Synthesis of 5-amino- and 5-hydroxy-3,3-difluoropiperidines

Article abstract of DOI:10.1039/c0ob00231c

Synthetic routes toward new 5-amino- and 5-hydroxy-3,3-difluoropiperidines, which are of high interest as building blocks in medicinal chemistry, are described. The key step involves the N-halosuccinimide-induced cyclization of 2,2-difluoro-4-pentenylamines toward 5-halo-3,3-difluoropiperidines, which were used to synthesize 5-amino-3,3-difluoropiperidine. In a second strategy, iodolactonization of 2,2-difluoro-4-pentenoic acid gave the corresponding ??-lactone, which was transformed into 5-hydroxy-3,3-difluoropiperidine.

Full text of DOI:10.1039/c0ob00231c

View Article Online / Journal Homepage / Table of Contents for this issue
COMMUNICATION
www.rsc.org/obc | Organic & Biomolecular Chemistry
Synthesis of 5-amino- and 5-hydroxy-3,3-difluoropiperidines†
Riccardo Surmont,‡a Guido Verniest,§a Jan Willem Thuring,b Peter ten Holte,b Frederik Derooseb and
Norbert De Kimpe*a
Received 8th June 2010, Accepted 5th July 2010
DOI: 10.1039/c0ob00231c
Synthetic routes toward new 5-amino- and 5-hydroxy-3,3-
difluoropiperidines, which are of high interest as build-
ing blocks in medicinal chemistry, are described. The
key step involves the N-halosuccinimide-induced cycliza-
tion of 2,2-difluoro-4-pentenylamines toward 5-halo-3,3-
difluoropiperidines, which were used to synthesize 5-amino-
3,3-difluoropiperidine. In a second strategy, iodolactoniza-
Fig. 1 Biologically active 5-amino- and 5-hydroxy-3,3-difluoropipe-
ridines.
tion of 2,2-difluoro-4-pentenoic acid gave the correspond-
ing c-lactone, which was transformed into 5-hydroxy-3,3-
difluoropiperidine.
kinase inhibitor and has been designed for cancer treatment.4 Ma-
trix metalloprotease inhibitors5 and excitatory amino acid recep-
The introduction of a fluorine substituent in pharmaceutical and
tor antagonists6 are found among 5-acyl-3,3-difluoropiperidine
agrochemical compounds is a small modification in their chem-
derivatives, which are also promising compounds for the use in
ical structure which can provoke highly advantageous changes
the treatment of nervous system disorders7 and diseases depending
in the chemical and biological properties of the compounds,
on renin activity.8 Substituted 5-hydroxy-3,3-difluoropiperidines,
including their lipophilicity, stability and bioavailability.1 Among
bearing additional functionalities at the piperidine ring (e.g. 2)
the wide range of organofluorinated compounds, substituted
difluorinated piperidines have received a lot of attention in the
have been synthesized as azasugar derivatives and displayed b-
glucosidase activities.9
patent literature as bifunctional building blocks for the use in
5-Substituted 3,3-difluoropiperidines are generally prepared
structure–activity relationship studies of bioactive compounds.2
by deoxofluorination of the corresponding functionalized 3-
Particularly in the case of 3,3-difluoropiperidines, the fluorine
piperidinones. Unfortunately, deoxofluorination reactions can
suffer from the formation of rearranged or dehydrofluorinated
atoms at the b-position of the nitrogen of the piperidine ring
dramatically lower the nucleophilicity and basicity of this nitrogen.
products,10 resulting in low yields of the targeted fluorohete-
Recently we published a new synthetic pathway toward valuable
rocycles. Therefore, the development of new strategies towards
4-substituted 3,3-difluoropiperidines consisting of a Cu-mediated
substituted difluoropiperidines, without the need for DAST or
1,4-addition of ethyl bromodifluoroacetate to 3-substituted acry-
analogous deoxofluorination reagents, is of current interest.
Recently, a 5-methyl-3,3-difluoropiperidine was synthesized via
a cyclization/fluorination reaction of chlorinated diallylamines in
superacid.10
lonitriles followed by d-lactamization and reduction.3 However,
this methodology could not be used for the synthesis of 5-
substituted 3,3-difluoropiperidines because the 1,4-addition of
ethyl bromodifluoroacetate to 2-substituted acrylonitriles turned
In order to establish a generally applicable and efficient large
out to be problematic. However, 5-substituted 3,3-difluorinated
scale synthesis of 5-substituted 3,3-difluoropiperidines, it was pro-
piperidines are promising compounds in medicinal chemistry
posed to investigate the electrophile-induced cyclization of difluo-
and a general method for the synthesis of these compounds is
rinated N-alkenylamines 6 towards 5-halo-3,3-difluoropiperidines
8, which were believed to be good precursors for 5-hydroxy- and
5-aminopiperidines.
At first, suitable starting protected amines 6 were synthesized
from the corresponding 2,2-difluoro-4-penten-1-ol 4, which was
prepared by reduction of 2,2-difluoro-4-pentenoic acid 3 (Scheme
1). This versatile synthetic building block is available from
fluorinated precursors such as tetrafluoroethylene gas11 or the
clearly lacking in the literature. It should be noted that only one
reference was found concerning 5-amino-3,3-difluoropiperidines
and no literature data were found concerning the corresponding
3,3-difluoro-5-hydroxypiperidines bearing no further substituents
at the piperidine ring. 5-Amino-3,3-difluoropiperidine 1 (Fig. 1)
is known to be a PIM (proto-oncogene serine/threonine-protein)
aDepartment of Organic Chemistry, Faculty of Bioscience Engineering,
commercially available and easily handled chlorodifluoroacetic
acid.12,13 The reduction of ethyl 2,2-difluoro-4-pentenoate to
Ghent University, Coupure Links 653, B-9000 Ghent, Belgium. E-mail:
norbert.dekimpe@ugent.be; Fax: +32 (0)9 264 62 43
bJohnson & Johnson, Pharmaceutical Research & Development, Division of
Janssen Pharmaceutica NV, Turnhoutseweg 30, B-2340 Beerse, Belgium
2,2-difluoropent-4-en-1-ol 4 has been described before using
NaBH4.14 However, to avoid the extra esterification step, it was
decided to directly reduce 2,2-difluoro-4-pentenoic acid 3 using 3
equivalents of LiAlH4 in diethyl ether. This sequence resulted in
difluoroalcohol 4 in 80% yield and was performed on a 20 gram
scale. Subsequently, 2,2-difluoropent-4-en-1-ol 4 was successfully
tosylated toward compound 5a in 64% yield by reaction of the
† Electronic supplementary information (ESI) available: Experimental
procedures and spectroscopic data for compounds 5a, 6b–d, 8–11 and
13–16 are available. See DOI: 10.1039/c0ob00231c
‡ Aspirant of the Research Foundation – Flanders (FWO-Vlaanderen)
§ Postdoctoral Fellow of the Research Foundation – Flanders (FWO-
Vlaanderen, Belgium)
4514 | Org. Biomol. Chem., 2010, 8, 4514–4517
This journal is
The Royal Society of Chemistry 2010
©

View Article Online
Scheme 2 Synthesis of 5-substituted difluoropiperidines 8–10.
resulted in an unsatisfactory 1 : 1 : 3 mixture of starting material,
amine 6c and pyrrolidine 7a (X = Cl), respectively. However, when
a catalytic amount (10 mol%) of tetrabutylammonium iodide
(TBAI) was added after reaction of 6c with NCS, cyclization
proceeded smoothly to give a 18 : 82 mixture of fluorinated
5-chloromethylpyrrolidine 7a and 5-chloropiperidine 8a after
heating for 15 h at 50 ◦C in CHCl3 (Table 1).19 A complete
conversion towards piperidine 8a was obtained after heating the
obtained mixture during 60 h at 83 ◦C in dichloroethane in the
presence of 1 equivalent of LiCl. Distinction between the isomeric
pyrrolidine 7a and piperidine 8a was made by means of mass
spectrometry. In the case of pyrrolidine 7a the presence of an ion
at m/z 196 accounted for the homolytic cleavage of a CH2Cl-
group. In contrast, the mass spectrum of 8a did not show a
Scheme 1 Synthesis of fluorinated protected amines 6a–c.
alcohol with tosyl chloride in pyridine. Unfortunately, tosylate 5a
could not be substituted by benzylamine or azide under various
reaction conditions. Alternatively, triflation of 2,2-difluoropent-4-
en-1-ol 4 via reaction with triflic anhydride yielded compound 5b,
which was directly transformed into the phthaloyl-protected amine
6a.14 Subsequently, reaction with 2 equiv. of NH2NH2·H2O in
ethanol at 50 ◦C gave the corresponding amine which was trapped
as its HCl salt. Because the obtained ammonium salt could not
be handled easily, it was decided to introduce the nitrogen via
an azide substitution of triflate 5b. Reaction of triflate 5b with
1.1 equivalents of sodium azide in DMSO gave the new 5-azido-
4,4-difluoropent-1-ene 6b after 24 h at room temperature in 67%
yield after distillation. Analogous to the synthesis of azide 6b,
the substitution of triflate 5b with benzylamine in THF at reflux
temperature proceeded very smoothly and yielded N-benzyl-2,2-
difluoro-4-pentenamine 6c in almost quantitative yield. Because
of the ease of preparation of amine 6c, even on a large scale,
and the simple purification of 6c via an acid–base extraction,
the cyclization of N-halo difluoroamines using this substrate was
investigated.
It is known that N-alkenyl-N-chloroamines can be cyclized
to pyrrolidines and piperidines via a Cu(I)-catalyzed radical
cyclization,15 via a catalytic hetero-Heck type reaction,16 or
via reaction with (Lewis) acids.17,18 To evaluate this cycliza-
tion method, amine 6c was treated with N-chlorosuccinimide
in dichloromethane (Scheme 2, Table 1), and subsequently, 2
equivalents of BF3.OEt2 as a Lewis acid were added to the formed
N-chloroamine 6d in dichloromethane at -78 ◦C. This reaction
1
m/z 196 fragment ion. In addition, in the H NMR spectrum
(CDCl3) of 5-chloropiperidine 8a the CHCl was assigned to the
well-resolved t ¥ t (11.3 Hz, 4.9 Hz) at d 4.04 ppm typically
for 5-halopiperidines, in comparison to the broad singlet of
the NCH of 5-(halomethyl)pyrrolidines.20 To evaluate the effect
of the halogen, fluorinated amine 6c was also treated with N-
iodosuccinimide in dichloromethane at room temperature. A fast
cyclization to 3,3-difluoro-5-iodopiperidine 8c occurred without
formation of the corresponding pyrrolidine. Also the reaction
of amine 6c with N-bromosuccinimide gave rise to a rapid
formation of a 72 : 18 mixture of 5-bromomethylpyrrolidine 7b
and 5-bromopiperidine 8b, respectively. The pyrrolidine inter-
mediate 7b was slowly, but completely converted into 1-benzyl-
5-bromo-3,3-difluoropiperidine 8b upon standing in acetone at
room temperature and was isolated in 77% yield after flash
chromatography. The three new fluorinated 5-halopiperidines 8a–
c (X = Cl, Br, I) show almost identical 1H NMR spectra. However,
in the 13C NMR spectrum (CDCl3) it is noted that the CHX peak
Table 1 Synthesis of 5-halogenated 3,3-difluoropiperidines 8a–c from N-benzyl-2,2-difluoro-4-pentenamine 6c (Scheme 2)
Difluoropiperidine 8
Entry
1
X
Reaction conditions A
Ratio 7 : 8
Reaction conditions B
(isolated yield)
Cl
1) 1 equiv. NCS, CH2Cl2,
0 ◦C, 2 h (6c → 6d)
18 : 82
1 equiv. LiCl, Cl(CH2)2Cl, D,
60 h
8a (86%)
2) 0.1 equiv. Bu4NI, CHCl3,
50 ◦C, 15 h (6d → 7a/8a)
1 equiv. NBS, CH2Cl2, rt, 2 h
(6c → 7b/8b)
2
3
Br
I
72 : 28
0 : 100
acetone, rt, 72 h
(not applicable)
8b (77%)
8c (82%)
1 equiv. NIS, CH2Cl2, rt, 2 h
(6c → 8c)
This journal is
The Royal Society of Chemistry 2010
Org. Biomol. Chem., 2010, 8, 4514–4517 | 4515
©

View Article Online
Table 2 Optimization of the synthesis of 3-azidopiperidine 9
enhanced acidity of the protons at the a-position of the piperidine
nitrogen of 11 in comparison to the 1-benzylderivative 8a.
Because the previous strategy was not applicable to synthesize
3,3-difluoro-5-hydroxypiperidines, which are of significant interest
as building blocks in medicinal chemistry, another synthetic
methodology was evaluated. For this purpose, the easily ac-
cessible 2,2-difluoro-4-pentenoic acid 3 was cyclized using a
iodolactonization reaction in acetonitrile toward a,a-difluoro-
g-iodomethyl-g-butyrolactone 13 in 90% yield (Scheme 3). The
iodine substituent was then displaced by azide in DMSO to give the
fluorinated g-azidomethyl-g-lactone 14 in 69% yield. Reduction of
14 by catalytic hydrogenation (H2, Pd/C) led to the intermediate
amine, which underwent a smooth rearrangement towards 3,3-
difluoro-5-hydroxy-2-piperidinone 15. However, due to the high
polarity of lactam 15, only 60% could be recovered after flash
chromatography and crystallization. Finally, 2-piperidinone 15
was successfully reduced using BH3·THF and the obtained 3,3-
difluor-5-hydroxypiperidine was trapped as the benzyloxycarbonyl
derivative 16 in 65% overall yield.23
X
R
Reaction conditions
Conversion 12 : 9
11 Cl Boc NaN3, rt to 100 ◦C, DMF 0%
/
or DMSO
8a Cl Bn
8b Br Bn
8b Br Bn
2 equiv. NaN3, DMSO,
100%
14 : 86
6 : 94
100 ◦C, 15 h
1.1 equiv. NaN3, DMF, rt, 29%
24 h
1.1 equiv. NaN3, DMF,
100%
10 : 90
7 : 93(71)a
70 ◦C, 15 h
8c
I
Bn
1 equiv. NaN3, DMF, rt,
24 h
100%
a Isolated yield of 3-azidopiperidine 9 in parentheses.
(determined by DEPT-135 analysis) is clearly shifted upfield from
the chlorine (50.3 ppm), over the bromine (40.1 ppm), to the iodine
derivative (15.4 ppm), which is another structural evidence of the
6-membered ring structure.
New 5-halogenated 3,3-difluoropiperidines 8 can thus be
prepared on a large scale and form good substrates for
further functionalization reactions toward 5-substituted 3,3-
difluoropiperidines. In particular, 5-amino-3,3-difluoropiperidine
is an interesting target molecule for the use as bifunctional building
block for incorporation in bioactive compounds. It is known that
piperidines having a leaving group at the 3-position react with
nucleophiles via intramolecular substitution of the leaving group
by nitrogen and subsequent opening of the bicyclic aziridinium
ion at the more substituted carbon by the nucleophile to give 3-
functionalized piperidines, which are sometimes accompanied by
small amounts of the corresponding pyrrolidines bearing a CH2Nu
substituent at the 5-position.21 Indeed, when 5-halopiperidines 8a–
c were reacted with sodium azide as a nucleophile in DMSO or
DMF, 5-azido-3,3-difluoropiperidine 9 was formed accompanied
by small amounts (about 6–14%) of the 5-azidomethylpyrrolidine
isomer 12 (Table 2). The best results were obtained by treating 5-
iodo-3,3-difluoropiperidine 8c with NaN3 in DMF at rt during
24 h, resulting in complete conversion towards 12 and 9 in a
ratio of 7 : 93, respectively. 5-Azido-3,3-difluoropiperidine 9 was
separated from the pyrrolidine isomer via flash chromatography
and was isolated in 71% yield. Catalytic hydrogenation of 5-azido-
3,3-difluoropiperidine 9 at atmospheric pressure gave 5-amino-
3,3-difluoropiperidine 10 in almost quantitative yield without
removal of the N-benzyl group (Scheme 2). Analogous reac-
tions of 3,3-difluoro-5-iodopiperidine 8c with nucleophiles such
as ammonia, sodium hydroxide, sodium acetate or potassium
cyanide in methanol or DMF gave complex mixtures. In order
to avoid the neighbouring group participation of the piperidine
nitrogen during the nucleophilic substitution reactions, 5-chloro-
3,3-difluoropiperidine 8a was converted to 1-Boc-5-chloro-3,3-
difluoropiperidine 11 via standard methods in almost quantitative
yield.22 However, when 1-Boc-5-chloropiperidine 11 was reacted
with sodium azide under various reaction conditions, only tetrahy-
dropyridine derivatives were observed, probably because of the
Scheme 3 Synthesis of 3,3-difluoro-5-hydroxypiperidine 16.
In conclusion, straightforward synthetic pathways toward 5-
amino- and 5-hydroxy-3,3-difluoropiperidines were developed,
starting from 2,2-difluoro-4-pentenoic acid. Cyclization of N-
benzyl-N-(2,2-difluoro-4-pentenyl)amine using NIS yielded a new
5-iodo-3,3-difluoropiperidine in high yield. This piperidine was
subsequently used for the efficient synthesis of 5-amino-3,3-
difluoropiperidine. Alternatively, iodocyclization of the starting
difluoropentenoic acid gave the corresponding g-lactone, which
proved to be a good starting material for the synthesis of
5-hydroxy-3,3-difluoropiperidine. The synthesized 5-substituted
difluoropiperidines were obtained on large scale and are of high
interest as building block for medicinal chemistry.
Acknowledgements
The authors are indebted to the Research Foundation – Flanders
(FWO-Flanders), Ghent University (GOA, BOF) and Johnson &
Johnson Pharmaceutical Research & Development, Division of
Janssen Pharmaceutica NV, for financial support.
Notes and references
1 (a) D. O’Hagan, Chem. Soc. Rev., 2008, 37, 308; (b) W. K. Hagmann,
J. Med. Chem., 2008, 51, 4359.
2 For examples, see; (a) M. G. Stanton, J. Hubbs, D. Sloman, C.
Hamblett, P. Andrade, M. Angagawc, G. Bi, R. M. Black, J. Crispino,
J. C. Cruz, E. Fan, G. Farris, B. L. Hughes, C. M. Kenific, R. E.
Middleton, G. Nikov, P. Sajonz, S. Shah, N. Shomer, A. A. Szewczak,
4516 | Org. Biomol. Chem., 2010, 8, 4514–4517
This journal is
The Royal Society of Chemistry 2010
©

View Article Online
F. Tanga, M. T. Tudge, M. Shearman, B. Munoz, Bioorg. Med. Chem.
Lett. 2010, 20, 755; (b) S. Ninkovic, J. F. Braganza, M. R. Collins, J.
C. Kath, H. Li, D. T. Richter PCT Int. Appl. WO 2010016005, 2010;
(c) S. Ninkovic, J. F. Braganza, M. R. Collins, J. C. Kath, H. Li and D.
T. Richter, Chem. Abstr., 2010, 152, 262750.
3 R. Surmont, G. Verniest, J. W. Thuring, G. Macdonald, F. Deroose and
N. De Kimpe, J. Org. Chem., 2010, 75, 929.
4 M. Burger, J. Lan, M. Lindvall, G. Nishiguchi, M. Tetalman PCT
Int. Appl. WO 2009109576, 2009; M. Burger, J. Lan, M. Lindvall, G.
Nishiguchi and M. Tetalman, Chem. Abstr., 2009, 151, 358734.
5 Y. Li, J. Zhou, D. Burns, W. Yao PCT Int. Appl. WO 2005037826, 2005;
Y. Li, J. Zhou, D. Burns and W. Yao, Chem. Abstr., 2005, 142, 430307.
6 S. A. Filla, K. J. Hudziak, B. M. Mathes, P. L. OrnsteinPCT Int. Appl.
WO 2002053555, 2002; S. A. Filla, K. J. Hudziak, B. M. Mathes and P.
L. Ornstein, Chem. Abstr., 2002, 137, 78866.
13 F. Xu, B. Simmons, J. Armstrong III and J. Murry, J. Org. Chem., 2005,
70, 6105.
14 D. A. Kendrick, C. Danzin and M. Kolb, J. Med. Chem., 1989, 32, 170.
15 (a) R. Go¨ttlich, Synthesis, 2000, 1561; (b) G. Heuger, S. Kalsow and R.
Go¨ttlich, Eur. J. Org. Chem., 2002, 1848; (c) M. Noack and R. Go¨ttlich,
Chem. Commun., 2002, 536.
16 J. Helaju and R. Go¨ttlich, Chem. Commun., 2002, 720.
17 (a) F. Minisci, Synthesis, 1973, 1; (b) L. Stella, Angew. Chem., Int.
Ed. Engl., 1983, 22, 337; (c) M. Hemmerling, A. Sjo¨holm and P.
Somfai, Tetrahedron: Asymmetry, 1999, 10, 4091; (d) A. Sjo¨holm, M.
Hemmerling, N. Pradeille and P. Somfai, J. Chem. Soc., Perkin Trans.
1, 2001, 891.
18 M. Noack, S. Kalsow and R. Go¨ttlich, Synlett, 2004, 1110.
19 (a) R. Go¨ttlich and M. Noack, Tetrahedron Lett., 2001, 42, 7771; (b) M.
Noack and R. Go¨ttlich, Eur. J. Org. Chem., 2002, 3171.
7 S. Gagliardi, E. Le Poul, L. Lingard, G. Palombi, S. M. Poli, J.-P. Rocher
PCT Int. Appl. WO 2006123257, 2006; S. Gagliardi, E. Le Poul, L.
Lingard, G. Palombi, S. M. Poli and J.-P. Rocher, Chem. Abstr., 2006,
145, 505460.
8 K. Masuya, F. Yokokawa, O. Irie, A. Nihonyanagi, A. Toyao, T.
Ehara, K. Konishi, T. Kanazawa, M. Suzuki PCT Int. Appl. 2006, WO
200609476, 2006; K. Masuya, F. Yokokawa, O. Irie, A. Nihonyanagi,
A. Toyao, T. Ehara, K. Konishi, T. Kanazawa and M. Suzuki, Chem.
Abstr., 2006, 145, 335938.
9 R.-W. Wang, X.-L. Qiu, M. Bols, F. Ortega-Caballero and F. L. Qing,
J. Med. Chem., 2006, 49, 2989.
10 E. Vardelle, A. Martin-Mingot, M.-P. Jouannetaud, C. Bachmann, J.
Marrot and S. Thibaudeau, J. Org. Chem., 2009, 74, 6025.
11 (a) J. F. Normant, O. Reboul, R. Sauveˆtre, H. Deshayes, D. Masure and
J. Villieras, Bull. Soc. Chim. Fr., 1974, 2072; (b) M. Kolb, F. Gerhart
and J.-P. Franc¸ois, Synthesis, 1988, 469.
20 (a) K. Abbaspour Tehrani, K. Van Syngel, M. Boelens, J. Contreras, N.
De Kimpe and D. W. Knight, Tetrahedron Lett., 2000, 41, 2507; (b) N.
De Kimpe and M. Boelens, J. Chem. Soc., Chem. Commun., 1993, 916;
(c) N. De Kimpe, M. Boelens, J. Piqueur and J. Baele, Tetrahedron Lett.,
1994, 35, 1925.
21 (a) D. Black and J. E. Doyle, Adv. Heterocycl. Chem., 1981, 27, 1; (b) R.
H. Reitsema, J. Am. Chem. Soc., 1949, 71, 2041; (c) C. F. Hammer
and J. D. Weber, Tetrahedron, 1981, 37, 2173; (d) J. Cossy, C. Dumas, P.
Michel and D. Gomez Pardo, Tetrahedron Lett., 1995, 36, 549; (e) N. De
Kimpe and M. Boelens, Tetrahedron Lett., 1996, 37, 3171; (f) E. Rosas
Alonso, K. Abbaspour Tehrani, M. Boelens, D. W. Knight, V. Yu and
N. De Kimpe, Tetrahedron Lett., 2001, 42, 3921; (g) C. V. Stevens, M.
Peristeropoulou and N. De Kimpe, Tetrahedron, 2001, 57, 7865.
22 V. D. Vitnik, M. D. Ivanovic, Z. J. Vitnik, J. B. Dordevic, Z. S. Zizak,
Z. D. Juranic and I. O. Juranic, Synth. Commun., 2009, 39, 1457.
23 (a) N. Nuh and C. M. Thompson, Tetrahedron, 1995, 51, 5935; (b) A.
Gross, D. R. Borcherdin, D. Friedrich and J. S. Sabol, Tetrahedron
Lett., 2001, 42, 1631.
12 H. Greuter, R. W. Lang and A. J. Romann, Tetrahedron Lett., 1988, 29,
3291.
This journal is
The Royal Society of Chemistry 2010
Org. Biomol. Chem., 2010, 8, 4514–4517 | 4517
©