Probing the interaction of HTI-286 with tubulin using a stilbene analogue

Article abstract of DOI:10.1021/ja048619e

HTI-286 is a synthetic analogue of the natural product hemiasterlin. HTI-286 is a potent antitumor agent that induces tubulin oligomerization. To investigate the binding stoichiometry and the binding site during this ligand-induced tubulin association, we

Full text of DOI:10.1021/ja048619e

Published on Web 07/21/2004
Probing the Interaction of HTI-286 with Tubulin Using a Stilbene Analogue
Mei-Chu Lo,*,† Ann Aulabaugh,† Girija Krishnamurthy,† Joshua Kaplan,‡ Arie Zask,‡
Robert P. Smith,§ and George Ellestad†
Biophysics/Enzymology-Chemical and Screening Sciences, Medicinal Chemistry-Chemical and Screening Sciences,
and Vaccines Research, Wyeth Research, 401 North Middletown Road, Pearl RiVer, New York 10965
Received March 10, 2004; E-mail: lom@wyeth.com
HTI-286 (1) (Figure 1) is a synthetic analogue of the natural
product hemiasterlin (2), a potent antitubulin agent.1 Hemiasterlins
are tripeptides with three highly modified amino acids, which give
rise to their in vivo stability and activity. The relative structural
simplicity of hemiasterlins allows for diverse structural manipulation
of the molecule via total synthesis. Among a series of hemiasterlin
analogues synthesized at Wyeth and UBC,2 HTI-286, containing a
phenyl group instead of an indole as in hemiasterlin, is a potent
inhibitor of cell proliferation and has substantially less interaction
with the multidrug resistance protein (P-glycoprotein) than currently
used antitumor agents.3 In our previous report, we found that HTI-
286 binds tubulin and induces the formation of a stable oligomer
with 13 tubulin units.4 The ability of HTI-286 to accommodate large
substituents on the phenyl ring and still maintain the antitubulin
potency2,5 prompted us to synthesize a derivative (3) containing
the chromophore stilbene (Figure 1) to further characterize this
ligand-induced association.6 The stilbene analogue has UV absor-
bance distinct from protein (Figure 2), which allows us to obtain
the binding stoichiometry and to probe the binding site. Herein we
describe our findings from analytical ultracentrifugation experiments
using this stilbene analogue.
Figure 1. Chemical structures of HTI-286, hemiasterlin, and the stilbene
analogue.
Figure 2. UV spectra of 50 µM HTI-286 (curve 1), 10 µM 3 (curve 2),
and 10 µM colchicine (curve 3).
Analytical ultracentrifugation is a solution-based method used
previously to study tubulin association induced by magnesium or
antitubulin agents including vinca alkaloids and cryptophycin.7-9
To investigate the effect of 3 on tubulin, we conducted sedimenta-
tion velocity experiments at different ratios of inhibitor to protein.
Sedimentation was monitored alternately at wavelengths dominated
by protein (262 nm, >90% of signal from protein) and by inhibitor
(312 nm, signal from 3 alone). The sedimentation profiles were
analyzed using SEDFIT10 to generate the sedimentation coefficient
distribution c(s) as shown in Figure 3.
In the absence of inhibitor (0 µM in Figure 3), the monomeric
tubulin (defined as an R/â heterodimer, 100 kDa) migrates with a
sedimentation coefficient of ca. 5.3 S. When more than 1 equiv of
3 (15.2 µM in Figure 3) was added to 10 µM tubulin, a major
species with a sedimentation coefficient of ca. 17.4 S appears at
262 and 312 nm. This 17.4 S species is a stable ring (Figure 4).
The outer diameter of the ring is 35-50 nm, similar to the
dimension reported for the ring structure of tubulin induced by
hemiasterlin.1 The amounts of tubulin and 3 in the ring were
estimated from the area under the 17.4 S peak detected at 262 and
312 nm, respectively. The analysis generates 10 µM tubulin and
10 µM of 3, indicating that the binding stoichiometry in the ring is
one inhibitor per tubulin monomer.
Figure 3. Sedimentation coefficient distribution c(s) derived from the
sedimentation velocity data monitored at 262 nm (left) and 312 nm (right).
The concentration of tubulin is 10 µM, and the concentrations of 3 are 0,
1.9, 3.8, 5.7, 7.6, and 15.2 µM. The ordinate is scaled for clarity of the c(s)
distributions of the intermediates. The full scale c(s) distribution of the 15.2
µM sample is shown in the inset.
Figure 4. Transmission electron micrograph of the tubulin ring induced
by 3.
In the sample containing 1.9 µM of 3 and 10 µM tubulin, two
peaks at 5.3 and 8.1 S appear at 262 nm, but only one peak at 8.1
S exists at 312 nm. It is likely that the 8.1 S species is tubulin
dimer (dimer of R/â-heterodimers) since it corresponds to the molar
mass of 200 kDa when converted to the mass distribution.
Moreover, 8.1 S is close to the predicted sedimentation coefficient
(7.95 S) for tubulin dimer using the Kirkwood-Riseman approach.11
† Biophysics/Enzymology-Chemical and Screening Sciences.
‡ Medicinal Chemistry-Chemical and Screening Sciences.
§ Vaccines Research.
9
9898
J. AM. CHEM. SOC. 2004, 126, 9898-9899
10.1021/ja048619e CCC: $27.50 © 2004 American Chemical Society

C O M M U N I C A T I O N S
The presence of monomer and dimer at this ratio was confirmed
using size-exclusion chromatography (SEC) coupled to multiangle
laser light scattering (MALLS) to obtain the molar mass (data not
shown). The ability to resolve monomer and dimer by SEC suggests
that the rate of interconversion of the species is slow. It is interesting
that the 5.3 S peak, corresponding to tubulin monomer, was not
observed for the detection at 312 nm, the wavelength at which 3
absorbs. The inability to observe the inhibitor-tubulin monomer
complex was also reported for dolastatin 10,12 a tubulin inhibitor
whose binding to tubulin is competitively inhibited by hemiasterlin.1
It is possible that 3 binds at the interface between two tubulin
monomers. Alternatively, the inability to observe the ligand-bound
tubulin monomer could be due to the reason that the concentration
of tubulin used in this study is well above the dissociation constant
of the liganded-tubulin dimer. The concentration of tubulin that
can be used is limited by the necessity to maintain a low ratio of
3 to tubulin and at the same time allow the reliable absorbance
measurement of 3. To estimate the binding stoichiometry in the
tubulin dimer, the overlapped peaks in the c(s) distribution at 262
nm were fitted to the Gaussian curves to resolve the individual
peaks. Analysis of the area under each peak generates the protein
concentration of 5.9 µM for the 5.3 S species and 4.1 µM for the
8.1 S species. The amount of inhibitor bound to the 8.1 S species
was estimated from the area under the peak detected at 312 nm.
This analysis yielded 1.9 µM inhibitor and the binding stoichiometry
of ca. 0.5, which corresponds to one inhibitor per tubulin dimer.
The binding stoichiometry of 0.5 is consistent with the binding at
the interface between two monomers. Alternatively, it could mean
that the liganded-tubulin monomer can associate with either ligand-
free or ligand-bound monomer to generate tubulin dimer.
When 3.8 µM of 3 was added to 10 µM tubulin, the resulting
c(s) distributions show two peaks at 5.3 and 9.4 S detected at 262
nm but only one asymmetric peak at 9.4 S detected at 312 nm.
The predicted sedimentation coefficient for tubulin trimer (trimer
of R/â-heterodimers) is 9.7 S in the linear arrangement and 10.6 S
in the triangular arrangement. It is likely that the 9.4 S peak
corresponds to tubulin trimer arranged in the linear fashion and
the asymmetry comes from the presence of the small amount of
dimer. Analysis of the area under the 9.4 S peak detected at 262
and 312 nm gives the binding stoichiometry of ca. 0.6, correspond-
ing to approximately two inhibitors per tubulin trimer. The presence
of monomer, dimer, and trimer at this ratio is also confirmed by
the SEC/MALLS method.
Figure 5. Sedimentation coefficient distribution c(s) derived from the
sedimentation velocity data. Curve 1 is from the sample containing 10 µM
tubulin and 10 µM colchicine, monitored at 353 nm; curves 2 and 3 are
from the sample containing 10 µM tubulin, 10 µM 3, and 10 µM colchicine,
monitored at 312 and 353 nm, respectively.
tubulin, the c(s) distribution shows a single peak at 17.4 S detected
at 312 nm (signal dominated by 3) and 353 nm (signal from
colchicine alone) (Figure 5). This suggests that colchicine binds to
the tubulin ring induced by the bound 3. Peak area analysis indicated
that equal amounts (10 µM) of 3 and colchicine bind the tubulin
ring simultaneously. The structure of the ternary complex of tubulin
with colchicine and the stathmin-like domain of RB3 was reported.13
In the complex, colchicine binds â-subunit at the intradimer surface
with R-subunit. Consistent with these results, recent photoaffinity
labeling studies at Wyeth5 indicated that a photoaffinity analogue
of HTI-286 cross-links to the R-tubulin within the residues 314-
339, which is distant from the colchicine binding site.
In summary, we obtained the binding stoichiometries in the
intermediate oligomers and the final ring structure of tubulin during
the ligand-induced association using the distinct absorbance of the
stilbene analogue. This analogue also allows us to compare the
binding site with other antitubulin agents.
Supporting Information Available: Synthesis procedures for the
stilbene analogue and the spectral characterization. This material is
available free of charge via the Internet at http://pubs.acs.org.
References
(1) Bai, R.; Durso, N. A.; Sackett, D. L.; Hamel, E. Biochemistry 1999, 38,
14302-14310.
(2) Zask, A.; Birnberg, G.; Cheung, K.; Kaplan, J.; Niu, C.; Norton, E.;
Suayan, R.; Yamashita, A.; Cole, D.; Tang, Z.; Krishnamurthy, G.;
Williamson, R.; Khafizova, G.; Musto, S.; Hernandez, R.; Annable, T.;
Yang, X.; Discafani, C.; Beyer, C.; Greenberger, L. M.; Loganzo, F.;
Ayral-Kaloustian, S. J. Med. Chem., in press.
(3) Loganzo, F.; Discafani, C. M.; Annable, T.; Beyer, C.; Musto, S.; Hari,
M.; Tan, X.; Hardy, C.; Hernandez, R.; Baxter, M.; Singanallore, T.;
Khafizova, G.; Poruchynsky, M. S.; Fojo, T.; Nieman, J. A.; Ayral-
Kaloustian, S.; Zask, A.; Andersen, R. J.; Greenberger, L. M. Cancer
Res. 2003, 63, 1838-1845.
(4) Krishnamurthy, G.; Cheng, W.; Lo, M.-C.; Aulabaugh, A.; Razinkov, V.;
Ding, W.; Loganzo, F.; Zask, A.; Ellestad, G. Biochemistry 2003, 42,
13484-13495.
At the ratios between 0.5 and 1, multiple intermediates exist. It
is difficult to assign unambiguously the identity of each intermediate
due to the peak overlap and possibly the small amount of it. These
peaks are not analyzed further.
(5) Nunes, M.; Kaplan, J.; Loganzo, F.; Zask, A.; Ayral-Kaloustian, S.;
Greenberger, L. M. Eur. J. Cancer 2002, 38, S119.
The distinct UV absorbance of 3 also allows us to probe its
binding site. To verify that 3 binds tubulin at the same site as the
parent inhibitor, HTI-286, different amounts of HTI-286 were added
to the sample containing 10 µM tubulin pre-incubated with 10 µM
of 3. Sedimentation was monitored alternately at 262 and 312 nm.
The amount of 3 bound to the ring structure of tubulin was estimated
from the peak area in the c(s) distribution detected at 312 nm. With
the addition of 10, 100, or 1000 µM of HTI-286, the bound 3 was
displaced by 27, 66, and 81%, respectively, suggesting that 3 binds
tubulin at the same site as HTI-286 and the binding is reversible.
We also compared the binding site of 3 with the classic antitubulin
agent, colchicine. Colchicine has a distinct UV absorbance at 353
nm at which stilbene does not absorb (Figure 2). Colchicine alone
does not induce tubulin oligomerization (Figure 5). However, in
the sample containing equal amounts (10 µM) of 3, colchicine, and
(6) This stilbene analogue binds tubulin with slightly higher affinity compared
to HTI-286 (apparent dissociation constant of 20 nM versus 100 nM) based
on isothermal titration calorimetry (ITC) studies. The ITC experiments
were conducted with the condition similar to that in ref 4 for the serial
titration.
(7) Frigon, R. P.; Timasheff, S. N. Biochemistry 1975, 14, 4559-4566.
(8) Lobert, S.; Frankfurter, A.; Correia, J. J. Biochemistry 1995, 34, 8050-
8060.
(9) Barbier, P.; Gregoire, C.; Devred, F.; Sarrazin, M.; Peyrot, V. Biochemistry
2001, 40, 13510-13519.
(10) Schuck, P. Biophys. J. 2000, 78, 1606-1619.
(11) Bloomfield, V.; Dalton, W. O.; Van Holde, K. E. Biopolymers 1967, 5,
135-148.
(12) Bai, R.; Taylor, G. F.; Schmidt, J. M.; Williams, M. D.; Kepler, J. A.;
Pettit, G. R.; Hamel, E. Mol. Pharmacol. 1995, 47, 965-976.
(13) Ravelli, R. B.; Gigant, B.; Curmi, P. A.; Jourdain, I.; Lachkar, S.; Sobel,
A.; Knossow, M. Nature 2004, 428, 198-202.
JA048619E
9
J. AM. CHEM. SOC. VOL. 126, NO. 32, 2004 9899