Discovery of a novel small molecule inhibitor targeting the frataxin/ubiquitin interaction via structure-based virtual screening and bioassays

Article abstract of DOI:10.1021/jm3017199

Friedreich's ataxia (FRDA) is an autosomal recessive neuro- and cardiodegenerative disorder for which there are no proven effective treatments. FRDA is caused by decreased expression and/or function of the mitochondrial protein frataxin. Here, we report f

Full text of DOI:10.1021/jm3017199

Subscriber access provided by DUESSELDORF LIBRARIES
Article
Discovery of a Novel Small Molecule Inhibitor Targeting the Frataxin/
Ubiquitin Interaction via Structure-Based Virtual Screening and Bioassays
Antonio Lavecchia
J. Med. Chem., Just Accepted Manuscript • DOI: 10.1021/jm3017199 • Publication Date (Web): 18 Mar 2013
Downloaded from http://pubs.acs.org on March 26, 2013
Just Accepted
“Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted
online prior to technical editing, formatting for publication and author proofing. The American Chemical
Society provides “Just Accepted” as a free service to the research community to expedite the
dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts
appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been
fully peer reviewed, but should not be considered the official version of record. They are accessible to all
readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered
to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published
in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just
Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor
changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers
and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors
or consequences arising from the use of information contained in these “Just Accepted” manuscripts.
Journal of Medicinal Chemistry is published by the American Chemical Society. 1155
Sixteenth Street N.W., Washington, DC 20036
Published by American Chemical Society. Copyright © American Chemical Society.
However, no copyright claim is made to original U.S. Government works, or works
produced by employees of any Commonwealth realm Crown government in the course
of their duties.

Page 1 of 54
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Discovery of a Novel Small Molecule Inhibitor Targeting the
Frataxin/Ubiquitin Interaction via Structure-Based Virtual
Screening and Bioassays
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
†
,
†
†
‡
Antonio Lavecchia, * Carmen Di Giovanni, Carmen Cerchia, Annapina Russo,
‡
†
Giulia Russo, Ettore Novellino
†
“
‡
Dipartimento di Chimica Farmaceutica e Tossicologica, “Drug Discovery” Laboratory, Università di Napoli
Federico II”, Via Domenico Montesano 49, 80131 Napoli, Italy
Dipartimento di Biochimica e Biotecnologie Mediche, Università di Napoli “Federico II”, Via Sergio Pansini 5,
8
0131 Napoli, Italy
1
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 2 of 54
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
ABSTRACT
Friedreich ataxia (FRDA) is an autosomal recessive neuroꢀ and cardiodegenerative disorder for
which there are no proven effective treatments. FRDA is caused by decreased expression and/or
function of the mitochondrial protein frataxin. Here, we report findings that frataxin is degraded via
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
47
the ubiquitin–proteasomal pathway and that it is ubiquitinated at residue K in Caluꢀ6 cells. A
1
47
theoretical model of the frataxinꢀK /Ub complex, constructed by combining bioinformatics
interface predictions with informationꢀdriven docking, revealed a hitherto unnoticed, potential
ubiquitinꢀbinding domain in frataxin. Through structureꢀbased virtual screening and cellꢀbased
assays, we discovered a novel small molecule (compound (+)ꢀ11) able to prevent frataxin
ubiquitination and degradation. (+)ꢀ11 was synthesized and tested for specific binding to frataxin by
an UFꢀLC/MS based ligandꢀbinding assay. Followꢀup scaffoldꢀbased searches resulted in the
identification of a lead series with micromolar activity in disrupting the frataxin/Ub interaction.
This study also suggests that frataxin could be a potential target for FRDA drug development.
2
ACS Paragon Plus Environment

Page 3 of 54
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
INTRODUCTION
Friedreich’s Ataxia (FRDA) is an inherited recessive neurodegenerative disorder caused by a
1,2
partial reduction in levels of the mitochondrial protein frataxin, which controls the iron
homeostasis. FRDA is the most common inherited form of ataxia with an incidence estimated
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
3
ꢀ5
between 1:30,000 to 1:50,000 in the US and Europe. The disease is characterized by a progressive
6
,7
neuropathy affecting the central and peripheral nervous systems, causing death of primary
sensory neurons of the dorsal root ganglia and a variable picture of accompanying neurological
4
symptoms. These symptoms tend to appear by the age of 20 in the majority of patients, who are
typically normal at birth and during early childhood, although the disease onset is highly variable
8
and can be incomplete and/or delayed for reasons as yet to be described. FRDA is caused, in the
majority of cases, by an abnormal GAA repeat expansion in the first intron of the human FRDA
2
9
gene, which inhibits transcription, leading in turn to multiple enzyme deficits, mitochondrial
1
0,11
dysfunction and oxidative damage.
At present there is no effective pharmacological treatment
available to slow down the progression of the disease.
Frataxin is encoded in the nucleus and synthesized in the cytoplasm as a precursor polypeptide
1–210
(frataxin
) that is transported to the mitochondrial matrix and proteolytically cleaved to the
81–210 12,13
42–210 14,15
mature form (frataxin
)
via a processing intermediate (frataxin
).
Maturation of the
frataxin precursor occurs within the mitochondrial matrix and no other intraꢀmitochondrial postꢀ
translational modifications have been identified. Recently, a pool of mature frataxin was detected in
1
6ꢀ18
the cytoplasm of several cell types of human origin
where participates in numerous biological
16,17,19
functions.
More recent findings have shown the direct regulation of frataxin precursor and
2
0
mature accumulation through the ubiquitinꢀproteasome system (UPS). Thus, smallꢀmolecule
inhibitors that promote frataxin stabilization are desirable and could potentially have therapeutic
value.
Here we report the discovery, synthesis, and characterization of a smallꢀmolecule (compound
(
+)ꢀ11) able to prevent the frataxin ubiquitination and degradation and promote the accumulation of
3
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 4 of 54
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
cellular frataxin. To reach this goal, we built a computational model of the frataxin/Ub interaction,
2
1
using HADDOCK (highꢀambiguityꢀdriven docking), a software of wide use in structural biology
22
for proteinꢀprotein docking. The putative frataxin Ubꢀbinding domain was chosen for in silico
targeting in a multiꢀstep structureꢀbased virtual screening approach using the National Cancer
Institute (NCI) Open Database. The most active compound (±)ꢀ11 was synthesized by a new
microwaveꢀassisted procedure and the resolution of its enantiomers was put forward by a salt
crystallization technique. Only the (+)ꢀisomer resulted active in accumulating both precursor and
mature frataxin in human lung adenocarcinoma (Caluꢀ6) cells. The binding affinity of (+)ꢀ11
toward frataxin was proved through an ultrafiltrationꢀliquid chromatography/mass spectrometry
(UFꢀLC/MS) based ligandꢀbinding assay.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
4
ACS Paragon Plus Environment

Page 5 of 54
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
RESULTS AND DISCUSSION
Frataxin Stability is Controlled by the UPS in Calu-6 cells. Previous results showed that
proteasomal degradation of both precursor and mature frataxin in human embryonic kidney 293
147
(HEKꢀ293) cells is mediated by a Ubꢀdependent mechanism and that K is the critical residue
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
2
0
responsible for frataxin ubiquitination. To validate these findings, we repeated these experiments
in Caluꢀ6 cells and analyzed frataxin turnover in the presence and absence of proteasome inhibition.
Caluꢀ6 cells, an anaplastic carcinoma of lung origin, were chosen as a model system in this study
because they are well characterized, provide a reasonably high transfection efficiency, and express
elevated levels of frataxin. Caluꢀ6 cells were transiently transfected with a construct expressing
frataxin fused to His tag and 32 hours after transfection cells were treated for 18 hours with either
1
0 ꢁM of the reversible proteasome inhibitor MG132 (NꢀcarbobenzoxylꢀLꢀleucinylꢀLꢀleucinylꢀLꢀ
leucinal) or 10 ꢁM Bortezomib, a peptide boronic acid proteasome inhibitor. Proteins from cell
lysates were then analysed by SDSꢀPAGE and Western blotting. As shown in Figure 1, the addition
of the reversible proteasome inhibitor MG132 and Bortezomib was found to cause a strong increase
in frataxin precursor compared to untreated samples.
Insert Figure 1
In order to verify whether the frataxin precursor is targeted to proteasomal degradation by Ub
labelling, we went on to perform an immunoprecipitation assay. Caluꢀ6 cells were transiently coꢀ
transfected with plasmids expressing both frataxin and Ub fused to His (6ꢀhistidine) and HA
(
haemagglutinin) tags, respectively. 24 hours after the transfection, cells were treated with 10 ꢁM of
MG132 for 18 hours. Frataxin was then specifically immunoprecipitated from untransfected and
transfected cells using antiꢀfrataxin, and the resulting frataxin immunocomplexes were analyzed by
SDSꢀPAGE and Western blotting with antibodies against Ub. As shown in Figure 2, both frataxin
and the monoUbꢀtagged precursor of frataxin accumulate by MG132 treatment. Interestingly, the
detection of a ladder of high molecular weight species characteristic of polyubiquitinated proteins
suggested that frataxin is also polyubiquitinated.
5
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 6 of 54
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
Insert Figure 2
1
47
To confirm that K is really the ubiquitination site of frataxin, we altered this residue to R by siteꢀ
147
directed mutagenesis, and compared the ability of the resulting mutant (HisꢀK R) to bind to
glutathione Sꢀtransferase (GST)ꢀfused Ub (GSTꢀmonoUb) with that of the frataxin wildꢀtype. Caluꢀ
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
6
cells were transiently transfected with the plasmid encoding either Hisꢀfrataxin or the mutant Hisꢀ
1
47
K
R. 24 hours after the transfection, cells were treated with MG132 and lysed. Hisꢀtagged
proteins were then tested for their ability to bind immobilized GSTꢀtagged Ub protein. The purified
GST and GSTꢀmonoUb were immobilized using glutathione (GSH)ꢀsepharose beads, and incubated
1
47
with cell lysates containing the recombinant Hisꢀfrataxin or the mutant HisꢀK R. After binding,
the proteins were eluted and analysed by Western blotting with antibodies against theꢀHis epitope.
Figure 3 shows that the signal for frataxin was present in the pullꢀdown sample containing the
147
147
wildeꢀtype frataxin, but not in that of the mutant HisꢀK R, thereby confirming that K
is
essential for binding with Ub. The absence of a signal for frataxin in the pullꢀdown preparation of
GST confirmed the specificity of the assay.
Insert Figure 3
1
47
Structural Model of the Frataxin-K /Ub Complex. A computational model of the frataxin/Ub
complex was constructed considering the formation of a covalent isopeptide bond between the
7
6
147
2
carboxyl group of the Cꢀterminal G of Ub and the εꢀNH group of frataxin K with the help of
21
the HADDOCK algorithm. The coordinates for the mature human frataxin (residues 90ꢀ208) and
23,24
Ub were taken from the current crystal structures
without modifications. The residues
147
interacting across the frataxinꢀK /Ub complex interface were predicted by two interface prediction
2
5
26
programs, WHISCY and ProMate, which generate the ambiguous interaction restraints (AIRs)
(Table 1).
Insert Table 1
8
44
47
70
Prediction programs identified L , I , G , and V as Ub active residues. Interestingly, they
coincide with the residues known to form the hydrophobic patch involved in most of the
6
ACS Paragon Plus Environment

Page 7 of 54
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
2
7,28
144
148
monoUb/Ubꢀbinding proteins interactions.
The predicted frataxin active residues V , Q ,
1
50
151
155
157
159
162
163
165
175
P
, N , W , S , P , G , P , R , Y , located on the external surface of the βꢀsheet,
match with the amino acids highly conserved in all eukaryote and prokaryote sequences of frataxin
Table 1). The majority of these residues have no charge, making a flat nearly neutral surface,
suitable to mediate proteinꢀprotein interactions.
(
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
76
147
The covalent isopeptide bond between the G COOH group of Ub and the εꢀNH
2
group of K
was introduced using unambiguous distance restraints based on typical interatomic distances for a
2
2
peptide bond in crystal structures, as described previously. HADDOCK generated 200 structures,
which were then subjected to clustering. Figure 4 shows a plot of the intermolecular energy, Einter
,
of the 200 complex structures as a function of their backbone rootꢀmeanꢀsquare deviation (rmsd)
from the lowest energy structure. After analysis, two clusters were obtained. Their statistical results
are summarized in Table 2.
Insert Figure 4 and Table 2
147
Cluster 1 of frataxinꢀK /Ub complex, containing 60 members, had the best average Einter (ꢀ
81.9±103.1 kcal/mol). Likewise, all cluster evaluation parameters showed a significantly better
2
performance of the bestꢀscoring cluster 1 versus cluster 2. For example, structures from the bestꢀ
scoring cluster 1 showed the largest buried surface area (BSA) at the interface (1325±94.8 Å),
suggesting that solutions of the bestꢀscoring cluster present sensible models for the Ub/frataxin
interaction. Accordingly, the five lowestꢀenergy structures of this cluster were selected for further
analysis. As depicted in Figure 5, the structure of the complex is rather well defined, with a
backbone rmsd on interface of 0.51±0.16 Å, and of good quality, with more than 96% of the
residues in the most favored permitted regions of the Ramachandran plot (see Experimental
Section).
Insert Figure 5
A large number of intermolecular hydrophobic contacts were identified, together with a few Hꢀ
74
92
155
bonds and saltꢀbridges, the major one being between Ub R and frataxin D . The W side chain
7
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 8 of 54
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
of frataxin, which points toward the centre of the interacting surface of Ub, is in close contact with
8
70
71
141
the Ub L , V and L side chains. Moreover, the frataxin G αꢀH makes intimate contact with Ub
47
G . Contribution to the stability of the complex also comes from the hydrophobic interactions
1
63
144
44
8
70
between P and V of frataxin and I , L and V of Ub. The three highly conserved frataxin
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
155
163
144
residues W , P and V seem to form a “trident” that anchors the frataxin domain to the
8
44
47
70
hydrophobic patch of Ub, including L , I , G , and V residues (Figure 6). Notably, frataxin
155
residue W and surrounding residues were shown to contribute to binding interactions with partner
2
9ꢀ32
proteins and also be important for frataxin function in Fe–S cluster biosynthesis.
Insert Figure 6
In Silico Screening of the Open NCI Library on Frataxin. With the goal of preventing
frataxin/Ub association, we went on to search for small molecules capable of directly targeting the
frataxin region that binds Ub. A multiꢀstep structureꢀbased virtual screening of more than 65,000
33
leadꢀlike compounds obtained from the NCI Open Database was carried out. This is a large public
database (260,071 compounds) from which sample compounds can be obtained through the
NCI/Developmental Therapeutics Program. In order to focus the virtual screening on compounds
suitable for further development, we selected a subset of leadꢀlike compounds (referred in this work
34,35
as the NCI leadꢀlike set).
The selection was made based on properties and functional groups
3
6
using the FILTER program that reduced the initial NCI Database to a subset of 65,375
compounds. After selecting leadꢀlike compounds, a fast docking protocol was employed to further
refine the NCI leadꢀlike set. The general workflow of the multiꢀstep docking approach implemented
in this work is presented in Figure 7. Structureꢀbased screening was conducted using two docking
3
7
38
programs, GLIDE and AUTODOCK. Consensus hits were selected and tested experimentally.
Details of the workflow, screening compound numbers and filters used for the virtual screening are
reported in the Experimental Section.
Insert Figure 7
8
ACS Paragon Plus Environment

Page 9 of 54
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
In total, 25 consensus compounds with favorable docking scores were identified, and of these, 13
were selected for experimental testing based on 3D visualization and assessment in the frataxin/Ub
interaction area, including the fit to the binding site shape, ligand conformation, and ability to form
Hꢀbonds with surrounding residues. The chemical structures of the selected hits are shown in Figure
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
8
, while their docking scores and leadꢀlike properties are available as Supporting Information.
Insert Figure 8
The 13 compounds were tested for their ability to prevent the frataxin ubiquitination. To this aim,
Caluꢀ6 cells were transiently transfected with Hisꢀfrataxin. 24 hours after the transfection,
untransfected and transfected cells were treated with either 10 ꢁM MG132 for 18 hours or 100 ꢁM
of each selected molecule for 72 hours. Untreated and treated cells were then harvested and protein
extracts were subjected to Western blotting analysis. All solutions were carefully monitored to
avoid artifacts due to precipitation or agglomeration of compounds. Under assay conditions,
aggregation of active compounds was not detectable. In addition, none of these compounds had
proteinꢀreactive groups. Among the tested compounds, the (±)ꢀ6ꢀ((2,3ꢀdimethoxyphenyl)(4ꢀ
morpholinyl)methyl)ꢀ1,3ꢀbenzodioxolꢀ5ꢀol ((±)ꢀ11) was found to be the most effective in
preventing the frataxin ubiquitination. Since compound 11 was tested as a racemic mixture, it was
synthesized and resolved in its enantiomers (+)ꢀ11 and (ꢀ)ꢀ11, which were assayed individually and
compared to the racemate.
Synthesis and Resolution of Racemic Mixture of 11. Reaction between morpholine (a), 2,3ꢀ
dimethoxybenzaldehyde (b) and benzo[d][1,3]dioxolꢀ5ꢀol (c) resulted in the synthesis of (±)ꢀ11
(Scheme 1). This reaction was studied under two conditions, as follows: pathway A) reflux in
ethanol for 72 h; pathway B) solventꢀfree microwave irradiation using CEM Discover S Class
o
microwave oven at 125 C for five minutes in absence of any catalyst.
Insert Scheme 1
The pathway A suffers from long reaction times up to several days and produces modest yields of
product (about 47%) with a moderate enantiomeric excess (57 %) for the (+)ꢀisomer. On the
9
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 10 of 54
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
contrary, the microwaveꢀassisted solventꢀfree reaction (pathway B), a most efficient synthetic
method in terms of energy and time consumption, furnishes better yields of the product (87%), with
an equimolar presence of the two (+)ꢀ and (ꢀ)ꢀenantiomers.
Based on these results, a hypothetical mechanism of reaction could be proposed to justify the
different trend of the two pathways A and B (Scheme 2). In the both synthetic methods, the
secondary amine a attacks the carbonyl group of aldehyde b yielding a racemic hemiaminale
intermediate H. In pathway A, the electron poor carbon of H undergoes a SN2 displacement by the
carbon in alpha to the phenol OH group of c to give a transition state with the geometry of a
pentagonal bipyramid, probably stabilized by an intramolecular Hꢀbond between the nitrogen of the
morpholine and the phenol OH group. The formation of this sixꢀmembered Hꢀbond could
determinate a transient intermediate responsible of the excess of the (+)ꢀisomer of 11.
Under microwave irradiation (pathway B), a trigonal planar carbocation intermediate can be formed
starting from H. The attack of phenol c, occuring onto either face of the carbocation, yields the
raceme mixture of 11.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Insert Scheme 2
The most practical method for the resolution of racemic amines is the preparation of diastereomeric
39
salts with an optically active acid, and then separation via crystallization. The resolution of (±)ꢀ11
was accomplished through the sequential use of Lꢀ and Dꢀtartaric acid (Scheme 3). When (±)ꢀ11
and Lꢀ(+)ꢀtartaric acid were used in a 1:0.5 ratio, (ꢀ)ꢀ11 and (+)ꢀ11 were isolated in 30% and 37%
yields with >99% and 97% ee, respectively. In an optimized protocol, (±)ꢀ11 and Lꢀ(+)ꢀtartaric acid
(0.5 equiv.) were mixed in ethanol and stirred overnight. The solid tartrate salt was separated from
the unreacted 11 through extraction with ether. Subsequent recrystallization of the salt from ethanol
and basification gave (ꢀ)ꢀ11. The (+)ꢀenantiomer was obtained from the mother liquor by similar
treatment with Dꢀ(ꢀ)ꢀtartaric acid. Both enantiomers were obtained in good yields and high
enantiomeric purity after a single crystallization of the corresponding tartrate salts. The
enantiomeric purity of both enantiomers was determined by chiral HPLC.
10
ACS Paragon Plus Environment

Page 11 of 54
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Insert Scheme 3
(
+)-11 Promotes the Accumulation of Frataxin Precursor in Calu-6 Cells. Figure 9 shows the
effect of 50 ꢀM of (±)ꢀ11, (+)ꢀ11 and (ꢀ)ꢀ11 on the accumulation of precursor and mature frataxin in
Caluꢀ6 cells. In comparison with the racemate (±)ꢀ11, the isomer (+)ꢀ11 indirectly induced a larger
increase in the cellular concentration of mature frataxin by significantly restoring the endogenous
level of frataxin precursor (Figure 9a). In contrast, (ꢀ)ꢀ11 displayed no increase in the cellular
concentration of frataxin, indicating a fine degree of selectivity in the binding site due to chiral
geometry.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
To verify that the increase of mature frataxin levels in cells is indeed due to the capacity of (+)ꢀ11
to prevent the conjugation of Ub molecule with frataxin precursor, Caluꢀ6 cells were transiently
transfected with the plasmid encoding Hisꢀfrataxin and 24 hours after the transfection, cells were
treated with either 10 ꢁM MG132 or 10 ꢁM MG132 and 50 ꢁM of (+)ꢀ11. Hisꢀfrataxin was then
analyzed for its ability to bind to immobilized GSTꢀmonoUb. The purified GST and GSTꢀmonoUb
were immobilized using GSHꢀsepharose beads and incubated with the cell lysates treated or not
with (+)ꢀ11. After the binding, the eluted proteins were analyzed by Western blotting with
antibodies antiꢀHis. Figure 9b shows that the signal for frataxin was present in the pullꢀdown
sample of untreated cells and absent in that of (+)ꢀ11 treated cells, indicating that this compound is
able to prevent the conjugation of Ub to frataxin precursor. Experiments conducted in Caluꢀ6 cells
with increasing compound concentrations showed a doseꢀdependent behavior of (+)ꢀ11 in
preventing the ubiquitination of frataxin (Figure S1 of Supporting Information). The IC50 for
inhibition of frataxin ubiquitination was determined to be 45 ꢀM. Cell viability was not impaired by
(+)ꢀ11 concentrations up to 100 ꢀM.
Insert Figure 9
Binding Affinity of Compound (+)-11 to Frataxin. To investigate the binding affinity of (+)ꢀ11
40,41
with frataxin, we employed an UFꢀLC/MS based ligandꢀbinding assay.
Initially in the binding
11
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 12 of 54
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
assay experiments compound (+)ꢀ11 (1 ꢀM) was incubated at 25 and 37 °C with 1 ꢀM for Hisꢀ
frataxin recombinant protein in a total volume of 200 ꢀL of assay buffer. After the incubation, the
unbound compound from the ligandꢀprotein complex was removed by washing with assay buffer.
Then, the ligand was dissociated from the complex and analyzed by LC/MS. In the positive
electrospray mass spectrum, the protonated molecule of compound (+)ꢀ11 was detected at m/z
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
+
374.1598 [M + H] . Figure 10 shows the LC/MS analysis of an ultrafiltrate obtained from the
incubation of test compound (+)ꢀ11 with frataxin protein at 25 °C. The peak area of compound (+)ꢀ
1
1 was enhanced 3.70ꢀfold in the chromatogram corresponding to the incubation with active
frataxin compared to the control experiment carried out with denatured protein, which showed that
compound (+)ꢀ11 possesses specific binding affinity to frataxin. The peak area enhancement of (+)ꢀ
1
1 was 3.32ꢀfold compared to control experiment when incubated at human body temperature of 37
°C. The relative binding affinity was ascertained by calculating the ratio between the average peak
area of compound incubated with active protein and the average peak area of compound incubated
with denatured protein. Our results show that (+)ꢀ11 has specific binding affinity to frataxin. We
attempted to investigate entropy and enthalpy changes by isothermal titration calorimetry, but these
experiments unfortunately did not yield interpretable data (data not shown).
Insert Figure 10
Although the absolute configuration of (+)ꢀ11 is not known at the moment, we assume that it has R
configuration, since docking calculations predicted the R enantiomer to bind more favorably to
frataxin than the corresponding S enantiomer. We are currently attempting to crystallize the (+)ꢀ
1
1/frataxin complex. Figure 11 illustrates a model of the (R)ꢀ11/frataxin complex as predicted by
144
146
155
157
162
163
165
GLIDE. According to this model, residues V , N , W , S , G , P and R form the
binding pocket of (R)ꢀ11, and a network of Hꢀbonds is predicted between the ligand and the side
1
46
157
165
chains of N , S and R . In particular, the metaꢁmethoxy oxygen of phenyl ring of 11 forms
165
bidentate Hꢀbonds to the R side chain, while the OH oxygen of the benzodioxole system accepts
146
a Hꢀbond from the NH carboxamide of N side chain. The morpholine oxygen atom is engaged in
12
ACS Paragon Plus Environment

Page 13 of 54
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
1
57
a Hꢀbond with the S OH group. Furthermore, the ligand makes hydrophobic contacts with
1
55
163
144
residues W , P and V . Specifically, the dimetoxyphenyl ring and the benzodioxole aromatic
ring of the ligand are oriented is such a way to establish parallelꢀdisplaced and Tꢀshaped πꢀπ
stacking interactions with W155 side chain.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Insert Figure 11
Scaffold Searching and SAR Development. Hierarchical screening is an efficient strategy
allowing an initial broad search over a chemically and pharmacologically diverse set of compounds,
followed by a focused search over a much larger database to find molecules related to potential lead
compounds. The most promising compound 11 was selected for scaffoldꢀsearching and SAR
expansion studies. Substructure and similarity followꢀup searches over the full NCI database
resulted in 38 compounds matching the scaffold, which were docked to the frataxin Ubꢀbinding site
and ranked according to the predicted GLIDE XP/AUTODOCK consensus scoring. The top 12
compounds with the lowest scoring value for frataxin Ubꢀbinding site were selected for
experimental testing (Table 3). It was gratifying that they all showed a doseꢀdependent behavior in
preventing the conjugation of Ub molecule to the precursor frataxin (Supporting Information Figure
S1) confirming the viability of the core structure from 11. Values of experimental IC50 (ꢂM) for
inhibition of frataxin ubiquitination were obtained as described in the Experimental Section. As
previously, aggregation effects were excluded, and compound identity and purity were confirmed
1
by elemental analysis and H NMR (see Supporting Information).
Insert Table 3
As can be seen from data shown in Table 3, replacement of the morpholine ring with a piperidine
(
compound 23) or pyrrolidine ring (compound 24) led to a considerable reduction of activity. This
is due to the loss of the Hꢀbond between the morpholine oxygen and the OH group of S157. We
observed a nearly 2ꢀfold increase in inhibitory activity (from 45 to 20 ꢁM) when the metoxy
substituents were in meta and para positions of phenyl ring (compound 15), with a slight preference
for the meta over the para position in this series (cfr. compound 18 and 14). Compounds 14, 16 and
13
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 14 of 54
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
17, lacking of the metaꢀmethoxy substituent on the aromatic ring, resulted less potent than
compound 11. Docking of the two most active compounds 15 and 18 into the frataxin Ubꢀbinding
domain indicated that the oxygen atoms of the meta and para methoxy groups are optimally
oriented to form two stable Hꢀbonds to R165 side chain (Supporting Information Figure S2).
We also observed that halogens (compounds 20 and 21) and basic (compound 22) substituents at
para position of the phenyl ring led to a nearly 1.5ꢀfold decrease in activity over compound 11.
Finally, replacement of substituted phenyl ring with a benzodioxole system led to an drastic
reduction in activity. From a visual inspection of compound 25 complexed to Ubꢀbinding domain of
frataxin, it seems clear that the presence of a second benzodioxole ring in the molecule increases the
steric hindrance inside the binding cavity and changes the optimal binding mode of the ligand, thus
decreasing the relative stability of the complex.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
14
ACS Paragon Plus Environment

Page 15 of 54
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
CONCLUSION
Here we describe our efforts in identifying novel smallꢀmolecule inhibitors that perturb
conjugation of Ub onto frataxin. To this aim, we verified that treatment of Caluꢀ6 cells with
proteasome inhibitors (MG132 and Bortezomib) led to a significant accumulation of frataxin
precursor and, indirectly, of its mature form, thereby confirming the role of the UPS in the
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
1
47
degradation pathway of frataxin. Siteꢀdirected mutagenesis experiments showed that K /R mutant
of frataxin completely failed to conjugate Ub, suggesting that this lysine residue is its target site for
Ub conjugation. With the help of the proteinꢀprotein docking program HADDOCK, we built a
1
47
computational model of the frataxinꢀK /Ub complex. A small molecule (compound (±)ꢀ11)
targeting the Ubꢀbinding domain of frataxin was identified by a multiꢀstep structureꢀbased virtual
screening. This chiral compound was synthesized and resolved in its optical isomers. (+)ꢀ11
resulted the active isomer capable to block the frataxin ubiquitination and to promote the
accumulation of both precursor and mature frataxin species in Caluꢀ6 cells. The binding affinity of
(+)ꢀ11 for frataxin was investigated through the UFꢀLC/MS based ligandꢀbinding assay. The peak
area of the ligand in the chromatogram corresponding to incubation with active frataxin increased
by 3.70ꢀfold compared to the chromatogram of test compound incubated with denatured protein,
revealing thus the existence of a specific interaction between (+)ꢀ11 and frataxin. A docking model
elucidating the putative interactions between compound 11 and frataxin Ubꢀbinding site is
proposed. Based on this model, it can be hypothesized that compound 11 could to mask the residues
required for Ub interaction or to mask a putative sequence containing some form of signal that is
recognized by Ub machinery. Substructure and similarity followꢀup searches on the most active hit
compound 11 yielded a series of morpholino analogues with a key metaꢀ and paraꢀmethoxy
substituted phenyl ring that possessed activity in the micromolar range.
In conclusion, by combining theoretical and experimental approaches, we identified a small
molecule that disrupts the frataxin/Ub interaction determining a halfꢀlife increase of cellular
15
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 16 of 54
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
frataxin. This finding is very significant because the increment in the residual levels of frataxin
could open therapeutic perspectives for the FRDA disease.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
16
ACS Paragon Plus Environment

Page 17 of 54
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
EXPERIMENTAL SECTION
Interface Prediction. The Xꢀray coordinate file of human frataxin (PDB ID: 1EKG; Chain A)
23
24
42
and Ub (PDB ID: 1UBQ) were downloaded from the PDB, while the multiple sequence
43
alignments were taken from the HSSP database (ftp://ftp.cmbi.kun.nl/pub/molbio/data/hssp). The
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
2
5
26
first aligned sequence in the HSSP file was taken as master sequence. WHISCY and ProMate
were used for interface predictions and combined for consensus scoring using WHISCYMATE.
2
5
Multiple sequence alignments were used for WHISCY prediction: if there was any disagreement
between structure and master sequences about a residue identity, the residue of the master sequence
was used. The parts of a structure that were not present in the alignment were not predicted and
ignored in the evaluation. ProMate predictions were obtained making use of the Web interface of
ProMate (http://bioinfo.weizmann.ac.il/promate/promate.html) using default settings. For
WHISCYMATE, a residue was predicted if its ProMate score was higher than or equal to 98.520,
its WHISCY score higher than or equal to 0.371, or if its ProMate and WHISCY scores were both
higher than or equal to 55.420 and 0.107, respectively. Interface predictions were used to generate
25
AIRs as discussed before: predicted residues were designated active residues and their surface
neighbors passive residues (Table 1).
147
147
Docking for the Frataxin-K /Ub Complex. A structural model of frataxinꢀK /Ub complex was
21,22
obtained following a WHISCYꢀHADDOCK docking approach.
The crystal structures of the
23
24
human frataxin and Ub were used for mapping the interfaces and docking the structures. The
restraints that were used in this study are listed in Table 1. Ub and frataxin were linked to one
76
147
another via a G ꢀK isopeptide bond. To account for the possible close contact between the
8
44
47
70
hydrophobic patch on Ub (residues L , I , G , and V ) and the highly conserved residues on the
frataxin external surface, we incorporated AIRs where active and passive residues on both Ub and
2
5
76
147
frataxin were defined exploiting the interface prediction WHISCY program. The G ꢀK
isopeptide bond was modelled by including a set of distance restraints based on typical interatomic
22
distances for a peptide bond in crystal structures, as described previously. Backbone and side
17
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 18 of 54
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
chain flexibility was also included to account for a possible conformational rearrangement that
could occur at the interface between both Ub and frataxin moieties (Table 1). Flexible segments
were defined as stretches of active and passive residues plus one sequential neighbor. The docking
calculations were performed with the standard HADDOCK protocol as described by Dominguez et
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
21
al. For each run, 1000 rigidꢀbody docking solutions were first generated by energy minimization.
The driving force for the docking at this stage comes mainly from the AIRs and from van der Waals
and electrostatic energy terms once the structures are within the nonbonded cutꢀoff (8.5 Å). The 200
best solutions according to the AIR restraint energy (as defined in Ref. 21) were subjected to
semiflexible simulated annealing in torsion angle space followed by a final refinement in explicit
4
4
water. During the simulated annealing and the water refinement, the amino acids at the interface
(side chains and backbone) are allowed to move to optimize the interface packing. The nonbonded
45
energies were calculated with the OPLS parameters using a 8.5 Å distance cutꢀoff. A dielectric
constant of 10 was chosen for the vacuum stages of the docking protocol (rigidꢀbody and semiꢀ
flexible refinement).
The resulting structures were finally subjected to a final refinement in explicit water, clustered
using a 2.0 Å backbone rmsd cutoff criterion and sorted according to the intermolecular energy
(sum of the van der Waals, electrostatic, and AIRs energy). BSA was calculated by taking the
difference between the sum of the solventꢀaccessible surface area for each partner separately and
the solventꢀaccessible area of the complex. The solventꢀaccessible area was calculated using a 1.4 Å
water probe radius. The five best structures of the lowest energy cluster of each HADDOCK run
were analyzed in terms of intermolecular contacts, and an average structure was calculated by
superimposing the structures on the backbone atoms of the flexible segments. The Ramachandran
plot of the generated model shows 71.7% of the residues in the most favored regions, 24.9% in the
additional allowed regions, 1.1% in the generously allowed region, and 2.3% in the disallowed
regions. Intermolecular contacts (Hꢀbonds and nonbonded contacts) were analyzed with DIMPLOT,
which is part of the LIGPLOT software, using the default settings (3.9 Å heavyꢀatoms distance cutꢀ
18
ACS Paragon Plus Environment

Page 19 of 54
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
off for nonbonded contacts; 2.7 Å and 3.35 Å protonꢀacceptor and donorꢀacceptor distance cutꢀoffs,
4
6
respectively, with minimum 90º angles [DꢀHꢀA, HꢀAꢀAA, DꢀAꢀAA] for hydrogen bonds).
23
Protein Preparation. The Xꢀray coordinates of human frataxin was prepared using the Protein
Preparation Wizard implemented in the Schrodinger package using default options: bond orders
were assigned, hydrogens were added and all water molecules were deleted. Hydrogens were then
optimized using the exhaustive sampling option, and the protein was minimized to an rmsd limit
from the starting structure of 0.3 Å using the Impref module of Impact with the OPLS_2005 force
field.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
33
Database: Lead-Like Selection and Preparation. The NCI Open Database with 260,071
47
36
compounds was obtained from ZINC. The compound database was processed with FILTER,
version 2.0.1, to select a subset of leadꢀlike compounds. We used the default parameters in the leadꢀ
like filter without further modifications. The resulting database contained 65,375 compounds. For
docking, different protonation states and tautomers were generated with the GLIDE module
LIGPREP. This yielded a total of 90,653 structures.
Docking with GLIDE and AUTODOCK. We employed the highꢀthroughput virtual screening
3
7
(
HTVS) docking mode in GLIDE, version 5.7, for a rapid structureꢀbased filtering of the NCI
leadꢀlike set. This data set was docked onto the putative frataxin Ubꢀbinding domain. The gridꢀ
155
enclosing box, which must contain the center of each ligand docked, was centered on the W side
1
55
chain and defined to enclose residues located within 14 Å from W , while the outer box, in which
all parts of the ligand must bind, was 35 Å in each direction. A van der Waals radius scaling factor
of 0.80 for atoms with a partial atomic charge (absolute value) less than 0.15 was used in order to
soften the potential for nonpolar parts of theprotein. Next, a flexible docking with default
parameters was employed for virtual screening. The general workflow of the multiꢀstep docking
approach implemented is depicted in Figure 7. The topꢀranked compounds with GLIDE HTVS
(4532 unique molecules) were docked flexibly in a stepwise manner with GLIDE Standard
Precision (SP) and Extra Precision (XP). We employed default parameters with the same receptor
19
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 20 of 54
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
grids used with GLIDE HTVS. Selected compounds docked with GLIDE XP (226 unique
3
8
molecules) were also docked with AUTODOCK , version 4.2. We used the same grid maps
155
(
centered on W residue) and default docking parameters. The only modification was the number
of docking runs that was set to 10 for faster virtual screening. A consensus score was generated by
combining both the AUTODOCK Binding Energy and the XP GLIDE score energy in order to
increase confidence in the resulting energy score. High ranking poses by two different scoring
functions represent by construction a more reliable prediction than any of the constituent scoring
functions alone. In practice, consensus scoring has been generally found to improve virtual
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
48,49
screening performance dramatically with respect to the individual scoring functions.
Out of the
final 50 compounds identified from the multistep docking approach, 13 (Figure 8) were selected for
50
experimental testing. The program QikProp, version 3.4, was used for calculating the ADME
proprieties.
5
1
Scaffold Search. An online search utility provided by the NCI (http://129.43.27.140/ncidb2/) was
used to search the entire NCI database for compounds similar to compound 11. Two methods were
used to judge compound similarity: search on the basis of substructure by SMILES string
5
2
53
(
http://daylight.com) and/or similarity by Tanimoto coefficient, with a cutoff of 0.85. From these
searches, selected compound structures were docked to the frataxin Ubꢀbinding site and ranked
according to predicted GLIDE XP/AUTODOCK consensus scoring. Compounds with the lowest
scoring value were requested and assayed for effect on frataxin ubiquitination.
General Procedures. Reagents, starting material, and solvents were purchased from commercial
suppliers and were used as received. Analytical TLC was performed on a 0.25 mm layer of silica
2
gel 60 F254 from Merck, and preparative TLC was performed on 20 × 20 cm glass plates coated
with a 2 mm layer of silica gel PF254 from Merck. Silica gel 60 (300ꢀ400 mesh), Merck, was used
for flash chromatography. Melting points were taken on a Kofler apparatus and are uncorrected.
1
Optical rotations were determined with a PerkinꢀElmerꢀ241 MC polarimeter. H NMR spectra were
recorded with a Brukerꢀ500 spectrometer. Chemical shifts are reported in δ relative to internal
20
ACS Paragon Plus Environment

Page 21 of 54
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Me
4
Si, and J values are reported in Hz. Electrospray mass spectra were recorded using a WATERS
ZꢀQ mass spectrometer equipped with an electrospray ionization (ESI) probe operating in positive
or negative ion mode. CEM Discover SꢀClass microwave reactor was used for the microwaveꢀ
assisted reactions.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Synthesis of 6-((2,3-dimethoxyphenyl)(4-morpholinyl)methyl)-1,3-benzodioxol-5-ol ((±)-11).
Morpholine (a), 2,3ꢀdimethoxybenzaldehyde (b) and benzo[d][1,3]dioxolꢀ5ꢀol (c) were combined
and two reaction conditions were employed. Pathway A: a (870 ꢀL, 10 mmol), b (1.66 g, 10 mmol)
and c (1.38 g, 10 mmol) were dissolved in ethanol (40 ml) and refluxed for 72 hours. After the
reaction proceeded for the stated period of time, a mass of colorless crystals separated. These were
collected and recrystallized from acetone/methanol to give 1.75 g of 11 (yield 47%). Mp (°C): 141ꢀ
1
1
42. H NMR (500 MHz, CHCl ꢀd ) δ 7.32ꢀ762 (m, 3H), 6.43 (s, 1H), 6.38 (s, 1H), 5.82 (d, J = 1.0
3
1
Hz, 1H), 5.78 (d, J = 1.0 Hz, 1H), 4.96 (s, 1H), 3.90 (s, 3H), 3.83 (s, 3H), 3.70 (m, 4H), 2.52 (m,
+
4
H). ESIꢀMS: 374 (M + H) . Anal. Calcd for C20
H
6
23NO : C 64.33, H 6.21, N 3.75. Found: C 64.41,
H 6.26, N 3.80. Pathway B: A mixture of a (87 ꢀL, 1 mmol), b (166 mg, 1 mmol) and c (138 mg, 1
mmol) were mixed in a 50 mL flask and irradiated in a CEM Discover S Class microwave oven at
o
1
25 C for ten minutes in absence of any catalyst. The progress of the reaction was monitored with
TLC. After cooling, the mixture was extracted with chloroform. Evaporation of the solvent under
reduced pressure gave the crude product, which was recrystallized from acetone/methanol to give
3
24 mg of 11 (yield 87%).
Resolution of (±)-11. To a solution of Lꢀ(+)ꢀtartaric acid (0.15 g, 1.0 mmol) in ethanol (3 mL) was
added a solution of (±)ꢀ11 (1.5 g, 4 mmol) in ethanol (15 mL) and the resulting mixture was stirred
overnight at room temperature. Ethanol was then removed on a rotary evaporator at 40 ºC. To the
residue, diethyl ether (10 mL) was added and the mixture was stirred for 1 h. Filtration of the
reaction mixture provided the tartrate salt, which was recrystallized from ethanol (10 mL) to obtain
white crystals. This solid was treated with aqueous NaOH solution, then extracted with CH Cl ,
2
2
dried over Na SO , and the solvent removed under reduced pressure to afford a white powder 0.21 g
2
4
21
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 22 of 54
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
(yield 30%), mp 141ꢀ142 ºC. The enantiomeric purity was determined by chiral HPLC (Daicel
Chiralcel OD‐H, hexane/isopropanol = 97.5:2.5, flow rate 1.0 mL/min) and was found to be >99%,
2
5
[α
]
D
2 2
= ꢀ171.4 (c 0.42, CH Cl ). The second isomer of 11 was isolated from the mother liquor and
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
treated with Dꢀ(ꢀ)ꢀtartaric acid in ethanol as described above, providing the corresponding (+)ꢀ
25
D
enantiomer, 0.27 g (yield 37%); mp 139ꢀ140 ºC;
[α
]
= +151.2 (c 0.39, CH
2
2
Cl ). The enantiomeric
purity was determined by chiral HPLC and was found to be 97%.
Cell Cultures, Constructs and Transfections. The human lung carcinoma Caluꢀ6 cell line
(
ATCC, HTB 56; ICLC, HTL97003) was cultured in Dulbecco’s Modified Eagle’s Medium
DMEM) with glutamax (Invitrogen, Carlsbad, California) supplemented with 10% fetal bovine
(
serum (FBS), 2mM Lꢀglutamine, PenicillinꢀStreptomycin 50U/ml and 0.1 mM nonꢀessential amino
acids. Plasmids were transfected in Caluꢀ6 cell line by using LIPOTAXI (Agilent) according to the
manufacturer’s instructions. Transfection efficiency was assessed by coꢀtransfecting a GFPꢀ
expressing vector and normalizing RNA levels against GFP mRNA levels (data not shown). Where
indicated, 24 h after transfection, cells were treated with either 10 ꢁM proteosome inhibitors
(MG132, Bortezomib) for 18 h or with 100 ꢀM of each of the 13 molecules selected by the multiꢀ
step docking approach for 72 h. Then, proteins were extracted and analyzed by western blotting.
DNA Constructs and Production of Recombinant Proteins. The cDNA of frataxin was obtained
by RTꢀPCR from cells using the primers: Forward 5’ꢀTATGTGGACTCTCGGGCGꢀ3’ and Reverse
5’ꢀTAGCATCTTTTCCGGAATAGGꢀ3’, and cloned into the eukaryotic expression vector
pcDNA3.1 H/Myc C (Invitrogen) and prokaryotic expression vector pRSETꢀA (Invitrogen). The
147
147
cDNA of mutated frataxin containing R for K was generated using the QuikChange Lightning
SiteꢀDirected Mutagenesis Kit (Stratagene, catalog #210518) using the primers: Forward 5’ꢀ
ATGTGATCAACAGGCAGACGCCAAACAAGꢀ3’
and
Reverse
5’ꢀ
CTTGTTTGGCGTCTGCCTGTTGATCACAꢀ3’ and cloned into the procaryotic expression vector
147
pRSETꢀA to obtain the HisꢀK R construct. The cDNA of human Ub (a gift from Dr V.
22
ACS Paragon Plus Environment

Page 23 of 54
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Colantuoni) was cloned in the prokaryotic expression vector pGEX4T3 (GE Healthcare, Waukesha,
Wi). The recombinant proteins GSTꢀmonoubiquitin and GST were expressed in Escherichia coli
and purified by using glutathione Sepharose 4B beads according to the manufacturer’s instructions
147
(
GE Healthcare). The recombinant proteins Hisꢀfrataxin and HisꢀK R were expressed in E.coli
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
and purified by the nickelꢀnitrilotriacetic acid (NiꢀNTA)ꢀAgarose chromatography according to the
manufacturer’s instructions (Qiagen, Valencia, California).
Chemicals. All compounds were kindly provided by NCI (http://dtp.cancer.gov). The identity and
1
purity of the assayed compounds was independently assessed by elemental analysis and H NMR
(see Supporting Information). Proteasome inhibitors MG132 and Bortezomib were from Sigmaꢀ
Aldrich and Santa Cruz Biotechnology, respectively.
Immunoprecipitation and Western Blotting. For immunoprecipitation assay, 1.5 mg of cell
extracts were incubated with 20 ꢀl of protein A/G agarose beads coated with 5 ꢀl of antibody
against frataxin (Millipore) at 4 °C for 12 h. The beads were washed and boiled in the SDS sample
buffer. The eluted proteins were loaded on 12% SDSꢀPAGE and detected by Western blotting.
Aliquots of protein samples (50 ꢀg) were resolved by 15% SDSꢀgel electrophoresis and transferred
into nitrocellulose filters. The membranes were blocked in PBS, 0.1% Triton and 5% dry milk for 2
h, and then challenged with antiꢀfrataxin (Millipore). The proteins were visualized with enhanced
chemiluminescence detection reagent according to the manufacturer’s instructions (Pierce,
Rockford, Illinois).
Ubiquitination Assay and GST Pull-Down. For Ub assay, Caluꢀ6 cells were transiently
1
47
transfected with plasmids encoding Hisꢀfrataxin or HisꢀK R. 24 h after transfection, cells were
treated with 10 ꢂM of the proteosomal inhibitor MG132 dissolved in DMSO for 18h or with 50 ꢀM
of compound (+)ꢀ11 for 3 days and MG132 in the last 18h. Then, cells were harvested and lysed for
1
0 min on ice in lysis buffer (50mM HEPES, 150 mM NaCl, 1 mM EDTA, 1mM EGTA, 10%
glycerol, 1% TritonꢀXꢀ100, 25 mM NaF, 10 ꢂM ZnCl , pH 7,5) containing protease inhibitor 1X
2
23
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 24 of 54
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
(Roche). Cell lysates were then collected and centrifuged for 15 min at 13,000g. 500 ꢂg of the
soluble fraction was incubated with 5 ꢂg GSTꢀmonoUb or GST coupled to Glutathione sepharose
4
B (Amersham Biosciences, Frieburg, Germany) in pullꢀdown buffer (50 mM TrisꢀHCl pH 7.5, 0.4
mM EDTA, 150 mM NaCl, 10% glicerol, 1% NPꢀ40, 1 mM sodiumꢀortovanadate, 50 mM NaF, 5
mM DTT and Protease Inhibitor Mix 1X) for 4 h at 4°C. After incubation, the beads were washed
three times with lysis buffer. Eluted proteins were loaded on 12% SDSꢀPAGE and analysed by
immunoblotting using antiꢀHis antibodies. Band intensities of ubiquitinated frataxin were quantified
by densitometry scanner analysis (BioꢀRad, Haercules, CA, USA), and relative values are shown as
percentage of control. Each data point represents mean±SEM of at least three independent
experiments. The IC50 values were calculated using a fourꢀparameter logistical model of the graph
of log dose against percentage of ubiquitinated frataxin values.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
UF-LC/MS Based frataxin Binding Assay. The test compound (+)ꢀ11 (1 ꢀM in 4 ꢀL DMSO) and
1
85 ꢀL of assay buffer consisting of 100 mM TrisꢀHCl (pH 7.5), 50 mM KCl, and 5 mM MgCl
2
were placed into a microcentrifuge tube and incubated for 2 min at 25 °C. The binding assay was
initiated by the addition of 1 ꢀM Hisꢀfrataxin recombinant protein and incubated further for 60 min
at 25 °C. After incubation the binding mixture was filtered through an ultramembrane filter
40
(Microcon YMꢀ10, Millipore, Billerica, MA) according to the modified method of Nikolic et al.
and centrifuged at 13 000 g for 20 min at 4 °C. The proteinꢀligand complex was treated with assay
buffer (200 ꢀL × 3) and centrifuged at 13 000 g for 20 min at 4 °C to remove the unbound ligand.
The ultramembrane filter was then placed into a new microcentrifuge tube, and the proteinꢀligand
complex was allowed to stand in 200 ꢀL of methanol for 20 min to dissociate the ligand completely.
The content was then centrifuged at 13 000 g for 20 min at 20 °C. The ultracentrifugate was dried
under nitrogen (NVAP 116 Nitrogen Evaporator, Organomation Associates, Inc. Berlin, MA). The
sample was reconstituted in 100 ꢀL of 50% methanol in deionized water (v/v) and analyzed by
LC/MS. A control experiment was carried out in a similar manner with denatured Hisꢀfrataxin
protein. Prior to the assay, the protein was denatured by heating at 90 °C for 1 h. The test compound
24
ACS Paragon Plus Environment

Page 25 of 54
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
(+)ꢀ11 was prepared in DMSO. The concentration of DMSO in the assay never exceeded 2% (v/v),
a concentration which was found not to influence the results of the assay. The binding experiments
for compound (+)ꢀ11 (1 ꢀM) against 1 ꢀM Hisꢀfrataxin protein were carried out at 25 and 37 °C to
evaluate the influence of temperature. All the binding assays were performed in duplicate and
analyzed twice.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
LC/MS Analysis of Compound (+)-11. Analysis of (+)ꢀ11 was carried out using an Agilent (Little
Falls, DE) 6520 AccurateꢀMass QꢀTOF mass spectrometer equipped with an Agilent 1200 RRLC
system. Chromatographic analysis of (+)ꢀ11 was carried out using ZORBAX Eclipse plus C18
column (2.1 mm × 100 mm, 1.8 ꢀm). The gradient solvent systems used were solvent A (95%
water/5% methanol/0.1% acetic acid, v/v) and solvent B (0.1% acetic acid in methanol, v/v). The
linear gradient was increased from 50% to 100% B in 5 min. The mobile phase was maintained with
1
00% B up to 6 min and then returned to the initial conditions of 50% B in 7 min. The column was
equilibrated with 50% B for 5 min. The flow rate was 0.2 mL/min, and the injection volume for all
samples was 2 ꢀL. Nitrogen was supplied as nebulizing and drying gas at flow rates of 30 and 480
L/h, respectively. The drying gas temperature was 350 °C. The ESI source was operated with a
capillary voltage of 3200 V. The fragmentor voltage was optimized to 175 eV. The released ligand
(2 ꢀL) was subjected to LC/MS analysis with a flow rate of 0.2 mL/min.
25
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 26 of 54
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
ASSOCIATED CONTENT
1
Supporting Information Available. Elemental analysis and H NMR data of compounds 14ꢀ25;
doseꢀresponse behavior of compound (+)ꢀ11 and its analogs 14ꢀ25; binding modes of compounds
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
15 and 18 into the frataxin Ubꢀbinding domain (PDF). Results of in silico screen for the 226 topꢀ
ranked structures from NCI Open Database (XLS). This material is available free of charge via the
Internet at http://pubs.acs.org.
AUTHOR INFORMATION
Corresponding Author
*
Eꢀmail: lavecchi@unina.it, Phone: +39ꢀ081ꢀ678613; Fax: +39ꢀ081ꢀ678012.
Author Contributions
All authors have given approval to the final version of the manuscript.
ACKNOWLEDGMENT
This work was financially supported by the Ministero dell’Istruzione, dell’Università e della
Ricerca Scientifica e Tecnologica (MIURꢀPRIN 2010ꢀ2011, prot. 2010W7YRLZ_003), Rome,
Italy. Gratitude is expressed to the National Cancer Institute for providing test compounds free of
charge. The authors are grateful to the Centro Interdipartimentale di Metodologie ChimicoꢀFisiche
CIMCF and the Centro Interdipartimentale di Analisi Strumentale dell’Università di Napoli
“
Federico II”. The authors would like to thank Prof. Adele Bolognese for thoughtful contributions
and critical reading of the manuscript.
ABBREVIATIONS
AIRs, ambiguous interaction restraints; BSA, buried surface area; Caluꢀ6, human pulmonary
carcinoma cell line; DMSO, dimethyl sulfoxide; DTT, dithiothreitol; EDTA, ethylene glycol
tetraacetic acid; EGTA, ethylenediaminetetraacetic acid; FRDA, Friedreich’s Ataxia; GSH,
26
ACS Paragon Plus Environment

Page 27 of 54
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
glutathione; GST, glutathione Sꢀtransferase; HA tag, hemagglutinin; HADDOCK, high ambiguity
driven proteinꢀprotein docking; His tag, 6ꢀHistidine; His, histidine; HPLC, highꢀperformance liquid
chromatography; HTVS, highꢀthroughput virtual screening; MG132, NꢀcarbobenzoxylꢀLꢀleucinylꢀ
LꢀleucinylꢀLꢀleucinal; NCI, National Cancer Institute; PBS, phosphate buffered saline; PDB,
Protein Data Bank; rmsd, rootꢀmeanꢀsquare deviation; SDSꢀPAGE, Sodium Dodecyl Sulphate ꢀ
PolyAcrylamide Gel Electrophoresis; Ub, ubiquitin; UFꢀLC/MS, ultrafiltrationꢀliquid
chromatography/mass spectrometry; UPS, Ubiquitin Proteasome system.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
27
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 28 of 54
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
REFERENCES
(1) Harding, A. E. Classification of the hereditary ataxias and paraplegias. Lancet 1983, 1, 1151–
1155.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
(
2) Campuzano, V.; Montermini, L.; Molto, M. D.; Pianese, L.; Cossee, M.; Cavalcanti, F.;
Monros, E.; Rodius, F.; Duclos, F.; Monticelli, A.; Zara, F.; Canizares, J.; Koutnikova, H.;
Bidichandani, S. I.; Gellera, C.; Brice, A.; Trouillas, P.; De Michele, G.; Filla, A.; De Frutos,
R.; Palau, F.; Patel, P. I.; Di Donato, S.; Mandel, J. L.; Cocozza, S.; Koenig, M.; Pandolfo, M.
Friedreich's ataxia: autosomal recessive disease caused by an intronic GAA triplet repeat
expansion. Science 1996, 271, 1423–1427.
(
(
(
(
3) Delatycki, M. B.; Williamson, R.; Forrest, S. M. Friedreich ataxia: an overview. J. Med.
Genet. 2000, 37, 1–8.
4) Pandolfo M. Friedreich ataxia: Detection of GAA repeat expansions and frataxin point
mutations. Methods Mol. Med. 2006, 126, 197–216.
5) Patel, P. I.; Isaya, G. Friedreich ataxia: from GAA tripletꢀrepeat expansion to frataxin
deficiency. Am. J. Hum. Genet. 2001, 69, 15–24.
6) Harding, A. E. Friedreich’s ataxia: a clinical and genetic study of 90 families with an analysis
of early diagnostic criteria and intrafamilial clustering of clinical features. Brain 1981, 104,
5
89–620.
7) Harding, A. E. Clinical features and classification of inherited ataxias. Adv. Neurol. 1993, 61,
–14.
(
1
(
8) De Michele, G.; Filla, A.; Cavalcanti, F.; Di Maio, L.; Pianese, L.; Castaldo, I.; Calabrese, O.;
Monticelli, A.; Varrone, S.; Campanella, G.; Leone, M.; Pandolfo, M.; Cocozza, S. Late onset
28
ACS Paragon Plus Environment

Page 29 of 54
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Friedreich's disease: clinical features and mapping of mutation to the FRDA locus. J. Neurol.
Neurosurg. Psychiatry 1994, 57, 977–979.
(
9) Sakamoto, N.; Ohshima, K.; Montermini, L.; Pandolfo, M.; Wells, R. D. Sticky DNA, a selfꢀ
associated complex formed at long GAA*TTC repeats in intron 1 of the frataxin gene, inhibits
transcription. J. Biol. Chem. 2001, 276, 27171–27177.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
(10) Pandolfo, M. Frataxin deficiency and mitochondrial dysfunction. Mitochondrion 2002, 2, 87–
3.
9
(11) Pandolfo, M. Friedreich ataxia and related RNA lossꢀofꢀfunction disorders. In Molecular
Neurology, 1th ed.; Waxman, S. Eds.; Elsevier Academic Press: San Diego, 2007; pp 277ꢀ
294.
(12) Condò, I.; Ventura, N.; Malisan, F.; Rufini, A.; Tomassini, B.; Testi, R. In vivo maturation of
human frataxin. Hum. Mol. Genet. 2007, 16, 1534–1540.
(13) Schmucker, S.; Argentini, M.; CarelleꢀCalmels, N.; Martelli, A.; Puccio, H. The in vivo
mitochondrial twoꢀstep maturation of human frataxin. Hum. Mol. Genet. 2008, 17, 3521–
3531.
(
14) Cavadini, P.; Adamec, J.; Taroni, F.; Gakh, O.; Isaya, G. Twoꢀstep processing of human
frataxin by mitochondrial processing peptidase. Precursor and intermediate forms are cleaved
at different rates. J. Biol. Chem. 2000, 275, 41469–41475.
(
15) Koutnikova, H.; Campuzano, V.; Koenig, M. Maturation of wildꢀtype and mutated frataxin by
the mitochondrial processing peptidase. Hum. Mol. Genet. 1998, 7, 1485–1489.
(
16) Acquaviva, F.; De Biase, I.; Nezi, L.; Ruggiero, G.; Tatangelo, F.; Pisano, C.; Monticelli, A.;
Garbi, C.; Acquaviva, A.M.; Cocozza, S. Extraꢀmitochondrial localisation of frataxin and its
association with IscU1 during enterocyteꢀlike differentiation of the human colon
adenocarcinoma cell line Cacoꢀ2. J. Cell. Sci. 2005, 118, 3917–3924.
29
ACS Paragon Plus Environment