Addressing the regioselectivity problem in organic synthesis

Article abstract of DOI:10.1039/b604129a

A screening process uncovered a heterogeneous catalytic system that hydrolyzes one of two nearly identical ketals in several diketals with a high selectivity. The Royal Society of Chemistry 2006.

Full text of DOI:10.1039/b604129a

View Article Online / Journal Homepage / Table of Contents for this issue
COMMUNICATION
www.rsc.org/chemcomm | ChemComm
Addressing the regioselectivity problem in organic synthesis
Fredric M. Menger* and Hao Lu
Received (in Bloomington, IN, USA) 21st March 2006, Accepted 11th June 2006
First published as an Advance Article on the web 5th July 2006
DOI: 10.1039/b604129a
to the tertiary amine of RCH(NHR)(NR2).8 In a stannylene
A screening process uncovered a heterogeneous catalytic system
that hydrolyzes one of two nearly identical ketals in several
diketals with a high selectivity.
derivative of a diol, one oxygen is acylated faster than the other.9
However, our definition of ‘‘enzyme-like regioselectivity’’ does not
encompass such cases because they arise mainly from inherent
reactivity and steric differences. Examples of true enzyme-like
regioselectivity are rather uncommon and would include the
epoxidation of squalene’s terminal double bond in preference to its
other identically substituted double bonds.10
Organic synthesis is winning the war against its challenges. The
reactions available to the synthetic organic chemist number in the
countless thousands. Natural products with over sixty chiral
centers have been constructed.1 Chiral catalysts and auxiliaries
giving enantiomeric excesses greater than 90% abound.2 We hear
occasionally that, ‘‘If you can draw it, you can make it.’’ Although
this statement is no doubt hyperbole, it does serve to drive home
the impressive successes in the field of organic synthesis.
Nevertheless, there remains a major problem in organic synthesis
that has not been solved. In fact, it has only rarely been addressed:
enzyme-like regioselectivity.
How do enzymes accomplish their remarkable regioselectivity?
A key factor lies in the binding of substrates to the enzyme prior to
reaction. Binding processes place only one of two nearly identical
substrate functionalities at the enzyme’s catalytic site, thereby
providing the necessary selectivity. Those wishing to emulate this
action must, therefore, construct a catalytic surface with specific
binding properties. One might conceivably do so by designing
catalytic host molecules that can discriminate between nearly
identical reactive groups in a guest (similar to enzyme action).11–15
But chemistry has not yet acquired an understanding of
intermolecular forces sufficient to predict the precise geometry
of most substrate/surface interactions. Consequently, synthesis of
such a host is a costly and chancy enterprise. We have, therefore,
adopted a screening program similar to that used in the discovery
of many enantioselective catalysts. A main difference is that we
searched for selectivity between two distant, non-interacting
functionalities rather than selectivity at a single locus.
We define ‘‘enzyme-like regioselectivity’’ as the ability to
distinguish between two functionalities in a molecule that (a) are
positionally distinct but otherwise virtually identical in their steric/
electronic environments; (b) are often located in quite different
parts of the molecule; and (c) are not affected by an intramolecular
catalysis that favors one of the functionalities. Chemists confronted
with two such groups are often unable to operate selectively upon
only one of them. For example, 3,6-diketodecane has never, to our
knowledge, been selectively mono-reduced. Enzymes, on the other
hand, are well known to react exclusively on the basis of location
alone.3 For example, stearyl CoA is regiospecifically dehydroge-
nated at the 9-position by the enzyme acyl CoA desaturase
(eqn (1))4 while the initial step in the conversion of cholesterol to
pregnenalone involves a regiospecific oxidation at the tertiary C-20
carbon, with the nearby tertiary C-25 carbon remaining untouched
(eqn (2)).5 No chemist nowadays could duplicate these feats non-
enzymatically, and hence the challenge.6
The reaction in question involved initially the selective
hydrolysis of only one of the two ketal groups in compound 1
of Scheme 1. Sterically and electronically the two ketal groups
are nearly identical because both of them are positioned at
secondary carbons in a saturated chain. Although the 2-ketal
might be considered more ‘‘exposed’’ than the 7-ketal, an
inherent reactivity difference (if such exists) was insufficient to
achieve useful selectivity with several classical homogeneous and
heterogeneous catalysts that were tested via literature procedures:
pyridinium p-toluenesulfonate (methanol, acetone, acetonitrile
or hexane);16 CeCl3?7H2O, NaI (methanol or acetonitrile);17
ð1Þ
ð2Þ
Regioselectivity, like diastereo- and enantioselectivity, is of
course commonplace in organic chemistry. For example, toluene is
nitrated faster at the para position than at the ortho and meta
positions. An aldehyde can be reduced in the presence of an
unhindered ketone.7 One can regioselectively cleave the C–N bond
Department of Chemistry, Emory University, 1515 Dickey Dr. Atlanta,
GA, USA 30322. E-mail: menger@emory.edu; Fax: +1 404 727 6586;
Tel: +1 404 727 6599
Scheme 1 Diketals examined in this work.
This journal is ß The Royal Society of Chemistry 2006
Chem. Commun., 2006, 3235–3237 | 3235

View Article Online
poly(4-vinylpyridinium p-toluenesulfonate) (methanol); oxalic acid
(methanol);18 TiCl4 (diethyl ether);19 HCl (acetone);20 AcOH
(acetone);21 PdCl2 (acetone);22 AlI3 (acetonitrile);23 Ph3CBF4
(dichloromethane);24 and Er(OTf)3 (acetonitrile).25 During our
screening, however, we did find one useful heterogeneous catalyst
that greatly favored deprotection of one ketone over the other:
MgSO4 in wet hydrocarbon solvents.26 A remarkable regioselective
preference for hydrolysis of the 2-ketal was evident as will be
described following the relevant experimental details below.
Diketal 1 itself was synthesized by adding n-C10H21MgBr to
2-methylcyclohexanone and dehydrating the resulting tertiary
alcohol to give 1-decyl-2-methylcyclohexene (Scheme 2).
Ozonolysis of the olefin followed by a dimethyl sulfide workup
gave 2,7,-diketoheptadecane from which the diketal was readily
prepared and identified spectroscopically (NMR and MS) and by
elemental analysis. Analogs of 1, mentioned below, were
synthesized in a like manner. Diketal 1 was ‘‘mono-hydrolyzed’’
as follows: in a 4 mL glass vial was added 2 mg of 1 in 2 mL wet
hexanes (prepared by repeatedly shaking hexanes with purified
water), 200 mg magnesium sulfate powder (EMD Chemicals, Inc.,
freshly opened), and a micro magnetic stirring bar. After stirring
for a measured time at room temperature, 100 mL were withdrawn,
filtered through a 0.2 mm-pore membrane, and injected into an
HPLC (C-18 reverse-phase column; methanol–water (85%/15%)
eluent; evaporative light scattering detector). Identifying the two
monoketone peaks HPLC peaks required the synthesis of the
7-monoketone compound (Scheme 3): 6-chloro-2-hexaone was
converted into the iodide derivative with NaI/acetone and then
treated with ethylene glycol/TSA in benzene. The resulting ketal
was reacted with Mg in THF to make the Grignard that was
added to undecylic aldehyde to give the 2-ketal-7-hydroxyhepta-
decane. A Dess–Martin oxidation of the secondary alcohol yielded
the desired 7-monoketone that has an HPLC retention time
Fig. 1 HPLC of the reaction between diketal 1 and MgSO4 in wet
hexanes at room temperature at 24 h. The peaks (in order of appearance)
are diketone, 7-monoketone, 2-monoketone and diketal.
identical to that of the earlier-emerging peak of the two
monoketone product peaks from the MgSO4-catalyzed hydrolysis.
A reaction of 1 with MgSO4 in wet hexanes after 24 h gave the
HPLC shown in Fig. 1. The mixture contained (in increasing order
of retention time) 21% diketone; 3% 7-monoketone; 58%
2-monoketone; and 18% diketal. Accordingly, the preference for
2-ketal hydrolysis over 7-ketal hydrolysis is 19 : 1. A repeat gave
70% 2-monoketone (typical of the experimental uncertainty).
These numbers were affirmed by two different experimenters. The
selective hydrolyses proceeded as well in wet benzene as in wet
hexanes, but wet ethyl acetate, chloroform and acetonitrile gave no
reaction. Nor was a reaction observed with a smaller amount
(50 mg) of MgSO4 or MgCl2 in wet benzene.
With these results in hand, the question came up, of course, as
to the regioselectivity with structural analogs of 1 and, as such, we
examined five additional diketals (Scheme 1). The HPLC data for
these MgSO4-catalyzed hydrolyses are given in Table 1. It can be
seen from the sixth column that very high regioselectivities
(ranging from 23 : 1 to .95 : 1) are also possible with these
analogs. The reasons for these regioselectivities are at the moment
obscure, but they seem almost certainly related to selective binding
of the diketals to the MgSO4 surface.
A key observation helped clarify the catalytic mechanism: The
ketals of 2- and 7-heptadecanone were hydrolyzed only very slowly
under our conditions (0–3% product after 24 h). We surmise that
effective binding to the MgSO4 interface requires chelation to two
ketals within the same molecule, after which only one is
hydrolyzed. Thereupon, the reactant dissociates from the MgSO4
surface, leading to the production of mainly mono-hydrolyzed
Scheme 2 Synthesis of diketal 1.
Table 1 HPLC-Determined percentages of compounds in the depro-
tection reaction with MgSO4 in wet hexanes after 24 h at room
temperature
2-Monoketone/
(6, 7 or 8)-
(6, 7 or 8)- 2-Mono-
Compound Diketone monoketone ketone Diketal monoketone
1
2
3
4
5
6
a
21
14
0
10
47
0
3
1
0
1
2
0
58
18
25
82
43
6
19 : 1
60 : 1
.95 : 1
46 : 1
23 : 1
60
18a
46
45
20a
80
.95 : 1
Little change was observed in another 24 h.
Scheme 3 Synthesis of the 7-monoketone corresponding to 1.
3236 | Chem. Commun., 2006, 3235–3237
This journal is ß The Royal Society of Chemistry 2006

View Article Online
5 L. Stryer, Biochemistry, W. H. Freeman, New York, 4th edn, 1995,
p. 705.
6 For simulations of such chemistry, see: R. Breslow, Acc. Chem. Res.,
1980, 13, 170.
7 G. H. Posner, A. W. Runquist and M. J. Chapdelaine, J. Org. Chem.,
1977, 42, 1202.
8 H. Yamamoto and K. Maruoka, J. Am. Chem. Soc., 1981, 103, 4186.
9 S. David and S. Hanessian, Tetrahedron, 1985, 41, 643.
10 E. E. van Tamelen, Acc. Chem. Res., 1968, 1, 111.
11 S. E. Bystro¨m, A. Bo¨rje and B. Akermark, J. Am. Chem. Soc., 1993,
115, 2081.
12 R. Breslow and P. Campbell, Bioorg. Chem., 1971, 1, 140.
13 D. A. Jaeger, M. D. Ward and C. A. Martin, Tetrahedron, 1984, 40,
2691.
14 C. J. Suckling, Ind. Eng. Chem. Prod. Res. Dev., 1981, 20, 434.
15 M. Yoshizawa, M. Tamura and M. Fujita, Science, 2006, 312, 251.
16 H. Hagiwara and H. Uda, J. Chem. Soc., Chem. Commun., 1987, 1351;
R. Sterozycki, Synthesis, 1979, 724.
17 E. Marcantoni, F. Nobili, G. Bartoli, M. Bosco and L. Sambri, J. Org.
Chem., 1997, 62, 4183.
product. It is suggested, therefore, that regioselectivity is predicated
upon tight bidentate binding of the reactant followed by loose or
non-existent monodentate binding of the product. The fact that
bis-hydrolysis product was observed (Table 1) implies that it was
formed prior to dissociation of the mono-product from the catalyst
surface.
An ability to operate on only one of two nearly identical,27
distant functional groups constitutes a synthetically important
technology. Our results in this field are encouraging but only an
initial confrontation with the problem. Thus, the presence of
double-hydrolyses limits the overall yield of the major mono-
ketone. Yet our data constitute a feasibility study that should
encourage further research into ‘‘surface-imposed regioselectivity’’.
It is to be hoped that with the appropriately designed catalysts,
molecular position, rather than classical stereoelectronic factors,
will someday play a bigger role in dictating the outcome of
synthetic reactions.
18 J. H. Babler, N. C. Malek and M. J. Coghlan, J. Org. Chem., 1978, 43,
1821.
This work was supported by the Petroleum Research
Foundation. We thank Dr Hailing Zhang for her cross-checking
a run.
19 G. Balme and J. Gore´, J. Org. Chem., 1983, 48, 3336.
20 P. A. Grieco, Y. Yokoyama, G. P. Withers, F. J. Okuniewicz and
C.-L. J. Wang, J. Org. Chem., 1978, 43, 4178.
21 P. A. Grieco, Y. Ohfune and G. Majetich, J. Am. Chem. Soc., 1977, 99,
7393.
22 B. H. Lipshutz, D. Pollart, J. Monteforte and H. Kotsuki, Tetrahedron
Lett., 1985, 26, 705.
Notes and references
23 P. Sarmah and N. C. Barua, Tetrahedron Lett., 1989, 30, 4703.
24 D. H. R. Barton, P. D. Magnus, G. Smith and D. Zurr, Chem.
Commun., 1971, 861.
25 R. Dalpozzo, A. De Nino, L. Maiuolo, M. Nardi, A. Procopsio and
A. Tagarelli, Synthesis, 2004, 496.
1 E. M. Suh and Y. Kishi, J. Am. Chem. Soc., 1994, 116, 11205, and ref. 3
therein.
2 (a) Acc. Chem. Res., 2004, 37, 487 (issue devoted to asymmetric
organocatalysis); (b) Chiral Reagents for Asymmetric Synthesis, ed. L. A.
Paquette, Wiley, Chichester, 2003 (see list of review articles and
monographs); (c) K. Ding, H. Du, Y. Yuan and J. Long, Chem. Eur. J.,
2004, 10, 2872 (see list in ref. 2b).
26 J. J. Brown, R. H. Lenhard and S. Bernstein, J. Am. Chem. Soc., 1964,
86, 2183.
3 A. C. Eliot and J. F. Kirsch, Acc. Chem. Res., 2003, 36, 757.
4 R. H. Abeles, P. A. Frey and W. P. Jencks, Biochemistry; Jones and
Bartlett, Boston, MA, 1992, p. 658.
27 Transforming only one of two identical groups is in fact an easier
problem than transforming only one of two nearly identical groups
because in the former case reaction at either site can give specificity.
This journal is ß The Royal Society of Chemistry 2006
Chem. Commun., 2006, 3235–3237 | 3237