Synthesis and evaluation of cell-permeable biotinylated PU-H71 derivatives as tumor Hsp90 probes

Article abstract of DOI:10.3762/bjoc.9.60

The attachment of biotin to a small molecule provides a powerful tool in biology. Here, we present a systematic approach to identify biotinylated analogues of the Hsp90 inhibitor PU-H71 that are capable of permeating cell membranes so as to enable the investigation of Hsp90 complexes in live cells. The identified derivative 2g can isolate Hsp90 through affinity purification and, as we show, represents a unique and useful tool to probe tumor Hsp90 biology in live cells by affinity capture, flow cytometry and confocal microscopy. To our knowledge, 2g is the only reported biotinylated Hsp90 probe to have such combined characteristics.

Full text of DOI:10.3762/bjoc.9.60

Synthesis and evaluation of cell-permeable
biotinylated PU-H71 derivatives as
tumor Hsp90 probes
Tony Taldone*,‡1, Anna Rodina‡1, Erica M. DaGama Gomes1,
Matthew Riolo1, Hardik J. Patel1, Raul Alonso-Sabadell1, Danuta Zatorska1,
Maulik R. Patel1, Sarah Kishinevsky1 and Gabriela Chiosis*1,2,3
Full Research Paper
Open Access
Address:
Beilstein J. Org. Chem. 2013, 9, 544–556.
1Molecular Pharmacology and Chemistry Program, Sloan-Kettering
Institute, 1275 York Avenue, New York, NY 10065, USA, 2Department
of Medicine, Memorial Sloan-Kettering Cancer Center, 1275 York
Avenue, New York, NY 10065, USA and 3Department of
Pharmacology, Weill Graduate School of Medical Sciences, 1300
York Avenue, New York, NY 10065, USA
doi:10.3762/bjoc.9.60
Received: 10 October 2012
Accepted: 20 February 2013
Published: 15 March 2013
This article is part of the Thematic Series "Synthetic probes for the study
of biological function".
Email:
Tony Taldone* - taldonet@mskcc.org; Gabriela Chiosis* -
chiosisg@mskcc.org
Guest Editor: J. Aubé
* Corresponding author ‡ Equal contributors
© 2013 Taldone et al; licensee Beilstein-Institut.
License and terms: see end of document.
Keywords:
affinity capture; biotin; flow cytometry; fluorescence microscopy;
PU-H71; tumor Hsp90
Abstract
The attachment of biotin to a small molecule provides a powerful tool in biology. Here, we present a systematic approach to iden-
tify biotinylated analogues of the Hsp90 inhibitor PU-H71 that are capable of permeating cell membranes so as to enable the
investigation of Hsp90 complexes in live cells. The identified derivative 2g can isolate Hsp90 through affinity purification and, as
we show, represents a unique and useful tool to probe tumor Hsp90 biology in live cells by affinity capture, flow cytometry and
confocal microscopy. To our knowledge, 2g is the only reported biotinylated Hsp90 probe to have such combined characteristics.
Introduction
Heat shock protein 90 (Hsp90) is a molecular chaperone that As a result of this, as well as the ability to block multiple
functions to properly fold proteins to their active conformation signaling pathways through inhibition of a single target, Hsp90
through its ATPase activity [1]. These client proteins include has become one of the most pursued molecular targets for anti-
many that are involved in malignant cell transformations (i.e., cancer therapy [2,3]. As a testament to this, there are numerous
HER2, EGFR, mutant ER, HIF1α, Raf-1, AKT, mutant p53). ongoing clinical trials evaluating Hsp90 inhibitors from a
544

Beilstein J. Org. Chem. 2013, 9, 544–556.
variety of chemotypes [4]. Although there are potentially In one application, biotinylated probes may be subjected to
numerous ways to block the activity of Hsp90, the most streptavidin-containing beads to identify potential direct and
successful to date, as exemplified by its exclusivity in mode of indirect interactors of the small molecule through affinity
action by those advanced to clinical trials, has been the ATP- capture. Streptavidin binds to biotin in the strongest noncova-
competitive inhibitors that bind to the N-terminal nucleotide lent interaction known, with Kd ~1 × 10−14 M. As we have
binding pocket [4,5].
already shown with PU-H71 attached directly onto beads, its
ability to bind to Hsp90 in client-protein-bound complexes may
Hsp90 belongs to the family of GHKL (G = DNA gyrase be used to identify and analyze the drivers of oncogenic trans-
subunit B; H = Hsp90; K = histidine kinases; L = MutL) formations on a tumor-by-tumor basis [13]. Therefore, we
ATPases, which is distinguished by a unique bent shape of its believe that there is considerable value in preparing biotiny-
nucleotide binding pocket [6]. This distinctive shape has lated analogues of PU-H71 (1a) with the ability to permeate cell
enabled for the design of highly selective ATP-competitive membranes so as to enable the investigation of oncogenic
inhibitors of Hsp90. Through the efforts of multiple drug- Hsp90 complexes in live cells. In contrast to PU-H71 beads,
discovery groups, many classes of inhibitors have been identi- which are limited to cell homogenates, these compounds may
fied [3,5,7,8]. While much is known about the general types of be further used to investigate Hsp90 complexes in live cells,
structures that inhibit Hsp90 and their structure–activity rela- which represents a more physiologically relevant state. These
tionship, less is understood about Hsp90 tumor biology. As a tools also have use in flow cytometry and microscopy, whereby
result we and others have been actively engaged in the syn- fluorescently labeled antibodies to biotin are used, as we
thesis of chemical tools designed to probe the function of Hsp90 describe below.
in transformed systems [9-11]. One class of Hsp90 inhibitors of
Results and Discussion
interest is the purine scaffold, including its representative
Design and synthesis of biotinylated purine
scaffold Hsp90 probes
PU-H71 (1a). This agent, currently in clinical investigation for
cancer, binds to the N-terminal nucleotide binding pocket of
Hsp90 [12].
Geldanamycin (GM) is a benzoquinone ansamycin first isolated
from a fermentation broth of Streptomyces hygroscopicus [15]
and was the first reported Hsp90 inhibitor [16]. It has played a
paramount role as a probe molecule to investigate Hsp90
biology, and in fact the attachment of GM to solid support
enabled the identification of Hsp90 as the target of its anti-
cancer activity through affinity purification [16]. Biotinylated
GM has also been synthesized and has been proposed as a tool
to identify proteins other than Hsp90 that GM may directly bind
to [17]. Since the available evidence suggests GM cannot effi-
ciently trap Hsp90 in client-bound complexes [13,18], it appears
that GM–biotin is of limited use beyond identifying potential
direct interactors. In contrast, PU-H71 (1a) is highly selective
for Hsp90 and furthermore can efficiently bind to and trap
Hsp90 in client-bound complexes allowing for the identifica-
tion of global tumor Hsp90 proteomes by mass spectrometry
[13].
We have shown that PU-H71 selects for tumor Hsp90 species,
and therefore labeled derivatives of PU-H71 may be used to
specifically dissect, in a tumor-by-tumor manner, the abun-
dance and the functions of the oncogenic Hsp90 [13,14].
Specifically, these tools, which may selectively retrieve only
those Hsp90 complexes that are “available” for inhibition, will
allow for a better characterization of the “oncogenic Hsp90”,
both with regards to its onco-client protein content and the
nature of its distinct post-translational modifications. This is in
contrast to immunoprecipitation of Hsp90, which we have
shown to identify and isolate both “oncogenic Hsp90” (i.e.,
PU-H71-binding) and “housekeeping Hsp90” (i.e., PU-H71
nonbinding) complexes.
The attachment of biotin to a small molecule provides a
powerful tool in biology. As research tools, biotin-labeled
chemical tools have the potential to extend the study of single We therefore set out here to design a series of biotinylated
targets to a particular class of molecules or even to an entire analogues derived from the purine scaffold Hsp90 inhibitor
proteome. In addition, the development of biotinylated chem- PU-H71 (1a) with the purpose of identifying compounds
ical tools that penetrate live cells and, thus, are designed both to capable of permeating cancer-cell membranes, binding selec-
probe and to modulate the activity of biomolecules in live bio- tively to intracellular oncogenic Hsp90 in live cancer cells, and
logical systems, allows for a type of “live biochemistry and able to trap and isolate Hsp90 bound to tumor-specific onco-
biology” that can complement traditional biochemical and bio- client proteins. Because the biotin tag enables pull-down experi-
logical approaches by promoting molecular characterization of ments through subsequent binding to streptavidin or avidin, a
biomolecules both in vitro and within their natural biological further requirement for our probe is that the linker be of suffi-
contexts.
cient length to enable the concomitant binding to Hsp90 and
545

Beilstein J. Org. Chem. 2013, 9, 544–556.
streptavidin. Thus, in the design of these probes, the type of A critical factor in the design of biotinylated purine-scaffold
linker, as well as its length, was systematically altered so as to Hsp90 probes is the site of attachment of biotin. From previous
identify compounds that demonstrate such combined properties work including X-ray crystal structure [19], extensive SAR
(Figure 1).
[20,21], and docking experiments [9], the N9-position of the
purine scaffold was shown to be an ideal site for attachment
As such we have prepared a number of biotinylated analogues, since it is directed towards the solvent. Furthermore, the amino
derived from 1a and 1b, containing linkers of various lengths (1 group of 1a or the desisopropyl analogue 1b provided a con-
to 17 atoms) and hydrophobicities (polyethylene-, amide- and/ venient handle with which to attach biotin directly or via a
or alkyl-containing). In addition, an amine-linked biotin linker through an amide bond. We chose to make analogues of
analogue 2a, which we have reported previously [9], was also both 1a and 1b because the isopropyl group in 1a may result in
prepared for comparison purposes. This differs from the others considerable effects on cell-permeability properties due to its
by the presence of an ionizable amine in the linker region. increased lipophilicity, while having little effect on the affinity
Although 2a is a potent Hsp90 binder, it is less effective at for Hsp90. While it is essential that the linker be of sufficient
capturing Hsp90 complexes and has poor cancer-cell perme- length to enable the concomitant binding to Hsp90 and strepta-
ability (Figure 2) and was therefore of limited use and served as vidin, it is also important that it is not exceedingly long for two
a further impetus for the synthesis of the novel probes described reasons. First, the possibility and extent of nonspecific binding
here.
increases with longer linkers. Second, longer linkers result in a
Figure 1: Design of the biotinylated Hsp90 probes based on PU-H71 (1a).
546

Beilstein J. Org. Chem. 2013, 9, 544–556.
Figure 2: Analysis of the affinity and selectivity of the biotinylated probes for Hsp90. (a) K562 cancer cells were treated for 24 h with DMSO (vehicle),
PU-H71 (1 µM) or indicated biotinylated probes (5, 10 and 20 µM) and the effect of these agents on Raf-1 steady-state levels was analyzed by
Western blot. β-Actin was used as a protein loading control, because its levels remain unchanged following Hsp90 inhibition. The affinity of these
agents for Hsp90 as present in a cancer-cell homogenate and their effect on K562 cell growth are presented under the immunoblot figure. Values
were determined as indicated in the Experimental section. (b) and (c) K562 lysates or (d) K562 cells were incubated for 4 h with the indicated concen-
trations of the indicated probes. Live cells were permeabilized prior to the affinity purification step. After washing with high-salt lysis buffer (b,c) or lysis
buffer (d), protein complexes purified on streptavidin agarose beads were visualized by Coomassie blue staining. (d) For the indicated experimental
conditions, protein complexes were also identified by immunobloting (see Raf-1). MW marker = molecular weight marker. D-Biotin was used to test for
background binding of the streptavidin agarose beads. (e) Binding of probes (5 µM) to Hsp90 is competitively blocked by cell pre-treatment with
PU-H71 (1a, 5 µM). (f) Experiment set-up as in (d). Affinity purified Hsp90 in complex with its regulatory co-chaperones Hsp70 and HOP was
analyzed by Western blot. (g) K562 lysates were incubated overnight with the indicated probes (50 µM). Affinity purified proteins were identified by
Western blot.
higher molecular weight of the compound, which can adversely 1b. 2b and 2c were prepared from 1a or 1b, respectively, in
affect their permeability across cell membranes.
99% and 56% yield by DCC coupling with D-biotin under soni-
cation (Scheme 1, step a). 2d–2i were prepared by reaction of
The synthesis of the biotinylated molecules is shown in 1a or 1b with three different commercially available
Scheme 1 and in each case occurs in a single step from 1a or N-hydroxysuccinimide (NHS) active ester containing biotin
547

Beilstein J. Org. Chem. 2013, 9, 544–556.
molecules (Scheme 1, steps b–d). Whereas reactions with 1b
occurred at rt and were complete after 1 h giving the desired
products in good yield (72–88%), reactions with 1a required
heating at 35 °C and were incomplete after 6 h as evidenced by
recovery of a significant amount of unreacted starting material.
The yields of isolated products ranged from 29–41%. 2d and 2e
were prepared from EZ-Link® NHS-LC-Biotin (Scheme 1, step
b). 2f and 2g were prepared from EZ-Link® NHS-LC-LC-
Biotin (Scheme 1, step c). 2h and 2i were prepared from
EZ-Link® NHS-PEG4-Biotin (Scheme 1, step d). 2a was
prepared as reported previously [9], by amination of the corres-
ponding bromide with EZ-Link® Amine-PEO3-Biotin.
Scheme 2: Reagents and conditions: (a) 6-Boc-aminocaproic acid,
DCC, DMAP, CH2Cl2, rt; (b) TFA, CH2Cl2, rt; (c) D-biotin, DCC, DMAP,
CH2Cl2, sonicate.
live cells. First, the probe should retain selective and tight
binding to tumor Hsp90. Second, it should permeate live cells
and while inside the cell, should bind to the oncogenic Hsp90.
Upon cell permeabilization, the probe should retain Hsp90
binding and concomitantly bind to streptavidin allowing for
subsequent isolation of Hsp90. Third, if isolation of oncogenic
Hsp90 in complex with its tumor-specific client proteins is the
desired outcome, the probe should also trap and lock the Hsp90/
protein complex, so that it is maintained throughout the subse-
Scheme 1: Reagents and conditions: (a) D-biotin, DCC, DMAP,
CH2Cl2, sonicate; (b) EZ-Link® NHS-LC-Biotin, DIEA, DMF, 35 °C or
rt; (c) EZ-Link® NHS-LC-LC-Biotin, DIEA, DMF, 35 °C or rt; (d)
EZ-Link® NHS-PEG4-Biotin, DIEA, DMF, 35 °C or rt.
It should be noted that in each of the products (2b, 2d, 2f, 2h) quent permeabilization and purification steps.
derived from 1a it was not immediately clear whether these
were a mixture of two compounds or rotamers, despite the Requirement 1: Retain tight binding to tumor Hsp90
seeming unambiguity in the synthesis. While HPLC showed a To ensure the biotinylated compounds still retained affinity for
single homogeneous peak, the NMR spectrum was very compli- tumor Hsp90, they were each evaluated in a fluorescence
cated. To settle this, 2d was prepared by an alternate synthesis polarization (FP) assay by using a cancer-cell homogenate (i.e.,
(Scheme 2). DCC coupling of 1a with 6-Boc-aminocaproic acid SKBr3 human breast cancer lysate). This assay measures
yielded 3 following removal of the Boc group, which was competitive binding to tumor cell Hsp90 complexes [23]. Each
further reacted with D-biotin to give a product with identical compound retained a good affinity for Hsp90, with values
NMR and HPLC profile to 2d, confirming that a mixture of two ranging from 30 to 150 nM (Figure 2a, Hsp90 binding). Two
rotamers was present and not a mixture of two compounds. general trends were observed. First, compared to the PU-H71
Additionally, intermediate 3 also demonstrates a complex NMR analogues, the desisopropyl analogues bound on average with
spectrum indicative of the presence of two rotamers. All of this approximately 2-fold greater affinity (i.e., 2c versus 2b, 2e
shows that, unlike the proton, the isopropyl group is bulky versus 2d, 2g versus 2f, 2i versus 2h), despite the fact that both
enough to hinder rotation of the tertiary amides and to enable 1a and 1b bound Hsp90 with similar affinity (24.5 versus
identification of two rotamers by NMR [22].
26 nM for 1a and 1b, respectively). This is likely a result of
increased steric crowding of the bulky isopropyl group in
analogues of 1a. Second, in terms of the linkers, the carbon
series appeared to have a somewhat higher Hsp90 affinity than
Biological evaluation of the biotinylated
Hsp90 probes
As indicated above, there are several requirements for a biotiny- the ethylene glycol series (i.e., 2d and 2f versus 2h; 2e and 2g
lated probe to be useful in dissecting Hsp90 tumor biology in versus 2i). In sum however, all of the compounds retained good
548

Beilstein J. Org. Chem. 2013, 9, 544–556.
affinity for Hsp90, supporting our notion for the ideal site of tions of biotinylated derivatives (10, 25 and 50 µM; Figure 2c).
biotinylation, and were thus suitable for further analysis.
Additionally, affinity-purified complexes were washed with
high-salt buffer to remove Hsp90-bound co-chaperones and
client proteins (Figure 2b and c).
Requirement 2: Permeate live cancer cells and bind
to oncogenic Hsp90
Having shown that each of the prepared biotinylated molecules Of the new biotin derivatives, only 2h, 2f, 2g and 2i performed
retained good affinity for Hsp90, we next evaluated these com- better than 2a and isolated substantial amounts of Hsp90
pounds in two functional read-outs that together measure that (Figure 2b and c). We were unable to affinity purify Hsp90 with
the probe has entered a live cancer cell and once inside the cell, derivatives 2b, 2c and 2e, indicating that while these com-
has bound to a substantial fraction of oncogenic Hsp90 mole- pounds entered the cancer cell and bound to intracellular Hsp90
cules. Specifically, K562 is a human leukemia cell line depen- (Figure 2a), the linker was of unfavorable length and did not
dent on Hsp90 for survival [13]. Thus, in such cells, occupancy allow for concomitant binding to streptavidin through the biotin
of Hsp90’s regulatory pocket by small molecules results in inhi- end of the probe. Consequently, isolation of Hsp90 from the cell
bition of its cancer-sheltering properties, leading to cell-growth homogenate failed with these biotinylated probes. As reported,
inhibition associated with degradation of Hsp90-chaperoned 2a containing a 13-atom linker was a modest probe for affinity
onco-proteins. These, such as is the case for Raf-1 in K562 purifications (Figure 2b and [9]), suggesting that for Hsp90, a
cells, become ubiquitinated and targeted for proteasomal de- linker longer than 13-atoms, and more exactly of 15-atoms or
gradation leading to a decrease in their steady-state levels longer, was needed to maintain concomitant Hsp90 and strepta-
(Figure 2a, Raf-1) [24].
vidin binding.
While the biotinylated analogues displayed decreased potency Requirement 4: Trap Hsp90 in an onco-client-bound
compared to 1a, it was clear from these results that some were conformation and isolate the endogenous Hsp90/
capable of entering live cells in such concentrations as to onco-client complexes from live cells
substantially occupy the oncogenic Hsp90 sites (Figure 2a, for To test for the probes’ ability to isolate Hsp90 in secondary and
derivatives 2b, 2c, 2e, 2g and 2d almost complete Raf-1 degrad- tertiary complexes, such as those containing onco-client
ation associated with cell-growth inhibition in a similar concen- proteins, affinity purifications were also performed from live
tration range). Other derivatives, such as 2i, 2f and 2h failed to K562 cancer cells (Figure 2d). In such a case, the biotinylated
exhibit such properties (Figure 2a).
tool is added to live cells where the compound binds to Hsp90
in an onco-client-bound conformation, locking and preserving
From the results, several conclusions concerning linker length the endogenous Hsp90/protein complexes throughout the subse-
and type can be drawn. In general, as the chain length increased, quent experimental steps (i.e., permeabilization). In contrast,
the ability to enter into the cancer cell decreased. With regards when adding a biotinylated tool to cell homogenates, one may
to linker nature, the polyethylene glycol linker containing encounter two potential limitations. First, due to the dynamic
derivatives (i.e., 2a, 2h and 2i) performed poorest by this nature of the Hsp90/client protein interactions, the endogenous
measure. Unexpectedly, derivatives 2g and 2f both containing complexes may be lost during the homogenization process and
the same 15-atom linker and differing only by the presence of H thus, pull-downs from homogenates may miss important inter-
(on derivative 2g) or iPr (on derivative 2f) exhibited distinct actors. Second, during homogenization, certain proteins may
behaviors, with only 2g appearing to be substantially taken up lose their well-regulated conformation and potentially aggre-
by the cancer cell.
gate. Such misfolded proteins are prone to be captured by chap-
erones resulting in “false positives” (i.e., nonendogenous Hsp90
client proteins). False positives increase the “background” on
the affinity resin, and the higher the background, the poorer the
Requirement 3: Bind concomitantly to Hsp90 and
streptavidin
Having identified which compounds were capable or not of identification of relevant endogenous Hsp90 complexes will be.
permeating live cancer cells, we next wanted to determine Therefore, while it is true that following the addition of the
whether the chain length was optimal to maintain concomitant biotinylated tool to cells, these are permeabilized or fixed/
binding to Hsp90 and streptavidin, so as to allow for isolation permeabilized and thus no longer alive, the capture of the onco-
and identification of Hsp90/onco-client complexes from cancer genic Hsp90 complexes takes place in the live cell.
cells. For this purpose, K562 lysates (Figure 2b and c) were in-
cubated with the biotinylated ligands and the complexes Consequently, in live-cell experiments, cells were first incu-
captured on streptavidin beads. To test for the probe’s selec- bated with the biotinylated PU-H71 derivatives to trap and
tivity, pull-downs were performed with increasing concentra- maintain the onco-client complexed to Hsp90. Next, cells were
549

Beilstein J. Org. Chem. 2013, 9, 544–556.
ruptured into a physiological buffer containing molybdate. The Interestingly, when tested for Hsp90 paralogue-selectivity, we
purpose of this step is to release the proteins from the cell yet noted for the 2g, 2i, 2f and 2h derivatives a substantial prefer-
maintain the Hsp90/onco-client protein complexes intact. ence for the affinity purification of the cytosolic Hsp90 over the
Following capture of these complexes on the streptavidin beads, endoplastic reticulum (ER) paralogue, Grp94 (Figure 2g). This
complexes were applied to a denaturing gel, then probed by is a surprising finding, because PU-H71 is a pan-Hsp90
both Coomassie stain (Figure 2d, top panel) and immunoblot inhibitor that binds equally well to the cytosolic and the ER
(Figure 2d, bottom panel Raf-1 blot). The Coomassie blue paralogues (Chiosis G, personal communication). We tested the
stained gels of these pull-downs showed a single band at affinity of 2g, 2i, 2f and 2h for the two paralogues, and identi-
approximately 90 kDa for derivatives 2g, 2i, 2d, 2f and 2h cally to the parent ligand PU-H71, we determined little prefer-
(Figure 2d), which was competitively blocked by pretreatment ence for Hsp90 over Grp94 (45 versus 451 nM for 2g; 83 versus
of cells with a soluble ligand (Figure 2e) indicating concomi- 226 nM for 2i; 98 versus 210 nM for 2f; 137 versus 313 nM for
tant binding to Hsp90 and streptavidin, and moreover 2h). These findings indicate that the selectivity profile was
confirming selective and strong binding between these probes unlikely imparted by the ligand. More likely, the ligand binds to
and Hsp90.
both Hsp90 and Grp94 in the cell extract; however, isolation of
the Grp94 complex on the streptavidin beads fails because of
Analogues of 1-atom linker (2b and 2c) and 8-atom linker (2d the inappropriate nature of the linker. Such was the case for
and 2e) showed a faint band or no band at 90 kDa, a finding probes 2b and 2c (see above), which, although they both bound
similar to experiments performed in cell homogenates, indi- effectively to Hsp90, could not concomitantly bind Hsp90 and
cating that the linker was of inadequate length for the purpose streptavidin, and thus isolation of Hsp90 from extracts failed
of affinity purification. Derivative 2d behaved erratically over with such probes.
several experiments, showing either faint or no isolation of
Hsp90 (Figure 2d and not shown). We potentially attribute such Potential uses of the biotinylated Hsp90 probes
behavior to interbatch variability in the loading capacity and Having shown that the probes bind to tumor Hsp90, we went on
nature of the streptavidin beads. 2d being of borderline charac- to demonstrate several potential uses for probe 2g. In addition
teristics with regards to chain (i.e., eight atoms in length and to affinity-purification of oncogenic Hsp90 from distinct
containing the sterically constraining iPr) and cancer-cell tumors, the biotinylated probes are useful to measure the drug-
permeability (Figure 2a) would fail to isolate Hsp90 in amounts accessible tumor Hsp90 by both flow cytometry and
visible by Coomassie staining when low-capacity streptavidin microscopy techniques. We exemplify here such use in the
beads are used. As such, we advise against the use of this K562 leukemia cells in the determination of cell-surface
derivative as a chemical tool. Most efficient at isolating Hsp90 (Figure 3a) and intracellular (Figure 3b and c) Hsp90 by flow
in complex with an onco-client protein such as Raf-1 were cytometry and by fluorescent microscopy (Figure 3c). For the
derivatives with 15-atom (2f and 2g) and 17-atom linkers (2h measurement of intracellular Hsp90, the use of digitonin was
and 2i) (Figure 2d, Hsp90 and Raf-1). From cells, 2g affinity effective in allowing the entry of the antibiotin antibody for
purifies Hsp90 in complex with its regulatory cochaperones, probe detection (Figure 3b). Both digitonin and saponin can be
Hsp70 and HSP-organizing protein (HOP) [1,2,13] (Figure 2f). used to reversibly open cellular pores and allow antibody entry,
thus allowing for retention of cell viability, if this is desired
It is important to note that the affinity purification strength of [25]. Staining with CD45, a plasma membrane protein, was
the biotinylated probes is weaker than that of directly solid- used as a positive control for detection of cell-surface Hsp90
support-linked PU-H71. This is likely a consequence of the (Figure 3a and b). The contribution to the signal of endogenous
solid-support loading capacity. While direct attachment of a levels of biotin in the cell was accounted for by the use of cells
ligand to the bead can result in high local concentrations of stained with a fluorescently labeled antibiotin antibody (control,
ligand, the attachment of ligand indirectly by means of biotin- Figure 3).
streptavidin is limited by the concentration of streptavidin avail-
able on the solid support. It is obvious that much lower numbers Conclusion
of bulky streptavidin molecules can be attached on any solid In our continuing efforts to develop tools that may be used to
support when compared to a low-molecular-weight ligand, such better understand tumor Hsp90 biology, we have prepared a
as PU-H71. Therefore, for isolation and identification of entire series of biotinylated analogues of the purine scaffold Hsp90
Hsp90 proteome isolations by mass spectrometry, as we inhibitor PU-H71 (1a) and its desisopropyl analogue 1b. The
recently reported [13], the beads containing PU-H71 directly at- goal of this study was to optimize probe 2a [9] and develop
tached by a covalent link remain the most efficient probe, and analogues capable of efficiently permeating the cancer cell
we continue to recommend their use for such purposes.
membrane so that they may be used as tools to investigate onco-
550

Beilstein J. Org. Chem. 2013, 9, 544–556.
Figure 3: Use of probe 2g to detect oncogenic Hsp90 by flow cytometry (a) and (b) and by microscopy (c). For (b) and (c) cells were permeabilized
with digitonin. DMSO, cells treated with vehicle only; control, cells treated with vehicle and stained with antibiotin-PE for flow cytometry and antibiotin-
FITC for fluorescent microscopy; 2g, cells treated with probe 2g and stained with antibiotin-PE for flow cytometry and antibiotin-FITC for fluorescent
microscopy. MFI, mean fluorescence intensity. CD45 is a plasma-membrane protein. For microscopy, nuclei were stained with DAPI. (a) and (b) right
panels; quantification of repeat experiments (n = 2).
genic Hsp90 and its complexes from live cells by affinity permeabilization of cells and complex capture on streptavidin
capture, flow cytometry and microscopy.
beads, Hsp90 bound to Raf-1 was however isolated with these
probes (Figure 2d). While apparently a paradoxical finding, one
Of all probes, we found only 2g to be very effective at both must note that 2h and 2i have in common the long 17-atom
permeating cancer cell membranes and binding to and isolating linker. It is possible that such compounds are prone to being
Hsp90 onco-protein complexes from live cells (Figure 2a and d; trapped in the lipid bilayers of the plasma membrane, and thus,
red boxes), and its use is thus indicated for such applications. significant amounts become available for binding to intracel-
Probes 2i, 2f and 2h remain of a yet uncharacterized category lular Hsp90 only after the cell-permeabilization step (such as is
(Figure 2a and d; green boxes). Unlike probe 2g, probes 2i and performed in Figure 2d). Alternatively, it is plausible that these
2h, and to some degree 2f, failed to substantially degrade Raf-1 compounds get into the cell and become available for binding to
at concentrations as high as 20 µM (Figure 2a). When incu- the oncogenic Hsp90 complex in the live cell. The long chain
bated at such concentrations with live cells, a step followed by however, characteristic of these probes, may interfere with
551

Beilstein J. Org. Chem. 2013, 9, 544–556.
recruitment of an E3 ligase to the Hsp90/onco-client complex, was concentrated under reduced pressure and the resulting
and proteasomal degradation of the client protein may be residue was purified by preparative TLC (CH2Cl2/MeOH-NH3
impeded as a result. We are conducting follow-up experiments (7 N), 10:1) to give 43.2 mg (99%) of 2b. 1H NMR (600 MHz,
to investigate such hypotheses; however, these studies are CDCl3, 2 rotamers) δ 8.22 (s, 1H), 7.22 (s, 0.6H), 7.21 (s,
outside the scope of this manuscript.
0.4H), 6.87 (s, 0.6H), 6.76 (s, 0.4H), 6.25 (br s, 0.6H), 6.16 (br
s, 0.4H), 5.96–5.88 (m, 2H), 5.85 (br s, 0.6H), 5.78 (br s, 0.4H),
In conclusion, our work identifies 2g as a probe for endogenous 4.63–4.54 (m, 0.6H), 4.45–4.32 (m, 1.6H), 4.25–4.21 (m, 0.4H),
oncogenic Hsp90 and its protein clientele in live cells. The 4.19–4.11 (m, 1.4H), 4.07–4.00 (m, 0.6H), 3.95–3.88 (m, 0.4H),
probe shows good affinity for tumor Hsp90 as demonstrated by 3.22–2.97 (m, 2.4H), 2.84–2.78 (m, 1H), 2.77–2.69 (m, 0.6H),
FP, and good permeability as demonstrated by the two pheno- 2.68–2.62 (m, 1H), 2.27–2.22 (m, 0.6H), 2.05–1.94 (m, 1.4H),
typic read-outs of oncogenic Hsp90 inhibition (i.e., cytotoxicity 1.89–1.74 (m, 1.4H), 1.72–1.43 (m, 3H), 1.40–1.16 (m, 3.6H),
and the ability to down regulate an Hsp90 onco-client protein in 1.06–1.00 (m, 4H), 0.97 (d, J = 6.7 Hz, 2H); MS (ESI) m/z:
the relevant cancer cell background). 2g can isolate Hsp90 739.2 [M + H]+; HRMS–ESI (m/z): [M + H]+ calcd for
through affinity purification from both cancer cell homogenates C28H36IN8O4S2, 739.1346; found, 739.1353; HPLC: tR = 9.83.
and live cells and is capable of trapping Hsp90 in an onco-
client-bound conformation facilitating the isolation of such (2c). 1b (9.1 mg, 0.0193 mmol), D-biotin (7.1 mg,
complexes and their analysis and identification through clas- 0.0290 mmol), DCC (8 mg, 0.0386 mmol) and a catalytic
sical biochemical techniques (i.e., Western blot). The probe, as amount of DMAP in CH2Cl2 (1 mL) was sonicated for 5 h. The
we demonstrate here, is also of use in detecting and analyzing reaction mixture was concentrated under reduced pressure and
tumor Hsp90 by flow cytometry and microscopy. To our knowl- the resulting residue was purified by preparative TLC (CH2Cl2/
edge, 2g is the only reported biotinylated Hsp90 probe to have MeOH-NH3 (7 N), 10:1) to give 7.5 mg (56%) of 2c. 1H NMR
such combined characteristics, and thus represents a unique (600 MHz, CDCl3/MeOH-d4) δ 7.97 (s, 1H), 7.17 (s, 1H), 6.86
useful tool to investigate Hsp90 tumor biology.
(s, 1H), 5.84 (s, 2H), 4.27–4.23 (m, 1H), 4.09–4.05 (m, 1H),
4.03 (t, J = 7.2 Hz, 2H), 3.02 (t, J = 6.4 Hz, 2H), 2.97–2.90 (m,
1H), 2.67 (dd, J = 4.9, 12.8 Hz, 1H), 2.49 (d, J = 12.8 Hz, 1H),
2.01 (t, J = 7.5 Hz, 2H), 1.83–1.75 (m, 2H), 1.54–1.34 (m, 4H),
Experimental
General
1H NMR spectra were recorded on a Bruker 500 or 600 MHz 1.27–1.18 (m, 2H); MS (ESI) m/z: 697.1 [M + H]+; HRMS–ESI
instrument. Chemical shifts were reported in δ values in parts (m/z): [M + H]+ calcd for C25H30IN8O4S2, 697.0876; found,
per million (ppm) downfield from TMS as the internal standard. 697.0904; HPLC: tR = 9.00.
1H data were reported as follows: chemical shift, multiplicity
(s = singlet, d = doublet, t = triplet, q = quartet, br = broad, m = (2d). 1a (15 mg, 0.0292 mmol), EZ-Link® NHS-LC-Biotin
multiplet), coupling constant (Hz), integration. High-resolution (14.6 mg, 0.0321 mmol) and DIEA (7.5 mg, 10.2 µL,
mass spectra were recorded on a Waters LCT Premier system. 0.0584 mmol) in DMF (0.5 mL) was heated at 35 °C for 6 h.
Low-resolution mass spectra were obtained on a Waters The reaction mixture was concentrated under reduced pressure
Acquity Ultra Performance LC with electrospray ionization and and the resulting residue was purified by preparative TLC
SQ detector. High-performance liquid chromatography analyses (CH2Cl2/MeOH-NH3 (7 N), 10:1) to give 10.3 mg (41%) of 2d.
were performed on a Waters Autopurification system with In addition, 6.9 mg of unreacted 1a was recovered to give an
PDA, MicroMass ZQ, and ELSD detector, and a reversed-phase actual yield of 77%. 1H NMR (500 MHz, CDCl3, 2 rotamers) δ
column (Waters X-Bridge C18, 4.6 × 150 mm, 5 µm) using a 8.29–8.26 (m, 1H), 7.29 (s, 0.4H), 7.28 (s, 0.6H), 6.87 (s, 0.4H),
gradient of (a) H2O + 0.1% TFA and (b) CH3CN + 0.1% TFA, 6.85 (s, 0.6H), 6.76 (br s, 0.4H), 6.74 (br s, 0.6H), 6.63–6.51 (br
5 to 95% b over 13 minutes at 1.2 mL/min. All reactions were s, 2H), 6.00–5.96 (m, 2H), 5.68 (br s, 0.4H), 5.58 (br s, 0.6H),
performed under argon protection. EZ-Link® NHS-LC-Biotin, 4.64–4.56 (m, 0.4H), 4.52–4.45 (m, 1H), 4.36–4.28 (m, 1H),
EZ-Link® NHS-LC-LC-Biotin, EZ-Link® NHS-PEG4-Biotin, 4.27–4.20 (m, 2H), 4.09–4.01 (m, 0.6H), 3.32–3.08 (m, 5H),
and EZ-Link® Amine-PEO3-Biotin were purchased from Pierce 2.94–2.86 (m, 1H), 2.76–2.69 (m, 1H), 2.37–2.31 (m, 1H),
(Rockford, Il). 1a [20], 1b [10] and biotinylated analogue 2a [9] 2.22–1.96 (m, 4H), 1.96–1.89 (m, 1H), 1.80–1.30 (m, 12H),
were prepared as previously described.
1.16–1.10 (m, 4H), 1.09–1.04 (m, 2H); MS (ESI) m/z: 852.3 [M
+ H]+; HRMS–ESI (m/z): [M + H]+ calcd for C34H47IN9O5S2,
852.2186; found, 852.2206; HPLC: tR = 8.82.
Synthesis of probes
(2b). 1a (30 mg, 0.059 mmol), D-biotin (19 mg, 0.078 mmol),
DCC (24 mg, 0.117 mmol) and a catalytic amount of DMAP in (2e). 1b (16.9 mg, 0.0359 mmol), EZ-Link® NHS-LC-Biotin
CH2Cl2 (1 mL) were sonicated for 9 h. The reaction mixture (17.9 mg, 0.0394 mmol) and DIEA (9.3 mg, 12.5 µL,
552

Beilstein J. Org. Chem. 2013, 9, 544–556.
0.0718 mmol) in DMF (0.5 mL) was stirred at rt for 1 h. The The reaction mixture was concentrated under reduced pressure,
reaction mixture was concentrated under reduced pressure and and the resulting residue was purified by preparatory TLC
the resulting residue was purified by preparative TLC (CH2Cl2/ (CH2Cl2/MeOH-NH3 (7 N), 10:1) to give 9.3 mg (32%) of 2h.
MeOH-NH3 (7 N), 10:1) to give 20.8 mg (72%) of 2e. 1H NMR In addition, 9.0 mg of unreacted 1a was recovered to give an
(500 MHz, CDCl3) δ 8.22 (s, 1H), 7.52 (t, J = 5.6 Hz, 1H), 7.36 actual yield of 81%. 1H NMR (500 MHz, CDCl3/MeOH-d4, 2
(s, 1H), 7.03 (s, 1H), 6.66 (t, J = 5.5 Hz, 1H), 6.25 (br s, 2H), rotamers) δ 8.18 (s, 0.4H), 8.16 (s, 0.6H), 7.32–7.30 (m, 1H),
6.03 (s, 2H), 4.52–4.47 (m, 1H), 4.33–4.28 (m, 1H), 4.25 (t, J = 6.98 (s, 0.6H), 6.96 (s, 0.4H), 5.98 (s, 2H), 4.56–4.49 (m, 0.4H),
6.8 Hz, 2H), 3.25–3.17 (m, 4H), 3.17–3.11 (m, 1H), 2.90 (dd, J 4.46–4.39 (m, 1H), 4.27–4.22 (m, 1H), 4.21–4.15 (m, 2H),
= 5.0, 12.9 Hz, 1H), 2.79–2.63 (m, 1H), 2.24 (t, J = 7.4 Hz, 4.07–3.99 (m, 0.6H), 3.71–3.66 (m, 2H), 3.61–3.51 (m, 12H),
2H), 2.19–2.13 (m, 2H), 2.02–1.94 (m, 2H), 1.74–1.58 (m, 6H), 3.50–3.45 (m, 2H), 3.38–3.29 (m, 2H), 3.25–3.16 (m, 2H),
1.56–1.48 (m, 2H), 1.46–1.31 (m, 4H); MS (ESI) m/z: 810.3 [M 3.12–3.07 (m, 1H), 2.88–2.81 (m, 1H), 2.68–2.63 (m, 1H),
+ H]+; HRMS–ESI (m/z): [M + H]+ calcd for C31H41IN9O5S2, 2.63–2.57 (m, 1.2H), 2.47–2.41 (m, 0.8H), 2.18–1.98 (m, 4H),
810.1717; found, 810.1703; HPLC: tR = 8.00.
1.70–1.52 (m, 4H), 1.41–1.32 (m, 2H), 1.08 (d, J = 6.7 Hz, 4H),
1.02 (d, J = 6.8 Hz, 2H); MS (ESI) m/z: 986.5 [M + H]+;
(2f). 1a (15 mg, 0.0292 mmol), EZ-Link® NHS-LC-LC-Biotin HRMS–ESI (m/z): [M + H]+ calcd for C39H57IN9O9S2,
(18.2 mg, 0.0321 mmol) and DIEA (7.5 mg, 10.2 µL, 986.2765; found, 986.2757; HPLC: tR = 8.53.
0.0584 mmol) in DMF (0.5 mL) was heated at 35 °C for 6 h.
The reaction mixture was concentrated under reduced pressure (2i). 1b (17.6 mg, 0.0374 mmol), EZ-Link® NHS-PEG4-Biotin
and the resulting residue was purified by preparatory TLC (24.2 mg, 0.0411 mmol) and DIEA (9.7 mg, 13 µL,
(CH2Cl2/MeOH-NH3 (7 N), 10:1) to give 8.2 mg (29%) of 2f. 0.0704 mmol) in DMF (0.5 mL) was stirred at rt for 1 h. The
In addition, 9.6 mg of unreacted 1a was recovered to give an reaction mixture was concentrated under reduced pressure and
actual yield of 81%. 1H NMR (500 MHz, CDCl3/MeOH-d4, 2 the resulting residue was purified by preparative TLC (CH2Cl2/
rotamers) δ 8.18 (s, 0.4H), 8.16 (s, 0.6H), 7.31 (s, 1H), 6.98 (s, MeOH-NH3 (7 N), 10:1) to give 31.0 mg (88%) of 2i. 1H NMR
0.6H), 6.95 (s, 0.4H), 6.90–6.80 (m, 2H), 5.98 (s, 2H), (500 MHz, CDCl3) δ 8.29 (s, 1H), 7.51 (t, J = 5.8 Hz, 1H), 7.32
4.55–4.47 (m, 0.4H), 4.47–4.41 (m, 1H), 4.27–4.23 (m, 1H), (s, 1H), 7.03 (t, J = 5.3 Hz, 1H), 6.90 (s, 1H), 6.79 (s, 1H), 6.57
4.22–4.16 (m, 2H), 4.03–3.95 (m, 0.6H), 3.34–3.31 (m, 0.6H), (br s, 2H), 6.01 (s, 2H), 5.97 (s, 1H), 4.53–4.48 (m, 1H),
3.24–3.19 (m, 1.4H), 3.17–3.07 (m, 5H), 2.89–2.82 (m, 1H), 4.35–4.25 (m, 3H), 3.79 (t, J = 6.1 Hz, 2H), 3.68–3.59 (m,
2.70–2.64 (m, 1H), 2.32–2.25 (m, 1H), 2.16–1.94 (m, 7H), 12H), 3.57 (t, J = 5.1 Hz, 2H), 3.46–3.40 (m, 2H), 3.24–3.18
1.70–1.18 (m, 18H), 1.09 (d, J = 6.7 Hz, 4H), 1.03 (d, J = 6.8 (m, 2H), 3.18–3.12 (m, 1H), 2.90 (dd, J = 5.0, 12.8 Hz, 1H),
Hz, 2H); MS (ESI) m/z: 965.5 [M + H]+; HRMS–ESI (m/z): [M 2.75 (d, J = 12.7 Hz, 1H), 2.54 (t, J = 6.0 Hz, 2H), 2.20 (t, J =
+ H]+ calcd. for C40H58IN10O6S2, 965.3027; found, 965.3010; 7.4 Hz, 2H), 2.01–1.40 (m, 2H), 1.79–1.59 (m, 4H), 1.48–1.38
HPLC: tR = 8.73.
(m, 2H); MS (ESI) m/z: 944.4 [M + H]+; HRMS–ESI (m/z): [M
+ H]+ calcd for C36H51IN9O9S2, 944.2296; found, 944.2307;
(2g). 1b (16.6 mg, 0.0352 mmol), EZ-Link® NHS-LC-LC- HPLC: tR = 7.82.
Biotin (22.0 mg, 0.0387 mmol) and DIEA (9.1 mg, 12.3 µL,
0.0704 mmol) in DMF (0.5 mL) was stirred at rt for 1 h. The 6-Amino-N-(3-(6-amino-8-(6-iodobenzo[d][1,3]dioxol-5-
reaction mixture was concentrated under reduced pressure and ylthio)-9H-purin-9-yl)propyl)-N-isopropylhexanamide (3).
the resulting residue was purified by preparative TLC (CH2Cl2/ 1a (50 mg, 0.0975 mmol), 6-Boc-aminocaproic acid (29 mg,
MeOH-NH3 (7 N), 10:1) to give 27.8 mg (86%) of 2g. 1H NMR 0.127 mmol), DCC (40.2 mg, 0.195 mmol) and a catalytic
(500 MHz, CDCl3/MeOH-d4) δ 8.12 (s, 1H), 7.60 (m, 1H), 7.30 amount of DMAP in CH2Cl2 (1.5 mL) was stirred at rt
(s, 1H), 7.09 (m, 1H), 6.98 (s, 1H), 5.97 (s, 2H), 4.44–4.38 (m, overnight. The reaction mixture was concentrated under
1H), 4.24–4.20 (m, 1H), 4.17 (t, J = 7.1 Hz, 2H), 3.18–3.04 (m, reduced pressure, and the resulting residue was partially puri-
7H), 2.83 (dd, J = 5.0, 12.9 Hz, 1H), 2.64 (d, J = 12.8 Hz, 1H), fied by preparative TLC (CH2Cl2/MeOH-NH3 (7 N), 12:1) to
2.16 (t, J = 7.5 Hz, 2H), 2.12–2.03 (m, 4H), 1.96–1.88 (m, 2H), give a residue, which was dissolved in TFA/CH2Cl2
1.66–1.18 (m, 18H); MS (ESI) m/z: 923.4 [M + H]+; (0.4:1.6 mL) and stirred for 20 min at rt. The reaction mixture
HRMS–ESI (m/z): [M + H]+ calcd for C37H52IN10O6S2, was concentrated under reduced pressure and the resulting
923.2558; found, 923.2595; HPLC: tR = 7.95.
residue was purified by preparative TLC (CH2Cl2/MeOH-NH3
(7 N), 10:1) to give 55 mg (90%) of 3. 1H NMR (500 MHz,
(2h). 1a (15 mg, 0.0292 mmol), EZ-Link® NHS-PEG4-Biotin CDCl3, 2 rotamers) δ 8.38–8.34 (m, 1H), 7.35 (s, 0.4H), 7.33 (s,
(18.9 mg, 0.0321 mmol) and DIEA (7.5 mg, 10.2 µL, 0.6H), 6.98 (s, 0.4H), 6.93 (s, 0.6H), 6.05–6.01 (m, 2H), 5.72
0.0584 mmol) in DMF (0.5 mL) was heated at 35 °C for 6 h. (br s, 2H), 4.69–4.63 (m, 0.4H), 4.29 (t, J = 7.2 Hz, 2H),
553

Beilstein J. Org. Chem. 2013, 9, 544–556.
4.10–4.02 (m, 0.6H), 3.31–3.25 (m, 1.2H), 3.20–3.14 (m, 0.8H), 50 mM Tris pH 7.4, 150 mM NaCl and 1% NP-40 lysis buffer.
2.80–2.70 (m, 2H), 2.37 (t, J = 7.5 Hz, 1.2H), 2.15–2.06 (m, Protein concentrations were measured by using the BCA kit
2H), 2.01–1.89 (m, 0.8H), 1.70–1.62 (m, 1.2H), 1.58–1.48 (m, (Pierce) according to the manufacturer's instructions. Protein
2H), 1.45–1.36 (m, 2H), 1.24–1.16 (m, 0.8H), 1.14 (d, J = 6.7 lysates (50 μg) were resolved by SDS-PAGE, transferred onto
Hz, 3.6H), 1.09 (d, J = 6.9 Hz, 2.4H); MS (ESI) m/z: 626.2 [M nitrocellulose membrane and incubated with an anti-Raf-1 anti-
+ H]+; HRMS–ESI (m/z): [M + H]+ calcd for C24H33IN7O3S, body from rabbit (1:500, sc-133, Santa Cruz) or anti-β-actin
626.1410; found, 626.1411; HPLC: tR = 7.92.
from mouse (1:2,500, A1978, Sigma-Aldrich). Membranes
were then incubated with the corresponding peroxidase-conju-
(2d). 3 (50 mg, 0.0798 mmol), D-biotin (25.3 mg, gated secondary antibody (1:3,000 dilution) and visualized by
0.1037 mmol), DCC (32.9 mg, 0.1596 mmol) and a catalytic the ECL detection reagent (Amersham).
amount of DMAP in CH2Cl2 (2 mL) was sonicated for 6 h. The
reaction mixture was concentrated under reduced pressure and Chemical precipitation from cells. K562 cells were treated
the resulting residue was purified by preparative TLC (CH2Cl2/ with the indicated compounds for 4 h, after which cells were
MeOH-NH3 (7 N), 10:1) to give 31.9 mg (47%) of 2d. MS collected and washed three times with PBS. Protein extracts
(ESI) m/z: 852.3 [M + H]+; HPLC: tR = 8.82.
were prepared by sonicating cells in 20 mM HEPES, pH 7.3,
50 mM KCl, 5 mM MgCl2, 20 mM Na2MoO4, 0.01% NP40
lysis buffer. Streptavidin agarose beads (40 μL) (Thermo Scien-
Biological evaluation of probes
Hsp90 competition assay. For the competition studies, fluores- tific) were washed three times with the lysis buffer and added to
cence polarization (FP) assays were performed as previously 500 μg of the total cellular protein extract diluted in lysis buffer
reported [9,23]. Briefly, FP measurements were performed on to a final volume of 120 μL. Samples were incubated at 4 °C for
an Analyst GT instrument (Molecular Devices, Sunnyvale, CA). 1 h, washed five times with the lysis buffer (or high salt buffer
Measurements were taken in black 96-well microtiter plates containing 1 M NaCl added to the lysis buffer) and applied to
(Corning # 3650) where both the excitation and the emission SDS-PAGE. Gels were stained with Coomassie blue (BioRad)
occurred from the top of the wells. A stock of 10 µM GM-cy3B according to the manufacturer's instructions.
was prepared in DMSO and diluted with Felts buffer (20 mM
Hepes (K), pH 7.3, 50 mM KCl, 2 mM DTT, 5 mM MgCl2, Competitive binding. K562 cells were pretreated with PU-H71
20 mM Na2MoO4, and 0.01% NP40 with 0.1 mg/mL BGG). To (5 μM) for 30 min, followed by treatment for 4 h with the indi-
each 96-well were added 6 nM fluorescent GM (GM-cy3B), cated biotinylated probe (5 μM). Cells were washed three times
3 µg SKBr3 lysate (total protein), and test compound (initial with PBS and sonicated in Felts buffer. Protein (500 μg) was
stock in DMSO) in a final volume of 100 µL Felts buffer. Com- added to streptavidin beads, and samples were incubated for 1 h
pounds were added in triplicate wells. For each assay, back- at 4 °C. Affinity-purified protein was washed and then applied
ground wells (buffer only), tracer controls (free, fluorescent GM to SDS-PAGE.
only) and bound GM controls (fluorescent GM in the presence
of SKBr3 lysate) were included on each assay plate. GM was Chemical precipitation from cell lysates. K562 cells were
used as positive control. The assay plate was incubated on a sonicated in 20 mM HEPES, pH 7.3, 50 mM KCl, 5 mM
shaker at 4 °C for 24 h and the FP values in mP were measured. MgCl2, 20 mM Na2MoO4, and 0.01% NP40 lysis buffer
The fraction of tracer bound to Hsp90 was correlated to containing added protease inhibitors. Affinity beads were
the mP value and plotted against values of competitor concen- prepared by addition of the biotinylated probes to the strepta-
trations. The inhibitor concentration at which 50% of bound vidin agarose resin (40 μL) (Thermo Scientific), which was first
GM was displaced was obtained by fitting the data. All experi- washed three times with the lysis buffer. Following incubation
mental data were analyzed using SOFTmax Pro 4.3.1 and at 4 °C for 1 h, the obtained Hsp90 affinity beads were washed
plotted using Prism 4.0 (Graphpad Software Inc., San Diego, three times with lysis buffer to remove any unbound materials.
CA).
The protein extract (500 μg) was then added to the probe-bound
beads, and samples were incubated at 4 °C overnight. Following
Western blotting. The K562 cell line was purchased from the five washes with lysis buffer, the protein isolates were subjected
American Type Culture Collection (Manassas, VA) and to SDS-PAGE.
cultured in Roswell Park Memorial Institute (RPMI) supple-
mented with 10% fetal bovine serum, 1% L-glutamine, 1% Growth inhibition assay. The effect of compounds on cell
penicillin and streptomycin. Cells were plated for 24 h prior to growth was evaluated with the Alamar Blue assay [26]. In
treatment for the indicated times with DMSO (vehicle) or with summary, K562 cells were plated at 20,000 cells/well on Costar
the indicated compounds. Protein extracts were prepared in 96-well plates. Treatment with the probes added at the indi-
554

Beilstein J. Org. Chem. 2013, 9, 544–556.
cated concentrations in triplicate wells was performed on the Acknowledgements
subsequent day and lasted for 72 h. The Alamar Blue reagent G.C. is funded by Leukemia and Lymphoma Society, Breast
resazurin (440 μM stock) was added at the end of the treatment Cancer Research Fund, W.H. Goodwin and A. Goodwin and the
to result in a final concentration of 50 μM. Plates were read 6 h Commonwealth Cancer Foundation for Research, The Experi-
later by using the Analyst GT instrument (Fluorescence inten- mental Therapeutics Center of Memorial Sloan-Kettering
sity mode, excitation 530 nm, emission 580 nm, with 560 nm Cancer Center (MSKCC), 1U01 AG032969-01A1, 1R21
dichroic mirror). Results were analyzed in SoftMax Pro. The CA158609-01A1, 1R21 AI090501 and 1R01 CA155226-01.
percentage of cell growth inhibition was calculated by T.T. is funded by Susan G. Komen for the Cure (KG091313)
comparing fluorescence readings obtained from treated versus and the Department of Defense, Breast Cancer Research
control cells, accounting for the initial cell population (time Program (PDF-BC093421). We also thank Dr. George
zero). The IC50 was calculated as the drug concentration that Sukenick and Dr. Hui Liu of the NMR Analytical Core Facility
inhibits cell growth by 50%.
at MSKCC for expert mass spectral analysis.
References
Flow cytometry analysis. Live cells. K562 cells were
pretreated for 4 h with the indicated biotinylated probe, washed
and stained on ice with CD45-Allophycocyanin (APC) (eBio-
science) in PBS/5% FBS for 30 min. Cells were then washed
and stained on ice with 0.125 µg of Anti-Biotin-PE in PBS/5%
FBS for 45 min, followed by 4',6-diamidino-2-phenylindole
(DAPI) (1 µg/mL) staining. Mean fluorescence intensity (MFI)
of phycoerythrin (PE) was determined in DAPI negative viable
cells. Digitonin permeabilized cells. K562 cells were
pretreated for 4 h with the indicated biotinylated probes, washed
and stained on ice with CD45-APC in PBS/5% FBS for 30 min.
CD45 is expressed on the cell surface of all hematopoietic cells
excluding mature erythrocytes and platelets. Cells were then
fixed for 30 min with Cytofix buffer (BD Biosciences), washed
and permeabilized with digitonin (10 µg/mL), followed by
washing and staining with 0.125 µg of anti-Biotin-PE in the
presence of digitonin for 30 min. Cells were then stained with
DAPI (1 µg/mL). Cells were washed and then analyzed by flow
cytometry (LSR-II, BD Biosciences).
1. Taipale, M.; Jarosz, D. F.; Lindquist, S. Nat. Rev. Mol. Cell Biol. 2010,
11, 515–528. doi:10.1038/nrm2918
2. Workman, P.; Burrows, F.; Neckers, L.; Rosen, N.
Ann. N. Y. Acad. Sci. 2007, 1113, 202–216.
doi:10.1196/annals.1391.012
3. Janin, Y. L. Drug Discovery Today 2010, 15, 342–353.
doi:10.1016/j.drudis.2010.03.002
4. Jhaveri, K.; Taldone, T.; Modi, S.; Chiosis, G. Biochim. Biophys. Acta
2012, 1823, 742–755. doi:10.1016/j.bbamcr.2011.10.008
5. Patel, H. J.; Modi, S.; Chiosis, G.; Taldone, T.
Expert Opin. Drug Discovery 2011, 6, 559–587.
doi:10.1517/17460441.2011.563296
6. Chène, P. Nat. Rev. Drug Discovery 2002, 1, 665–673.
doi:10.1038/nrd894
7. Porter, J. R.; Fritz, C. C.; Depew, K. M. Curr. Opin. Chem. Biol. 2010,
14, 412–420. doi:10.1016/j.cbpa.2010.03.019
8. Donnelly, A.; Blagg, B. S. J. Curr. Med. Chem. 2008, 15, 2702–2717.
doi:10.2174/092986708786242895
9. Taldone, T.; Zatorska, D.; Patel, P. D.; Zong, H.; Rodina, A.; Ahn, J. H.;
Moulick, K.; Guzman, M. L.; Chiosis, G. Bioorg. Med. Chem. 2011, 19,
2603–2614. doi:10.1016/j.bmc.2011.03.013
10.Taldone, T.; Gomes-DaGama, E. M.; Zong, H.; Sen, S.; Alpaugh, M. L.;
Zatorska, D.; Alonso-Sabadell, R.; Guzman, M. L.; Chiosis, G.
Bioorg. Med. Chem. Lett. 2011, 21, 5347–5352.
Fluorescence microscopy. K562 cells were treated with 10 μM
2g or DMSO (control) at 37 °C for 4 h. Cells were then
collected, washed twice with PBS and attached to a chamber
slide by centrifugation at 1,000 rpm at 4 °C for 5 min. Cells
were then fixed with 4% paraformaldehyde in PBS at room
temperature for 15 min and then washed twice with PBS. Cells
were permeabilized with 50 μg/µL digitonin (Gold Biotech-
nology special grade Cat# D-180-250) in PBS at room tempera-
ture for 15 min. Cells were washed twice with PBS and incu-
bated in 10% BSA in PBS at room temperature for 1.5 h. Cells
were then washed twice with PBS and incubated with Anti-
Biotin-FITC antibody (Sigma cat# F6762), diluted 1:50 in PBS,
at room temperature for 1 h. Cells were then washed twice with
PBS and stained with DAPI in ProLong Gold anti-fade reagent
(Life Technologies cat# P36935) at which point a cover slip
was attached to the chamber slide. Slides were visualized using
a Leica SP5 Upright point-scanning confocal microscope at an
objective of 40× oil (x = 2048, y = 2048, z = 1).
doi:10.1016/j.bmcl.2011.07.026
11.Hughes, P. F.; Barrott, J. J.; Carlson, D. A.; Loiselle, D. R.;
Speer, B. L.; Bodoor, K.; Rund, L. A.; Haystead, T. A. J.
Bioorg. Med. Chem. 2012, 20, 3298–3305.
doi:10.1016/j.bmc.2012.03.043
12.Taldone, T.; Chiosis, G. Curr. Top. Med. Chem. 2009, 9, 1436–1446.
doi:10.2174/156802609789895737
13.Moulick, K.; Ahn, J. H.; Zong, H.; Rodina, A.; Cerchietti, L.;
Gomes-DaGama, E. M.; Caldas-Lopes, E.; Beebe, K.; Perna, F.;
Hatzi, K.; Vu, L. P.; Zhao, X.; Zatorska, D.; Taldone, T.;
Smith-Jones, P.; Alpaugh, M.; Gross, S. S.; Pillarsetty, N.; Ku, T.;
Lewis, J. S.; Larson, S. M.; Levine, R.; Erdjument-Bromage, H.;
Guzman, M. L.; Nimer, S. D.; Melnick, A.; Neckers, L.; Chiosis, G.
Nat. Chem. Biol. 2011, 7, 818–826. doi:10.1038/nchembio.670
14.Darby, J. F.; Workman, P. Nature 2011, 478, 334–335.
doi:10.1038/478334b
15.DeBoer, C.; Meulman, P. A.; Wnuk, R. J.; Peterson, D. H. J. Antibiot.
1970, 23, 442–447. doi:10.7164/antibiotics.23.442
555

Beilstein J. Org. Chem. 2013, 9, 544–556.
16.Whitesell, L.; Mimnaugh, E. G.; De Costa, B.; Myers, C. E.;
Neckers, L. M. Proc. Natl. Acad. Sci. U. S. A. 1994, 91, 8324–8328.
doi:10.1073/pnas.91.18.8324
17.Clevenger, R. C.; Raibel, J. M.; Peck, A. M.; Blagg, B. S. J.
J. Org. Chem. 2004, 69, 4375–4380. doi:10.1021/jo049848m
18.Tsaytler, P. A.; Krijgsveld, J.; Goerdayal, S. S.; Rüdiger, S.;
Egmond, M. R. Cell Stress Chaperones 2009, 14, 629–638.
doi:10.1007/s12192-009-0115-z
19.Immormino, R. M.; Kang, Y.; Chiosis, G.; Gewirth, D. T. J. Med. Chem.
2006, 49, 4953–4960. doi:10.1021/jm060297x
20.He, H.; Zatorska, D.; Kim, J.; Aguirre, J.; Llauger, L.; She, Y.; Wu, N.;
Immormino, R. M.; Gewirth, D. T.; Chiosis, G. J. Med. Chem. 2006, 49,
381–390. doi:10.1021/jm0508078
21.Biamonte, M. A.; Shi, J.; Hong, K.; Hurst, D. C.; Zhang, L.; Fan, J.;
Busch, D. J.; Karjian, P. L.; Maldonado, A. A.; Sensintaffar, J. L.;
Yang, Y.-C.; Kamal, A.; Lough, R. E.; Lundgren, K.; Burrows, F. J.;
Timony, G. A.; Boehm, M. F.; Kasibhatla, S. R. J. Med. Chem. 2006,
49, 817–828. doi:10.1021/jm0503087
22.Andrés, J. M.; Manzano, R.; Pedrosa, R. Chem.–Eur. J. 2008, 14,
5116–5119. doi:10.1002/chem.200800633
23.Du, Y.; Moulick, K.; Rodina, A.; Aguirre, J.; Felts, S.; Dingledine, R.;
Fu, H.; Chiosis, G. J. Biomol. Screening 2007, 12, 915–924.
doi:10.1177/1087057107306067
24.Schulte, T. W.; An, W. G.; Neckers, L. M.
Biochem. Biophys. Res. Commun. 1997, 239, 655–659.
doi:10.1006/bbrc.1997.7527
25.Hawley, T. S.; Hawley, R. G., Eds. Methods in Molecular Biology: Flow
Cytometry Protocols, 2nd ed.; Humana Press Inc.: Totowa, NJ, 2004.
doi:10.1385/1592597734
26.O'Brien, J.; Wilson, I.; Orton, T.; Pognan, F. Eur. J. Biochem. 2000,
267, 5421–5426. doi:10.1046/j.1432-1327.2000.01606.x
License and Terms
This is an Open Access article under the terms of the
Creative Commons Attribution License
(http://creativecommons.org/licenses/by/2.0), which
permits unrestricted use, distribution, and reproduction in
any medium, provided the original work is properly cited.
The license is subject to the Beilstein Journal of Organic
Chemistry terms and conditions:
(http://www.beilstein-journals.org/bjoc)
The definitive version of this article is the electronic one
which can be found at:
doi:10.3762/bjoc.9.60
556