Synthesis, biological characterization and molecular modeling insights of spirochromanes as potent HDAC inhibitors

Article abstract of DOI:10.1016/j.ejmech.2015.11.010

In the last decades, inhibitors of histone deacetylases (HDAC) have become an important class of anti-cancer agents. In a previous study we described the synthesis of spiro[chromane-2,4′-piperidine]hydroxamic acid derivatives able to inhibit histone deacetylase enzymes. Herein, we present our exploration for new derivatives by replacing the piperidine moiety with various cycloamines. The goal was to obtain highly potent compounds with a good in vitro ADME profile. In addition, molecular modeling studies unravelled the binding mode of these inhibitors.

Full text of DOI:10.1016/j.ejmech.2015.11.010

Accepted Manuscript
Synthesis, Biological Characterization and Molecular Modeling Insights of
Spirochromanes as Potent HDAC Inhibitors
Florian Thaler, Loris Moretti, Raffaella Amici, Agnese Abate, Andrea Colombo,
Giacomo Carenzi, Maria Carmela Fulco, Roberto Boggio, Giulio Dondio, Stefania
Gagliardi, Saverio Minucci, Luca Sartori, Mario Varasi, Ciro Mercurio
PII:
S0223-5234(15)30347-0
10.1016/j.ejmech.2015.11.010
EJMECH 8200
DOI:
Reference:
To appear in: European Journal of Medicinal Chemistry
Received Date: 4 August 2015
Revised Date: 7 October 2015
Accepted Date: 5 November 2015
Please cite this article as: F. Thaler, L. Moretti, R. Amici, A. Abate, A. Colombo, G. Carenzi, M.C. Fulco,
R. Boggio, G. Dondio, S. Gagliardi, S. Minucci, L. Sartori, M. Varasi, C. Mercurio, Synthesis, Biological
Characterization and Molecular Modeling Insights of Spirochromanes as Potent HDAC Inhibitors,
European Journal of Medicinal Chemistry (2015), doi: 10.1016/j.ejmech.2015.11.010.
This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to
our customers we are providing this early version of the manuscript. The manuscript will undergo
copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please
note that during the production process errors may be discovered which could affect the content, and all
legal disclaimers that apply to the journal pertain.

ACCEPTED MANUSCRIPT
Graphical abstract
Synthesis, biological characterization and molecular modeling insights of spirochromanes
as potent HDAC inhibitors
Florian Thaler*, Loris Moretti, Raffaella Amici, Agnese Abate, Andrea Colombo, Giacomo
Carenzi, Maria Carmela Fulco, Roberto Boggio, Giulio Dondio, Stefania Gagliardi, Saverio
Minucci, Luca Sartori, Mario Varasi, and Ciro Mercurio*

ACCEPTED MANUSCRIPT
Synthesis, Biological Characterization and
Molecular Modeling Insights of Spirochromanes
as Potent HDAC Inhibitors
a,b*
b
a,b
c,b,d
e,f
Florian Thaler, Loris Moretti, Raffaella Amici, Agnese Abate,
Andrea Colombo,
c,b,g
c,b
a,h
e,i
Giacomo Carenzi,
Maria Carmela Fulco, Roberto Boggio, Giulio Dondio, Stefania
e,j
k,l
c,b
c,b
c*
Gagliardi, Saverio Minucci, Luca Sartori, Mario Varasi, and Ciro Mercurio
a
Genextra Group, Congenia srl, Genextra Group, Via Adamello 16, 20139 Milan, Italy;
Drug Discovery Program, Department of Experimental Oncology, European Institute of
b
Oncology, Via Adamello 16, 20139 Milan, Italy;
c
Genextra Group, DAC srl, Genextra Group, Via Adamello 16, 20139 Milan, Italy;
d
current address: Sbarro Health Research Organization, Bio Life Science Building-Temple
University, 1900 North 12th Street, Philadelphia, PA 19122-6099 (USA);
e
NiKem Research srl, Via Zambeletti 25, 20021 Baranzate (MI), Italy;
f
current address: Teva Pharmaceuticals Fine Chemicals, Strada Statale Briantea, 23892 Bulciago
(
Italy);
g
current address: Fondazione Istituto Insubrico di Ricerca per la Vita, Via Lepetit, 34, 21040
Gerenzano (Italy);
h
current address: NBE Bioanalytics, Merck Serono, Italy;
i
current address: Aphad srl, Via della Resistenza 65, 20090 Buccinasco (MI), Italy;
j
current address: Flamma S.p.A., Via Bedeschi 22, 24040 Chignolo d’Isola (BG), Italy;
k
European Institute of Oncology, Via Adamello 16, 20139 Milan, Italy;
l
University of Milan, University of Milan, Via Celoria 26, 20133 Milan, Italy

ACCEPTED MANUSCRIPT
*
2
Correspondance to F.Thaler, C. Mercurio European Institute of Oncology, Via Adamello 16,
0139 Milan (Italy), T +390294375130; F +390294375138; Email: florian.thaler@ieo.eu,
ciro.mercurio@ieo.eu.
Abbreviations: HDAC, histone deacetylases; HAT, histone acetyltransferases; SQM, Semi-
empirical Quantum Mechanics; CYP, cytochrome P450.

ACCEPTED MANUSCRIPT
Abstract
In the last decades, inhibitors of histone deacetylases (HDAC) have become an important class
of anti-cancer agents. In a previous study we described the synthesis of spiro[chromane-2,4'-
piperidine]hydroxamic acid derivatives able to inhibit histone deacetylase enzymes. Herein, we
present our exploration for new derivatives by replacing the piperidine moiety with various
cycloamines. The goal was to obtain highly potent compounds with a good in vitro ADME
profile. In addition, molecular modeling studies unravelled the binding mode of these inhibitors.
Keywords
Epigenetics, histone deacetylases, antiproliferation, privileged structures, molecular modeling
1
. Introduction
Post-translational modifications on histones have been demonstrated to be essential regulators of
the gene expression.[1] Among these, acetylation and deacetylation of the ε-amino groups of
lysine residues represent crucial modifications of these proteins. The positively charged lysines
present at the N-terminal tail of the histones allow tight binding of negatively charged DNA
around them. This strong histone-DNA interaction blocks the binding sites of promoters and thus
inhibits gene transcription. The neutralization of the charge through the acetylation of the amino
groups of the lysines reduces the electrostatic interactions with the DNA leading to the
nucleosome unwrapping.[2] This unfolding of the chromatin allows the transcriptional factors to
access the gene promoter regions with the consequent gene expression. The balance between
hyperacetylation and hypoacetylation is maintained by two counteracting enzyme families,
namely histone acetyltransferases (HATs) and histone deacetylases (HDACs). Currently, 18
different human HDACs have been identified and are grouped into two families: the classical
HDACs and the silent information regulator (Sir)-related proteins (sirtuins). Classical HDACs
2
+
are Zn dependent metalloenzymes and based on their sequence homology to the yeast
analogues can be further divided into 4 classes. Class I comprises HDACs 1, 2, 3, and 8, class IIa
HDACs 4, 5, 7, 9, class IIb HDACs 6 and 10 and class IV HDAC11.[3-9] Sirtuins, also known
as class III HDACs, form a structurally and mechanistically distinct group and comprise Sirt1 to

ACCEPTED MANUSCRIPT
+
Sirt7. Class III HDACs are characterized by their dependency on NAD as cofactor.[10] Since
the acetylation status of histones is critical for the gene expression modulation and cell fate, there
is no surprise that its dysregulation is involved in the development of several cancers. Recent
studies revealed that HDACs are responsible for the deacetylation of several non-histones too,
such as proteins relevant for tumorigenesis, for cancer cell proliferation, and for immune
functions.[11, 12]
In the past two decades the inhibition of HDACs has emerged as an attractive therapy to reverse
aberrant epigenetic changes associated with cancer, inflammation and neurodegenerative
diseases.[13-18] Several compounds reached clinical investigations and they demonstrated to be
particularly efficacious in the treatment of different hematological malignancies.[19-21] So far,
the clinical studies have culminated in the market approval of four distinct chemical entities (see
Fig. 1). The hydroxamic acid vorinostat, known also as SAHA or Zolinza®[22], and the cyclic
depsipeptide romidepsin, known FK228 or Istodax®,[23, 24] have been approved for the
treatment of the cutaneous T–cell lymphoma (CTCL). The disulfide romidepsin, which is
converted in cells into the active (reduced) form,[24, 25] has also been approved for the
treatment of peripheral T-cell lymphoma (PTCL) as well as the hydroxamate belinostat, known
also as PXD101 or Beleodaq®[26, 27]. Finally, on February 2015 the FDA approved oral
panobinostat (known also as LBH589 or Farydak®) in combination with bortezomib and
dexamethasone in patients with recurrent multiple myeloma.[28] Despite these successes, the
known HDAC inhibitors are less efficacious in other tumors, in particular in solid malignancies.
Therefore, there is still a need to develop HDAC inhibitors with a better clinical profile.[19, 20,
2
9]
(
Insert Figure 1 here)
In previous studies we described a spiropiperidine hydroxamic acid scaffold with potent HDAC
inhibitory activity.[30] Structure-activity relationship studies (SAR) showed that modifications
on the spirochromane core moiety had a substantial influence on the biological activities. The
shift of the N-hydroxyacrylamide moiety or the replacement of the chromane ring by

ACCEPTED MANUSCRIPT
spirobenzofuran generally furnished less active compounds. Furthermore, the exploration at the
4
-oxo moiety demonstrated that the ketone group is critical for the target inhibition. In fact,
modification of this group was detrimental for HDAC and antiproliferative activity. On the other
hand, various N-substitutions on the piperidine group were well tolerated, and no major
differences in the activities between these compounds were found. Considering that HDAC
2
+
inhibitors are composed by a Zn binding moiety, a spacer, which can be further divided into
linker and connecting unit, and a cap group,[31, 32] this minor effect of the different
modifications could be ascribed to an outside orientation of the N-substituents as the cap
group.[33] Based on these observations we further explored this scaffold in order to find more
active inhibitors maintaining their good ADME properties. For this purpose, various cycloamines
were studied as replacement of the piperidine moiety. In the present study, we describe the
synthesis of spirochromane azetidines, pyrrolidines, azepanes and piperidin-3yl derivatives with
various N-substitutents as well as their biological activity in detail. At the same time, molecular
modeling investigations were performed on a number of spirochromane inhibitors. Hence,
molecular docking combined with Semi-empirical Quantum Mechanics (SQM) optimization
allowed us to elucidate the ligand-binding mode of these molecules and to understand their SAR
at the atomic level.
2
. Results and Discussion
The synthesis of compounds 1a to 1c has been previously described.[30, 33]
The secondary amines 15–18 were prepared starting from commercially available 2-hydroxy-5-
bromoacetophenone (2) as illustrated in Scheme 1. Cyclization of compound 2 with N-BOC
protected cycloamine-ones in methanol in presence of pyrrolidine gave the spirocycles 3–6,
which were then coupled with methylacrylate in presence of palladium acetate [Pd(OAc) ], tri(2-
2
methylphenyl)phosphine (P(o-tol) ), triethylamine (TEA) affording the spirocyclic acrylic
3
methyl esters 7–10. The methyl esters 7 and 9 were hydrolyzed with sodium hydroxide (NaOH)
in a water/dioxane mixture, while intermediates 8 and 10 were treated with hydrogen chloride
(
HCl) in acetic acid (AcOH). The resulting carboxylic acids were coupled with NH OTHP (O-
2
(tetrahydropyran-2-yl)hydroxylamine) in presence of EDC (N-(3-dimethylaminopropyl)-N’-

ACCEPTED MANUSCRIPT
ethylcarbodiimide hydrochloride) and HOBt (N-hydroxybenzotriazole). Cleavage of the THP
protecting group with HCl in diethylether or dioxane gave the requisite hydroxamic acids 15–18.
(
Insert Scheme 1 here)
N-substituted spirocycles 31a–h to 34a–h were prepared starting from the spirocyclic acrylic
methyl esters 7–10 by cleaving the tert-butyloxycarbonyl (BOC) protecting group with HCl in
dioxane as outlined in Scheme 2. The free piperidines 19–22 were then either directly alkylated,
acylated or reductively alkylated using sodium triacetoxyborohydride (NaBH(OAc) ) as
3
reducing agent. Hydrolysis of methyl esters 23a–h to 26a–h with NaOH in a water/dioxane
mixture or HCl in acetic acid and subsequent introduction of the hydroxamic acid according to
the procedure described above for compounds 15–18 (Scheme 1) furnished the requisite N-
hydroxy-spirochromaneacrylamides 31-h to 34a-h.
(
Insert Scheme 2 here)
The synthesis of the pure enantiomers of the N-benzyl-spiro(chromane-2,3’-pyrrolidine)
analogues 32i and 32j was accomplished by attaching (S)-(6-methoxynaphthalen-2yl)propionic
acid to the free amine 20. The resulting two diastereoisomers were separated by normal phase
column chromatography and subsequent re-crystallisation from i-PrOH (Scheme 3).
Subsequently, the chiral auxiliary group was cleaved in HCl in acetic acid under microwave
(MW) irradiation providing the corresponding amino acids 35 and 36, which were then
converted into the methyl esters 37 and 38. The following steps to obtain the requisite
enantiomeric hydroxamic acids 32i and 32j were carried out according to the procedures of the
racemic 32d as outlined in Scheme 2.
(
Insert Scheme 3 here)

ACCEPTED MANUSCRIPT
Scheme 4 summarizes the preparation of the [1'-benzyl-4-oxo-spiro(chromane-2,3’-piperidine)-
6
-yl]acrylhydroxamide enantiomers 33i and 33j. Spiropiperidine 5 was first deprotected in acidic
conditions. The coupling with (R)-2-acetoxy-2-phenylacetic acid, separation of the two resulting
diastereoisomers by column chromatography and removal of the chiral auxiliary were performed
following the previously described method for the single enantiomer of 4-oxo-spiro(chromane-
2
,3’-piperidine).[34] The free piperidines of the separated enantiomers 39 and 40 were then
protected with BOC anhydride. Heck reaction of the BOC protected compounds 41 and 42 with
methyl acrylate in presence of Pd(OAc) , P(o-tol) and TEA afforded the acrylic acid methyl
2
3
esters, which were converted to the desired hydroxamic acids 33i and 33j following the
procedures described for the racemic spiropiperidine 33d as outlined in Scheme 2.
(
Insert Scheme 4 here)
The putative HDAC inhibitors were profiled in vitro for their activity against the enzymes and
the ability to block proliferation of tumor cells. The inhibitory effect of the compounds was
assessed using a commercially available fluorescent based assay with partially purified HeLa
nuclear extracts as enzyme source, which is according to the manufacturer’s instruction a rich
source of HDACs 1 and 2 (http://www.enzolifesciences.com/BML-AK500/fluor-de-lys-hdac-
fluorometric-activity-assay-kit/). A comparative protein expression analysis of some HDACs
between HeLA nuclear extract and total HeLa extract was performed in house and the Western
blot results are in line with the manufacturer statements (see Fig. 2). Furthermore, reducing the
protein expression of HDACs 1, 2 and 3 by sequential immunoprecipitation in the HeLa nuclear
extract demonstrated that the observed histone deacetylase activity of HeLa nuclear extract is
mainly imputable to HDAC1 and HDAC2 and not to HDAC3.
(
Insert Figure 2 here)
The results are summarized in Table 1, which includes also the biological data of vorinostat as
well as the 4-oxo-spiro(chromane-2,4’-piperidine) series 1a to 1c as references[30, 33]. The most

ACCEPTED MANUSCRIPT
potent spiro(chromane-2,3’-azetidines) were the 1-methyl-indol-3-yl and the benzyl analogues
3
1f and 31d with IC values of 0.062 and 0.067 µM, respectively. On the other end, the acetyl
50
derivative 31b had with an IC value of 0.485 µM the lowest activity within this subseries.
5
0
Major variations in activity were observed for the spiro(chromane-2,3’-pyrrolidine) and
spiro(chromane-2,3’-piperidine) subseries. The 2-phenylethyl derivative 32g exhibited an IC50
value of 0.027 µM, and its spiro(chromane-2,3’-piperidine) analogue 33g was even more active
inhibiting at 1 nM. The acetyl substituent emerged also in these two sub-series to be detrimental:
compound 32b (IC = 0.955 µM) was substantially less active than all other examples within the
5
0
subseries of the spiropyrrolidines. Furthermore, the acetyl analogue 33b had an IC value of
5
0
0
.207 µM and was equipotent to the methyl (0.237 µM) and to the ethyloxycarbonyl (0.232 µM)
inhibitors 33a and 33c. The 2-phenylethyl derivative 34g was, with an IC value 0.056 µM, the
5
0
most potent compound among the spiro(chromane-2,4’-azepane) analogues, and almost six times
more active than the acetyl and the ethyloxycarbonyl analogues 34b and 34c (IC values of
5
0
0
.323 and 0.300 µM, respectively). In summary, several compounds within all four cycloamine
subseries emerged to be potent HDAC inhibitors, some of them were more potent than their
corresponding spiro(chromane-2,4’-piperidine) analogues.[30, 33]
(
Insert Table 1 here)
Furthermore, we evaluated the biochemical selectivity profile of some selected inhibitors
considering HDAC2 and HDAC3 as representatives of HDACs of class I and HDAC6 as
representative of class II HDACs. As shown in Table 3, the potency of the tested spirochromanes
as well as for vorinostat was in the in the nanomolar range. Neither vorinostat nor any of the
spiro derivatives showed preferential activity versus any of the isoforms, and thus the studied
inhibitors can be classified as typical pan HDAC inhibitors.
(
Insert Table 2 here)

ACCEPTED MANUSCRIPT
The antiproliferative activities of the HDAC inhibitors were determined against three tumor cell
lines of different tissue origin: K562 (chronic myelogenous leukaemia cell line), A549
(carcinomic human alveolar basal epithelial cells) and HCT116 (human colon cancer cells). As
shown in Table 1, several compounds had a significantly higher antiproliferative activity than
their spiro-2,4’-piperidine analogues. In specific, the indolyl derivative 31f was the most potent
compound in K562 cell line among the spiro(chromane-2,3’-azetidines) with an IC value of
5
0
0
.046 µM, and showed together with the benzyl analogue 31d a superior cellular activity in the
A549 and HCT116 cell line than the other spiro(chromane-2,3’-azetidines). On the contrary, the
acetyl spirocycle 31b resulted substantially less potent than the other examples in the
antiproliferative assays with IC values more than 10 µM in all three cell lines. The indolyl
5
0
derivatives 32f and 34f were the most potent compounds also within the spiro(chromane-2,3’-
pyrrolidine) and spiro(chromane-2,4’-azepane) series in the K562 cell line. Furthermore, these
two compounds, the benzyl and the 2-phenylethyl analogues 32d, 34d, 32g and 34g as well as
the methyl and para fluorobenzyl spiroazepanes 34a and 34e exhibited IC values lower than
5
0
1
00 nM in HCT116 cells. The methyl analogue 33a emerged as the most active compound
within the spiro(chromane-2,3’-piperidine) subseries in the K562 and A549 cell lines with IC50
values of 0.056 and 0.103 µM, respectively. Compound 33a, the 2-phenylethyl 33g and the
indolyl analogues 33f had a superior antiproliferative activity in HCT116 cells than the other
spirochromane-2,3’-piperidines. Furthermore, we found that the unsubstituted spiropiperidine 17
was substantially more potent than the other free amines 15, 16, and 18. In specific, compound
1
7 exhibited antiproliferative activity in the three cell lines with IC values of 0.555 µM (K562),
50
0
.714 µM (A549) and 0.228 µM (HCT116), respectively. Overall, the antiproliferative activity
of the spirocyclic inhibitors was generally higher in K562 and HCT116 cells than in the A549
cell line. In particular, thirteen compounds exhibited cellular growth inhibition with IC values
5
0
ranging between 0.040 and 0.100 µM in HCT116 cells and resulted to be substantially more
potent than vorinostat (IC =0.757 µM) and the previously disclosed spiropiperidine reference
50
compounds 1a-c (IC values of 0.477 µM, 0.777 µM and 0.201 µM, respectively). All four
5
0
cycloamine subclasses are present in this list and we find indolyl, (4-fluoro)benzyl, 2-
phenylethyl and methyl analogues for the various N-substitutions. On the other hand,
unsubstituted and acyl spirocycles, in particular the acetyl derivatives, exhibited in general a
lower cellular potency in all three cell lines.

ACCEPTED MANUSCRIPT
With exception of the achiral spiro(chromane-2,3’-azetidine) and spiro(chromane-2,4’-
piperidine) subseries, the other compounds discussed above were prepared and biologically
tested as racemic mixtures. Thus, in order to gain insights if there are differences in the
biological activities of stereoisomers, the pure enantiomers of the benzyl substituted
spiro(chromane-2,3’-pyrrolidine) 32d (32i and 32j, respectively) and spiro(chromane-2,3’-
piperidine) hydroxamic acid derivatives 33d (33i and 33j, respectively) were prepared and
tested. The (–)-spiropyrrolidine 32i exhibited an IC value of 26 nM in the enzymatic assay,
5
0
which is virtually identical to that of the (+)-enantiomer (37 nM). Both enantiomers were
equipotent as cell growth inhibitors in the three tested cell lines. Nonetheless compound 32i was
around 5 times more active than 32a in the enzyme inhibitory assay, the two compounds
exhibited comparable antiproliferative activity in K562 and A549 cells. It is often difficult to
rationalize a link between the biochemical activity and the cellular potency on the sole basis of
the pharmacodynamic properties of a molecule. In fact, several factors contribute to the activity
in cells, such as cellular membrane permeability, sub-cellular compartmentalization, and cellular
mechanisms for efficiently extruding xenobiotica. All these contributions could justify an
observed discrepancy between biochemical and cellular potency.
A similar behavior was found for the two enantiomerically pure spiro(chromane-2,3’-piperidine)
derivatives 33i and 33j. Also in this case, no significant differences in their biochemical
inhibitory activity with IC50 values of around 100 nM were observed between the two
stereoisomers. In addition, both compounds had practically the same antiproliferative potency in
the K562, A549 and HCT116 cell lines.
To elucidate the mode of interaction of spirochromanes inhibitors with class I HDAC enzymes
molecular modeling studies were performed. The binding conformations were predicted with
molecular docking and later optimized at the Semi-empirical Quantum Mechanics level as
described in the Experimental section. For this study, the structure of HDAC1 in complex with
an acetate molecule[35] and the structure of HDAC2 in complex with the inhibitor vorinostat[36]
were prepared and used. These two structures were chosen to account for the conformation
variability of the L2 loop, which is involved in the binding of the substrate and of the inhibitors
as shown for HDAC8.[37] The ligand-binding determination procedure was applied first on three
reference inhibitors: vorinostat, trichostatin A and panobinostat and the results (data not shown)

ACCEPTED MANUSCRIPT
were in good agreement with the experiments (X-ray, binding affinities, etc.). Then, for the
exploration of the binding mode of spirochromanes, several representative molecules were
docked against the two protein structures. In particular, the N-methyl and the N-benzyl
derivatives of all studied spiro-ring sizes (4, 5, 6, and 7 membered rings) were fitted into the
binding sites and optimized. The well-accepted pharmacophoric model of Zn-binding HDAC
inhibitors identifies four parts of the ligand based on its interactions with the target.[32] The 'zinc
binding group' (ZBG) that in the case of vorinostat and the spirochromanes is the hydroxamic
acid, the 'linker' making hydrophobic contacts along the channel (connecting the binding site to
the outside), the 'cap' which is the part interacting with the outer surface of the enzyme and the
'connecting unit' between the linker and the cap, corresponding to the amide in vorinostat. In the
case of the herein described inhibitors, the resulting binding mode with the HDAC proteins, in
line with the pharmacophore, is presented in Fig. 3 and 4. In Fig. 3 the interaction of the common
substructure of the spirochromanes is depicted and as expected the metal is coordinated through
the hydroxamate group.
(
Insert Figure 3 here)
In addition, this group, as described by Vanommeslaeghe et al.,[38] makes three polar
interactions with residues histidine 140 and 141, and tyrosine 303. The next portion of the ligand
structure, the propenyl-benzene, is the linker that extends toward the outside and fits into the
channel making hydrophobic contacts with the glycine 149 and leucine 271. But the main feature
of this interaction is the π−π stacking in between the conserved phenylalanine residues 150 and
2
05 as investigated by Zhou et al..[39] The connecting unit of the studied inhibitors is
represented by the second ring of the chromane, specifically its carbonyl group is found in the
same region as the amide’s one in vorinostat (see Fig. 4).
(
Insert Figure 4 here)

ACCEPTED MANUSCRIPT
In general, this part of the pharmacophore is characterized by an sp2 group.[32] It appears to
serve more as a geometric constraint, to direct properly the cap group, instead of interacting with
the target. This observation would explain the loss of activity observed by the replacement of this
carbonyl with various non-sp2 atoms in ligands of the same chemotype previously
published.[30] Some of these modifications were also submitted to the docking protocol and
showed a lower SQM energy after optimization (data not shown). Lastly, the variable part of this
series of ligands is found outside of the protein, as seen in Fig. 4, and it is the cap group of the
pharmacophore interacting with the outer surface of the protein. To understand the role of the
cap group in the HDAC ligand-protein complexes, the hydrophobic favorable interaction areas
were visualized as shown in Fig. 4 (see the Experimental section for details). These regions,
close to the protein surface, specify the areas where favorable interactions can be made with
aliphatic atoms of the ligand. In the case of the studied inhibitors, the ring of variable size (4 to 7
atoms) attached to the chromanone moiety does not make any additional interaction while its N-
derivative can to a certain extend. About these substitutions, a trend of the enzymatic values is
observed for the various spirochromane series, in specific acetyl > methyl > benzyl >
phenylethyl with the best activity for the phenylethyl and the worst for the acetyl (see Table 1).
The explanation is likely to be related to the inhibitor’s portion approaching the favorable
hydrophobic interaction regions. When the acetyl is present, a sub-optimal binding with the
target would be achieved because of the polarity of the group. While growing the size of the
hydrophobic N-substituent from methyl to phenylethylene the inhibitor structure can fill a larger
area enhancing the binding with the target. Furthermore, the hydrophobic interaction area
extends in various directions around the opening of the channel (see Fig. 4), explaining the same
level of activity for the different stereoisomers observed for the spiro(chromane-2,3'-pirrolidine)
and spiro(chromane-2,3'-piperidine) (32i, 32j, 33i and 33j in Table 1) and for different
conformations of the symmetric ligands as resulting from docking solutions.
As already outlined our efforts were directed towards compounds with an excellent potency and
retaining the ADME profile compared to vorinostat and the spiropiperidine references 1a, 1b and
1
c. Thus, we selected the hydroxamic acid derivatives, which demonstrated to be highly active in
the biochemical assay and having IC values of ≤ 100 nM in at least one cell line for ADME
5
0
studies. The chosen compounds were first tested for their metabolic oxidative Phase I stability in
mouse and human. For this purpose the inhibitors were incubated in mouse and human hepatic

ACCEPTED MANUSCRIPT
microsomal preparations at 37 °C and the percentage of non-metabolized product was assessed
after 30 min.[33, 40, 41] As shown in Table 3, the most evident result is that all indolyl
analogues were found to be highly metabolically instable in applied conditions, with the only
exception of the azetidine derivative 31f in human microsomes. In specific, more than 70% of
compounds 32f, 33f, and 34f were metabolized after 30 min of incubation. The benzyl derivative
3
1d with 44% unmetabolized product after 30 min incubation in mouse and human hepatic
microsomes resulted to be more stable than the pyrrolidine analogues 32d, 32i, and 32j (more
than 70% of the product was metabolized in mouse and human microsomes), and the azepane
derivative 34d (25% remaining in mouse and 19% in human microsomes). The methyl analogues
3
3a and 34a emerged as the most stable compounds with a comparable or in some cases even
superior stability than that of the spiropiperidine reference compounds 1a, 1b, and 1c. For further
in vitro characterizations, we decided not to investigate those compounds, which were
metabolized over 80% in the microsome preparations at the given experimental conditions and
thus resulted to be less stable than the lead spiropiperidine compounds. The additionally
characterized inhibitors 31d, 31f, 32d, 32g, 33a, 34a and 34g exhibited a good solubility, being
in particular ≥ 500 ꢀM at a pH value of 4.5. Furthermore, the same inhibitors did not exhibit any
relevant inhibition of human cytochrome CYP1A2: the enzyme activity was blocked less than
1
0% at 10 µM concentration. More interaction was found for the other three tested human
isoenzymes CYP 2C19, CYP 2D6 and CYP 3A4. In particular, the indolyl derivative 31f
blocked 84% of the CYP 3A4 enzyme activity at 10 µM concentration and the phenylethylene
analogue 34g 81% of the CYP 2C19 enzyme activity. On the other hand, the methyl derivative
3
4a showed virtually no interference with the CYP 3A4 (12% inhibition at 10 µM), and the other
two cytochromes (CYP 2C19 and CYP 2D6 activities less than 10% at at 10 µM). Eventually,
the methyl derivatives 33a and 34a, the benzyl derivative 31d and the phenethylene 32g
exhibited human cytochrome inhibitory activities comparable to the lead spiropiperidines and
these results provided us a pool of compounds with a superior activity and with good in vitro
ADME properties.
(
Insert Table 3 here)

ACCEPTED MANUSCRIPT
3
. Conclusions
In previous studies we had identified a HDAC inhibitory scaffold based on privileged
spiropiperidine structures.[30] Chemical modifications of the spirochromane moiety resulted in a
number of compounds with good HDAC inhibitory potency as well as in vitro and in vivo
ADME properties.[33, 42] Herein, we describe our further exploration of the scaffold through
the replacement of the piperidine moiety by various cycloamines. The main goal was to achieve
derivatives with good in vitro potency while maintaining their good ADME properties. For this
purpose, we prepared and evaluated the biological activity of a series of spirochromane
azetidines, pyrrolidines, azepanes and piperidin-3yl derivatives with various N-substitutents.
Enzyme inhibition experiments revealed that the variation of the ring size ranging from 4 to 7
atoms attached to the chromanone moiety does not result in major changes in the biochemical
activity. These observations are in agreement with molecular modeling studies, which showed
that these moieties do not make any specific interaction with the protein, while the N-substitutent
does interact to a various extend. In specific, the type of the group found at this position can
contribute, positively or negatively, to the binding based on its complementarity with the
identified favorable hydrophobic interaction regions at the outer surface of the enzyme. We
found several potent inhibitors and selected 12 compounds, which exhibited a potent
antiproliferative activity in tumor cell lines with IC values of less than 100 nM, for further in
5
0
vitro ADME characterization. Hepatic microsomal stability and inhibition studies of selected
human CYP450 isoenzymes allowed the identification of several candidates for further
development. In specific, the methyl derivatives 33a and 34a, the benzyl derivative 31d and the
phenylethylen 32g exhibited stability to Phase I oxidative conditions and cytochrome inhibitory
activities comparable to the lead spiropiperidines.
4
4
. Experimental
.1 Chemistry. All reagents and solvents were of commercially available reagent grade quality
and were used without further purification. Flash chromatography purifications were performed
1
13
on Merck silica gel 60 (0.04–0.063 mm). Nuclear magnetic resonance spectra ( H NMR and C
NMR) were recorded on a Bruker spectrometer at 300 MHz and 75 MHz, respectively, with
TMS as internal standard, and, unless stated otherwise, at 300 K. The spectra are referenced in

ACCEPTED MANUSCRIPT
ppm (δ) and coupling constants (J) are expressed in hertz (Hz). Standard abbreviations indicating
multiplicity were used as follows: s=singlet, d=doublet, t=triplet, q=quartet, m=multiplet and
bs=broad signal. HPLC purity of the target compounds was assessed either using a Waters 2777
Sample Manager and a Waters 1525 Binary HPLC Pump, equipped with a Waters 2996 diode
array and a Micromass ZQ 2000 Single quadrupole (Waters) (Method 1, 2, and 5) or on an
Acquity UPLC apparatus, equipped with a diode array and a Micromass ZQ single quadruple
(Waters) (all other methods). The flow rate was adjusted to 2.0 mL/min for Method 1 and
Method 5, 0.3 mL/min for Method 2, and 0.6 mL/min for the other methods. Mobile phase A
was composed by a mixture of water and acetonitrile (95:5) containing 0.1% TFA and mobile
phase B by a mixture of water and acetonitrile (5:95) containing 0.1% TFA (Methods 1, 3, and
7
). Methods 2 and 5 were run with a mobile phase A composed by water containing 0.1% TFA
and a mobile phase B composed by acetonitrile containing 0.1% TFA. Mobile phase A was
composed by a mixture of water and methanol (95:5) containing 0.1% formic acid for Method 4
and by water and acetonitrile (95:5) containing 0.07% formic acid for Method 6. Mobile phase
B was composed by a mixture of water and methanol (5:95) containing 0.1% formic acid for
Method 4 and by water and acetonitrile (5:95) containing 0.05% formic acid for Method 6.
Purity refers to UV detection at 254 nm.
METHOD 1: XBridge C8 (3.5 µm, 4.6 × 50 mm) column; gradient: 0.00-1.00 min (A: 95%, B:
5
9
%), 1.00–8.00 min (A: 0%, B: 100%), 8.00–8.10 min (A: 90%, B: 10%), 8.10–8.50 min (A:
5%, B: 5%), 8.50–9.50 min (A: 95%, B: 5%).
METHOD 2: Atlantis dC18 (3 µm, 2.1 × 50 mm) column; gradient: 0.00–0.20 min (A: 95%, B:
5
9
%), 0.20–5.00 min (A: 0%, B: 100%), 5.00–6.00 min (A: 0%, B: 100%), 6.00–6.10 min (A:
5%, B: 5%); 6.10–7.00 min (A: 95%, B: 5%).
METHOD 3: BEH C18 (1.7 µm, 2.1 × 50 mm) column; gradient: 0.00–0.25 min (A: 95%, B:
5
9
%), 0.25–3.30 min (A: 0%, B: 100%), 3.30–4.00 min (A: 0%, B: 100%), 4.00–4.10 min (A:
5%, B: 5%); 4.10–5.00 min (A: 95%, B: 5%).

ACCEPTED MANUSCRIPT
METHOD 4: BEH C18 (1.7 µm, 2.1 × 50 mm) column; gradient: 0.00–0.25 min (A: 95%, B:
5
9
%), 0.25–3.30 min (A: 0%, B: 100%), 3.30–4.00 min (A: 0%, B: 100%), 4.00–4.10 min (A:
5%, B: 5%); 4.10–5.00 min (A: 95%, B: 5%).
METHOD 5: XBridge C8 (3.5 µm, 4.6 × 50 mm) column; gradient: 0.00-1.00 min (A: 95%, B:
5
9
%), 1.00–7.50 min (A: 0%, B: 100%), 7.50–8.50 min (A: 0%, B: 100%), 8.10–8.50 min (A:
5%, B: 5%), 8.50–9.50 min (A: 95%, B: 5%).
METHOD 6: BEH C18 (1.7 µm, 2.1 × 50 mm) column; gradient: 0.00 (A: 98%, B: 2%), 0.00–
.00 min (A: 0%, B: 100%), 3.00–3.50 min (A: 0%, B: 100%), 3.50–4.50 min (A: 98%, B: 2%).
3
METHOD 7: BEH C18 (1.7 µm, 2.1 × 50 mm) column; gradient: 0.00–0.25 min (A: 95%, B:
5
9
%), 0.25–3.30 min (A: 0%, B: 100%), 3.30–4.00 min (A: 0%, B: 100%), 4.00–4.10 min (A:
5%, B: 5%); 4.10–5.00 min (A: 95%, B: 5%).
(±)-6-bromo-4-oxo-spiro(chromane-2,3'-pyrrolidine)-1'-carboxylic acid tert-butyl ester (4)
A mixture of 2-hydroxy-5-bromoacetophenone (2, 5.00 g, 23.3 mmol), N-BOC pyrrolidin-3-one
4.30 g, 23.3 mmol) and pyrrolidine (3.87 mL, 46.5 mmol) in MeOH (20 mL) was heated under
(
microwave irradiation for 4 h at 70 °C. The solution was concentrated and the crude mixture was
purified by column chromatography (eluent: petroleum ether/EtOAc 95:5 to 7:3) to give the tert-
1
butyl ester 4 (6.00 g, 68%) as a yellow solid. H NMR (CDCl ) δ (ppm): 8.00 (d, J = 2.64 Hz,
3
1
H), 7.58 (dd, J = 8.80, 2.05 Hz, 1H), 6.89 (d, J = 8.80 Hz, 1H), 3.68–3.85 (m, 1H), 3.50–3.66
(m, 2H), 3.23–3.46 (m, 1H), 2.88–3.04 (m, 1H), 2.84 (d, J = 16.73 Hz, 1H), 2.23–2.39 (m, 1H),
+
1
.84–2.02 (m, 1H), 1.47 (s, 9H). LC-MS: (ES+) MH -56: 326.
The bromospirochromanes 3, 5 and 6 were prepared analogously starting from compound 2 and
the appropriate amine following the procedure for intermediate 4 (see SI).
(±)-(E)-3-[1'-tert-Butoxycarbonyl-4-oxo-spiro(chromane-2,3’-pyrrolidine)-6-yl]-acrylic acid
methyl ester (8)
A mixture of compound 4 (4.70 g, 12.3 mmol), Pd(OAc) (55.1 mg, 0.246 mmol), P(o-tol) (149
2
3
mg, 0.490 mmol), TEA (5.13 mL, 36.8 mmol), methyl acrylate (3.32 mL, 36.9 mmol) in dry

ACCEPTED MANUSCRIPT
DMF (10 mL) was heated under N atmosphere to 100 °C for 3 h. After cooling down to RT, the
2
solution was poured into water (100 mL) and extracted with EtOAc (3 times 100 mL). The
organic phase was dried over Na SO , filtered, and evaporated. The crude residue was
2
4
recrystallized in petroleum ether:diisopropylether (1:1) to give the acrylic acid methyl ester 8
1
(
3.85 g, 81%) as a light yellow solid. H NMR (CDCl ) δ (ppm): 8.06 (d, J = 2.05 Hz, 1H), 7.69
3
(dd, J = 8.80, 2.05 Hz, 1H), 7.66 (d, J = 15.85 Hz, 1H), 7.03 (d, J = 8.80 Hz, 1H), 6.41 (d, J =
1
3
6.14 Hz, 1H), 3.82 (s, 3H), 3.69–4.00 (m, 1H), 3.53–3.69 (m, 2H), 3.26–3.48 (m, 1H), 2.81–
+
.07 (m, 2H), 2.25–2.35 (m, 1H),1.81–2.05 (m, 1H), 1.48 (bs, 9H). LC-MS: (ES+) MH -56: 332.
The acrylic acid methyl esters 7, 9 and 10 were prepared analogously starting from the
corresponding bromo derivatives following the procedure for intermediate 8 (see SI).
(E)-3-[1'-tert-Butoxycarbonyl-4-oxo-spiro(chromane-2,3'-azetidine)-6-yl]-acrylic acid (11)
1
M NaOH (1.60 ml) was added to a solution of methyl ester 7 (450 mg, 1.20 mmol) in water (11
ml) and dioxane (22 ml). The mixture was stirred at RT overnight, then aqueous HCl 10% was
added until reaching pH 7. The dioxane was removed and the residue was diluted with water.
The aqueous layer was acidified to pH 5 with 10% aqueous HCl and the product was extracted
with CH Cl . The collected organic phases were dried over Na SO and concentrated to give the
2
2
2
4
1
acrylic acid (430 mg, 100%) as a yellow solid. H NMR (DMSO–d ) δ (ppm): 2.33 (bs, 1H),
6
8
.00 (dd, J = 8.51, 2.05 Hz, 1H), 7.96 (d, J = 2.35 Hz, 1H), 7.60 (d, J = 16.14 Hz, 1H), 7.20 (d,
J = 8.51 Hz, 1H), 6.48 (d, J = 16.14 Hz, 1H), 4.01 (d, J = 9.68 Hz, 2H), 3.91 (d, J = 9.39 Hz,
+
2
H), 3.20 (s, 2H), 1.39 (s, 9H); LC–MS: (ES+) MH –56: 304.
(±)-(E)-3-[1'-tert-Butoxycarbonyl-4-oxo-spiro(chromane-2,3’-pyrrolidine)-6-yl]acrylic acid
(12)
5
mL of an 20% aqueous HCl (5 mL, 30 mmol) was added to a suspension of 8 (464 mg, 1.2
mmol) in glacial AcOH (5 mL). The mixture was stirred at 85 °C for 3 h and was then
evaporated to give (±)-(E)-3-[4-oxo-spiro(chromane-2,3’-pyrrolidine)-6-yl]acrylic acid (370 mg,
1
quantitative) as a white solid (hydrochloride salt). H NMR (DMSO-d ) δ (ppm): 12.30 (bs, 1 H),
6
9
.79 (bs, 1 H), 9.67 (bs, 1 H), 7.87-8.19 (m, 2 H), 7.61 (d, J=15.85 Hz, 1 H), 7.08 (d, J=9.39 Hz,

ACCEPTED MANUSCRIPT
1
H), 6.48 (d, J=15.85 Hz, 1 H), 3.24-3.66 (m, 4 H), 3.21 (d, J=17.02 Hz, 1 H), 3.11 (d, J=17.02
+
Hz, 1 H), 2.22-2.43 (m, 1 H), 2.10 (dt, J=13.86, 9.94 Hz, 1 H). LC-MS: (ES+) MH : 274.
TEA (0.50 mL, 3.6 mmol) was added to the suspension of (±)-(E)-3-[4-oxo-spiro(chromane-
2
,3’-pyrrolidine)-6-yl]acrylic acid (370 mg, 1.20 mmol) in 10 mL of CH Cl . After complete
2 2
dissolution of the acid, BOC anhydride (314 mg, 1.44 mmol) was added and the mixture was
stirred at RT for 1 h. The solution was poured into water (10 mL), neutralized with aqueous 5%
citric acid solution and extracted with CH Cl . The organic phase was dried over Na SO ,
2
2
2
4
evaporated resulting in a white solid (406 mg, 91%) of acrylic acid 12, which was used in the
1
next step without purification. H NMR (DMSO–d ) δ (ppm): 7.78–8.20 (m, 2H), 7.57 (d, J =
6
1
3
1
5.85 Hz, 1H), 7.11 (d, J = 9.10 Hz, 1H), 6.46 (d, J = 15.85 Hz, 1H), 3.61 (d, J = 12.32 Hz, 1H),
.25–3.55 (m, 3H), 3.15 (d, J = 16.73 Hz, 1H), 2.98 (d, J = 17.02 Hz, 1H), 2.11–2.28 (m, 1H),
+
.88–2.11 (m, 1H), 1.27–1.52 (m, 9H); LC–MS: (ES+) MH –56: 318.
(±)-(E)-3-[1'-tert-Butoxycarbonyl-4-oxo-spiro(chromane-2,3’-piperidine)-6-yl]acrylic acid
(13)
The methyl ester 9 (500 mg, 1.25 mmol) was suspended in 20 mL of a dioxane/water mixture
1:1). 1.62 mL 1 M NaOH was added and the solution was stirred at RT overnight. After
(
acidification with an aqueous solution containing 5% citric acid the product was collected by
1
filtration and dried giving the acrylic acid 13 (490 mg, quantitative) as a white solid. H NMR
(
DMSO–d ) δ (ppm): 12.30 (bs, 1H), 7.80–8.05 (m, 2H), 7.59 (d, J = 15.85 Hz, 1H), 6.83–7.11
6
(m, 1H), 6.44 (d, J = 16.14 Hz, 1H), 3.69–4.02 (m, 2H), 3.02–3.16 (m, 1H), 2.68–3.00 (m, 3H),
+
1
.92–2.15 (m, 1H), 1.58–1.90 (m, 2H), 0.90–1.59 (m, 10H); LC–MS: (ES+) MH : 388.
The acrylic acid 14 was prepared following the procedure for intermediate 12 (see SI).
E)-3-[4-Oxo-spiro(chromane-2,3’-azetidine)-6-yl]-N-hydroxy-acrylamide (15)
(
Acrylic acid 11 (430 mg, 1.20 mmol) was converted into the (E)-3-[1'-tert-butoxycarbonyl-4-
oxo-spiro(chromane-2,3'-azetidine)-6-yl]-N-(tetrahydro-pyran-2-yloxy)-acrylamide (0.40 g,
1
7
3%, pale yellow solid) following the procedures for compound 16. H NMR (DMSO–d ) δ
6
(ppm): 11.13 (bs, 1 H), 7.93 (d, J=2.35 Hz, 1 H), 7.84 (d, J=8.80 Hz, 1 H), 7.50 (d, J=15.85 Hz, 1

ACCEPTED MANUSCRIPT
H), 7.21 (d, J=8.51 Hz, 1 H), 6.50 (d, J=15.85 Hz, 1 H), 4.91 (bs, 1 H), 4.00 (d, J=9.39 Hz, 2 H),
3
.93-3.98 (m, 1 H), 3.90 (d, J=9.39 Hz, 2 H), 3.44-3.74 (m, 1 H), 3.20 (s, 2 H), 1.48-1.84 (m, 6
+
H), 1.39 (s, 9 H); LC–MS: (ES+) MH –83: 375. Cleavage of the THP-protecting group (360 mg,
0
.78 mmol) in CH Cl (14 mL) was carried out by adding dropwise 1.90 mL 4 M HCl in dioxane
2 2
and stirring the solution overnight at RT. The formed precipitate was filtered off, washed with
CH Cl , dried to give the requisite hydroxamic acid 15 (154 mg, 65%, hydrochloride salt) as a
2
2
1
white solid. H NMR (DMSO–d ) δ (ppm): 10.71 (bs, 1H), 9.42 (bs, 1H), 9.24 (bs, 1H), 7.92 (d,
6
J = 2.05 Hz, 1H), 7.86 (dd, J = 8.80, 2.05 Hz, 1H), 7.46 (d, J = 15.85 Hz, 1H), 7.21 (d, J = 8.80
Hz, 1H), 6.46 (d, J = 15.55 Hz, 1H), 4.03 J = 4.33 (m, 4H), 3.31 (s, 2H); LC–MS: Method 1,
+
rt=1.27; (ES+) MH : 275.
(±)-(E)-3-[4-Oxo-spiro(chromane-2,3’-pyrrolidine)-6-yl]-N-hydroxy-acrylamide (16)
TEA (0.271 mL, 1.95 mmol) was added to solution of the acrylic acid 12 (484 mg, 1.29 mmol)
in CH Cl (10 mL). The mixture was cooled down to 0 °C, and EDC (372 mg, 1.95 mmol) and
2
2
HOBt (263 mg, 1.95 mmol) were added. The mixture was stirred at 0 °C for 3 h, then NH OTHP
2
(182 mg, 1.56 mmol) was added and the mixture was stirred at RT overnight. The solution was
washed with aqueous 5% NaHCO solution and brine. The organic layer was then dried over
3
Na SO and evaporated. The crude residue was purified by column chromatography (eluent:
2
4
CH Cl /MeOH 98:2) to give the N-(tetrahydro-pyran-2-yloxy)-acrylamide intermediate (499 mg,
2
2
1
8
1
9%) as a light yellow solid. H NMR (DMSO-d ) δ (ppm): 11.13 (bs, 1 H), 7.94 (d, J=2.05 Hz,
6
H), 7.81 (d, J=9.10 Hz, 1 H), 7.49 (d, J=15.55 Hz, 1 H), 7.12 (d, J=8.80 Hz, 1 H), 6.48 (d,
J=16.43 Hz, 1 H), 4.91 (bs, 1 H), 3.81-4.03 (m, 1 H), 3.52-3.69 (m, 2 H), 3.32-3.52 (m, 3 H),
3
.15 (d, J=17.02 Hz, 1 H), 2.98 (d, J=17.31 Hz, 1 H), 2.09-2.24 (m, 1 H), 1.90-2.09 (m, 1 H),
+
1
.70 (bs, 3 H), 1.54 (bs, 3 H), 1.40 (m, 9 H). LC–MS: (ES+) MH : 473.
1
M HCl in Et O (5 mL, 5 mmol) was added dropwise to a solution of N-(tetrahydro-pyran-2-
2
yloxy)-acrylamide intermediate (280 mg, 0.59 mmol) in CH Cl (2 mL). The mixture was stirred
2
2
at RT for 1 h. The precipitate was filtered off, washed with CH Cl , dried and collected giving
2
2
1
the requisite hydroxamate 16 (108 mg, 57%) as a white solid (hydrochloride salt). H NMR
DMSO–d ) δ (ppm): 9.74 (bs, 1H), 9.61 (bs, 1H), 7.93 (d, J = 2.05 Hz, 1H), 7.82 (dd, J = 8.80,
(
6
2
.05 Hz, 1H), 7.45 (d, J = 15.55 Hz, 1H), 7.09 (d, J = 8.80 Hz, 1H), 6.46 (d, J = 15.55 Hz, 1H),

ACCEPTED MANUSCRIPT
3
3
.55 (dd, J = 12.47, 5.72 Hz, 1H), 3.42 (bs, 2H), 3.24–3.37 (m, 1H), 3.20 (d, J = 17.02 Hz, 1H),
.10 (d, J = 17.02 Hz, 1H), 2.21–2.43 (m, 1H), 2.09 (dt, J = 14.09, 9.83 Hz, 1H). LC–MS:
+
Method 7, rt=0.91; (ES+) MH : 289.
The N-hydroxy-acrylamides 17 and 18 were prepared following the procedure for compound 16
(see SI).
(±)-(E)-3-[4-Oxo-spiro(chromane-2,3'-pyrrolidine)-6-yl]-acrylic acid methyl ester (20)
4
M HCl in dioxane (10 mL, 80 mmol) was added to a solution of spirocycle 8 (4.30 g, 11.2
mmol) in 15 mL CH Cl and the mixture was stirred at RT for 1 h. The precipitate was filtered
2
2
off, washed with CH Cl , dried and collected (3.3 g, 92%) giving the methyl ester 20 as a white
2
2
1
solid (hydrochloride salt). H NMR (DMSO–d ) δ (ppm): 9.65 (bs, 2H), 7.97–8.11 (m, 2H), 7.70
6
(d, J = 15.85 Hz, 1H), 7.09 (d, J = 9.39 Hz, 1H), 6.60 (d, J = 16.14 Hz, 1H), 3.73 (s, 3H), 3.50–
3
.67 (m, 1H), 3.35–3.49 (m, 2H), 3.25–3.31 (m, 1H), 3.21 (d, J = 17.02 Hz, 1H), 3.11 (d, J =
+
1
7.02 Hz, 1H), 2.20–2.43 (m, 1H), 2.10 (dt, J = 14.01, 9.87 Hz, 1H). LC–MS: (ES+) MH : 288.
The acrylic acid methyl esters 19, 21 and 22 were prepared following the procedure for
intermediate 20 (see SI).
(
±)-(E)-3-[1'-Benzyl-4-oxo-spiro(chromane-2,3’-pyrrolidine)-6-yl]-acrylic acid methyl ester
24d)
(
A suspension of spirocycle 20 (440 mg, 1.36 mmol) in aqueous 10% NaHCO solution was
3
extracted with CH Cl . The organic phase was dried over Na SO and evaporated. The resulting
2
2
2
4
oil was dissolved in CH Cl (15 mL), treated with benzaldehyde (0.166 mL, 1.63 mmol) and
2
2
NaBH(OAc) (433 mg, 2.04 mmol), and the resulting clear solution was stirred at RT for 2 h.
3
Water was added to the mixture and the pH was brought to basic conditions with NH . The
3
mixture was extracted with CH Cl . The organic layer was dried, evaporated, and the crude
2
2
residue was purified by column chromatography (eluent: petroleum ether/EtOAc 7:3) to give
1
compound 24d (482 mg, 94%) as a light yellow solid. H NMR (CDCl ) δ (ppm): 8.02 (d, J =
3
2.35 Hz, 1H), 7.65 (dd, J = 8.80, 2.35 Hz, 1H), 7.65 (d, J = 15.55 Hz, 1H), 7.13–7.39 (m, 5H),
7.03 (d, J = 8.80 Hz, 1H), 6.39 (d, J = 16.14 Hz, 1H), 3.82 (s, 3H), 3.71 (d, J = 12.91 Hz, 1H),
3.66 (d, J = 13.20 Hz, 1H), 2.95 (d, J = 16.43 Hz, 1H), 2.89 (d, J = 16.43 Hz, 1H), 2.80–2.97

ACCEPTED MANUSCRIPT
(m, 2H), 2.63–2.80 (m, 2H), 2.26 (ddd, J = 13.35, 7.78, 5.28 Hz, 1H), 1.92–2.12 (m, 1H). LC–
+
MS: (ES+) MH : 378.
The alkylated methyl esters 23a, 23d–g, 24a, 24e–f, 25a, 25d–g, 26a, and 26d–f were prepared
analogously starting from the appropriate amine and aldehyde following the procedure for
intermediate 24d (see SI).
(
±)-(E)-3-[1'-Acetyl-4-oxo-spiro(chromane-2,3’-pyrrolidine)-6-yl]-acrylic acid methyl ester
24b)
(
A suspension of methyl ester 20 (400 mg, 1.23 mmol) in CH Cl (15 mL) was treated with acetyl
2
2
chloride (0.105 mL, 1.49 mmol) and DIPEA (0.264 mL, 1.49 mmol) and stirred at RT for 1 h.
The solution was then evaporated and the crude product was purified by chromatography column
(
eluent: CH Cl /MeOH 95:5) to give the 1'-acetyl derivative 24b (240 mg, 59%) as a white solid.
2 2
1
H NMR (CDCl ) δ (ppm): 8.06 (d, J = 1.76 Hz, 1H), 7.65–7.78 (m, 1H), 7.65 (d, J = 16.14 Hz,
3
1
2
3
H), 7.00 (d, J = 8.51 Hz, 1H), 6.40 (d, J = 16.14 Hz, 1H), 3.82 (s, 3H), 3.34–4.13 (m, 4H),
+
.82–3.12 (m, 2H), 2.29–2.57 (m, 1H), 2.11 (s, 3H), 1.87–2.09 (m, 1H); LC–MS: (ES+) MH :
30.
The acyl derivatives 23b–c, 23h, 24c, 24h, 25b–c, 25h, 26b–c and 26h were prepared
analogously starting from the appropriate amine and acyl chloride following the procedure for
intermediate 24b (see SI).
(±)-(E)-3-[1'-(2-Phenyl-ethyl)-4-oxo-spiro[chromane-2,4'-pyrrolidine]-6-yl]-acrylic
acid
methyl ester (24g)
A suspension of methyl ester 20 (500 mg, 1.55 mmol) in 15 mL CH Cl was treated with 0.33
2
2
mL DIPEA (1.86 mmol) and 0.251 mL phenethyl bromide (1.86 mmol), and stirred at RT for 8
days. The mixture was concentrated and the crude residue was purified by column
chromatography (eluent: petroleum ether/EtOAc 7:3) to give methyl ester 24g as a yellow solid
1
(
280 mg, 46%). H NMR (CDCl ) δ (ppm): 8.05 (d, J = 2.35 Hz, 1H), 7.67 (dd, J = 8.80, 2.35
3
Hz, 1H), 7.66 (d, J = 15.85 Hz, 1H), 7.12–7.39 (m, 5H), 7.06 (d, J = 8.51 Hz, 1H), 6.40 (d, J =
1
2
6.14 Hz, 1H), 3.82 (s, 3H), 3.09 (d, J = 10.27 Hz, 1H), 2.84–3.03 (m, 4H), 2.52–2.84 (m, 5H),
+
.18–2.42 (m, 1H), 1.84–2.16 (m, 1H); LC–MS: (ES+) MH : 392.

ACCEPTED MANUSCRIPT
The 2-phenyl-ethyl derivative 26g was prepared following the procedure for intermediate 24g
(see SI).
(±)-(E)-3-[1'-Benzyl-4-oxo-spiro(chromane-2,3’-pyrrolidine)-6-yl]-acrylic
acid
hydrochloride (28d)
The methyl ester 24d (472 mg, 1.25 mmol) was then hydrolyzed in aqueous 20% HCl solution
and AcOH following the procedure described for the hydrolysis of 8 giving the acrylic acid 28d
1
(
450 mg, 90%) as a white solid (hydrochloride salt). H NMR (DMSO–d , 353K) δ (ppm): 7.96
6
(d, J = 2.35 Hz, 1H), 7.93 (dd, J = 8.51, 2.35 Hz, 1H), 7.53–7.71 (m, 3H), 7.36–7.52 (m, 3H),
7
1
2
.14 (d, J = 8.51 Hz, 1H), 6.44 (d, J = 16.14 Hz, 1H), 4.44 (d, J = 13.50 Hz, 1H), 4.40 (d, J =
3.79 Hz, 1H), 3.36–3.68 (m, 4H), 3.22 (d, J = 17.02 Hz, 1H), 3.10 (d, J = 17.02 Hz, 1H), 2.22–
+
.47 (m, 2H). LC–MS: (ES+) MH : 364.
The methyl esters 27a, 27d–e, 27g, 28a–c, 28e–h, 29a, 29c–d, 30a–d and 30g–h were
hydrolysed following the procedure for the acrylic acid 28d (see SI), while the methyl esters
2
7b–c, 27f, 27h, 29b, 29e–h and 30e–f were prepared as described for the acrylic acid 11 (see
SI).
The N-hydroxy-acrylamides 31a, 31d, 31h, 32a-h, 33a, 33c, 33d, 33e, 33h, 34a-e, and 34g-h
were prepared following the procedure for compound 16 (see SI), while the N-hydroxy-
acrylamides 31b-c, 31e-g, 33b, 33f, 33g, and 34f were prepared as described for compound 15
(see SI).
(-)-(E)-3-[1'-Benzyl-4-oxo-spiro(chromane-2,3’-pyrrolidine)-6-yl]-N-hydroxy-acrylamide
(32i)
1
.89 mL (13.6 mmol) of TEA was added to solution of 1.57 g (6.81 mmol) (S)-2-(6-
methoxynaphthalen-2-yl)propanoic acid in 80 mL of CH Cl . The mixture was cooled down to 0
2
2
°
C, and 1.95 g (10.2 mmol) EDC and 1.38 g (10.2 mmol) HOBt were added. The mixture was
stirred at 0 °C for 2 h, then the 2.20 g (6.81 mmol) acrylic acid methyl ester hydrochloric salt 20,
previously treated with 0.947 mL, (6.81 mmol) of TEA, in 20 mL CH Cl was added and the
2
2
mixture was stirred at RT overnight. The solution was washed with aqueous 5% NaHCO3
solution and brine. The organic layer was then dried over Na SO and evaporated. The crude
2
4

ACCEPTED MANUSCRIPT
residue was purified by column chromatography (eluent: CH Cl /i-PrOH 99:1) and the 6-
2
2
methoxynaphthalen-2-yl diastereoisomers 35 and 36 were separated. The less polar isomer was
re-crystallized from i-PrOH to give 980 mg (29%) of (+)-acrylic acid methyl ester 35 as a white
1
solid. H NMR (CDCl ) δ (ppm): 8.01 (d, J=2.35 Hz, 1 H), 7.48-7.82 (m, 5 H), 7.38 (ddd,
3
J=10.78, 8.73, 1.61 Hz, 1 H), 7.10-7.25 (m, 2 H), 7.01 and 6.59 (d, J=8.51 Hz, 1 H), 6.41 (d,
J=16.14 Hz, 1 H), 4.02-4.17 (m, 1 H), 3.96 (s, 3 H), 3.82 (s, 3 H), 3.65-3.91 (m, 3 H), 3.28-3.57
(m, 1 H), 2.95 and 2.79 (d, J=16.73 Hz, 1 H), 2.86 (d, J=16.73 Hz, 1 H), 2.15-2.39 (m, 1 H),
+
1
.69-2.15 (m, 1 H), 1.56 (d, J=6.16 Hz, 3 H); LC-MS: (ES+) MH : 500; optical rotation: α =
D
+
5.9°, c=1.025 g/100 mL in CH Cl . 926 mg (1.85 mmol) of the methyl ester 35 suspended in 5
2 2
mL of glacial AcOH, was then treated with 5 mL (30 mmol) of 20% aqueous HCl. The mixture
was stirred in the microwave oven at 120°C for 3 h. 5 mL of water was added and the reaction
mixture was extracted three times with CH Cl . The aqueous phase was evaporated, then 25 mL
2
2
of methanol and a catalytic amount of H SO were added and the mixture was heated to reflux
2
4
overnight. The solution was poured into water and NH was added to adjust the pH to 9-10. The
3
aqueous phase was extracted with CH Cl , dried over Na SO , filtered, and the solvent was
2
2
2
4
evaporated to give the (+)-oxo-spiro(chromane-2,3’-pyrrolidine) methyl ester 37 (210 mg,
1
3
9.5%) as a yellow oil (free base). H NMR (DMSO-d ) δ (ppm): 7.99 (d, J=2.05 Hz, 1H), 7.96
6
(dd, J=8.51, 2.35 Hz, 1 H), 7.67 (d, J=16.14 Hz, 1 H), 7.06 (d, J=8.51 Hz, 1 H), 6.56 (d, J=16.14
Hz, 1 H), 3.72 (s, 3 H), 2.74-3.10 (m, 6 H), 1.88-2.10 (m, 1 H), 1.65-1.85 (m, 1 H); LC-MS:
+
(
ES+) MH : 288; optical rotation: α = +1.97°, c=0.72 g/100 mL in CH Cl .
D
2
2
1
80 mg (0.63 mmol) of the methyl ester 37 was then treated with benzaldehyde and
NaBH(OAc) according to the procedure used for compound 24d to give 180 mg (76%) of the
3
1
(
+)-(E)-3-[1'-benzyl-4-oxo-spiro(chromane-2,3’-pyrrolidine)-6-yl]-acrylic acid methyl ester [ H
NMR (DMSO-d ) δ (ppm): 7.97 (d, J=2.35 Hz, 1 H), 7.97 (dd, J=9.39, 2.35 Hz, 1 H), 7.66 (d,
6
J=16.14 Hz, 1 H), 7.17-7.35 (m, 5 H), 7.10 (d, J=9.10 Hz, 1 H), 6.56 (d, J=16.14 Hz, 1 H), 3.72
(s, 3 H), 3.61 (s, 2 H), 2.91-3.04 (m, 2 H), 2.69-2.89 (m, 2 H), 2.53-2.67 (m, 2 H), 1.86-2.21 (m,
+
2
H); LC-MS: (ES+) MH : 378; Optical rotation: α =+1.68°, c=0.96 g/100 mL in CH Cl )],
D
2
2
which was first hydrolyzed using aqueous HCl (20%) solution and AcOH following the
procedure described for compound 12 giving 120 mg (88%) of the corresponding acrylic acid as
1
its hydrochloride salt and as a white solid [ H NMR (DMSO-d ) δ (ppm): 11.01-11.75 (m, 1 H),
6
7
.91-8.12 (m, 2 H), 7.53-7.71 (m, 3 H), 7.37-7.53 (m, 3 H), 6.95-7.26 (m, 1 H), 6.30-6.57 (m, 1

ACCEPTED MANUSCRIPT
+
H), 4.29-4.58 (m, 2 H), 2.91-3.67 (m, 6 H), 1.98-2.45 (m, 2 H); LC-MS: (ES+) MH : 364;
Optical rotation: α = -3.4°, c=0.43 g/100 mL in MeOH)], and then treated with NH OTHP
D
2
following the procedure described for compound 16 giving 80 mg (63%) of the N-(tetrahydro-
1
pyran-2-yloxy)-acrylamide as a light yellow solid [ H NMR (DMSO-d ) δ (ppm): 11.13 (bs, 1
6
H), 7.90 (d, J=2.05 Hz, 1 H), 7.67-7.84 (m, 1 H), 7.47 (d, J=15.85 Hz, 1 H), 7.16-7.39 (m, 5 H),
7
3
2
.10 (d, J=8.80 Hz, 1 H), 6.46 (d, J=16.14 Hz, 1 H), 4.79-5.00 (m, 1 H), 3.86-4.10 (m, 1 H),
.62 (s, 2 H), 3.47-3.58 (m, 1 H), 3.01 (d, J=17.02 Hz, 1 H), 2.95 (d, J=17.02 Hz, 1 H), 2.55-
+
.88 (m, 4 H), 1.88-2.18 (m, 2 H), 1.47-1.83 (m, 6 H); LC-MS: (ES+) MH : 463]. The THP
protected hydroxamic acid was then hydrolyzed according to the procedure for compound 16
1
giving 31.3 mg (44%) of the requisite hydroxamic acid 32i as its hydrochloride salt. H NMR
(
DMSO-d ) δ (ppm): 7.92 (d, J=2.35 Hz, 1 H), 7.80 (dd, J=8.36, 2.49 Hz, 1 H), 7.34-7.61 (m, 6
6
H), 7.12 (d, J=8.22 Hz, 1 H), 6.51 (d, J=16.14 Hz, 1 H), 4.34 (bs, 2 H), 3.11-3.53 (m, 6 H),
2
.29-2.44 (m, 2 H); optical rotation: α = -4.4°, c=0.27 g/100 mL in MeOH; LC-MS: Method F,
D
+
rt=1.22; (ES+) MH : 379.
The synthesis of (+)-(E)-3-[1'-benzyl-4-oxo-spiro(chromane-2,3’-pyrrolidine)-6-yl]-N-hydroxy-
acrylamide (32j) is disclosed in the supplementary information.
(-)-(E)-3-[1'-Benzyl-4-oxo-spiro(chromane-2,3’-piperidine)-6-yl]-N-hydroxy-acrylamide
(33i)
5
mL of 4 M HCl in dioxane was added to a solution of 2.50 g (6.31 mmol) of intermediate 5
dissolved in 20 mL of CH Cl . The mixture was stirred for 5 h at RT and the resulting precipitate
2
2
was collected by filtration to give 2.079 g (89%) of (±)-6-bromospiro[chroman-2,3'-piperidin]-4-
1
one hydrochloride. H NMR (DMSO-d ) δ (ppm): 9.69 (bs, 1H), 8.83 (bs, 1 H), 7.82 (d, J=2.93
6
Hz, 1H), 7.80 (dd, J=6.75, 2.64 Hz, 1H), 7.11 (d, J=9.10 Hz, 1H), 3.39-3.54 (m, 1H), 3.10-3.26
(m, 2H), 3.00 (d, J=17.02 Hz, 1H), 2.86 (d, J=17.31 Hz, 1H), 2.62-2.86 (m, 1 H), 1.98-2.18 (m,
1
H), 1.78-1.98 (m, 1H), 1.56-1.78 (m, 2H); LC-MS: 297. 1.577 mL (11.32 mmol) of TEA, 1.085
g (5.66 mmol) of EDC and 0.866 g (5.62 mmol) of HOBt·H O were added to a solution of 0.732
2
g (3.77 mmol) of (R)-2-acetoxy-2-phenylacetic acid in 50 mL CH Cl and the mixture was
2
2
stirred at RT for 10 min. Then, 1.39 g (4.12 mmol) of the racemic spiropiperidine intermediate
was added and the mixture was stirred at RT for 18 h. After washing with 5% NaHCO and with
3

ACCEPTED MANUSCRIPT
brine the organic layer was dried over Na SO and evaporated to dryness. The crude mixture of
2
4
diastereoisomers was purified by silica gel chromatography (eluent: petroleum ether/EtOAc from
1
6
7
:4 to 3:7). The less polar isomer that eluted first (430 mg, 22%, H NMR (DMSO-d ) δ (ppm):
6
.68-7.83 (m, 2H), 7.19-7.65 (m, 5H), 6.94 (d, J=8.80 Hz, 1H), 6.38 (s, 1H), 4.11 (d, J=13.50
Hz, 1H), 3.45-3.87 (m, 1H), 2.88-3.19 (m, 2H), 2.77 (d, J=17.02 Hz, 1H), 2.67 (d, J=17.02 Hz,
H), 2.10 (s, 3H), 1.36-2.02 (m, 3H), 1.00-1.32 (m, 1H)) was then dissolved in 25 mL of EtOH
1
and 25 mL of 12 M HCl. The resulting mixture was stirred at reflux overnight. Then EtOH was
evaporated and the residue was brought to basic pH with 20% NaOH and extracted with EtOAc.
The organic phase was dried over Na SO and evaporated to dryness to give 216 mg (80%) of
2
4
1
the (-)-6-bromospiropiperidin intermediate 39. H NMR (DMSO-d ) δ (ppm): 7.76 (d, J=2.35 Hz,
6
1
H), 7.71 (dd, J=8.51, 2.64 Hz, 1H), 7.04 (d, J=8.80 Hz, 1H), 2.91 (d, J=16.73 Hz, 1H), 2.78 (d,
J=17.02 Hz, 1H), 2.74-2.90 (m, 2H), 2.65-2.74 (m, 1H), 2.54-2.64 (m, 1H), 1.80-1.90 (m, 1H),
.52-1.79 (m, 2H), 1.31-1.47 (m, 1H); optical rotation: α = -19.28°, c=0.5 g/100 mL in MeOH
1
D
(
hydrochloride salt).
98 mg (2.69 mmol) of the (-)-6-bromospirochroman intermediate 39 was dissolved in 30 mL of
CH Cl . 0.939 mL (6.74 mmol) TEA and 588 mg (2.69 mmol) BOC anhydride were added and
7
2
2
the solution was stirred overnight at RT and was then washed with 5% citric acid followed by
% NaHCO . The organic layer was dried over Na SO and evaporated to dryness. The crude
5
3
2
4
product was purified by column chromatography (eluent: petroleum ether/EtOAc 6:4, v:v) to
1
give 1.00 g (94%) of the tert-butyl ester 41. H NMR (DMSO-d ) δ (ppm): 7.79 (d, J=2.35 Hz,
6
1
2
H), 7.74 (d, J=8.22 Hz, 1H), 6.67-7.10 (m, 1H), 3.59-3.99 (m, 2H), 2.61-3.18 (m, 4H), 1.90-
.07 (m, 1H), 1.63-1.83 (m, 2H), 1.48-1.55 (m, 1H), 1.47 (s, 9H); LC-MS: (ES+) 395; optical
rotation: α = +20.67°, c=0.505 g/100 mL in MeOH.
D
5
87 mg (6.81 mmol) of methyl acrylate and 0.950 mL (6.81 mmol) of TEA were added to a
solution of 900 mg (2.27 mmol) of intermediate 41 in 3 mL DMF and the mixture was degassed
with nitrogen. 27.7 mg (0.091 mmol) P(o-tol) and 10.2 mg (0.045 mmol) Pd(OAc) were added
3
2
and the mixture was heated at 100 °C under nitrogen for 3 h. Further Pd(OAc) (10.2 mg, 0.045
2
mmol) was added and the solution was heated for additional 5 h and then partitioned between
Et O and water. The aqueous phase was washed with Et O and the collected organic layer were
2
2
washed with water, dried over Na SO and evaporated to dryness. The crude product was
2
4
purified by column chromatography (eluent: petroleum ether: EtOAc 9:1 to 7:3) to give 734 mg

ACCEPTED MANUSCRIPT
(81%) of (+)-tert-butyl 6-[(E)-3-methoxy-3-oxo-prop-1-enyl]-4-oxo-spiro[chromane-2,3'-
1
piperidine]-1'-carboxylate. H NMR (DMSO-d ) δ (ppm): 7.89-8.09 (m, 2H), 7.68 (d, J=16.14
6
Hz, 1H), 6.97-7.09 (m, 1H), 6.57 (d, J=16.14 Hz, 1H), 3.75-4.03 (m, 2H), 3.72 (s, 2H), 2.62-
3
.21 (m, 3H), 2.03 (bs, 1H), 1.65-1.87 (m, 2H), 1.28-1.62 (m, 3H), 1.00-1.26 (m, 9H); LC-MS:
(
ES+) 402; optical rotation: α = +22.40 c=0.25 g/100 mL in MeOH. 710 mg (1.77 mmol) of the
D
acrylic acid methyl ester was dissolved in 50 mL CH Cl . 2 mL of 4 M HCl in dioxane was
2
2
added and the mixture was stirred at RT for 3 h. The precipitate was filtered, washed with
CH Cl and dried to give 516 mg (86%) of the free amine intermediate as its hydrochloride salt [
2
2
1
H NMR (DMSO-d ) δ (ppm): 9.08 (bs, 2H), 8.05 (dd, J=8.80, 2.35 Hz, 1H), 8.02 (d, J=2.35
6
Hz, 1H), 7.69 (d, J=16.14 Hz, 1H), 7.15 (d, J=8.51 Hz, 1H), 6.59 (d, J=16.14 Hz, 1H), 3.73 (s,
3
H), 3.42-3.56 (m, 1H), 3.13-3.27 (m, 2H), 3.02 (d, J=17.02 Hz, 1H), 2.86 (d, J=16.73 Hz, 1H),
2
.78-2.88 (m, 1H), 1.59-2.11 (m, 4H); LC-MS: (ES+) 302; optical rotation: α = -43.67 c=0.24
D
g/100 mL in MeOH], which was then converted with benzaldehyde and NaBH(OAc) according
3
1
to the procedure used for compound 24d into the N-benzyl analogue (547 mg, 93%). H NMR
(
DMSO-d ) δ (ppm): 7.98 (dd, J=8.80, 2.35 Hz, 1H), 7.90 (d, J=2.05 Hz, 1H), 7.65 (d, J=16.14
6
Hz, 1H), 7.14-7.28 (m, 5H), 7.10 (d, J=8.51 Hz, 1H), 6.54 (d, J=16.14 Hz, 1H), 3.72 (s, 3H),
.53 (d, J=13.79 Hz, 1H), 3.45 (d, J=13.79 Hz, 1H), 2.96 (d, J=16.73 Hz, 1H), 2.89 (d, J=16.73
Hz, 1H), 2.53-2.59 (m, 1H), 2.38-2.46 (m, 3H), 1.65-1.91 (m, 3H), 1.41-1.61 (m, 1H); LC-MS:
ES+) 392; optical rotation: α = -29.92 c=0.26 g/100 mL in MeOH. Hydrolysis of 525 mg (1.34
3
(
D
mmol) of the methyl ester according to the procedure for compound 11 gave 435 mg (86%) of
1
the corresponding acrylic acid [ H NMR (DMSO-d ) δ (ppm): 12.29 (s, 1H), 7.94 (dd, J=8.66,
6
2
.20 Hz, 1H), 7.86 (d, J=2.05 Hz, 1H), 7.57 (d, J=16.14 Hz, 1H), 7.15-7.29 (m, 5H), 7.09 (d,
J=8.80 Hz, 1H), 6.43 (d, J=15.85 Hz, 1H), 3.53 (d, J=14.09 Hz, 1H), 3.45 (d, J=14.09 Hz, 1H),
2
1
.96 (d, J=16.73 Hz, 1H), 2.89 (d, J=17.02 Hz, 1H), 2.54-2.61 (m, 1H), 2.33-2.46 (m, 3H), 1.69-
.93 (m, 3H), 1.48-1.65 (m, 1H); LC-MS: (ES+) 378; optical rotation: α = -18.96 c=0.25 g/100
D
mL in MeOH], which was then converted following the procedure described for compound 16
into the requisite hydroxamic acid 33i, that was purified by trituration in i-PrOH (81.5 mg, 18%,
1
hydrochloride salt). H NMR (DMSO-d ) δ (ppm): 10.71 (bs, 1H), 10.21 (bs, 1H), 7.78-8.12 (m,
6
2H), 7.35-7.76 (m, 6H), 7.25 (d, J=8.51 Hz, 1H), 6.44 (d, J=15.55 Hz, 1H), 4.36-4.53 (m, 1H),
4.26 (dd, J=12.91, 5.58 Hz, 1H), 3.46-3.59 (m, 2H), 3.10-3.28 (m, 1H), 2.99 (d, J=17.02 Hz,
1H), 2.86-2.95 (m, 1H), 2.79 (d, J=17.02 Hz, 1H), 1.87-2.18 (m, 2 H), 1.71-1.87 (m, 1H), 1.51-

ACCEPTED MANUSCRIPT
+
1
.71 (m, 1H); LC-MS: Method C, rt=3.11; (ES+) MH : 393; optical rotation: α = -12.30 c= 1
D
g/100 mL in MeOH.
The synthesis of (+)-(E)-3-[1'-benzyl-4-oxo-spiro(chromane-2,3’-piperidine)-6-yl]-N-hydroxy-
acrylamide (33j) is disclosed in the supplementary information.
4
4
.2 Biological Assays.
.2.1 HDACs Inhibition Assay
The in vitro activity of HDAC inhibitors was assayed using the HDAC fluorescent histone
deacetylase activity assay kit (Biomol Research Laboratories, Plymouth Meeting, PA). In
specific, 15 µL of a nuclear extract from HeLa cells was diluted to 50 µL with the assay buffer
containing the HDAC inhibitor and a peptide containing acetylated lysine as substrate at a
concentration of 200 µM. The samples were incubated for 15 min at room temperature and then
exposed to a developer for a further 10 min. Substrate fluorescence was measured using an
excitation at 355 nm and emission at 460 nm, and the % inhibition was calculated relative to
untreated samples. In order to eliminate possible background interferences the fluorescence was
measured for each compound at the test concentrations in the assay buffer in absence of the
enzyme. IC50 values were determined using a non-linear regression analysis of the
concentration/inhibition data.
The in vitro activities of recombinant human HDACs 2, 3 and 6 (BPS Bioscience, San Diego,
CA) were assessed using a fluorescent substrate (AK500, Biomol Research Laboratories,
Plymouth Meeting, PA). 6.5 µL of the recombinant HDAC proteins (which corresponds to 20
µg/well, 6 µg/well and 500 µg/well for HDACs 2, 3, and 6, respectively) were diluted to 25 ꢀL
with the assay buffer containing the HDAC inhibitors and 25 ꢀM of the substrate. The samples
were incubated for 15 min at room temperature and then exposed to a developer for another 10
min.
4
.2.2 Immunoblots and Antibodies

ACCEPTED MANUSCRIPT
Whole cell extracts were obtained by lysis in buffer supplemented with proteases inhibitors (20
mM Tris HCl pH 7.5, 25% glycerol, 420 mM NaCl, 15 mM MgCl , 0.2 mM EDTA, 0.1%
2
NP40). Proteins were separated by SDS-PAGE, blotted onto PVDF membranes and probed with
the indicated antibodies. The employed antibodies were anti-HDAC1, anti-HDAC3, anti-
HDAC4, anti-HDAC5 and anti-HDAC6 (Cell Signaling, Leiden, The Netherlands), anti-HDAC2
(Abcam, Cambridge, UK) and anti α-tubulin (Sigma-Aldrich, Milan, Italy). Horseradish
peroxidase–conjugated secondary antibodies (Sigma-Aldrich, Milan, Italy) were used with
enhanced chemiluminescence (Amersham Pharmacia Biotech) for detection.
4
.2.3 Cell Growth Assay
The effect of the HDAC inhibitors on cell proliferation was evaluated against K562 (chronic
myeloid carcinoma), A549 (non-small-cell lung cancer) and HCT-116 (human colon cancer)
tumor cell lines using the CellTiter–Glo Luminescent Cell Viability Assay (Promega, Madison,
WI, USA) according to the manufacturer's instructions. K562, A549 and HCT–116 cells were
first incubated for 72 h with the inhibitors at different concentrations. Then, an equivalent of the
CellTiter–Glo reagent was added, the solution was mixed for 2 min in order to induce cell lysis,
and the luminescence was recorded after another 10 min. The IC50 was calculated using
GraphPad Software.
4
. 2.4 Molecular Modeling
Target Preparation
The protein structures of the catalytic domain of HDAC1 and HDAC2, pdb identifiers 4BKX
[
35] and 4LXZ[36], respectively, were prepared for docking with Autodock[43] and Plants[44]
as well as a for a final optimization with Mopac[45]. The pdb files of the structures were
downloaded from the Protein Data Bank (www.rcsb.org) and prepared with same multi-step
procedure here described. In Chimera software[46] the HDAC proteins, the zinc ions and the
ligands (vorinostat in 4LXZ and the acetate in 4BKX) were kept while the other molecules were
removed. Hydrogen atoms were added to the residues and optimized for the H-bonding networks