to RabGGTase. Analysis of the crystal structure of the
RabGGTase:PA complex revealed that PA forms largely hydrophobic interactions with the isoprenoid binding site of
RabGGTase and that it attaches covalently to the N-terminus of the α subunit. We found that in contrast to other protein
prenyltransferases, RabGGTase is autoinhibited through N-terminal His2 coordination with the catalytic zinc ion. Mutation of
α
αHis dramatically enhances the reaction rate, indicating that the activity of RabGGTase is likely regulated in vivo. The covalent
binding of PA to the N-terminus of the RabGGTase α subunit seems to potentiate its interaction with the active site and explains
the selectivity of PA for RabGGTase. Therefore, psoromic acid provides a new starting point for the development of selective
RabGGTase inhibitors.
two (in a few cases one) geranylgeranyl moieties to two
cysteine residues at the C-terminus of Rab proteins. Unlike the
other protein prenyltransferases, farnesyltransferase (FTase)
and geranylgeranyltransferase I (GGTase I), which recognize a
C-terminal CAAX box, RabGGTase requires an accessory Rab
escort protein (REP) to recruit Rab proteins with diversified C-
terminal sequences.
Inhibition of RabGGTase has been proposed to be
exploitable for the treatment of several human diseases. A
weak but specific RabGGTase inhibiting phosphonocarboxylate
(PC) has been identified and was considered as a lead drug
candidate for the treatment of thrombotic disorders and
osteoporosis and for antitumor therapy.3,4 More recent studies
have shown that inhibition of RabGGTase induces p53-
independent apoptosis in tumor cells and validated this enzyme
as a promising target for cancer chemotherapy.5 It was
INTRODUCTION
■
Post-translational modification of proteins with isoprenoids,
known as prenylation, is a vital process affecting hundreds of
proteins controlling signal transduction and cargo transport in
eukaryotic cells. This process involves the covalent attachment
of either a 15-carbon farnesyl or a 20-carbon geranylgeranyl
group to conserved C-terminal cysteine residues. The lipid
modification enables prenylated proteins to reversibly associate
with intracellular membranes.1 The major substrates of protein
prenylation include the signal transducing Ras superfamily of
small GTPases. Rab GTPases, with over 60 members in
humans, constitute the largest subgroup of the Ras superfamily.
Rab GTPases play key roles in the regulation of intracellular
membrane trafficking through association with specific
membranes and the recruitment of an array of specific effector
proteins responsible for transport, tethering/docking, and
fusion of vesicles.2
Prenylation of Rab proteins is mediated by Rab geranylger-
anyltransferase (RabGGTase). RabGGTase covalently attaches
Received: December 7, 2011
Published: April 5, 2012
© 2012 American Chemical Society
7384
dx.doi.org/10.1021/ja211305j | J. Am. Chem. Soc. 2012, 134, 7384−7391

Journal of the American Chemical Society
Article
Figure 1. Structures of PA family members identified as RabGGTase inhibitors.
Table 1. IC50 Values and Selectivity for the Inhibition of RabGGTase by PA and Analogues*
*
IC50 values were determined by at least three independent measurements by means of (a) a fluorometric assay (plate assay) for RabGGTase and
(b) a SDS-PAGE assay (gel assay) for the three protein prenyltransferases. (c) Dose-dependent inhibition was observed in this case but was found to
be saturated at 50% of inhibition.
7385
dx.doi.org/10.1021/ja211305j | J. Am. Chem. Soc. 2012, 134, 7384−7391

Journal of the American Chemical Society
Article
Figure 2. Inhibition of RabGGTase by PA 1. (A) Lineweaver−Burk plot with NBD-FPP as varied substrate at indicated PA concentrations. The
■
●
velocity, v, is expressed in arbitrary units. (B) Titration of 400 nM NBD-FPP in the absence ( ) and the presence of 2 μM PA ( ) with
RabGGTase. The data were fitted with a quadratic equation to give a Kd of 73 22 nM for the binding of NBD-FPP to RabGGTase and an apparent
Kadpp value of 1.9 0.1 μM for the binding in the presence of PA. (C) The same data as in (B) in the presence of PA were fitted by numerical
simulation to a competitive titration model to give a genuine Kd value of 122 6 nM for the RabGGTase:PA interaction.
Scheme 1. Synthesis of PA analogues via modification of the natural product (A) and by de novo synthesis in the lactone series
(B) or in the lactam series (C)*
*
(a) NaClO2, NaH2PO4, DMSO/H2O, rt, 72 h, 53%; (b) RONH2·HCl, pyridine, rt, 4−24 h, 47% to quant.; (c) (COCl)2, CH2Cl2, rt, 16 h; (d) N-
methylpropan-1-amine, CH2Cl2, rt, 1 h, 29% over 2 steps; (e) CuO, K2CO3, pyridine, 130 °C, 48 h, 54−77%; (f) H2, Pd/C, EtOAc, rt, 6 h; (g)
SOCl2, Et2O, rt, 24 h, 64−96% over 2 steps; (h) NBS, AIBN, CCl4, 70 °C, 8 h, 59−90%; (i) dioxane/H2O, 100 °C, 6 h, 51% to quant.; (j) PCC,
CH2Cl2, rt, 7 h, 64−92%; (k) LiCl, DMF, 130 °C, 8 h, 72%; (l) K2CO3, DMF, 60 °C, 16 h, 65%; (m) Pd(OH)2/C, HCOONH4, EtOH 60 °C, 2 h,
79%; (n) H2SO4 conc., DMF, 150 °C, 16 h, 69%; (o) BuLi, TMEDA, THF, 65 °C, 2 h, then DMF, 65 °C, 2 h, 42%; (p) BBr3, CH2Cl2, rt, 1 h, 22%;
(q) NaBH4, EtOH, rt, 16 h, 37%; (r) NaH, MeI, DMF, rt, 16 h, 71%;(s) BBr3, CH2Cl2, rt, 1 h, 23%. NBS, N-bromosuccinimide; AIBN,
azobisisobutyronitrile; PCC, pyridinium chlorochromate; DMF, dimethylformamide; TMEDA, tetramethylenediamine; and THF, tetrahydrofuran.
proposed that the inhibition of RabGGTase may be at least
partially responsible for the proapoptotic action of several
FTase inhibitors that are currently in late-stage clinical trials.5
In addition, overexpression of RabGGTase and a few Rab
proteins has been reported in several cancer types.6
Furthermore, overexpression of Rab25 in breast and ovarian
cancer cells was reported to increase the aggressiveness of these
cancers.6,7 Potent and selective RabGGTase inhibitors could
provide invaluable tools for the study of Rab prenylation and its
role in skeletal disorders and cancer and offer new starting
points for the development of drug candidates. However, the
availability of such compounds is rather limited.8−12 Herein we
7386
dx.doi.org/10.1021/ja211305j | J. Am. Chem. Soc. 2012, 134, 7384−7391

Journal of the American Chemical Society
Article
Figure 3. Structural analysis of RabGGTase complexed with PA and apo-RabGGTase. (A) Binding of PA to the active site of RabGGTase. The β
subunit of RabGGTase is displayed as gray ribbons. The N-terminus of RabGGTase covalently linked to PA is displayed as an orange worm. PA and
the interacting side chains of the β subunit are displayed as sticks colored by atom type. The side chains of F289 and W244 involved in
β
β
hydrophobic interaction with PA are colored in green. Blue dashed lines indicate hydrogen bonds. The water molecules involved in hydrogen
bonding are displayed as red spheres. The zinc ion is shown as a cyan sphere. (B) Superimposition of the structures of RabGGTase in complex with
PA and GGPP (PDB code: 3DST). The β subunit of RabGGTase is displayed in gray surface representation. In both figures, the α subunit and the
side chain of αMet1 have been omitted for clarity. PA is displayed as sticks colored by atom type. The geranylgernyl chain of GGPP is shown in blue,
and the pyrophosphate head is shown in red. (C) Superimposition of the structures of apo-RabGGTase (PDB code: 1DCE) and RabGGTase
complexed with PA. The β subunit of RabGGTase is displayed in gray surface representation. The α subunits of apo-RabGGTase and RabGGTase
complexed with PA are displayed as blue and orange ribbons. The side chain of apo-RabGGTase αH2 that coordinates the zinc ion is shown as blue
sticks. Because the electron density of αH2 side chain in the RabGGTase:PA complex was not visible, it is represented as an alanine. (D)
Superimposition of the structures of apo-RabGGTase, FTase and GGTase-I (PDB code: 1DCE, 1FT1, and 1N4P). The β subunit of RabGGTase is
displayed in gray surface representation. The α subunits of RabGGTase, FTase, and GGTase-I are displayed as blue, golden, and red ribbons,
respectively. The side chain of RabGGTase H2 that coordinates the zinc ion is shown as sticks colored by atom type.
α
report the identification of the natural product psoromic acid
(PA), and analogues thereof, as a new class of potent and
selective RabGGTase inhibitors. The crystal structure of the
RabGGTase:PA complex provided insights into the binding
mode and explained the selectivity of the compound.
compounds embodying the PA core structure displayed
appreciable inhibitory activity, suggesting that the PA core
defines the scaffold of a new RabGGTase inhibitor class (Figure
1). These compounds belong to the depsidones, a family of
lichen secondary metabolites. Depsidones exhibit a range of
biological activities including antioxidant, inhibition of 5-
lipoxygenase, and inhibition of HIV-1 integrase.15−17 The
apoptotic activity of PA was documented.18
We further evaluated the inhibitory activity of PA toward the
other protein prenyltransferases, FTase and GGTase-I, using a
SDS-PAGE in vitro prenylation assay. Interestingly, PA showed
selective inhibition of RabGGTase over FTase and GGTase-I
by a factor of more than 100 and 60, respectively (Table1).
The inhibition mode of PA and its affinity for RabGGTase
were then examined using the fluorometric assay using NBD-
FPP. PA was found to be a competitive inhibitor with respect to
the lipid substrate (Figure 2A). The fluorescence change upon
binding of NBD-FPP to RabGGTase was used to perform
titrations to determine binding affinities. An inhibition of NBD-
RESULTS AND DISCUSSION
■
We screened a library of 13 030 compounds including 3967
natural products, 1418 small molecules, and 7645 semisynthetic
compounds using a fluorometric Rab prenylation assay that
monitors the change of fluorescence observed upon transfer of
NBD-farnesyl pyrophophate (NBD-FPP), a fluorescent ana-
logue of the lipid substrate geranylgeranyl pyrophosphate
(GGPP).13,14 A prescreen at a concentration of 100 μM was
used to identify compounds that inhibit RabGGTase, and the
compounds that displayed more than 70% inhibition were
selected for concentration-dependent measurements. PA 1
(Figure 1) was found to be a relatively potent inhibitor of
RabGGTase, with an IC50 value of 1.4 0.1 μM. Additional
7387
dx.doi.org/10.1021/ja211305j | J. Am. Chem. Soc. 2012, 134, 7384−7391

Journal of the American Chemical Society
Article
FPP binding to RabGGTase was observed in the presence of
PA. The affinity of RabGGTase:NBD-FPP was weakened by a
factor of ca. 30 in the presence of 2 μM PA (Figure 2B). This is
in line with competitive binding of PA versus NBD-FPP. Fitting
of the data by numerical simulation with a competitive model
(Figure 3A). The N-terminal residues of the RabGGTase α
subunit (αHis2 to Glu22) are flexible, and only the electron
α
density of the main chain spanning His2 to Val6 located near
α
the active site could be traced (Figure 3C). The interaction of
the N-terminal region of the α subunit with the β subunit in the
RabGGTase:PA structure involves hydrogen bonding of the
main-chain carbonyl of Leu5 with the amide of Met285, the
led to a genuine Kd of 122
6 nM for binding of PA to
RabGGTase (Figure 2C). These results suggest that PA is a
potent and selective inhibitor of RabGGTase. In particular, the
chemotype represented by the PA core provides a promising
starting point for the development of RabGGTase inhibitors.
Preliminary structure−activity relationship (SAR) studies
indicate that the aldehyde function is required for the activity of
the natural product (Figure 1). The carboxylic acid group is
dispensable, since compounds 3 and 4 showed good inhibitory
activity toward RabGGTase.
To refine the SAR profile, a series of PA analogues was
synthesized either by modification of the natural product or by
de novo synthesis. The reactivity of the aldehyde function of
PA was used to obtain access to the 4,6-dicarboxylic acid 5 or to
obtain a series of oximes 6. The carboxylic group at the 6-
position was converted to the amide 7 (Scheme 1A). In
addition, PA analogues were synthesized via a route derived
from a previous total synthesis of PA (Scheme 1B).19,20 The
key formation of the diaryl ether bond was achieved by means
of a modified Ullmann procedure using copper(II) oxide as
catalyst. Substantial variations of the reaction conditions
(copper or palladium catalysts, base, coligands and solvent)
did not lead to general improvement of yields. Removal of the
benzyl protecting groups by palladium/carbon-catalyzed hydro-
genation and subsequent conversion of the acid into the acid
chloride led to the formation of seven-membered ring lactones.
The aldehyde function at the 4-position was then introduced by
oxidation of the methyl group through monobromination,
hydrolysis, and oxidation of the resulting primary alcohol,
followed by deprotection reactions when required.
Compounds 5 and 6 displayed significantly lower inhibitory
activity (Table 1), suggesting that the aldehyde function at the
4-position plays a decisive role in generating the potency of PA-
derived inhibitors. This prompted us to examine whether PA
could form a Schiff base with the enzyme. Subsequent ESI-MS
analysis indeed indicated that one PA molecule was attached to
each subunit of RabGGTase as well as those of FTase and
GGTase-I when these enzymes were incubated with PA
(Supporting Information, Figure S1). The single modification
of each subunit of the proteins by PA suggests that it may
selectively form a Schiff base with the N-terminal amine or a
lysine residue.
α
β
main-chain carbonyl of Gly3 with the water molecule that
α
coordinates the zinc ion, and the side chain of Met1 with
α
βTyr44 (Figure 3A).
In the apo-RabGGTase structure, the N-terminal region of
the α subunit binds to the β subunit in an extended
conformation with His2 coordinating the catalytic zinc ion
α
(Figure 3C,D).22 Such an interaction was not observed in
FTase and GGTase-I structures (Figure 3D).23 Together with
the fact that the aldehyde group is indispensable for the
inhibitory activity of PA-derived compounds (Table 1), it is
conceivable that the covalent binding of PA to the N-terminus
of RabGGTase α subunit enhances the binding of PA to the
active site. Because the observed interaction of the N-terminal
region of the α subunit was significantly altered in the
RabGGTase:PA structure compared with the apo-RabGGTase
structure (Figure 3C), the PA binding to the active site plays a
predominant role in the RabGGTase:PA interaction. Thus, it
appears that PA binding with the active site governs the
specificity of the interaction with RabGGTase. Superimposition
of the RabGGTase:PA structure with RabGGTase in complex
with the lipid substrate GGPP revealed that the phenyl ring B
of PA occupies the binding site of the second isoprenyl unit of
GGPP (Figure 3B). The structure is consistent with the
competitive binding mode as determined above.
Comparing to other so far identified RabGGTase inhibitors,
PA is the only one that displays such a binding mode, i.e.,
associates with the lipid binding site and covalently binds to the
N-terminal amine of the α subunit. The tetrahydrobenzodia-
zepine (THB) and peptide inhibitors of RabGGTase
coordinate the catalytic zinc ion, whereas PA does not. THB
and peptide inhibitors achieve selectivity toward RabGGTase
by targeting the TAG tunnel (tunnel adjacent to the GGPP
binding site), which is unique in RabGGTase.8,9,11 In contrast,
PA does not associate with the TAG tunnel. PA is much more
potent than the first selective RabGGTase inhibitor (IC50 = 560
μM), NE10790, a PC analogue of bisphosphonate risedronate
(RIS).3 Recently, a new series of PC analogue of RIS has been
developed with IC50s of 1−50 μM toward RabGGTase and no
obvious inhibition of GGTase-I at milimolar concentration.10
Allenoate-derived compounds showed low micromolar inhib-
ition of both RabGGTase and GGTase-I, but no apparent
inhibition of FTase at 100 μM.12 The submicromolar THB
inhibitors, low micromolar peptide inhibitors, and PA display
similar selectivity for RabGGTase over FTase and GGTase-
I,8,11 suggesting it is possible to develop potent and selective
RabGGTase inhibitors combining rational design and high-
throughput screening.
To gain more insight into the principles governing the
binding of PA, we attempted to determine the crystal structure
of PA-bound RabGGTase with IGG and LRR domains
truncated. In our previous work, the engineered RabGGTase
has dramatically improved crystallization properties and
displayed identical enzymatic activity to the wild-type
enzyme.9,21 Since the imine bond is not stable under X-ray
exposure, RabGGTase crystals were soaked with 1 mM PA and
were subsequently treated with 80 mM NaCNBH3 to reduce
the imine to a stable amine. The structure was determined at
2.2 Å resolution. The crystal structure revealed that PA binds
covalently to the N-terminal amine of the RabGGTase α
subunit, and the interaction of PA with RabGGTase is largely
hydrophobic, involving hydrophobic interaction of two phenyl
rings with Trp244 and Phe289 and a weak hydrogen bond
Among other compounds, only 19 and 14 were identified as
potent inhibitors of RabGGTase with IC50 values of 4.2 0.3
and 6.1
0.8 μM, respectively. These findings demonstrate
that the removal of the carboxylic acid and both methyl groups
from PA led only to a slight decrease of activity (compound 19,
Table 1 entry 23), suggesting that these functions are not
essential for RabGGTase inhibition. Similarly, the methoxy
group at the 8-position can be omitted without pronounced
influence on the activity (compound 14, Table 1 entry 18).
β
β
between the lactone carbonyl and Trp52 via a water molecule
β
7388
dx.doi.org/10.1021/ja211305j | J. Am. Chem. Soc. 2012, 134, 7384−7391

Journal of the American Chemical Society
Article
Figure 4. Single-turnover prenylation by RabGGTase variants. (A) SDS-PAGE assay using NBD-FPP. The solid lines show single exponential fits,
3
giving kobs of 0.15, 0.27, and 3.4 min−1 for wild-type RabGGTase ( ), ΔM1 ( ), and H2A ( ), respectively. (B) Prenylation assay using H -GGPP.
■
●
△
●
△
The solid lines show single exponential fits, giving kobs of 0.22 and 0.4 for wild type RabGGTase ( ) and ΔM1 ( ), respectively, and a double
exponential fit for H2A mutant ( ), giving the kobs of 13 and 0.2 min−1 for the fast and slow phases, respectively.
■
These results are in line with the structural analysis, since these
groups were not involved in apparent interactions and the
contact of PA with RabGGTase arose mainly from hydrophobic
interaction of two phenyl rings. Comparison of the results for
18 and its precursors 15−17 (Table 1, entries 19−22) again
stresses the crucial role of the aldehyde function for
RabGGTase inhibition by this class of compounds. The
hydroxyl group at the 3-position is required, since its omission
led to loss of the inhibitory activity (compound 12, Table entry
16). Substitution of the 3-hydroxyl with a methoxyl group led
to a moderate reduction of the inhibitory activity (compare 4
with 3 and 20 with 21), whereas substitution with a larger
isopropoxyl group resulted in substantial reduction of the
inhibitory activity (compare 13 with 14 and 18 with 19).
Structural analysis indicated that the 3-hydroxyl may be
involved in hydrogen bonding with the Schiff base, which
could stabilize the Schiff base (Figure 3A). These results
suggest that 14, which incorporates the depsidone core
structure and a 3-hydroxyl and 4-aldehyde motif, represents a
minimal structure in the PA series for RabGGTase inhibition.
Interestingly, simplified PA analogues 14 and 19 retained the
selectivity for RabGGTase with respect to FTase and GGTase-I
(Table 1, entries 18 and 23). Replacement of the lactone of the
depsidone core structure by a more stable lactam function was
also investigated. A series of lactam analogues was prepared on
the basis of the minimal scaffold represented by compound 14
(Scheme 1C) and was evaluated for RabGGTase inhibition
(Table 1). Such a replacement did not affect the inhibitory
potency of the compounds, (compare 14 with its lactam
analogue 21, Table 1 entries 18 and 25). As expected, the
aldehyde function is crucial for the binding (22, Table 1 entry
26). N-methylation of the lactam led to a three-fold decrease in
the inhibitory activity (23, Table 1 entry 27). Coherently the
crystal structure suggested that such a modification leads to a
lated Rab proteins (Figure S4, Supporting Information). The
low efficiency of PA to affect cellular prenylation might be
partially due to its moderate membrane permeability. As
determined by the parallel artificial membrane permeation assay
(PAMPA) using an artificial lipid bilayer, PA displayed
dramatically lower membrane permeability than the control
compound, BMS3, which has been proved to be active in cells
(Figure S5, Supporting Information).8 The THB and peptide
inhibitors were shown to inhibit RabGGTase in cells with IC50s
of 10−400 nM and 10−50 μM, respectively.8,11 The first
selective RabGGTase inhibitor, NE10790, inhibited RabGG-
Tase activity in cells with an IC50 of ∼600 μM.3 The most
potent PC analogue of bisphosphonate risedronate displayed
the lowest effective dose of 3 μM for inhibition of Rab11
prenylation in cells.10 Slight inhibition of Rab5b prenylation in
cells was observed with the treatment of 20 μM one of
allenoate-derived compounds.12 Further alterations of PA
structure to improve its cellular activity will be carried out in
future studies.
The association of the N-terminal region of the α subunit
with the active site suggests an autoinhibitory mechanism in
RabGGTase.22 Observation of the coordination of His2 with
α
the catalytic zinc ion in the apo-RabGGTase structure
prompted us to ask what the consequence of such an
interaction is. His2 is conserved in RabGGTase in mammals
α
and yeast (Supporting Information, Figure S3). However, this
histidine residue is not present in FTase and GGTase-I. To this
end, RabGGTases with N-terminal mutations were subjected to
the in vitro prenylation assay. Remarkably, the H2A mutant
α
displayed an observed rate of 3.4 min−1 in a single turnover
reaction using NBD-FPP, which was more than 20-fold faster
than the wild-type enzyme (kobs = 0.15 min−1). In contrast,
truncation of αMet1 (ΔM1) only led to a slight increase of the
reaction rate (kobs = 0.27 min−1) (Figure 4A). Similar results
were obtained in the prenylation assay using tritium-labeled
GGPP (Figure 4B). These results suggested that RabGGTase is
autoinhibited through the coordination of αHis2 with the
catalytic zinc ion. The functional role of RabGGTase
autoinhibition remains elusive. However, it appears that this
feature is exploited by PA to achieve selectivity for RabGGTase.
The question of the function or purpose of RabGGTase
autoinhibition is an intriguing one. RabGGTase is a much
steric clash with Trp244.
β
RabGGTase inhibition in HeLa cells was determined by
reprenylation assay using biotin-GPP, where unprenylated Rab
proteins in cells were detected.11 Neither cytotoxicity nor
inhibition of RabGGTase was observed when the cells were
treated with PA in various concentrations (1−300 μM). In
contrast, treatment with compactin, which inhibits the
biosynthesis of isoprenoids, led to accumulation of unpreny-
7389
dx.doi.org/10.1021/ja211305j | J. Am. Chem. Soc. 2012, 134, 7384−7391

Journal of the American Chemical Society
Article
parts of the progress curve. Compounds causing a residual
RabGGTase activity of less than 30% were selected for IC50
determination. Concentration-dependent inhibition of RabGGTase
was measured using different concentrations of inhibitors resulting
from a two-fold dilution series. Reaction progress at every
concentration of inhibitor was measured in triplicate, and 0% activity
controls were measured for every inhibitor at every concentration. IC50
values were calculated using a four-parameter equation.14
slower enzyme than FTase or GGTase-I, but the results
described here demonstrate that a single amino acid
substitution dramatically increases the reaction rate. It therefore
appears likely that the low intrinsic activity is of physiological
significance. One possibility is that maximal activity is only
developed on interaction with an as yet unidentified activation
factor in vivo.
SDS-PAGE Assay (Gel Assay) To Investigate Selectivity of
Inhibitors. In the prenylation buffer (50 mM HEPES, pH 7.2, 50 mM
NaCl, 2 mM MgCl2, 2 mM DTE, and 0.01% Triton X-100), a 10 μL
mixture of FTase with NBD-GPP, GGTase-I with NBD-FPP, and
RabGGTase with NBD-FPP in the presence and the absence of
various concentrations of inhibitor was incubated for 5 min at 37 °C,
the same volume of protein substrate GST-KRas, GST-RhoA, and
Rab7 with REP-1 was added to initiate the reaction. For the FTase
assay, 3 μM GST-KiRas, 2 μM NBD-GPP, and 20 nM FTase, for the
GGTase-I assay 6 μM GST-RhoA, 6 μM NBD-FPP, and 1 μM
GGTase-I, for the RabGGTase assay 3 μM Rab, 3 μM REP-1, 2 μM
NBD-FPP, and 40 nM RabGGTase were used, respectively. In the
negative controls, the enzymes were omitted. Reactions were run for
the same period of time (10 min) and quenched by adding cold 2 ×
SDS sample buffer. The reaction mixtures were resolved by 15% SDS-
PAGE, and the gel was scanned using a Fluorescent Image Reader
FLA-5000 (Fujifilm) (excitation laser at 473 nm and emission cutoff
filter at 510 nm). The fluorescent bands corresponding to the
formation of the prenylated product were quantified using the AIDA
densitometry software. The data were processed in the same way as in
the plate assay.
CONCLUSIONS
■
In conclusion, we have identified the natural product PA as a
new potent and selective inhibitor of RabGGTase. SAR analysis
with analogues of PA led to identification of a minimal
structure, involving the depsidone core structure and a 3-
hydroxyl and 4-aldehyde motif, for selective inhibition of
RabGGTase. The crystal structure of RabGGTase in complex
with PA revealed that the interaction involves hydrophobic
interactions of PA with the isoprenoid binding site and
hydrogen bonding between the N-terminal region of the α and
β subunits. We conclude that the covalent binding of PA to the
N-terminus of the α subunit enhances the binding of PA to the
active site. However, neither cytotoxicity nor inhibition of
RabGGTase was observed in cells treated with PA up to 300
μM. Biochemical experiments confirmed that the autoinhibition
of RabGGTase is resulted from coordination of N-terminal
αHis2 with the catalytic zinc ion. Taken together, the SAR
profile and mapping of the interactions between PA and the
enzyme in the crystal structure of the complex provide essential
guidance for the design of new RabGGTase inhibitors.
In vitro Prenylation Assay of RabGGTase Variants. A mixture
of 10 μM Rab7 and RabGGTase variants with 100 μM NBD-FPP in
prenylation buffer (50 mM HEPES, pH 7.2, 50 mM NaCl, 5 mM
DTE, 2 mM MgCl2, and 20 μM GDP) was added to 10 μM REP-1 to
start the reaction. In the negative control, the REP-1 was omitted. At
defined time intervals, 10 μL of reaction mixture were withdrawn, and
the reactions were quenched by adding 10 μL 2 × SDS-loading buffer.
After boiling at 95 °C for 5 min, the samples were resolved by 15%
SDS-PAGE, and fluorescence intensities of the bands corresponding to
the NBD-farnesylated Rab7 were measured and quantified as
described above.
METHODS
■
Synthesis and Characterization of the Compounds. For full
details, refer to the Supporting Information.
Protein Expression and Purification. Rat REP-1 was cloned into
pFastBacHTb (Invitrogen) with N-terminal His tag and was expressed
in SF21 cells. The α and β subunits of rat FTase, GGTase-I, and
RabGGTase variants were cloned into pGATEV and pET vectors,
respectively. Both subunits were coexpressed in Escherichia coli.
Human K-Ras, human RhoA, and canine Rab7 were cloned into
pGATEV vector as N-terminal His-GST fusions and were expressed in
E. coli. All proteins were purified by nickel-nitrilotriacetic acid affinity
chromatography. After cleavage by tobacco etch virus (TEV) protease,
the proteins were further purified by size exclusion chromatography.
Fluorometric Assay (Plate Assay). All solution-based enzyme
assays were performed by means of an automated system consisting of
a Zymark SciClone ALH 500 in conjunction with a Tecan 18 Genius
Pro reader, using black 384-well plates. All inhibition measurements
were performed in triplicate and measured under the following
conditions: 37 °C, buffer: 50 mM HEPES, pH 7.2, 50 mM NaCl, 2
mM MgCl2, 2 mM DTE, 20 μM GDP, and 0.01% Triton X-100. Final
concentrations: 3 μM Rab, 3 μM REP, 400 nM RabGGTase, 9 μM
NBD-FPP 60, and 100 μM test compound (from a 10 mM solution in
DMSO). For the 100% activity control, instead of the test compound,
1% of DMSO was added. A 0% activity control was measured for every
screened compound to enable correction of potential fluorophore
bleaching and nonspecific effects of the test compound on the
fluorescent signal. The 0% activity control wells contained 400 nM
RabGGTase, 9 μM NBD-FPP, and 100 μM test compound in 10 μL of
assay buffer. On every plate, a reference inhibitor was used for
interplate control. After mixing of the fluorescent substrate, the
enzyme, and the compound solutions (or DMSO), the reaction
mixtures were incubated for 5 min at 37 °C. Then, the protein
substrates (Rab7 and REP1) were added and fluorescence intensity
was recorded for 30 min at intervals of 14 s (λex = 479 nm, λem = 520
nm). Data were evaluated using Microsoft Excel. The fluorescence
intensity at individual time points was normalized with respect to the
0% activity control, and reaction rates were determined from the linear
The prenylation assay using radioactive H3-GGPP, which is the
native lipid substrate, was performed with these RabGGTase variants.
In the prenylation buffer, 2 μM of REP-1 was added to a mixture of 2
μM Rab7, RabGGTase,and H3-GGPP and 8 μM unlabeled GGPP to
start the reaction. At each time point, 20 μL of the reaction mixture
was subjected to nitrocellulose paper. After 3× washing with 0.75%
H3PO4, the intensity of radioactivity on the paper was measured under
a Wallac MicroBeta TriLux 1450 LSC and Luminiscence Counter
(PerkinElmer) and was quantified by a MicroBeta Windows
Workstation (version V4.70.05).
Data Collection and Structural Determination of the
RabGGTase:PA Complex. RabGGTaseΔLLRΔIG, in which the
LRR and IG domain were deleted, was crystallized as described
before.9 The RabGGTase crystals were soaked in cryo-protecting
buffer containing 1 mM PA for 1 h and were subsequently treated with
80 mM NaCNBH3 overnight. Diffraction data were collected at 100 K
at station X10SA of the Swiss Light Source (SLS, Villigen,
Switzerland) or at a copper rotating X-ray anode with an osmic
mirror (λ = 1.5419 Å, 50 kV, 100 mA, 0.1 mm collimator) (in-house).
Structures were determined at 2.2 Å resolution using in-house X-ray
sources. Full data collection and refinement statistics are summarized
in Supporting Information Table S1. |FOFC| difference electron density
is shown in Supporting Information Figure S2. The coordinate has
been deposited in the Protein Data Bank with accession number
4EHM.
7390
dx.doi.org/10.1021/ja211305j | J. Am. Chem. Soc. 2012, 134, 7384−7391

Journal of the American Chemical Society
Article
Blankenfeldt, W.; Alexandrov, K.; Waldmann, H. Angew. Chem., Int. Ed.
2008, 47, 3747−3750.
(10) McKenna, C. E.; Kashemirov, B. A.; Blazewska, K. M.; Mallard-
Favier, I.; Stewart, C. A.; Rojas, J.; Lundy, M. W.; Ebetino, F. H.;
Baron, R. A.; Dunford, J. E.; Kirsten, M. L.; Seabra, M. C.; Bala, J. L.;
Marma, M. S.; Rogers, M. J.; Coxon, F. P. J. Med. Chem. 2010, 53,
3454−3464.
(11) Tan, K. T.; Guiu-Rozas, E.; Bon, R. S.; Guo, Z.; Delon, C.;
Wetzel, S.; Arndt, S.; Alexandrov, K.; Waldmann, H.; Goody, R. S.;
Wu, Y. W.; Blankenfeldt, W. J. Med. Chem. 2009, 52, 8025−8037.
(12) Watanabe, M.; Fiji, H. D.; Guo, L.; Chan, L.; Kinderman, S. S.;
Slamon, D. J.; Kwon, O.; Tamanoi, F. J. Biol. Chem. 2008, 283, 9571−
9579.
ASSOCIATED CONTENT
* Supporting Information
Full experimental detail, synthesis and characterization of
compounds, ESI-MS measurements of covalent modifications,
X-ray data collection, and refinement statistics, and |FOFC|
difference electron density map. This material is available free
of charge via the Internet at http://pubs.acs.org.
■
S
AUTHOR INFORMATION
Corresponding Author
yaowen.wu@mpi-dortmund.mpg.de
■
Present Addresses
(13) Wu, Y. W.; Waldmann, H.; Reents, R.; Ebetino, F. H.; Goody, R.
S.; Alexandrova, K. ChemBioChem 2006, 7, 1859−1861.
(14) Wu, Y. W.; Alexandrov, K.; Brunsveld, L. Nat. Protoc. 2007, 2,
2704−2711.
#Institute for Molecular Bioscience The University of Queens-
land, Brisbane, 306 Carmody Road, St Lucia, QLD,
4072, Australia
○Lehrstuhl fur Biochemie, Universitat Bayreuth, Uni-
(15) Neamati, N.; Hong, H.; Mazumder, A.; Wang, S.; Sunder, S.;
Nicklaus, M. C.; Milne, G. W.; Proksa, B.; Pommier, Y. J. Med. Chem.
1997, 40, 942−951.
̈
̈
versitatsstraße 30, 95447 Bayreuth, Germany
̈
◆School of Chemistry and Biomedical and Health Research
Centre, University of Leeds, Leeds, LS2 9JT, United Kingdom
¶Laboratoire de Chimie de Coordination du CNRS, 205 route
de Narbonne, 31077 Toulouse, France
(16) Hidalgo, M. E.; Fernandez, E.; Quilhot, W.; Lissi, E.
Phytochemistry 1994, 37, 1585−1587.
(17) Muller, K. Appl. Microbiol. Biotechnol. 2001, 56, 9−16.
(18) Correche, E. R.; Enriz, R. D.; Piovano, M.; Garbarino, J.;
Gomez-Lechon, M. J. Altern. Lab Anim. 2004, 32, 605−615.
(19) Sala, T.; Sargent, M. V. J. Chem. Soc., Perkin Trans. 1 1979,
2593−2598.
Author Contributions
⊥These authors contributed equally.
Notes
(20) Kulkarni, N. N.; Kulkarni, V. S.; Lele, S. R.; Hosangadi, B. D.
Tetrahedron 1988, 44, 5145−5149.
The authors declare no competing financial interest.
(21) Guo, Z.; Wu, Y. W.; Das, D.; Delon, C.; Cramer, J.; Yu, S.;
Thuns, S.; Lupilova, N.; Waldmann, H.; Brunsveld, L.; Goody, R. S.;
Alexandrov, K.; Blankenfeldt, W. EMBO J. 2008, 27, 2444−2456.
(22) Zhang, H.; Seabra, M. C.; Deisenhofer, J. Structure 2000, 8,
241−251.
ACKNOWLEDGMENTS
This work was supported by the Deutsche Forschungsgemein-
schaft, the Max-Planck-Gesellschaft, and the Zentrum fur
Angewandte Chemische Genomik. C.D. and R.S.B. thank the
Alexander von Humboldt-Stiftung for fellowships. We thank
the X-ray communities of the Max Planck Institute of
Molecular Physiology (Dortmund, Germany) and the Max
Planck Institute of Medical Research (Heidelberg, Germany)
for collecting diffraction data at the Swiss Light Source of the
Paul Scherrer Institute (Villigen, Switzerland) and for giving us
generous access to and support for station X10SA.
■
̈
(23) Lane, K. T.; Beese, L. S. J. Lipid Res. 2006, 47, 681−699.
REFERENCES
■
(1) Konstantinopoulos, P. A.; Karamouzis, M. V.; Papavassiliou, A. G.
Nat. Rev. Drug Discovery 2007, 6, 541−555.
(2) Stenmark, H. Nat. Rev. Mol. Cell Biol. 2009, 10, 513−525.
(3) Coxon, F. P.; Helfrich, M. H.; Larijani, B.; Muzylak, M.; Dunford,
J. E.; Marshall, D.; McKinnon, A. D.; Nesbitt, S. A.; Horton, M. A.;
Seabra, M. C.; Ebetino, F. H.; Rogers, M. J. J. Biol. Chem. 2001, 276,
48213−48222.
(4) Roelofs, A. J.; Hulley, P. A.; Meijer, A.; Ebetino, F. H.; Russell, R.
G.; Shipman, C. M. Int. J. Cancer 2006, 119, 1254−1261.
(5) Lackner, M. R.; Kindt, R. M.; Carroll, P. M.; Brown, K.; Cancilla,
M. R.; Chen, C.; de, S. H.; Franke, Y.; Guan, B.; Heuer, T.; Hung, T.;
Keegan, K.; Lee, J. M.; Manne, V.; O’Brien, C.; Parry, D.; Perez-Villar,
J. J.; Reddy, R. K.; Xiao, H.; Zhan, H.; Cockett, M.; Plowman, G.;
Fitzgerald, K.; Costa, M.; Ross-Macdonald, P. Cancer Cell 2005, 7,
325−336.
(6) Mitra, S.; Cheng, K. W.; Mills, G. B. Semin. Cell Dev. Biol. 2011,
22, 57−68.
(7) Cheng, K. W.; Lahad, J. P.; Kuo, W. L.; Lapuk, A.; Yamada, K.;
Auersperg, N.; Liu, J.; Smith-McCune, K.; Lu, K. H.; Fishman, D.;
Gray, J. W.; Mills, G. B. Nat. Med. 2004, 10, 1251−1256.
(8) Bon, R. S.; Guo, Z.; Stigter, E. A.; Wetzel, S.; Menninger, S.;
Wolf, A.; Choidas, A.; Alexandrov, K.; Blankenfeldt, W.; Goody, R. S.;
Waldmann, H. Angew. Chem., Int. Ed. 2011, 50, 4957−4961.
(9) Guo, Z.; Wu, Y. W.; Tan, K. T.; Bon, R. S.; Guiu-Rozas, E.;
Delon, C.; Nguyen, T. U.; Wetzel, S.; Arndt, S.; Goody, R. S.;
7391
dx.doi.org/10.1021/ja211305j | J. Am. Chem. Soc. 2012, 134, 7384−7391