Sequential copper-catalyzed rearrangement-allylic substitution of bicyclic hydrazines with grignard reagents

Article abstract of DOI:10.1002/adsc.200900037

The problem of the synthesis of trans-1,4-disubstituted hydrazino- and aminocyclopentenes has been resolved by a sequential copper-catalyzed rearrangement-allylic alkylation of 2,3-diazabieyelo-[2.2.1]heptenes. The cis-fused 5,5-membered allylic carbazate

Full text of DOI:10.1002/adsc.200900037

COMMUNICATIONS
DOI: 10.1002/adsc.200900037
Sequential Copper-Catalyzed Rearrangement–Allylic Substitution
of Bicyclic Hydrazines with Grignard Reagents
Stefano Crotti,a Ferruccio Bertolini,a Franco Macchia,a and Mauro Pineschia,*
a
Dipartimento di Scienze Farmaceutiche, Universitꢀ di Pisa, Via Bonanno 33, 56126 Pisa, Italy
Fax : (+39)-(0)50-221-9660; e-mail: pineschi@farm.unipi.it
Received: January 19, 2009; Published online: April 6, 2009
Supporting information for this article is available on the WWW under
http://dx.doi.org/10.1002/adsc.200900037.
Abstract: The problem of the synthesis of trans-1,4-
disubstituted hydrazino- and aminocyclopentenes
has been resolved by a sequential copper-catalyzed
rearrangement–allylic alkylation of 2,3-diazabicyclo-
ACHTUNGTRENNUNG
(OTf)2]/(Æ)-BINAP-catalyzed re-
arrangement, can be alkylated with a broad spec-
trum of Grignard reagents with a predominant SN2’-
regioselectivity. The N-Boc protecting group proved
to be optimal as regards yields, reaction times and
regioselectivities.
Keywords: allylic alkylation; copper; Grignard re-
agents; rearrangement; sequential reactions
Figure 1. Possible approach to 1,4-substituted azido- or hy-
drazinocyclopentenes.
ported methods gave invariably 1,2-disubstituted hy-
drazinocyclopentenes.[5] It should be noted that
Substituted aminocyclopentenes hold a tremendous Grignard reagents, which are the most readily avail-
amount of synthetic potential as intermediates for the able organometallic reagents, were never considered.
preparation of a variety of biologically interesting
Furthermore, it is known that the direct nucleophil-
molecules.[1] For example, 4-amino-2-cyclopentene-1- ic displacement at the bridgehead position of
methanol is a widely used chiral building block for [2.2.1]bicyclic systems, which can be a reasonable
the preparation of carbocyclic adenosine derivatives pathway to these compounds, is reported to be ex-
which possess antiviral properties.[2] Installation of a tremely difficult [Eq. (3), Figure 1].[6]
carbon-based nucleophile onto a pre-existing cyclo-
We report here a novel regioselective synthesis of
pentene represents a viable strategy to introduce sub- trans-1,4-disubstituted hydrazino- and aminocyclopen-
stitution on this ring. For example, the allylic displace- tenes by a sequential copper-catalyzed rearrange-
ment of 4-cyclopentene-1,3-diol monoacetate with ment–allylic
alkylation
of
2,3-diazabicyclo-
aryl- and alkenyl-Grignard reagents followed by sub-
ACHTUNGTRENNUNG
stitution of the hydroxy group with (PhO)2P(=O)N3
gave cis-1,4-azidoarylcyclopentenes [Eq. (1), Fig- of bicyclic [2.2.1] diacylhydrazines, such as 1b, to give
ure 1].[3a] cis-1,4-Disubstituted bis-Cbz-protected hy- oxadiazines has been extensively studied and ex-
drazinocyclopentenes can be obtained by a diastereo- plained on the basis of a [3,3]-sigmatropic (hetero-
selective palladium-catalyzed ring opening of Cbz- Cope) rearrangement [Eq. (4), Scheme 1].[7] More re-
protected bicyclic hydrazine (1a) with soft nucleo- cently, the corresponding protic or Lewis acid-cata-
philes [Eq. (2), Figure 1).[4] Quite recently, several lyzed (LA) rearrangement of carbobenzyloxy-protect-
methods for the ring opening of bicyclic hydrazines ed bicyclic [2.2.1] hydrazine 1a was explained by
with carbon nucleophiles were developed, but all re- means of a transient allylic cation to give 5,6-bicyclic
Adv. Synth. Catal. 2009, 351, 869 – 873
ꢁ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
869

Stefano Crotti et al.
COMMUNICATIONS
Scheme 1. Modes of rearrangement of differently protected [2.2.1] bicyclic hydrazines.
intermediate 4a [Eq. (5), Scheme 1].[4,5b] As com- However, for the copper-catalyzed reaction of the N-
pound 4a has an allylic leaving group, we considered Boc derivative 1c, the rather small solvent effect to-
this compound as a possible candidate for an allylic gether with the failure of detection of cationic inter-
alkylation with Grignard reagents. However, in our mediates, point to a concerted cyclic mechanism
hands, when bicyclic hydrazine 1a was allowed to which can formally be interpreted as a [3,4]-sigma-
react with H2SO4 in CF3CH2OH in accordance with a tropic rearrangement. In this scenario, the azaphilic
previously reported procedure,[5b] a remarkably differ-
ent reaction outcome was obtained. To our surprise,
Table 1. Screening of Lewis acids (LA) for the rearrange-
the main product was not the expected 5,6-bicyclic
carbazate 4a but the 5,5-analogue 3a (65% isolated
yield).[8]
ment of [2.2.1] bicyclic hydrazines.[a]
In order to improve the yields of allylic carbazate
of type 3, the rearrangement of differently carbamoyl
protected bicyclic hydrazines 1a, 1c, 1d was investigat-
ed under the catalysis of different Lewis acids
(Table 1). Using catalytic amounts (3.0 mol%) of
CuACHTUNGTRENNUNG(OTf)2 we soon realized that the N-tert-butoxycar-
bonyl (Boc) protected hydrazine 1c was much more
prone to the rearrangement reaction than 1a
(entry 1), and 1d (entry 2), and afforded cleanly the
corresponding carbazate 3c as a white solid (entry 3).
Copper(I) and copper(II) salts with a coordinating
ligand (entries 4–6), proved to be ineffective for this
Entry Substrate Lewis Acid
Conversion
[%]
Yield
[%][b]
1[c]
2[c]
3
4
5
1a
1d
1c
1c
1c
1c
1c
1c
1c
Cu
Cu
Cu
N
95
58 (3a)
complex mixture
100
<5
97 (3c)
N.d.
N.d.
CuCl
Cu
CuCl2
Sc(OTf)3
FeCl3
rearrangement. On the other hand, also Sc
FeCl3 promoted this kind of transformation in high
yields (entries 7 and 8), whereas Zn(OTf)2 gave a
ACHTUNGRTEN(NUNG OTf)3 and
6
N.d.
AHCTUNGTRENNUNG
7
G
100
95
<10
100
82 (3c)
80 (3c)
N.d.
very low conversion (entry 9). The effect of adding
phosphorus-based ligands to the reaction catalyzed by
CuACHTUNGTRENNUNG(OTf)2 was also examined. Very interestingly we
8[d]
9
Zn
Cu
E
A
10[e] 1c
97 (3c)
found that the reaction performed in the presence of
(Æ)-BINAP (6 mol%) was particularly efficient and
complete conversion was observed in 10 min at room
temperature (entry 10).[9] A similar result can be ob-
tained by the use of non-coordinating solvents such as
CH2Cl2 or toluene (data not shown in Table 1).
[a]
Conditions: 1a or 1c, Lewis acid (3 mol%) in THF at
room temperature for 16 h, unless stated otherwise.
Isolated yields after chromatographic purification on
silica gel.
The reaction has been carried out in CH2Cl2 because in
THF extensive polymerization of the solvent was ob-
served.
8 mol% of this salt was used.
The reaction was carried out in the presence of 6 mol%
of BINAP for 10 min at room temperature.
[b]
[c]
The detection by MS of dimeric structures,[5b] and
the polymerization of THF suggested a stepwise
mechanism with the involvement of open cationic
species at least for N-Cbz protected compound 1a.[8b]
[d]
[e]
870
asc.wiley-vch.de
ꢁ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Adv. Synth. Catal. 2009, 351, 869 – 873

Sequential Copper-Catalyzed Rearrangement–Allylic Substitution
COMMUNICATIONS
Table 2. Copper-catalyzed one-pot rearrangement–alkylation
of 1c with Grignard reagents.[a]
Figure 2. Plausible mechanism for the formation of 5,5-bicy-
clic carbazate 3c.
Entry
Grignard
Yield [%][b]
5/6[c]
1
2
3
4
5
6
7
MeMgBr
EtMgBr
allylMgBr
PhCH2MgBr
AHCTUNGTREG(NUNN i-PrO)Me2SiMeMgCl
i-PrMgCl
75 (5a)
70 (5b)
80 (5c)
75 (5d)
75 (5e)
80 (5f)
72 (5g)
87/13
86/14
83/17
81/19
82/18
86/14
83/17
Cu(I)-phosphine catalyst might activate the nitrogen
of the hydrazine promoting the rearrangement with
the subsequent formation of volatile 2-methylpro-
pene, as shown in Figure 2.[10]
Copper-catalyzed allylic alkylation with organome-
tallic reagents represents a widely used synthetic
strategy to introduce a carbon substituent on a mole-
cule.[11] To our delight, after ensuring by TLC analysis
PhMgBr
8
9
85 (5h)
92/8
that the CuACHTUNGTRENNUNG(OTf)2-BINAP-catalyzed rearrangement
58 (5i)
85/15
had taken place (after ca. 10 min for 1c), the subse-
quent one-pot addition of MeMgBr and EtMgBr to
1c gave the corresponding monoprotected 1,4-adducts
5a and 5b with good SN2’-regioselectivity and com-
10
vinylMgBr
complex mixture
[a]
Conditions: 1c (1 equiv.), CuAHCTUNGTERG(NNUN OTf)2 (3 mol%), rac-
BINAP (6 mol%) in THF at room temperature for
15 min, then RMgBr (3.0 equiv.) 08C for 1 h.
Isolated yields of compound of type 5 after chromato-
graphic purification on silica gel.
plete
anti-stereoselectivity
(entries 1
and
2,
Table 2).[12] In order to verify the scope of the reac-
tion, we next examined the addition of other
Grignard reagents to 1c. Interestingly, the use of allyl-
magnesium bromide, which is rarely used in copper-
catalyzed allylic alkylations, gave the corresponding
allylated cyclopentene 5c very cleanly (entry 3). Also
the use of benzylmagnesium bromide was satisfactory
[b]
[c]
1
Determined by H NMR of the crude reaction mixture
(see Supporting Information for details).
(entry 4). More sterically demanding reagents, such as silica gel, whereas regioisomeric adducts of type 6,
i-PrMgCl, and the synthetic equivalent of the could not be obtained in a pure form. The trans-ste-
“CH2OH” anion [i.e., ClMgCH2SiMe
equally good yields of the corresponding adducts (en- by H NMR analysis and comparison with related 1,4-
tries 5 and 6).
substitued cyclopentene derivatives.[3,6] Evidently, an
2ACHTUNGETRN(NUNG i-PrO)], gave reochemistry of adducts of type 5 was demonstrated
1
Addition of aromatic Grignard reagents proceeded allylic carbazate of this kind is not able to direct the
in good yields (entries 7 and 8), and also a heteroaro- trajectory of the incoming organocopper nucleophile
matic Grignard reagent was successfully employed, by coordination, and the displacement follows the
albeit with a not complete conversion of 3c (entry 9). canonical anti-stereoselective pathway.[14] All ring-
On the other hand, the addition of vinylmagnesium opened adducts contained only one protecting group
bromide was inefficient and delivered a complex mix- on the nitrogen distal to the cyclopentene ring, indi-
ture of products with a modest conversion (ca. 30% cating that the decarboxylation of the 3-carbo-tert-bu-
after 1 h at 08C , entry 10). It is worth mentioning tyloxy-carbazic acid (A), which is formed in situ after
that all compounds of type 5 were easily isolated in the allylic displacement of the carbazate, occurs spon-
the pure state by chromatographic purification on taneously (Scheme 2).
Scheme 2. Plausible mechanism for the formation of monoprotected adducts.
Adv. Synth. Catal. 2009, 351, 869 – 873
ꢁ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
asc.wiley-vch.de
871

Stefano Crotti et al.
COMMUNICATIONS
3H), 1.49 (s, 9H), 1.59 (ddd, J1 =5.5 Hz, J2 =7.8 Hz, J3 =
13.5 Hz, 1H), 1.93 (ddd, J1 =3.3 Hz, J2 =7.8 Hz, J3 =11.1 Hz,
1H), 2.78–2.91 (m, 1H), 4.08–4.17 (m, 1H), 5.51–5.58 (m,
1H), 5.80–5.89 (m, 1H), 6.06–6.19 (m, 1H); 13C NMR
(62.5 MHz, CDCl3): d=21.0, 28.3, 37.4, 38.7, 66.7, 80.6,
129.3, 141.6, 156.5; anal. calcd. for C11H20N2O2: C 62.23%, H
9.50%; found: C 62.30%, H 9.48%.
It is known that the hydrazine moiety can be con-
verted by standard methods into valuable products,
such as pyrazole derivatives,[15] 1,2,4-triazolo deriva-
tives,[16] or into the corresponding amine by several
reduction methods.[5i,17] In conclusion, replacement of
the amide by urethane protecting groups, in particular
with PG=Boc, inhibited the classical [3,3]-sigmatrop-
ic Lewis acid-catalyzed rearrangement of 2,3-diaza-
1-(tert-Butoxycarbonyl)-2-[(1R*,4R*)-4-allylcyclo-
pent-2-en-1-yl]hydrazine (5c) (Table 2, entry 3)
ACHTUNGTRENNUNG
Using the typical procedure, CuACTHNUTRGENUN(G OTf)2 (2.65 mg,
(OTf)2/(Æ)-
0.006 mmol), rac-BINAP (7.50 mg, 0.012 mmol), hydrazine
1c (59.2 mg, 0.2 mmol) were used. The mixture was cooled
at 08C and then allylMgBr (1.0M in Et2O; 0.6 mL,
0.6 mmol) was added. The mixture was allowed to react for
1 hour at 08C (100% conversion). The product was isolated
by column chromatography eluting with hexanes/AcOEt
(8:2), as an oil; yield: 38 mg (80%). 1H NMR (250 MHz,
CDCl3): d=1.48 (s, 9H), 1.67 (ddd, J1 =5.5 Hz, J2 =7.8 Hz,
J3 =13.2 Hz, 1H), 1.88 (ddd, J1 =3.5 Hz, J2 =7.8 Hz, J3 =
11.5 Hz, 1H), 2.01–2.11 (m, 2H), 2.82–2.92 (m, 1H), 3.82–
4.01 (br s, 1H), 4.02–4.11 (m, 1H), 4.93–5.07 (m, 2H), 5.62–
5.80 (m, 2H), 5.87–5.91 (m, 1H), 6.02–6.13 (br s, 1H);
13C NMR (62.5 MHz, CDCl3): d=28.3, 34.8, 39.9, 43.9, 66.5,
80.4, 115.8, 130.4, 136.8, 139.9, 147.6, 156.6 (w); anal. calcd.
for C13H22N2O2: C 65.51%, H 9.30%; found: C 65.65%, H
9.25%.
Experimental Section
Typical Procedure: 1-(tert-Butoxycarbonyl)-2-{(1R*,
4R*)-4-[4-(ethoxycarbonyl)phenyl]cyclopent-2-en-1-
yl)}hydrazine (5h)
Supporting Information
Under argon atmosphere, a mixture of CuACTHNURTGNENG(U OTf)2 (2.65 mg,
Additional experimental procedures and spectral data are
available as Supporting Information.
0.006 mmol) and racemic BINAP (7.50 mg, 0.012 mmol) in
anhydrous THF (1.0 mL) was stirred for 10 min at room
temperature. Cyclic hydrazine 1c (59.2 mg, 0.2 mmol) in an-
hydrous THF (1.0 mL) was added, and the mixture was
stirred at room temperature for ca. 15 min (TLC detection).
The mixture was cooled at 08C, and then freshly prepared
ArMgBr (ca. 1.0M; 3 equiv, 0.6 mmol) was added. The mix-
ture was allowed to react for 1 hour at 08C (100% conver-
sion). The title compound was isolated by column chroma-
tography eluting with hexanes/AcOEt (8:2), as an oil; yield:
Acknowledgements
This work was supported by Ministero dellꢀ Istruzione, dellꢀ
Universitꢁ e della Ricerca (M.I.U.R. Rome, PRIN 2006, “Cat-
alysts, methodologies and new regio- and stereoselective pro-
cesses in organic synthesis”) and by the University of Pisa.
1
59 mg (85%). H NMR (250 MHz, CDCl3): d=1.31–1.40 (m,
3H), 1.48 (s, 9H), 1.87 (ddd, J1 =5.0 Hz, J2 =7.5 Hz, J3 =
12.5 Hz, 1H), 2.26 (ddd, J1 =3.3 Hz, J2 =11.5 Hz, J3 =
11.7 Hz, 1H), 4.02–4.10 (m, 2H), 4.30–4.39 (m, 2H), 5.89–
5.96 (m, 2H), 6.09–6.20 (br s, 1H), 7.12–7.19 (m, 2H); 7.90–
7.97 (m, 2H); 13C NMR (62.5 MHz, CDCl3): d=14.3, 28.3,
39.1, 50.2, 60.8, 66.9, 80.8, 127.1, 128.5, 124.8, 132.3, 138.1,
150.6, 156.8 (w), 166.5; anal. calcd. for C19H26N2O4: C
65.87%, H 7.56%; found: C 65.89%, H 7.48%.
References
[1] M. T. Crimmins, Tetrahedron 1998, 54, 9229–9272.
[2] S. M. Daluge, M. T. Martin, B. R. Sickles, D. A. Living-
stone, Nucleosides Nucleotides Nucleic Acids 2000, 19,
297–327.
[3] a) T. Ainai, Y.-G. Wang, Y. Tokoro, Y. Kobayashi, J.
Org. Chem. 2004, 69, 655–659; see also: b) Y. Kobaya-
shi, K. Nakata, T. Ainai, Org. Lett. 2005, 7, 183–186,
and references cited therein.
1-(tert-Butoxycarbonyl)-2-[(1R*,4R*)-4-methyl-
cyclopent-2-en-1-yl]hydrazine (5a) (Table 2, entry 1)
[4] A. Pꢂrez. Luna, M. Cesario, M. Bonin, L. Micouin,
Org. Lett. 2003, 5, 4771–4774.
Using the typical procedure, CuACTHNUTRGENUN(G OTf)2 (2.65 mg,
0.006 mmol), rac-BINAP (7.50 mg, 0.012 mmol), hydrazine
1c (59.2 mg, 0.2 mmol) were used. After cooling at 08C
MeMgBr (3.0M in Et2O; 0.2 mL, 0.6 mmol) was added. The
mixture was allowed to stir for 1 hour at 08C (100% conver-
sion). The product was isolated as an oil by column chroma-
tography eluting with hexanes/AcOEt (8:2); yield: 32 mg
[5] a) V. S. Sajisha, K. V. Radhakrishnan, Adv. Synth.
Catal. 2006, 348, 924–930; b) C. Bournard, D. Robic,
M. Bonin, L. Micouin, J. Org. Chem. 2005, 70, 3316–
3317; c) M. Pineschi, F. Del Moro, P. Crotti, F. Macchia,
Org. Lett. 2005, 7, 3605–3608; d) K. V. Radhakrishnan,
S. V. Sajisha, S. Anas, K. S. Krishnan, . Synlett 2005,
2273–2276; e) C. Bournaud, C. Falciola, C. Lecourt, S.
1
(75%). H NMR (250 MHz, CDCl3): d=1.13 (d, J=7.0 Hz,
872
asc.wiley-vch.de
ꢁ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Adv. Synth. Catal. 2009, 351, 869 – 873

Sequential Copper-Catalyzed Rearrangement–Allylic Substitution
COMMUNICATIONS
Rosset, A. Alexakis, L. Micouin, Org. Lett. 2006, 8,
3581–3584; f) J. John, V. S. Sajisha, S. Mohanlal, K. V.
Radhakrishnan, Chem. Commun. 2006, 3510–3512;
g) F. Bertolini, F. Macchia, M. Pineschi, Tetrahedron
Lett. 2006, 47, 9173–9176; h) F. Menard, C. F. Weise,
M. Lautens, Org. Lett. 2007, 9, 5365–5367; i) F.
Menard, M. Lautens, Angew. Chem. 2008, 120, 2115–
2118; Angew. Chem. Int. Ed. 2008, 47, 2085–2088; j) S.
Crotti, F. Bertolini, F. Macchia, M. Pineschi, Chem.
Commun. 2008, 3127–3129.
temperature. PPh3 and Josiphos-type ligands gave a
similar acceleration of the rearrangement. On the other
hand, Monophos and triethyl phosphite delayed the re-
arrangement.
[10] However, an alternative stepwise mechanism in which
the N-Boc protecting group acts as a nucleophile thus
trapping an allylic cation intermediate cannot be ruled
out. For a review on the reactivity of N-Boc protecting
group, see: C. Agami, F. Couty, Tetrahedron 2002, 58,
2701–2724.
[6] The reaction of acylnitroso Diels–Alder cycloadducts
with Grignard reagents catalyzed by copper salts gave
only minor amounts of trans-1,4-hydroxamic acid deriv-
atives: M. D. Surman, M. J. Mulvihill, M. J. Miller, J.
Org. Chem. 2002, 67, 4115–4121.
[7] a) C.-Y. Chung, D. Mackay, T. D. Sauer, Can. J. Chem.
1972, 50, 1568–1573; b) D. Mackay, C. W. Pilger, L. L.
Wong, J. Org. Chem. 1973, 38, 2043–2049; c) J. J. Tufar-
iello, J. J. Spadaro, Jr., Tetrahedron Lett. 1969, 3935–
3938.
[11] For reviews, see: a) B. Breit, P. Demel, in: Modern Or-
ganocopper Chemistry, (Ed.: N. Krause) Wiley-VCH,
Weinheim, 2002, pp 188–223; b) A. S. E. Karlstrom, J.-
E. Bꢃckvall, in: Modern Organocopper Chemistry,
(Eds.: N. Krause) Wiley-VCH, Weinheim, 2002,
pp. 259–288.
[12] The use of a reversed order of addition (i.e., the sub-
strate 1c was added to the in situ formed magnesium
cuprate) gave a completely different product distribu-
tion. In this case an inseparable mixture of regioisomer-
ic adducts bearing a double protection of the hydrazine
moiety was obtained.
[13] M. Matsuumi, M. Ito, Y. Kobayashi, Synlett 2002,
1508–1510.
[14] For copper directed allylic substitutions, see: B. Breit,
P. Demel, Adv. Synth. Catal. 2001, 343, 429–432, and
pertinent references cited therein.
[8] a) The presence of rotamers in the 1H and 13C NMR
analysis of p-Br-benzoyl- and p-nitrobenzoyl deriva-
tives of compound 3c (see compounds 7–9 in Support-
ing Information), and the formation of addition prod-
ucts of types 5 and 6 (not bearing the carbamate pro-
tection on the nitrogen linked to the cyclopentene
framework, see Table 2 and Scheme 2), indicate the
necessary occurrence of a 5,5-fused ring system. After
the preparation of this manuscript, X-ray crystallo-
graphic proof on related derivatives was independently
obtained by Mark Lautens and co-workers: b) A. Mar-
tins, S. Lemouzy, M. Lautens, Org. Lett. 2009, 11, 181–
183.
[15] M. Ge, E. Cline, L. Yang, Tetrahedron Lett. 2006, 47,
5797–5799.
[16] R. Kuang, A. K. Ganguly, T.-M. Chan, B. N. Pramanik,
D. J. Blythin, A. T. Maphail, A. K. Saksena, Tetrahe-
dron Lett. 2000, 41, 9575–9579.
[17] For example, see: D. Enders, R. Lochtman, Eur. J. Org.
[9] The reaction catalyzed by 3.0 mol% of Cu
ACHTUNGRTENUN(NG OTf)2, with-
Chem. 1998, 689–696.
out added phosphine, was complete in ca. 3 h at room
Adv. Synth. Catal. 2009, 351, 869 – 873
ꢁ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
asc.wiley-vch.de
873