The preparation of several 1,2,3,4,5-functionalized cyclopentane derivatives

Article abstract of DOI:10.3762/bjoc.9.195

With the goal of eventually synthesizing [5]radialene (3), the still missing member of the parent radialene hydrocarbons, we have prepared the pentaacetates 21 and 31, the pentabromide 29 and the hexabromide 32. In principle these should be convertible by elimination reactions to the desired target molecule.

Full text of DOI:10.3762/bjoc.9.195

The preparation of several 1,2,3,4,5-functionalized
cyclopentane derivatives
André S. Kelch1, Peter G. Jones2, Ina Dix1,3 and Henning Hopf*1
Full Research Paper
Open Access
Address:
Beilstein J. Org. Chem. 2013, 9, 1705–1712.
1Institute of Organic Chemistry, Technical University of
Braunschweig, Hagenring 30, D-38106 Braunschweig, Germany, Fax
+49-(0)531-391-5288, 2Institute of Inorganic and Analytical Chemistry,
Technical University of Braunschweig, Hagenring 30, D-38106
Braunschweig, Germany and 3Novartis Pharma AG, CH-4056 Basel,
Switzerland
doi:10.3762/bjoc.9.195
Received: 04 June 2013
Accepted: 26 July 2013
Published: 19 August 2013
Associate Editor: T. P. Yoon
Email:
Henning Hopf* - h.hopf@tu-bs.de
© 2013 Kelch et al; licensee Beilstein-Institut.
License and terms: see end of document.
* Corresponding author
Keywords:
bromination; cyclopentanes; esterification; permethylcyclopentadiene;
polyfunctional compounds; radialenes
Abstract
With the goal of eventually synthesizing [5]radialene (3), the still missing member of the parent radialene hydrocarbons, we have
prepared the pentaacetates 21 and 31, the pentabromide 29 and the hexabromide 32. In principle these should be convertible by
elimination reactions to the desired target molecule.
Introduction
Radialenes are cyclic cross-conjugated hydrocarbons that One reason for this absence is certainly the expected low
consist solely of semicyclic double bonds. Beginning with stability of 3: all its lower and higher analogs are highly reac-
[3]radialene (1), prepared first by Dorko in 1965 [1], this class tive hydrocarbons that under normal laboratory conditions poly-
of basic oligoenes has been the subject of constantly growing merize easily and/or readily undergo addition reactions with
attention by preparative and structural chemists, and – more other compounds – oxygen [2,3], dienophiles [2,3], carbenes
recently – by material scientists [2-5]. Although derivatives [6], and other trapping agents.
bearing alkyl and aryl and also functionalized substituents,
beginning with 1 and ending with [6]radialene (4, Scheme 1) Another reason is probably the practical consideration that there
are known for the whole series [2-5], among the parent hydro- are more and higher-yielding preparative routes to three-, four-
carbons one representative is still missing: [5]radialene (3).
and six-membered rings than to five-membered ones.
1705

Beilstein J. Org. Chem. 2013, 9, 1705–1712.
groups (see below), and describe the reaction of 8 with various
bromination reagents. Whenever possible we confirm the struc-
tures of intermediate compounds or byproducts by X-ray struc-
tural analysis.
Results and Discussion
cis,cis,cis,cis-1,2,3,4,5-Pentakis(hydroxy-
methyl)cyclopentane (16) and derivatives
The pentaalcohol 16 is a known compound, having first been
described by Tolbert and co-workers in 1985 in a short commu-
nication [7]. Although in a later contribution these authors
reported on the all-cis geometry (X-ray analysis) of a derivative
of this compound (the pentamesylate, see below) [8]. The
experimental and spectroscopic details of the original synthesis
were never fully published [9]. This missing information will be
presented in this publication (see Supporting Information
File 1). The Tolbert synthesis of 16 is summarized in Scheme 2.
Scheme 1: The first members of the [n]radialene series and retrosyn-
thesis for [5]radialene (3).
The crucial intermediate 5-[(benzyloxy)methyl]cyclopenta-
diene (12) was first prepared from benzyl alcohol (9) via the
chloro ether 10 and cyclopentadienylthallium (11) according
Looking at possible precursors for 3, it seems reasonable to start to a route developed by Corey and co-workers in their
with a substrate that contains at least a complete five-membered prostaglandin work [10]. Cycloaddition of maleic anhydride
carbocycle or – even better – a starting material that already (MA) to 12 subsequently afforded the endo-adduct 13, which
displays all ten carbon atoms required in the product. For the after esterification provided the diester 14 in excellent yield [7].
first protocol, derivatives such as 6 are possible candidates, not This bicyclic adduct was converted by ozonolysis followed by
least for the simple reason that in all other reported syntheses of LAH reduction into the all-cis-derivative 15, whose benzyl
the radialene hydrocarbons [2,3] final steps are involved that ether protection group was finally removed by catalytic hydro-
employ β-elimination processes – be it of halides, aminoxide genation over palladium on carbon [7,8]. The full experimental
derivatives (Cope elimination), acetates (ester pyrolysis), etc. details and the complete set of the spectroscopic data of 16 are
[2,3].
collected in Supporting Information File 1.
When starting with derivatives of type 6 it should be noted that The first successful attempt to convert the pentaalcohol into a
these compounds can exist in the form of different diastereo- derivative bearing better leaving groups is Tolbert´s prepar-
mers; for example – as shown in 5 – all five substituents could ation of the pentamesylate 18 (Scheme 3) [8].
point in the same direction (cis,cis,cis,cis-configuration). Alter-
natively, one (or more) of these functionalized groups could be All our attempts to prepare the all-cis-pentabromide 17 from 16
directed in opposite directions (e.g. cis,cis,cis,trans-configur- failed so far, as is also shown in Scheme 3. Only in one case
ation, as in 7). Whether this stereochemical difference will have was a halogenated derivative obtained in poor yield: treatment
an influence on the ultimate outcome of the multiple elimina- of 16 with triphenylphosphine–bromine in carbon tetrachloride
tion process is an open question; but it is likely to influence the provided the tribromo ether derivative 19 as a minor compo-
rate of this multi-step transformation, which will involve nent. Surprisingly, this had the cis,cis,cis,trans-configuration of
increasingly reactive (unstable) intermediates.
the substituents at the five-membered ring, as was shown by
single crystal X-ray analysis (Figure 1).
As far as substrates are concerned that possess the complete set
of necessary carbon atoms, the commercially available It is unclear at the present time whether the correspondingly
pentamethylcyclo-1,3-pentadiene (8) offers itself as the best configurated pentaalcohol was already present in the starting
starting material.
material 16 as a side-product, which had been overlooked in
previous studies, or whether epimerization took place during the
In the present publication we describe the synthesis of the two attempted conversion of the pentaalcohol to the pentabromide
diastereomeric derivatives 5 and 7 bearing various functional 17. The complete, structure-confirming spectroscopic and
1706

Beilstein J. Org. Chem. 2013, 9, 1705–1712.
Scheme 2: Preparation of cis,cis,cis,cis-1,2,3,4,5-pentakis(hydroxymethyl)cyclopentane (16) according to Tolbert [7].
Scheme 3: The preparation of derivatives of 16 better suited for nucleophilic substitution and elimination.
analytical data of 19 can be found in Supporting Information these experiments complex product mixtures resulted, the
File 1.
GC–MS analysis of which demonstrated that only partially
halogenated derivatives of 16 had been produced; their
In closing this section on the preparation of bromine deriva- chromatographic separation failed.
tives of 16, it should also be mentioned that attempts to prepare
the corresponding pentachloride (by treatment with either In order to avoid working in solution and subjecting the
SOCl2 or PCl3) and pentaiodide (red P/iodine) also failed. In all presumably reactive 5 to extended work-up and purification
1707

Beilstein J. Org. Chem. 2013, 9, 1705–1712.
The first route dates back to studies by Diels, Alder and
co-workers [11,12], carried out in the early 1930s and 40s,
which were taken up later by Le Goff and his students [13].
These pioneers demonstrated that, on treatment of dimethyl
acetylenedicarboxylate (22) with dimethyl malonate (23) under
the catalytic influence of pyridinium acetate, the fully substi-
tuted cycloheptadienes 24 and 25 (Scheme 5) were produced;
both authors also proposed reaction mechanisms for this com-
plex transformation that can be considered as reasonable even
today; for the details the reader is referred to the original litera-
ture [12,13].
The reproduction of this protocol posed no problems (see
Supporting Information File 1). The only difference we noted
between our results and those reported in the chemical litera-
ture concerns the ratio of 24 to 25. Whereas Diels reported a
ratio of 3:1, and in Le Goff´s hands it was as much as 15:1, we
observed the formation of 24 as a minor product (24:25 = 1:8).
Figure 1: Structure of 19 in the crystal; ellipsoids represent 50% prob-
ability levels.
conditions, we next tried to prepare the xanthate 20 and the One reason for this difference could be a pericyclic 1,5-
acetate 21 from 16, derivatives that could be subjected to flash hydrogen shift between the two valence isomers generating
vacuum pyrolysis (Scheme 4).
different ratios under different preparation and work-up condi-
tions. Another equilibration mechanism could involve a base-
Whereas all our attempts to synthesize 20 failed (see Supporting induced 1,5-hydrogen shift. Since the mechanism of formation
Information File 1), the pentaester 21 was obtained under stan- of 24 and 25 was not the main concern of our investigations, we
dard conditions from 16 by treatment with excess acetyl chlo- did not clarify these points. The differentiation between the
ride in excellent yield; its spectroscopic data can be found in the symmetrical 24 and the non-symmetrical 25 by 1H NMR spec-
Supporting Information File 1.
troscopy is easy and was achieved by Le Goff. More complete
and up-to-date spectroscopic data of the two compounds are
summarized in Supporting Information File 1. Still, the early
workers [11-13] did not distinguish between a possible meso-
structure (cis-orientation of allylic hydrogen atoms) and the
chiral (C2-symmetrical) configuration (trans-orientation of
hydrogen atoms) of 24. Fortunately, recrystallization of a 24/25-
cis,cis,cis,trans-1,2,3,4,5- (26),
cis,cis,trans,cis-1,2,3,4,5-Pentakis(methoxy-
carbonyl)cyclopentane (27) and several of
their derivatives
For the other two routes to precursors of [5]radialene (3) mixture from diethyl ether provided single crystals of 24 suit-
discussed here, the ten carbon atoms of the desired final prod- able for an X-ray structure investigation which proved the chiral
uct were available already at a very early stage of the synthesis. configuration (Figure 2).
Scheme 4: Preparation of the pentaacetate 21 from 16.
1708

Beilstein J. Org. Chem. 2013, 9, 1705–1712.
Scheme 5: Preparation of the cycloheptadiene octaesters 24/25 according to Diels [11] and Le Goff [13], respectively, followed by ring contraction to
the cyclopentadiene pentaester 28.
(98%). Although we were unable to separate the two compo-
nents quantitatively by column chromatography on silica gel,
the main isomer began to crystallize partially from the highly
viscous hydrogenation mixture after standing for several
months. The crystals obtained were suitable for X-ray struc-
tural analysis, which showed that this isomer has the
cis,cis,cis,trans-geometry, i.e. 26 (Figure 3).
Figure 2: Structure of 24 in the crystal; ellipsoids represent 30% prob-
ability levels.
The ring contraction of a mixture of 24/25, following the proce-
dure given by Le Goff [13], proceeded as expected and
provided the potassium salt of 28 in acceptable yields (60%).
From this, free 1,2,3,4,5-pentakis(methoxycarbonyl)cyclopenta-
Figure 3: Structure of 26 in the crystal; ellipsoids represent 30% prob-
ability levels.
1,3-diene 28 was liberated by hydrochloric acid treatment in
excellent yield (>90%, Scheme 5).
Finally, the hydrogenation of 28 over Pd/C in methanol The structural assignment of the minor hydrogenation product
furnished a mixture of the saturated pentaesters 26 and 27 (ratio as 27, which could only be obtained in enriched form (ratio
4:1, capillary GC–MS-analysis) in practically quantitative yield 26:27 ca. 1:1) is based on the following considerations. Alto-
1709

Beilstein J. Org. Chem. 2013, 9, 1705–1712.
gether there exist three hydrogenation products of 28: the two 1.63–2.52 in 2:2:1 ratio for the ring protons. In the 13C NMR
shown (26, 27) and also the all-cis-configurated pentaester, spectrum the ring methine carbon atoms absorb at 45.5, 46.0,
presumably the most highly strained of the three isomers. For and 46.7 ppm, whereas the primary carbon atoms are registered
symmetry reasons this compound should display two sharp at δ = 60.0, 63.2, and 65.5 ppm. The signals of the minor
signals in its 1H NMR spectrum: one for the tertiary hydrogen isomer(s) can hardly be differentiated from the noise.
atom at the ring and one for the methyl group of its five iden-
tical ester groups (ratio 1:3). It should not be difficult to iden- The bromination of 30 proceeded with varying results; most of
tify these signals in the proton spectrum of a mixture of hydro- the methods that were unsuccessfully used above in the case of
genation products because of their relative intensities. Although 16 also fail for its isomer (HBr, PBr3, SOBr2 in different
a large number of signals can indeed be recognized in the halogenated solvents). However, with the triphenylphos-
corresponding chemical shift regions, none displays an extraor- phine–bromine complex, a mixture of the pentabromides 29
dinary intensity. It is hence unlikely that the all-cis-pentaester is was obtained in very good yield (73%). The superior results in
produced in significant amounts on hydrogenation.
comparison with 16 probably can be rationalized by a decreased
steric shielding of the hydroxymethyl groups in 30, but are
As far as derivatives of 26 and 27 are concerned, we always probably also associated with the different solubility of the
worked with the mixture of the two diastereomers (Scheme 6). pentaalcohol mixture. Likewise, the pentaacetate 31 can be
obtained from 30 in excellent yield (Scheme 5). As in the case
Reduction of 26/27 with lithium aluminum hydride in either of 16, no xanthate (see above) could be prepared from 30.
diethyl ether or THF provided the diastereomeric pentaalcohols
Bromination of 1,2,3,4,5-pentamethylcyclo-
30. The yield of the process is probably higher than the finally
isolated 50%. Both isomers are readily soluble in water, and to penta-1,3-diene (8)
obtain the pure mixture the crude reduction product was first The other substrate that has all the carbon atoms of the final
extracted with diethyl ether or THF. This leaves the inorganic [5]radialene (3) already in place is 1,2,3,4,5-pentamethylcyclo-
salts and the pentaalcohols as a residual solid. When this is penta-1,3-diene (8). Formally this hydrocarbon merely has to be
extracted with refluxing methanol or (preferably) isopropanol dehydrogenated three times to provide the target molecule. To
the desired reduction product 30 slowly dissolves, leaving the accomplish this goal we first had to replace these hydrogen
salts behind. The removal of the solvent from 30 is difficult and atoms by a better leaving group, e.g. bromine as in Scheme 7.
requires extended pumping under high vacuum (2 weeks). Even
then the residue shows traces of the originally added solvent (by After trying numerous different protocols (Br2 addition, NBS in
NMR analysis).
different concentrations and in different solvents, 1,3-dibromo-
5,5-dimethylhydantoin) [14] we finally found that excess NBS
The exact structure assignment of 30 is hampered by the fact (6 equivalents) and refluxing in tetrachloromethane at 80 °C
that no pure isomer was available. We assume, however, that gave the best results. Under these conditions the hexabromide
the isomeric composition is not changed by the LAH reduction 32 was produced in 45% yield. The structure of this cyclopenta-
(i.e. 4:1 as in the case of the esters 26/27). The signals for the diene derivative was derived from its spectroscopic data (see
cis,cis,cis,trans-pentaalcohol are clearly discernible. The Supporting Information File 1) and also by an X-ray structural
1H NMR spectrum displays three signal groups from δ = analysis (Figure 4).
Scheme 6: Derivatives derived from the pentaester mixture 26/27.
1710

Beilstein J. Org. Chem. 2013, 9, 1705–1712.
which the final ring hydrogen atom has been substituted, but
also to a mixture of hexabromides among which the hexabro-
mide 32 is the major reaction product.
With several derivatives (18 [7,8], 21, 29, 31, and 32) that could
lead to [5]radialene (3) in hand, a synthesis of this long missing
hydrocarbon is now within sight.
Supporting Information
Supporting Information File 1
Experimental procedures and characterization data.
[http://www.beilstein-journals.org/bjoc/content/
supplementary/1860-5397-9-195-S1.pdf]
Acknowledgments
We thank Professor Laren M. Tolbert (Georgia Institute of
Technology) for helpful discussions. This publication is based
on the Ph. D. thesis of André Stefan Kelch, Braunschweig,
2002.
Scheme 7: Bromination of 1,2,3,4,5-pentamethylcyclopenta-1,3-diene
(8).
References
1. Dorko, E. A. J. Am. Chem. Soc. 1965, 87, 5518–5520.
doi:10.1021/ja00951a067
2. Hopf, H.; Maas, G. Angew. Chem. 1992, 104, 953–977.
doi:10.1002/ange.19921040804
Angew. Chem. Int., Ed. Engl. 1992, 31, 931–954.
doi:10.1002/anie.199209313. The first comprehensive review on
[n]radialenes and their derivatives.
3. Maas, G.; Hopf, H. In The Chemistry of Functional Groups; Patai, S.;
Rappoport, Z., Eds.; The Chemistry of Dienes and Polyenes, Vol. 1; J.
Wiley & Sons: Chichester, 1997; pp 927 ff.
4. Gholami, M.; Tykwinski, R. R. Chem. Rev. 2006, 106, 4997–5027.
doi:10.1021/cr0505573
5. Hopf, H. Angew. Chem. 2012, 124, 12111–12113.
doi:10.1002/ange.201206101
Angew. Chem., Int. Ed. 2012, 51, 11945–11947.
doi:10.1002/anie.201206101. The number of known compounds
possessing a radialene-type distribution of their π-electrons increases
quickly once heteroatoms (O, S,-NR etc.) are introduced into the
carbon skeleton; for a short review on recent developments in
heteroradialene chemistry.
6. Trabert, L.; Hopf, H.; Schomburg, D. Chem. Ber. 1981, 114,
2405–2414. doi:10.1002/cber.19811140707
Figure 4: Structure of 32 in the crystal; ellipsoids represent 50% prob-
ability levels.
7. Tolbert, L. M.; Gregory, J. C.; Brock, C. P. J. Org. Chem. 1985, 50,
548. doi:10.1021/jo00204a030
8. Brock, C. P.; Gregory, J. C.; Tolbert, L. M. Acta Crystallogr., Sect. C
1986, 42, 1063–1066. doi:10.1107/S0108270186093460
9. L. M. Tolbert (Georgia Institute of Technology), private communication
to H. Hopf.
A rationalization of the route from 8 to 32 is also summarized in
Scheme 7. Since in all our bromination experiments we isolated
small amounts of a hydrocarbon that, according to mass spec-
tral analysis, has the formula C20H30, i.e. is a dimer of the
starting material, we assume that the permethylcyclopentadi-
enyl radical 33 plays an important role in the bromination reac-
tion. This could not only lead to the bromo derivative 34 in
10.Corey, E. J.; Koelliker, U.; Neuffer, J. J. Am. Chem. Soc. 1971, 93,
1489–1490. doi:10.1021/ja00735a031
11.Diels, O.; Alder, K. Justus Liebigs Ann. Chem. 1932, 498, 16–49.
doi:10.1002/jlac.19324980103
1711

Beilstein J. Org. Chem. 2013, 9, 1705–1712.
12.Diels, O. Ber. Dtsch. Chem. Ges. 1942, 75, 1452–1467.
doi:10.1002/cber.19420751206
13.Le Goff, E.; LaCount, R. B. J. Org. Chem. 1964, 29, 423–427.
doi:10.1021/jo01025a043
14.Oakes, V.; Rydon, H. N.; Undheim, K. J. Chem. Soc. 1962,
4678–4685. doi:10.1039/jr9620004678
License and Terms
This is an Open Access article under the terms of the
Creative Commons Attribution License
(http://creativecommons.org/licenses/by/2.0), which
permits unrestricted use, distribution, and reproduction in
any medium, provided the original work is properly cited.
The license is subject to the Beilstein Journal of Organic
Chemistry terms and conditions:
(http://www.beilstein-journals.org/bjoc)
The definitive version of this article is the electronic one
which can be found at:
doi:10.3762/bjoc.9.195
1712