Development of selective RabGGTase inhibitors and crystal structure of a RabGGTase-inhibitor complex

Article abstract of DOI:10.1002/anie.200705795

Stopping the transfer: Based on the structure of pepticinnamin E, specific inhibitors of Rab geranylgeranyl transferase (RabGGTase) with activity in cells were developed, and the first crystal structure of the enzyme in complex with an inhibitor is reported (see inhibitor structure and positioning in the active site of the enzyme). The findings may have implications for the chemical-biological study of Rab prenylation and vesicular transport and the involvement of RabGGTase in the establishment of disease. (Chemical Equation Presented)

Full text of DOI:10.1002/anie.200705795

Angewandte
Chemie
DOI: 10.1002/anie.200705795
Enzyme Inhibitors
Development of Selective RabGGTase Inhibitors and Crystal Structure
of a RabGGTase–Inhibitor Complex**
Zhong Guo, Yao-Wen Wu, Kui-Thong Tan, Robin S. Bon, Ester Guiu-Rozas, Christine Delon,
Uyen T. Nguyen, Stefan Wetzel, Sabine Arndt, Roger S. Goody, Wulf Blankenfeldt,
Kirill Alexandrov,* and Herbert Waldmann*
Rab guanosine triphosphatases (GTPases) constitute the
most prominent branch of the Ras superfamily of GTPases
and are responsible for a broad range of intracellular
trafficking events such as vesicle formation, vesicle and
organelle motility, and tethering of vesicles to their target
compartments.[1,2] They require the covalent attachment of
two geranylgeranyl groups to their C-terminal cysteine
residues for biological activity. This modification is catalyzed
by Rab geranylgeranyl transferase (RabGGTase, GGTase II)
which modifies all 60 mammalian RabGTPases. In contrast to
other protein prenyltransferases, such as farnesyltransferase
(FTase) and geranylgeranyl transferase I (GGTase I),
RabGGTase does not recognize its protein substrate directly
but functions in concert with a protein named Rab escort
protein (REP).[3] Although numerous effective and specific
inhibitors of FTase[4] and GGTase I[5] are known, only one
specific but weak phosphonocarboxylate inhibitor of
RabGGTase has been reported so far.[6] This is despite the
biological importance of RabGGTase and its involvement in
the establishment of several diseases. Thus, the aforemen-
tioned phosphonocarboxylate was considered to be a lead
compound for the development of new treatments for
thrombotic disorders and excessive osteoclast-mediated
bone resorption, which can cause tumor-induced osteolysis
and postmenopausal osteoporosis.[6] More recent studies
indicate that inhibition of RabGGTase induces p53-inde-
pendent apoptosis and additionally validate this enzyme as a
promising target for anticancer therapy.[7] According to the
findings reported by Ross-Macdonald and co-workers,[7]
RabGGTase may be responsible for the proapoptotic activity
of the farnesyltransferase inhibitors[4] currently in late-stage
clinical trials.[8]
To study the role of RabGGTase in skeletal disorders and
cancer and, in general, the biological function of Rab proteins,
potent and selective inhibitors of the enzyme with activity in
cells would be invaluable. However, such compounds are
currently not accessible.[7] Their development would be
greatly facilitated by the availability of a crystal structure of
RabGGTase in complex with such an inhibitor. However,
such a structure is also currently not available. Here we report
the identification of such compounds and the first crystal
structure of RabGGTase in complex with an inhibitor.
For the development of inhibitors displaying the proper-
ties detailed above, we have drawn on our earlier findings that
peptides modeled on the naturally occurring FTase inhibitor
[*] Dr. K.-T. Tan, Dr. R. S. Bon, Dr. E. Guiu-Rozas, Dipl.-Chem. S. Wetzel,
Dr. S. Arndt, Prof. Dr. H. Waldmann
Max-Planck-Institut für molekulare Physiologie
Abt. Chemische Biologie
Otto-Hahn-Strasse 11, 44227 Dortmund (Germany)
and
TU Dortmund, Fachbereich Chemie
44227 Dortmund (Germany)
Fax: (+49)231-133-2499
E-mail: herbert.waldmann@mpi-dortmund.mpg.de
pepticinnamin E
may
also
inhibit
RabGGTase
(Scheme 1).[9–11] In order to gain insight into the structural
parameters determining inhibition by these tripeptide deriv-
atives, we assembled a library of 469 further peptides and
characterized them in an in vitro Rab prenylation assay.
The tripeptide library was synthesized according to
methods reported previously (Scheme 1; see also the Sup-
porting Information).[9] Structure variation included different
amino acids (for example, Ala, Leu, His, Tyr, Ser, Thr, Lys,
Glu, Gln), various long- and short-chain aliphatic, olefinic, or
(hetero)aromatic amides at the N terminus, and carboxylic
acid, several esters, or various amides at the C terminus; the
structural diversity of the collection was thereby guaranteed
(see Table 1 for examples).
Z. Guo,[+] Y.-W. Wu,[+] Dr. C. Delon, Dipl.-Chem. U. T. Nguyen,
Prof. Dr. R. S. Goody, Dr. W. Blankenfeldt, Dr. K. Alexandrov
Max-Planck-Institut für molekulare Physiologie, Abt. Physikalische
Biochemie, Otto-Hahn-Strasse 11, 44227 Dortmund (Germany)
Fax: (+49)231-133-2399
E-mail: kirill.alexandrov@mpi-dortmund.mpg.de
[+] These authors contributed equally to this work.
[**] RabGGTase: Rab geranylgeranyl transferase. We thank the X-ray
communities of the Max-Planck-Institut für molekulare Physiologie
(Dortmund, Germany) and the Max-Planck-Institut für medizini-
sche Forschung (Heidelberg, Germany) for collecting diffraction
data at the Swiss Light Source of the Paul Scherrer Institute
(Villigen, Switzerland) and for giving us generous access and
support for the station X10SA. K.A. was supported by a Heisenberg
Award of the Deutsche Forschungsgemeinschaft (DFG). This work
was supported in part by DFG grants to K.A. (grant no.: AL 484/7-2)
and to K.A., R.S.G, and H.W. (grant no.: SFB642), and by the
Zentrum für Angewandte Chemische Genomik. R.S.B. thanks the
Alexander von Humboldt Stiftung for a scholarship.
After release from the solid support, the compounds were
purified by HPLC and isolated in analytically pure form (see
the Supporting Information for representative HPLC traces).
For the evaluation of the tripeptides as RabGGTase inhib-
itors, we employed a recently developed fluorometric in vitro
Rab prenylation assay[11] that monitors the change in fluores-
Supporting information for this article is available on the WWW
under http://www.angewandte.org or from the author.
Angew. Chem. Int. Ed. 2008, 47, 3747 –3750
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
3747

Communications
histidines and tyrosines, especially as AA1 and AA3, are the
best RabGGTase inhibitors. Except for compound 4, all of the
most potent compounds contain a tyrosine residue as AA3. In
most of the potent inhibitors, AA2is an ( N-methylated)phe-
nylalanine or a histidine. The data indicate that polar C
termini like free carboxylic acids (entries 1 and 2), hydroxa-
mic acids (entries 3 and 4), N-methyl-N-(pyridine-2-yl)ethyl-
amides (entries 5–8), or N-aminobutylamides (entries 9 and
10) are beneficial for RabGGTase inhibition. Furthermore, a
Cbz group (entries 1–4) or long lipophilic chains (entries 5–
10) at the N terminus improve binding to RabGGTase. In
general, the introduction of simple alkyl esters at the C
terminus, shorter alkyl chains at the N terminus, or small (Gly,
Ala), hydrophobic (Val, Leu, Pro), hydrophilic (Ser, Thr,
Gln), or charged (Lys, Glu) amino acids leads to a substantial
decrease in inhibitory potency (data not shown).
Scheme 1. Solid-phase approach for the synthesis of a library of
potential peptide-based RabGGTase inhibitors. Fmoc: 9-fluorenyl-
methoxycarbonyl; SPPS: solid-phase peptide synthesis.
The identified compounds were evaluated for their ability
to inhibit all three prenyltransferases in a sodium dodecyl-
sulfate (SDS) PAGE end-point in vitro prenylation assay
(Table 2; see also the Supporting Information).[10] These
Table 1: Results of the solid-phase synthesis and inhibition of RabGG-
Tase for a selected group of peptides.
Table 2: IC50 values determined by means of SDS-PAGE end-point
assays.
Entry Cmpd.
R6–AA1–AA2–AA3–R1
R6[b] AA1 AA3 R1[b]
AA2
IC50 [mm][a]
Entry
Inhibitor
RabGGTase
FTase
GGTase I
1
2
3
4
5
6
7
8
9
10
1
2
3
4
5
6
7
8
9
10
Cbz d-Tyr Phe
Tyr
OH
OH
NHOH
NHOH
A
A
A
A
B
B
4.1Æ0.3
22.7Æ1.8
9.0Æ1.0
5.2Æ0.7
2.8Æ0.4
4.7Æ1.0
6.3Æ0.7
11.0Æ1.2
5.2Æ0.8
7.2Æ0.3
IC50 [mm]
IC50 [mm]
IC50 [mm]
CbzHis
CbzHis
CbzHis
(
(
(
NMe)Phe Tyr
NMe)Phe Tyr
NMe)Phe Trp
1
2
3
4
5
6
7
8
8.8Æ0.7
2.8Æ0.1
4.3Æ0.4
10.0Æ0.9
98Æ4.7
97Æ31
>100
>100
60Æ5.3
–
–
I
His
His
Tyr
His
His
Tyr
Phe
Phe
His
His
His
His
Tyr
Tyr
Tyr
Tyr
Tyr
Tyr
35Æ5.8
II
II
II
II
II
experiments demonstrated that compounds 5–7 are 10–30-
fold selective for RabGGTase, while the selectivity of 8 was
3.5–6-fold (see Table 2). In addition, compound 5 is selective
for RabGGTase over FTase by a factor of at least 10.
Unfortunately, the IC50 values of 6 and 7 for FTase were not
extractable from the assay. These compounds showed dose-
dependent inhibition of FTase, but the inhibition was
saturated at less than 50% (see the Supporting Information,
Figure S2), which was not observed in the case of the other
two prenyltransferases. This finding suggests that these
compounds partially inhibit FTase by a mixed competitive
and noncompetitive mechanism. Thus, inhibitors 5–8 show
moderate to fairly good selectivity for RabGGTase over
FTase and GGTase I and can be regarded as the first known
selective low-micromolar RabGGTase inhibitors.
[a] All IC50 values were determined by at least three independent
measurements. [b] Cbz: benzyloxycarbonyl.
cence of a fluorescent analogue of geranylgeranylpyrophos-
phate (GGPP) upon transfer to a Rab protein. Briefly, in this
analogue, the terminal isoprene unit of the geranylgeranyl
group is replaced by a nitrobenzoxadiazole (NBD) fluoro-
phore
to
yield
{3,7,11-trimethyl-12-(7-nitrobenzo-
[1,2,5]oxadiazo-4-ylamino)-dodeca-2,6,10-trien-1} pyrophos-
phate (NBD-FPP; see the Supporting Information for the
structure).
The IC50 values provide an indication of the inhibitory
activity of the identified tripeptides under the chosen
conditions but do not provide direct information on either
their mode of action or their affinity for RabGGTase.
Therefore, a series of fluorometric titrations was performed
with 8 by using the fluorescence of NBD-FPP bound to
RabGGTase as a reporter of the enzymeꢀs interaction with
the inhibitor (see the Supporting Information, Figure S3).[10]
An example of the results of RabGGTase–NBD-complex
titration with 8 is shown in Figure 1. Fitting of the data to the
competitive model with 1:1 stoichiometry of the complex led
to a Kd value of 1.0 Æ 0.08 mm (Figure 1).
During the final step of the catalysis, binding of the NBD-
farnesylated C terminus of Rab to REP leads to a dramatic
increase in the fluorescence of NBD, thereby providing a
convenient readout of the reaction. This assay was adapted
for automated screening (for details, see the Supporting
Information). After an initial prescreen at a fixed concen-
tration to identify potential inhibitors, the most potent
compounds were selected for concentration-dependent inhib-
ition measurements (Table 1). In general, tripeptides rich in
3748
www.angewandte.org
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2008, 47, 3747 –3750

Angewandte
Chemie
Figure 1. Analysis of the interaction of compound 8 with RabGGTase
by using the fluorescence of bound NBD-FPP as a reporter. Titration of
NBD-FPP-modified RabGGTase with 8. The data were fitted by
numerical simulation to a competitive model to give a dissociation
constant of 8 from RabGGTase of Kd =(1.0Æ0.08) mm. (The Kd value
for NBD-FPP was fixed at 163 nm). The concentration of NBD-FPP was
400 nm and of RabGGTase was 650 nm (see Figure S3 in the
Supporting Information for more experimental details). NBD fluores-
cence was excited at 479 nm and the data were collected at 547 nm.
Figure 2. Structural analysis of truncated RabGGTase in complex with
2. A) Surface representation of RabGGTase. The a and b subunits are
shown in gray and yellow, respectively. Compound 2, shown as a
Corey–Pauling–Koltun model in red, binds to the substrate-binding
site. B) Interactions of compound 2 with the active site of RabGGTase.
The b subunit of RabGGTase is displayed as gray tubes. Compound 2
and interacting side chains of the b subunit are displayed as sticks
colored by atom type. Blue dotted lines indicate hydrogen bonds. The
Zn2+ ion is shown as a purple sphere. C) Superimposition of the
structure of RabGGTase in complex with 2 and the structure of
GGTase I in complex with substrate peptide KICVIL and a nonhydrolyz-
able analogue of GGPP (Protein DataBank (PDB) access code: 1N4Q).
RabGGTase is displayed as in (B), while the ligands are displayed as
ball-and-stick models. The GGTase I substrate peptide is colored
green, compound 2 is colored according to atom type, and the GGPP
analogue is shown in dark orange. D) Surface representation of the
active site of RabGGTase in complex with 2. Surface residues of the
active site are colored as follows: hydrophobic in yellow, polar in pink,
positively charged in blue, and negatively charged in red. Highlighted
areas represent sites that could be used for additional anchoring of
the inhibitor.
In order to gain insight into the molecular details of the
interaction between RabGGTase and the identified inhib-
itors, we used a cocrystallization approach to obtain an
enzyme–inhibitor complex structure. After substantial exper-
imentation, crystals of the complex of compound 2 with a
truncated version of RabGGTase were obtained and the
structure was determined with a resolution of 2.3 Š (Figure 2;
for details, see the Supporting Information).
Compound 2 binds in a deep bifurcated cavity containing
the active center at the interface of the a and b subunits of
RabGGTase (Figure 2A) and adopts an extended conforma-
tion with the C terminus pointing outward (Figure 2B). The
interaction between RabGGTase and compound 2 is mainly
hydrophobic. Only the inhibitorꢀs carbamate group forms a
strong hydrogen bond to the protein, namely with the side
chain of Arg144 of the b subunit. The imidazole moiety of
compound 2 appears to be part of a hydrogen-bonding
network involving several water molecules and possibly also
Tyr97 of the b subunit. A weak hydrogen bond to Tyr241 in
the b subunit anchors the C-terminal carboxylate group of the
inhibitor. Superimposition of the structure of the related
GGTase I in complex with a monoprenylated peptide and an
analogue of geranylgeranyl pyrophosphate (PDB access
code: 1N4Q)[12] reveals that the backbone of compound 2
and the histidyl and tyrosyl side chains occupy the proteinꢀs
substrate-binding site (Figure 2C). The side chain of the C-
terminal tyrosine is highly flexible, thereby indicating that it
may act as a general gatekeeper to block access of the protein
substrate and that its exact nature may not be important. The
phenylalanine protrudes into the hydrophobic isoprenoid-
binding pocket providing three additional potential anchoring
sites for the inhibitor (Figure 2C). Analysis of the complex
structure reveals more attachment points in the immediate
vicinity of the bound inhibitor that could be employed to
improve the affinity and specificity of the inhibitors (Fig-
ure 2D). Site 1 is composed of residues Asp287, Pro288, and
Phe289, which are located on the tip of helix 12 that ends at
is positively charged and, by homology, is expected to anchor
the phosphate groups of GGPP. Taken together, these
observations open a route to further inhibitor improvement
by structure-guided ligand design.
Finally, we wanted to establish whether the identified
compounds inhibit Rab prenylation in cellular systems. To
this end, COS-7 cells overexpressing an EYFP-Rab7 fusion
protein were incubated with the compounds for 24 h and then
lysed. The lysate was subjected to in vitro prenylation with
recombinant RabGGTase and REP and a biotin-containing
analogue of GGPP, biotin–GPP (compound 12, see the
Supporting Information). Prenylated cell lysates were sub-
jected to Western blotting with steptavidin–horseradish
peroxidase to detect Rab–geranylbiotin conjugates
(Figure 3).
Compounds 2–4 and 6–8 all caused increased labeling of
EYFP-Rab7, a result indicating inhibition of prenylation to
an extent comparable with that of compactin (which prevents
formation of GGPP by inhibiting the mevalonate pathway).
2+
the active site. Site 2is represented by the Zn ion and the
coordinating His290, while site 3, built of Arg232 and Lys235,
Angew. Chem. Int. Ed. 2008, 47, 3747 –3750
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
3749

Communications
[5] F. E. Oualid, L. H. Cohen, G. A. van der Marel, M. Overhand,
Curr. Med. Chem. 2006, 13, 2385 – 2427.
[6] F. P. Coxon, M. H. Helfrich, B. Larijani, M. Muzylak, J. E.
Dunford, D. Marshall, A. D. McKinnon, S. A. Nesbitt, M. A.
Horton, M. C. Seabra, F. H. Ebetino, M. J. Rogers, J. Biol. Chem.
2001, 276, 48213 – 48222. The IC50 value of the phosphonocar-
boxylate NE10790 was determined to be 502 Æ 113 mm: see
reference [11].
[7] M. R. Lackner, R. M. Kindt, P. M. Carroll, K. Brown, M. R.
Cancilla, C. Chen, H. de Silva, Y. Franke, B. Guan, T. Heuer, T.
Hung, K. Keegan, J. M. Lee, V. Manne, C. OꢀBrien, D. Parry, J. J.
Perez-Villar, R. K. Reddy, H. Xiao, H. Zhan, M. Cockett, G.
Plowman, K. Fitzgerald, M. Costa, P. Ross-Macdonald, Cancer
Cell 2005, 7, 325 – 336. In this paper, nanomolar inhibitors of
RabGGTase are described. However, they are even better
inhibitors of FTase, that is, they are nonselective.
Figure 3. Inhibition of RabGGTase by compounds 6–8. A) COS-7 cells
transiently coexpressing EYFP-Rab7 and ECFP-RabGDI were incubated
with compounds at the concentrations indicated. Control cells were
treated with 20 mm compactin, a known prenylation inhibitor, or with
1% DMSO, or remained untreated. B) Cells treated with KT81682, a
compound that is not a RabGGTase inhibitor in cells but displays an
in vitro IC50 value of (11.6Æ1.2) mm (not shown) show a signal
comparable to the background (DMSO, untreated).
[8] P. A. Konstantinopoulos, M. V. Karamanzis, A. G. Papavassiliou,
Nat. Rev. Drug Discovery 2007, 6, 541 – 555.
In conclusion, we have developed potent RabGGTase
inhibitors selective for this enzyme over FTase and GGTase I
and endowed with cellular activity.[13] The availability of these
inhibitors may open up new opportunities for the study of
RabGGTase and its involvement in the establishment of
diseases. In addition, we have determined the first crystal
structure of RabGGTase in complex with an inhibitor. This
structure provides an unprecedented opportunity for struc-
ture-guided design of more potent and selective RabGGTase
inhibitors.
[9] a) K. Hinterding, P. Hagebuch, J. Retey, H. Waldmann, Angew.
Chem. 1998, 110, 1298 – 1301; Angew. Chem. Int. Ed. 1998, 37,
1236 – 1239; b) M. Thutewohl, L. Kissau, B. Popkirova, I. M.
Karagumi, T. Nowak, M. Bate, J. Kuhlmann, O. Müller, H.
Waldmann, Angew. Chem. 2002, 114, 3768 – 3772; Angew. Chem.
Int. Ed. 2002, 41, 3616 – 3620; c) M. Thutewohl, H. Waldmann,
Bioorg. Med. Chem. 2003, 11, 2591 – 2615; d) M. Thutewohl, L.
Kissau, B. Popkirova, I. M. Karagumi, T. Nowak, M. Bate, J.
Kuhlmann, O. Müller, H. Waldmann, Bioorg. Med. Chem. 2003,
11, 2617 – 2626.
[10] B. Dursina, R. Reents, C. Delon, Y.-W. Wu, M. Kulharia, M.
Thutewohl, A. Veligodsky, A. Kalinin, V. Evstifeev, D. Ciobanu,
S. E. Szedlacsek, H. Waldmann, R. S. Goody, K. Alexandrov, J.
Am. Chem. Soc. 2006, 128, 2822 – 2835.
Received: December 18, 2007
Published online: April 10, 2008
[11] a) Y.-W. Wu, H. Waldmann, R. Reents, F. H. Ebetino, R. S.
Goody, K. Alexandrov, ChemBioChem 2006, 7, 1859 – 1861;
b) Y.-W. Wu, K. Alexandrov, L. Brunsveld, Nat. Protocols 2007,
2, 2704 – 2711.
[12] J. S. Taylor, T. S. Reid, K. L. Terry, P. J. Casey, L. S. Beese,
EMBO J. 2003, 22, 5963 – 5974.
[13] While our data were in press different selective inhibitors of
RabGGTase were reported: M. Watanabe, H. D. G. Fiji, L. Guo,
L. Chan, S. S. Kinderman, D. J. Slamon, O. Kwon, F. Tamanoi, J.
Biol. Chem., in press, http://www.jbc.org/cgi/doi/10.1074/
jbc.M706229200.
Keywords: inhibitors · prenylation · protein modifications ·
protein structures · transferases
.
[1] K. F. Leung, R. Baron, M. C. Seabra, J. Lipid Res. 2005, 47, 467 –
475.
[2] M. Miaczynska, S. Christoforidis, A. Giner, A. Shevchenko, S.
Uttenweiler-Joseph, B. Habermann, M. Wilm, R. G. Parton, M.
Zerial, Cell 2004, 116, 445 – 456, and references therein.
[3] P. J. Casey, M. C. Seabra, J. Biol. Chem. 1996, 271, 5289 – 5292.
[4] a) A. D. Hamilton, S. M. Sebti, Oncogene 2000, 19, 6584 – 6593;
b) R. A. Gibbs, T. J. Zahn, J. S. Sebolt-Leopold, Curr. Med.
Chem. 2001, 8, 1437 – 1465.
3750
www.angewandte.org
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2008, 47, 3747 –3750