Albumin-directed stereoselective reduction of 1,3-diketones and β-hydroxyketones to anti diols

Article abstract of DOI:10.1039/c0ob00648c

The reduction of 1,3-diketones and β-hydroxyketones with NaBH ₄ in aqueous acetonitrile is highly stereoselective in the presence of stoichiometric amounts of bovine or human albumin, giving anti 1,3-diols with d.e. up to 96%. The same reaction, without albumin, gives syn and anti 1,3-diols in approximately 1:1 ratio. The presence of an aromatic carbonyl group is essential for diastereoselectivity in the NaBH₄/albumin reduction of both 1,3-diketones and β-hydroxyketones. Thus, 3-hydroxy-1-(p-tolyl)-1- butanone is stereoselectively reduced in the presence of albumin, while reduction of its isomer 4-(p-tolyl)-4-hydroxy-2-butanone is not stereoselective. The albumin-controlled reduction is not stereospecific as both enantiomers of 1-aryl-3-hydroxy-1-butanones are reduced to diols with identical stereoselectivities. Circular dichroism of the bound substrates confirms that aromatic ketones are recognized by the protein's IIA binding site. Binding studies also suggest that 1,3-diketones are recognized in their enol form. From the effect of pH on binding of a diketone it is concluded that, in the complex with the substrate, ionizable residues His242 and Lys199 are in the neutral and protonated forms, respectively. A homology model of BSA was obtained and docking of model substrates confirms the preference of the protein for aromatic ketones. Modelling of the complexes with the substrates also allows us to propose a mechanism for the reduction of 1,3-diketones in which the chemoselective reduction of the first (aliphatic) carbonyl is followed by the diastereoselective reduction of the second (aromatic) carbonyl. The role of albumin is thus a combination of chemo- and stereocontrol.

Full text of DOI:10.1039/c0ob00648c

View Online / Journal Homepage / Table of Contents for this issue
Organic &
Biomolecular
Chemistry
Dynamic Article Links
Cite this: Org. Biomol. Chem., 2011, 9, 1987
www.rsc.org/obc
PAPER
Albumin-directed stereoselective reduction of 1,3-diketones and
b-hydroxyketones to anti diols†‡
Federico Berti,*a Simone Bincoletto,a Ivan Donati,b Giampaolo Fontanive,a Massimo Fregonesea and
Fabio Benedetti*a
Received 30th August 2010, Accepted 23rd November 2010
DOI: 10.1039/c0ob00648c
The reduction of 1,3-diketones and b-hydroxyketones with NaBH4 in aqueous acetonitrile is highly
stereoselective in the presence of stoichiometric amounts of bovine or human albumin, giving anti
1,3-diols with d.e. up to 96%. The same reaction, without albumin, gives syn and anti 1,3-diols in
approximately 1 : 1 ratio. The presence of an aromatic carbonyl group is essential for diastereoselectivity
in the NaBH4/albumin reduction of both 1,3-diketones and b-hydroxyketones. Thus,
3-hydroxy-1-(p-tolyl)-1-butanone is stereoselectively reduced in the presence of albumin, while
reduction of its isomer 4-(p-tolyl)-4-hydroxy-2-butanone is not stereoselective. The albumin-controlled
reduction is not stereospecific as both enantiomers of 1-aryl-3-hydroxy-1-butanones are reduced to diols
with identical stereoselectivities. Circular dichroism of the bound substrates confirms that aromatic
ketones are recognized by the protein’s IIA binding site. Binding studies also suggest that 1,3-diketones
are recognized in their enol form. From the effect of pH on binding of a diketone it is concluded that, in
the complex with the substrate, ionizable residues His242 and Lys199 are in the neutral and protonated
forms, respectively. A homology model of BSA was obtained and docking of model substrates confirms
the preference of the protein for aromatic ketones. Modelling of the complexes with the substrates also
allows us to propose a mechanism for the reduction of 1,3-diketones in which the chemoselective
reduction of the first (aliphatic) carbonyl is followed by the diastereoselective reduction of the second
(aromatic) carbonyl. The role of albumin is thus a combination of chemo- and stereocontrol.
Binding4 and crystallographic2a,5 studies have disclosed the
three-dimensional structure and binding motifs of human serum
albumin (HSA). This 585 a.a. protein (MW = 65 kDa) consists of
three highly helical domains (Fig. 1) held together by seventeen
Introduction
Albumin is the most abundant protein in the blood system where
it serves several functions, including the control of osmotic blood
pressure and the maintenance of blood pH.1 However, albumin’s
disulfide bonds. At least five binding sites for long chain fatty acids,
most striking property is its ability to bind and transport an
with different affinities, are located in domains I and III and at the
interfaces between domains I–II and II–III.5c Two binding sites
for small aromatic and heteroaromatic molecules, with different
specificities, are present in subdomains IIA (drug site 1) and IIIA
(drug site 2)2,5a (Fig. 1). Cystein-34 and the N terminus are the
main binding sites for metal ions.2a
Primary structures of mammalian albumins are highly con-
served. In particular, 76% homology is found between the amino
acid sequences of HSA and bovine serum albumin (BSA).6 Thus,
a similar three-dimensional structure and binding site architecture
can be expected for the two proteins.
A broad chemical reactivity is associated with the small-
molecule binding site located in albumin’s IIA subdomain. A
lysine residue is present in this site in both HSA (Lys-199)
and BSA (Lys-222). In addition to being a site for covalent
interactions with many drugs such as, for example, aspirin2 and
benzylpenicillin,7 this basic residue, surrounded by a hydrophobic
environment, is responsible for albumin’s ability to behave as
extraordinary large variety of small molecules. Albumin is the
main carrier of unesterified fatty acids and other endogenous
ligands such as hemin and bilirubin; it possesses a broad specificity
for small aromatic and heteroaromatic compounds, including
many drugs, and is able to bind metal cations and several inorganic
ions.1,2 A possible role as carrier of nitric oxide has also been
proposed.3
aDepartment of Chemical Sciences, University of Trieste, via Giorgieri 1,
34127 Trieste, Italy. E-mail: fberti@units.it, benedett@units.it; Fax: +39
040 5582402
bDepartment of Life Sciences, University of Trieste, via Giorgieri 1, 34127
Trieste, Italy
† Dedicated to Professor Charles J. M. Stirling on his 80th birthday.
1
‡ Electronic supplementary information (ESI) available: H NMR of the
Mosher esters of racemic and enantiomerically pure hydroxy ketone 1b;
Scatchard plot of the data from Fig. 4b; superimposition of the IIa binding
sites of HSA and BSA. See DOI: 10.1039/c0ob00648c
This journal is
The Royal Society of Chemistry 2011
Org. Biomol. Chem., 2011, 9, 1987–1999 | 1987
©

View Online
Table 1 Baker’s yeast reduction of diketones 3
Diketone
Ar
Hydroxyketone
Yield [%]
e.e. [%]a
3a
3b
3c
phenyl
p-tolyl
2-naphthyl
(S)-(+)-1a
(S)-(+)-1b
(S)-(+)-1c
30
83
35
96
>99
95
a e.e. was determined from the NMR spectra of the corresponding Mosher’s
esters.
Fig. 1 3-D structure of HSA. The three domains are colour-coded: I,
blue; II, green; III, red. The small-molecule binding sites in domains IIa
(drug site 1) and IIIa (drug site 2) are shown.
an enzyme-like catalyst in reactions such as b-eliminations,8 the
decomposition of Meisenheimer adducts9 and the eliminative ring
fission of 5-nitrobenzisoxazole to 2-hydroxy-5-nitrobenzonitrile
(Kemp elimination).10 In the latter reaction, in particular, the
catalytic activity of BSA and HSA is similar to that of specifically
designed catalytic antibodies.10c
Due to its low cost, large scale availability, and exceptional
ability to bind a wide range of substrates in an asymmetric
environment, BSA has found many applications as a stationary
phase in chiral chromatography,2c,11 and as a chiral auxiliary and
catalyst in organic reactions.12 Asymmetric oxidations of sulfides,13
disulfides,14 thioacetals,15 and tertiary amines,16 reduction of aro-
matic ketones,17 Diels–Alder cycloadditions,18 epoxidation19 and
dihydroxylation20 of alkenes, and the enantioselective hydrolysis
of esters21 have been reported to proceed with varying levels of
stereoselectivity in the presence of BSA.
Recently,22 we have reported on the remarkable ability of
albumin to direct the borohydride reduction of b-diketones to
the formation of anti-1,3-diols. Here we give a full account of this
work and present new data on the albumin controlled reduction
of b-hydroxyketones; we also describe the results of a study on the
binding of diketones and hydroxyketones to BSA that allows us
to identify the site of interaction. Finally, we present a model for
the BSA-substrate complex that fully accounts for the selectivity
observed in the albumin directed borohydride reduction.
Scheme
1,3-diketones.
1 NaBH4/albumin reduction of b-hydroxyketones and
((+)-1a–c). The hydroxyketone 2b, in which the aromatic group is
remote from the carbonyl, was also included in this series for
comparison with the corresponding aromatic ketone 1b.
Substrates were obtained by conventional literature methods,
as described in the experimental part. The enantiomerically pure
hydroxy ketones (+)-1a–c were obtained by bakers’ yeast reduction
of the corresponding diketones 3a–c (Table 1), as originally
described by Cheˆnevert23 for the phenyl derivative (+)-1a. The
(S) configuration was established for this compound,23 in agree-
ment with Prelog’s rule for the reduction of prochiral ketones.24
Considering the well known preference of yeast dehydrogenase for
Prelog-like reduction of methyl ketones,23b it seems safe to assign
the same configuration also to hydroxyketones 1b,c. Enantiomeric
1
excesses have been calculated from the H NMR spectra of the
corresponding Mosher esters‡ and are reported in Table 1.
Reduction of hydroxyketones
The reduction of hydroxyketones 1a–e with six equivalents of
sodium borohydride in a 9 : 1 water–acetonitrile mixture proceeds
with no appreciable stereoselectivity (Table 2). However, when the
same reaction is carried out in the presence of BSA, formation
of the anti isomers 4 is favoured, with diastereoselectivities up
to 94 : 6 for 1a, 1b and 1d.25§ The BSA-induced anti selectivity is
Results and discussion
Preliminary results22 indicated that albumin provides efficient anti
control in the reduction of aromatic b-diketones, but has no
effect in the reduction of aliphatic substrates. Thus, in the present
study, we focused on the reduction of aromatic 3-hydroxyketones
(1a–e) and 1,3-diketones (3a–j), to the corresponding anti and
syn 1,3-diols 4 and 5 (Scheme 1). In order to determine whether
albumin discriminates between enantiomers, hydroxyketones
1a–c were studied both in racemic and enantiomerically pure form
§ The stereochemistry of the novel diols 4/5b,e,f,i was confirmed by: a)
13C spectral analysis of the corresponding acetonides with characteristic
resonances for the geminal methyl groups at around 20 and 30 ppm
1988 | Org. Biomol. Chem., 2011, 9, 1987–1999
This journal is
The Royal Society of Chemistry 2011
©

View Online
Table 2 NaBH4/BSA reduction of hydroxyketones 1, 2 and diketones 3 to anti and syn diols 4, 5a
anti/syn ratiosb,c
Chemoselectivityb,d
NaBH4
Substrate
NaBH4
NaBH4/BSA
NaBH4/BSA
1a
(+)-1a
47 : 53
45 : 55
93 : 7
94 : 6
1b
(+)-1b
45 : 55
44 : 56
94 : 6
94 : 6
1c
(+)-1c
44 : 56
40 : 60
80 : 20
80 : 20
1d
1e
60 : 40
35 : 65
93 : 7
75 : 25
2b
3a
3b
3c
3d
50 : 50
47 : 53
45 : 55
45 : 55
60 : 40
50 : 50
93 : 7
98 : 2
80 : 20
94 : 6
60 : 40
66 : 33
65 : 35
60 : 40
>98 : 2
>98 : 2
90 : 10
90 : 10
3e
3f
35 : 65
43 : 57
75 : 25
75 : 25
n.d.
n.d.
65 : 35
80 : 20
3g
3h
30 : 70
33 : 66
70 : 30
56 : 44
55 : 45
40 : 60
80 : 20
>90 : 10
3i
40 : 60
52 : 48
50 : 50
>90 : 10
96 : 4
60 : 40
>90 : 10
3j
3k
50 : 50
a All reactions were carried out in 9 : 1 H2O–CH3CN with 6 equiv NaBH4 and 1 equiv BSA. b By HPLC. c Ratios of anti and syn diols 4 and 5. d Ratios of
hydroxyketones 1 and 2 at 50% conversion.
also good with the 2-naphthyl and p-nitrophenyl ketones 1c
and 1e that are preferentially reduced to syn diols in the absence
for the syn acetonides and 20–22 ppm for the anti acetonides;26 b)
of protein. Thus, the directing effect of BSA is little influenced by
comparison with authentic samples of trans diols 5b,e prepared by
the steric and electronic properties of the aromatic group. Clearly,
the stereoselective reduction of aldols 2b,e with tetramethylammonium
triacetoxyborohydride (Evans’ reagent).27
the origin of this directing effect must be a binding interaction
This journal is
The Royal Society of Chemistry 2011
Org. Biomol. Chem., 2011, 9, 1987–1999 | 1989
©

View Online
Table 3 Effect of the pH on the stereoselectivity of the NaBH4/BSA
reduction of hydroxyketone 1b and diketone 3ba
between the protein and the hydroxyketones, resulting in the
selective presentation of one diastereotopic face of the carbonyl
of the bound substrate to the reducing agent.
Data in Table 2 also show that, in the presence of BSA, the
reduction of enantiomerically pure and racemic hydroxyketones 1
proceeds with the same diastereoselectivity. Thus, the protein does
not discriminate between enantiomers.
The reduction of hydroxyketone 2b is not stereoselective in
the presence of BSA, indicating that an aromatic carbonyl is
essential for stereoselection. This conclusion is confirmed by the
lack of selectivity observed in the reduction of 2,4-pentanedione
3k (Table 2).
anti/syn ratios
pH
1b → 4b,5b
3b → 4b,5b
5.0b
6.0c
7.0c
8.0c
9.0d
94 : 6
95 : 5
90 : 10
82 : 18
67 : 33
94 : 6
86 : 14
85 : 15
77 : 23
n.d.
a Conditions: 1 mM substrate, 1 mM BSA, 5 mM NaBH4, 1 : 9 acetoni-
trile/buffer (10 mM, containing 50 mM NaCl). b Sodium citrate buffer.
c Sodium phosphate buffer. d Sodium borate buffer.
Reduction of diketones
introduction of a bulky t-butyl group in this position (3h) is not
tolerated by the protein, resulting in a complete loss of selectivity.
The lack of selectivity observed in the reduction of acetylacetone
(3k) and 1,3-cyclopentanedione (not shown) confirms that the
presence of an aromatic carbonyl is necessary for stereoselection
control by BSA.
The NaBH4 reduction of 1,3-diketones 3 in 9 : 1 H2O–CH3CN
gives the corresponding anti and syn diols in ratios close to
1 : 1 (Table 2). The reaction also fails to display any appreciable
chemoselectivity: a typical profile for the reduction of diketone 3b
(Fig. 2) shows that the intermediate isomeric hydroxyketones 1b
and 2b are formed (and react) at comparable rates. The ratios of
hydroxyketones 1 and 2 at 50% conversion (Table 2) confirm this
behaviour for all the diketones studied.
pH dependence
The reduction of hydroxyketone ( )-1b and diketone 3b, in the
presence of BSA, was studied between pH 5 and 9 (Table 3). Data
in the table show that the stereoselectivity depends strongly on
the pH. In the reduction of hydroxyketone ( )-1b the anti/syn
ratio decreases from 95 : 5 at pH 5–6 to 67 : 33 at pH 9 (Table
3). The reduction of diketone 3b follows a similar trend with
anti/syn ratios dropping from 94 : 6 at pH 5 to 77 : 23 at pH 8,
indicating the probable involvement of protonated sites in binding
of both hydroxyketones and diketones. In consideration of this
pH-dependence all synthetic reactions were generally carried out
at pH 6 even if, under these conditions, an excess of sodium
borohydride (6 equiv) is required for the reaction to be complete,
due to the partial decomposition of borohydride.
Other albumins
Fig. 2 Time profile for the NaBH4 reduction of diketone 3b in 9 : 1
H2O–CH3CN. Triangles: diketone 3b; empty squares: hydroxyketone 2b;
filled squares: hydroxyketone 1b; empty circles: anti diol 4b; filled circles:
syn diol 5b.
To determine whether other albumins possess the same ability
to control the selectivity of the conversion of 1,3-diketones into
diols, reduction of diketone 3b was carried out in the presence of
Human Serum Albumin (HSA) and Ovalbumin (OVA). HSA has
a high (76%) sequence homology with BSA6 and, not surprisingly,
a good anti/syn stereoselectivity (88 : 12) was observed when the
reaction was carried out in the presence of this protein. OVA,
belonging to the serpin family, is the main component of chicken
egg’s albumen: its function is similar to that of BSA and HSA
but the sequence and 3-D structure are completely unrelated to
those of mammalian albumins.28 Nevertheless, the stereoselectivity
observed in the presence of this protein is comparable to that
obtained with HSA and only slightly lower than with BSA. This
finding might be interesting for practical applications, as egg
albumin is very cheap.
A completely different pattern emerges when the reduction is
carried out in the presence of albumin (Table 2). Again formation
of the anti diols 4 is preferred, with stereoselectivities comparable
to those already observed in the reduction of hydroxyketones 1.
The NaBH4/BSA reduction of diketones is not enantioselective
and all diols are formed as racemic mixtures. Anti/syn Ratios
higher than 9 : 1 are observed for phenyl (3a,j), p-tolyl (3b,i) and
ferrocenyl derivatives (3d). Overall stereoselectivities are lower
with para-substituted compounds 3e,f or when the aromatic group
is 2-naphthyl (3c) or 2-furyl (3g). Even with these substrates,
however, the directing effect of BSA is clearly evident, resulting
in an inversion in the stereoselectivity from the syn diol, which is
favoured without protein, to the anti isomer; a similar behaviour
was observed in the reduction of hydroxyketones 1c (p-NO2) and
1e (2-naphthyl). Replacement of the methyl group of 3a,b with
a longer and functionalized chain, as in 3i, or with a phenyl
group (3j) has little effect on the stereoselectivity. On the contrary,
All the reactions described so far were carried out in a water–
acetonitrile solution containing equimolar amounts of substrate
and BSA. To test experimental conditions more amenable to large
scale reactions, reduction of diketone 3b was also carried out in a
heterogeneous system; lyophilized BSA and sodium borohydride
were suspended in a toluene solution of the ketone and the
1990 | Org. Biomol. Chem., 2011, 9, 1987–1999
This journal is
The Royal Society of Chemistry 2011
©

View Online
mixture was stirred at room temperature for 24 h. The results
were encouraging as, without optimization, the stereoselectivity
was still 80% in favour of the anti diol 4b, while the protein could
be simply recovered by filtration and recycled.
Binding of diketones and hydroxyketones by albumin
To gain more insight on the mechanism by which albumins control
the regio- and stereoselectivity of the reduction of diketones and
hydroxyketones, the interaction of substrates 1b, 2b and 3b with
BSA was investigated by UV and CD spectroscopy. The analysis
was restricted to the region above 300 nm, in which the protein is
transparent. The broad band at 316 nm (e = 5000), present in the
UV spectrum of diketone 3b, recorded in 10% aqueous acetonitrile
(Fig. 3a), indicates partial enolization of the aromatic carbonyl
in this solvent system.29 The intensity of the enol band strongly
increases upon addition of one equivalent of BSA, indicating that
the tautomeric equilibrium is further shifted toward the enol form
by the interaction with the protein. The same interaction also
Fig. 4 Ellipticities at 324 nm of albumin solutions in 9 : 1 water–ace-
tonitrile containing diketone 3b. Circles: BSA (57 mM); squares: BSA
(57 mM) and hydroxyketone ( )-1b (570 mM); triangles: BSA (57 mM) and
(+)-1b (570 mM); diamonds: fluoresceinated BSA (57 mM). All values are
corrected for the ellipticities at 330 nm of BSA and fluoresceinated BSA,
respectively.
causes a shift of the enol band to higher wavelengths (lmax
=
effects observed for (+)-1b and ( )-1b indicates that BSA does not
discriminate between enantiomeric hydroxyketones; a similar lack
of enantiospecificity has been observed in the binding of warfarin
(3-(a-acetonylbenzyl)-4-hydroxycoumarin) by HSA.5e,31
The isomeric hydroxyketone 2b, and the diol products 4b, 5b,
on the contrary, do not inhibit binding of 3b to BSA, indicating
that these compounds are not recognized by the protein. This may
appear surprising considering the structural similarity between
molecules of this set and the broad specificity of albumins, but is
consistent with the absence of selectivity observed in the reduction
of 2b. Clearly, in these small ligands possessing a limited number of
functional groups, the presence of an aromatic carbonyl is essential
for efficient recognition by albumin.
324 nm, e = 13500). Binding of this enol by BSA is confirmed
by CD. Being achiral species, neither the diketone 3b, nor its
enol display any CD. In the complex with the protein, however,
the enol is surrounded by an asymmetric environment, resulting
in the appearance, in the CD spectrum of a mixture of 3b and
BSA, of an induced band30 at the wavelength characteristic of the
enol absorption (Fig. 3b). Titration of the protein with increasing
amounts of 3b clearly shows saturation by the ligand (Fig. 4) and
standard treatment of the data‡ gives a value of 5.2 0.2 ¥ 104 L
mol-1 for the association constant between 3b and BSA. Similar
values are found for the binding constants of albumin with other
small aromatic molecules.2
In conclusion, data presented in this section strongly suggest
that specific binding of 1b and 3b to BSA is at the origin of the high
selectivities observed in the reduction of these substrates (Table
2), while for 2b lack of binding results in complete absence of
selectivity.
Identification of the binding site
Binding and crystallographic studies indicate that drugs and other
small aromatic molecules containing anionic or electronegative
groups bind to the IIA site of HSA (drug site 1), a sock
shaped hydrophobic cavity lined at its entrance and bottom with
positively charged residues.2a This site, highly conserved in BSA,
is therefore a good candidate for binding of hydroxyketones 1 and
diketones 3.
A single lysine residue is present in this site both in HSA
(Lys199) and BSA (Lys222); no lysines are present in the other
small molecule binding site (IIIA site; drug site 2) of either
albumin. This lysine can be selectively modified with fluorescein
isothiocyanate (FITC)9c and the modification prevents this residue
from participating in binding and catalysis.9c,10c Fluoresceinated
BSA (FITC-BSA) was thus prepared and its interaction with
3b was studied by CD (Fig. 4). It can be clearly seen that
fluoresceination of Lys199 inhibits over 90% of the binding of the
diketone, confirming that recognition takes place in the IIA site.
Accordingly, the anti/syn diastereoselectivity in the borohydride
reduction of 3b drops from 94 : 6 with BSA to 75 : 25 with
Fig. 3 a) UV spectra of BSA, diketone 3b and of a 1 : 1 mixture of BSA
and 3b. Spectra were recorded in 9 : 1 H2O–CH3CN with all concentrations
equal to 1.67 ¥ 10-4 M. b) CD spectra of BSA and of a 1 : 1 mixture of
BSA and 3b in 9 : 1 water–acetonitrile. [BSA] = [3b] = 57 mM.
The interaction of hydroxyketones 1b, 2b and diols 4b, 5b
with BSA could not be studied directly, as these species have no
UV absorption bands above 300 nm. Nevertheless, their binding
could be investigated indirectly by competition experiments. Fig. 4
shows that binding of diketone 3b to BSA is inhibited in the
presence of a ten-fold excess of both racemic hydroxyketone 1b
and enantiomerically pure (+)-1b; it can thus be concluded that
1b and 3b bind to the same protein site. A comparison of the
This journal is
The Royal Society of Chemistry 2011
Org. Biomol. Chem., 2011, 9, 1987–1999 | 1991
©

View Online
FITC-BSA. The residual binding and diastereoselectivity ob-
served in the presence of FITC-BSA can be accounted for by
the presence of a small amount of non-fluoresceinated protein in
the sample, as indicated by the modification index of 0.9 calculated
for FITC-BSA.
Further insight into the interactions of diketone 3b with BSA
is provided by a study of the pH-dependence of the induced band
in the CD spectrum of the diketone-protein complex (Fig. 5). A
bell-shaped curve is found between pH 3 and 11, with a broad
maximum between pH 6 and 8, in general agreement with the
variation of selectivity observed in the reduction of 3b (Table 3).
It would seem reasonable to attribute the changes in binding with
the pH to the titration of the only ionizable residues present in
BSA’s IIA site, namely His242 and Lys199. In this hypothesis, pKa
values of around 4.5 and 9.5 for His242 and Lys199, respectively,
can be derived from the shape of the curve, in agreement with
an estimate of the pKa’s of the protein’s protonated residues,
carried out with the PROPKA programme,32 which indicates these
two amino acids as the protein’s most acidic histidine and lysine,
respectively. However, care should be taken with this interpretation
as the curve of Fig. 5 might simply reflect the intrinsic pH stability
of the protein.33 In either case, from the location of the maximum
we can conclude that in the complex with the bound enol His242
and Lys199 are in the neutral and protonated forms, respectively.
Fig. 5 Ellipticities at 324 nm of a 1 : 1 solution of diketone 3b and BSA
(28.5 mM; water–acetonitrile 9 : 1) as a function of pH. The following
buffers were used (40 mM, containing 100 mM NaCl): sodium acetate
(pH < 6); sodium phosphate (pH 6–8); sodium borate (pH 8–10); sodium
carbonate (pH > 10).
The results show that, in vacuo, diketone 3a is present as
a mixture of intramolecularly hydrogen bonded enols, while a
significant amount of the diketone form is present in water, in
agreement with experimental data.29 A strong UV absorption at
321 nm is calculated for the aryl enol 6a, shifted to 302.1 nm in
the alkyl enol 7a. This band is absent in the calculated electronic
spectrum of 3a. Breaking the intramolecular hydrogen bond of
the enols by rotating the hydrogen away from the carbonyl causes
the bands to shift down to 307 and 273 nm, respectively. These
results support the hypothesis that diketones 3 are recognized by
albumin as enols, the tautomeric equilibrium being shifted to this
form by the hydrophobic environment of albumin’s binding site.
The observed value of 324 nm for the UV maximum of diketone
3a bound to BSA is in good agreement with the calculated value
of 321 nm for 6; we can thus conclude that the aryl enol is most
likely present in the BSA complex, and that a strong hydrogen
bond is established by this enol, either intramolecularly or with
donor/acceptor groups in the protein binding site.
Modelling
A molecular modelling study was carried out to test the mecha-
nistic hypothesis suggested by the experimental data. Diketone 3a
and hydroxyketones 1a and 2a were chosen as reference ligands
for this study. The geometries of diketone 3a and of the s-cis aryl
and alkyl enols 6a and 7a, respectively, were initially optimized
in vacuo at the B3LYP/6.31G(d,p) level of the DF theory,34
and then re-optimized in water applying the IEF polarizable
continuum model for solvation;35 results are in Table 4. The
corresponding s-trans enols were also considered initially, but their
energies were so high that their contribution to the tautomeric
population of 3a at room temperature can be considered negligible.
The electronic spectra of the three isomers were also calculated,
at the CI-S/B3LYP/6.31G+(d,p) level and selected absorption
wavelengths are reported in Table 4.
No crystal structure of BSA has ever been reported, to the
best of our knowledge, and we were unable to obtain crystals
Table 4 Calculated energies (B3LYP/6.31G(d,p)) and electronic transitions in the 300–330 nm range (CI-S/B3LYP/6.31G+(d,p)) for diketone 3a and
enols 6a, 7a
Rel. energya [kcal mol-1]
(% population)b
Rel. energyc [kcal mol-1]
(% population)b
Wavelength [nm] (energy) [eV]
[oscillator strength]
Tautomer
3a
6a
1.78 (2.7)
0.19 (29.3)
—
0.17 (41.6)
0.17 (30.3)
321.0 (3.8685) [0.6158]
7a
0 (55.4)
0 (40.4)
302.1 (4.1036) [0.4913]
a Gas phase. b At 300 K. c In water.
1992 | Org. Biomol. Chem., 2011, 9, 1987–1999
This journal is
The Royal Society of Chemistry 2011
©

View Online
Table 5 Calculated relative complexation energies of BSA with both
enantiomers of hydroxyketones 1a and 2a
of the protein or its complexes with diketones. BSA, however, is
highly homologous with HSA, for which many structures have
been described;5 in particular, in the region corresponding to the
IIa domain (residues 190–300) the two proteins differ for only
twelve residues. We have thus build an homology model of BSA
starting from the crystallographic structure of the complex of
HSA with the R-(+) enantiomer of warfarin (pdb file 1H9Z).5e
This structure was chosen for the structural similarity between
the 4-hydroxycoumarin ring of warfarin and the aryl enols 6
(highlighted).
Hydroxyketone
DDEa [kcal mol-1]
(R)-1
(S)-1
(R)-2
(S)-2
0
0.1
7.2
10.6
a Relative Amber complexation energies DE = Ecomplex - (Eligand + Eprotein).
It can be seen from Fig. 6 that the aromatic ring of the diketone is
deeply buried into the hydrophobic cavity lined by Ala291, Ile290,
Ile264, Ser287, Leu260, Leu238, Leu219. The polar region of the
enol is oriented toward the polar, positively charged rim of the
binding site with a water molecule at hydrogen bond distance
from the enol oxygen on one side and from the Ne2 of His242
on the other. The carbonyl oxygen of the substrate is at hydrogen
bond distance from Arg199, with the cationic head of Lys222
lying nearby. The model suggests a simple explanation for the
regioselectivity observed in the reduction of the diketone (Fig. 6b):
while the aromatic carbonyl is inaccessible, completely buried into
the site, the alkyl-carbonyl is exposed to the solvent at the top of
the site leaving both its faces available to the incoming hydride.
The B3LYP/6.31G(d,p) geometries of hydroxyketones 1a and
2a were similarly obtained, following a systematic Amber confor-
mational search,38 which yielded four conformations for 1a and
two conformations for 2a. These structures were then docked
to the IIa binding site of the BSA homology model and the
process was repeated for both enantiomers of each compound. The
relative complexation energies for the four complexes obtained
after optimization of the best docking solutions are reported in
Table 5.
The structure of the homology model was obtained by mutating
all the necessary amino acid side chains in the HSA structure,
and by submitting the mutated structure to a series of molecular
dynamics and optimization cycles carried out with the Amber
force field as implemented in Sybyl 7.3.36 The two sites are almost
perfectly superimposable, the only exception being HSA’s Arg222
and Lys199 whose positions are exchanged in BSA.‡ The diketone
3a and its tautomers 6a and 7a were then docked to the IIA site
of the BSA model thus obtained, with the docking suite Grid
22b,37 and model geometries of the complexes were obtained
by submitting again the complexes to molecular dynamics and
mechanics optimizations. Very similar docking solutions were
obtained for the three tautomers, with the phenyl ring and
the polar groups similarly oriented in the three complexes and
occupying the same region that hosts the aromatic group of
the benzocoumarin system in the warfarin-HSA complex. The
structure of the model complex of BSA with the aryl enol 6a is
shown in Fig. 6.
Data in Table 5 show that binding of hydroxyketone 2a is
less favourable than that of 1a by 7–10 kcal mol-1, matching
the experimental observation that 2a is not recognized by BSA.
In agreement with the experiments, comparable affinities are
Fig. 6 a) Model of the complex between enol 6a and BSA’s IIa site. b) Connolly surface showing the alkyl carbonyl of 6a emerging from the hydrophobic
cavity of the IIa site; the surface is coloured according to the electrostatic potential.
This journal is
The Royal Society of Chemistry 2011
Org. Biomol. Chem., 2011, 9, 1987–1999 | 1993
©

View Online
Fig. 7 Superimposition of the calculated structures of the complexes of BSA with hydroxyketones R-1a (yellow) and S-1a (pink). a) The side view shows
the main interactions with the binding site residues; b) a front view shows that the methyl groups block the access to opposite diastereotopic faces of the
carbonyl.
Scheme 2 Mechanism for the BSA-directed reduction of 1,3-diketones and 3-hydroxyketones. The aromatic diketone is bound by BSA to give a
ketoenol-BSA complex that is chemoselectively reduced by borohydride in the first step. The resulting second complex (BSA-hydroxyketone) is then
stereoselectively reduced to anti diol in the second step.
calculated for the enantiomers of hydroxyketone 1a, which are
recognized by the protein with a similar pattern of interactions
and in the same gauche conformation, with respect to the carbon
chain (Fig. 7a). In the complex, the stereogenic carbon is located
near the entrance of the binding site, and thus there is enough
space to host the methyl group of both enantiomers.
The model clearly shows (Fig. 7b) that, in the best solutions
found for both the complexes with R-1a and S-1a, the carbonyl
is deeply buried inside the inner cavity and access to one face is
blocked by the methyl group. Reduction of the hydroxyketone may
then take place either in an intermolecular mode, with the hydride
approaching the carbonyl from the less hindered side, or by the
classical internal delivery mode in which hydride is transferred
from a boron species complexed with the substrate’s hydroxy
group.39 Inspection of the model indicates that enough space is
available in the binding site for 120◦ rotation around the CH2–
CH(OH) bond of the hydroxyketone (anticlockwise for R-1a and
clockwise for S-1a) leading to the gauche conformation which is
necessary for the latter mechanism to operate. Either mechanism
would lead to the formation of the anti diol.
the stereoselectivity (anti/syn ratios) of the overall conversion
to diols 4 and 5 and the chemoselectivity observed in the first
reduction. This correlation, combined with the observation that
only hydroxyketones 1, and not the isomeric ketones 2, are
stereoselectively reduced in the presence of BSA, suggests a
possible model for the stereoselective reduction of hydroxyketones
and 1,3 diketones that has been validated by binding studies
and by docking simulations. According to this model (Scheme
2) a complex is formed between the diketone, in its enol form
6, and BSA, in which the aliphatic carbonyl is more available
and is reduced first. The model predicts that in the complex both
faces of the carbonyl are sufficiently exposed to the incoming
nucleophile. Accordingly, no enantioselectivity is observed in
this first reduction. The carbonyl group of the hydroxyketone
1 thus formed is then stereoselectively reduced in the second
step. Complexation of this substrate by the protein controls the
anti diastereoselectivity of this step (Scheme 2, Fig. 7), either
by directing the path of the incoming external nucleophile or
by favouring an alternative intramolecularly assisted mechanism,
in which the hydroxy group directs the delivery of hydride to
the syn face of the carbonyl. The overall diastereoselectivity, in
this hypothesis, derives from a combination of chemoselectivity,
favouring the formation of hydroxyketones 1 in the first step of the
reaction, and albumin-controlled stereoselectivity in the second
step.
Conclusions
The data presented show that, in the reduction of diketones 3
in the presence of BSA, there is a strong correlation between
1994 | Org. Biomol. Chem., 2011, 9, 1987–1999
This journal is
The Royal Society of Chemistry 2011
©

View Online
Unlike other reactions promoted by albumins, in this case there
is no catalytic effect associated with the protein. No multiple
turnover is observed and a stoichiometric amount of protein is
required for efficient control of the regio- and stereoselectivity.
While inhibition by the reaction products can prevent multiple
turnover in enzyme-catalyzed reactions, this is not the case here,
as we have demonstrated that 1,3-diols 4,5 are much poorer binders
with respect to substrates 1 and 3.
In the mechanism of Scheme 2 albumin’s function is to
present the substrate to an external reducing agent so as to
differentiate between the carbonyls in the first step and between
the diastereotopic faces of the aromatic ketone in the second step.
This role is reminiscent of that of dirigent proteins (DP) that
control the enantioselective oxidative coupling of phenols in plant
secondary metabolism, but do not possess catalytic activity.40 Our
work suggests that the DP concept is not restricted to biosynthetic
pathways, but may also be applied to organic synthesis, offering
new possibilities to the control of regio- and stereoselectivity in
organic reactions.
known compounds. Spectral data for aldols 1c and 1e are given
below.
3-Hydroxy-1-(2-naphthalenyl)-1-butanone (1c). (65% from 2-
◦
1
acetyl-naphthalene): m.p. 46–47 C, from diisopropyl ether; H
NMR: d = 1.34 (d, J = 6.1 Hz, 3H), 3.15 (dd, J = 17.5, 3.1 Hz,
1H), 3.25 (dd, J = 17.5, 8.6 Hz, 1H), 3.5 (1H, OH), 4.44 (m, 1H),
7.50–8.41 (m, 7H); 13C NMR: d = 22.4, 46.5, 64.0, 123.4, 126.7,
127.6, 128.4, 128.5, 129.5, 132.3, 133.9, 135.6, 200.5, one peak is
missing due to overlapping; EI-MS: m/z (%): 214 (16) [M+], 170
(22) [M+ - C2H4O], 155 (100) [C10H7CO+]; elemental analysis calcd
(%) for C14H14O2: C 78.5, H 6.59; found C 78.4, H 6.60.
3-Hydroxy-1-(4-nitrophenyl)-1-butanone (1e). (oil, 33% from
1
p-nitroacetophenone): H NMR: d = 1.33 (d, J = 6.2 Hz, 3H),
3.05 (1H, OH), 3.15 (m, 2H), 4.46 (m, 1H), 8.11–8.34 (m, 4H);
13C NMR: d = 22.4, 47.2, 63.8, 123.8, 129.1, 141.0, 150.5, 198.9;
EI-MS: m/z (%): 209 (1) [M+], 165 (28) [M+ - C2H4O], 155 (100)
[ArCO+].
Experimental
General procedure for the synthesis of diketones 3 by the Claisen
condensation
Optical rotations were measured on a Perkin–Elmer 241 polarime-
ter fitted with a 10 cm cell. IR spectra were recorded using a
1
Thermo-Nicolet FT-IR Avatar 320. H and 13C NMR spectra
Diketones 3a–d,f,g,i were obtained by the Claisen condensation
of aromatic ketones with the appropriate ester; the procedure is
described for the synthesis of 1-(2-furanyl)butane-1,3-dione (3g).
2-Acetylfuran (3.29 g, 30 mmol) in dry ethyl acetate (15 mL) was
added dropwise, at 0 ◦C, to a suspension of NaH (0.79 g of a 60%
suspension in mineral oil, 33 mmol) in dry ethyl acetate (15 mL)
and the reaction mixture was stirred at 0 ◦C for 2 h and at 25 ◦C
for further 12 h (TLC). 10% Aqueous NH4Cl was then carefully
added (30 mL) and the mixture was acidified to pH 5 with HCl. The
aqueous phase was separated and extracted with ethyl acetate (2 ¥
15 mL). The combined organic phases were dried over anhydrous
sodium sulfate and evaporated under reduced pressure; the crude
product was purified by FC, with ethyl acetate/petroleum ether
1 : 1 as eluant, to give the oily diketone 3g (0.90 g, 20%). 1H NMR:
d = 2.14 (s, 3H), 6.07 (s, 1H, CH, enol form), 6.54–7.60 (m, 3H),
16.3 (s, 1H, OH, enol form); 13C NMR: d = 24.4, 96.0 ( CH,
enol), 112.4, 115.5, 145.9, 154.7, 176.1, 189.5; EI-MS: m/z (%):
152 (43) [M+], 137 (28) [M+ - CH3], 95 (100) [C4H5OCO+]. With
the same procedure◦were obtained the known diketones 3b (65%;
m.p. 55 ◦C, lit.45 56 C), 3c (54%; m.p. 97.5 ◦C, lit.46 98–98.5 ◦C),
3d (58%, m.p. 79 ◦C, lit.44 80 ◦C), 3f (33%, m.p. 113–114 ◦C, lit.47
116–117 ◦C). Compound 3i was obtained similarly from equimolar
amounts of p-methyl acetophenone and g-butyrolactone, in THF
(45%; m.p. 38 ◦C, lit.48 38–39 ◦C).
were recorded in CDCl3, unless otherwise stated, on a Jeol EX400
spectrometer operating at 400 MHz and 104 MHz, respectively;
chemical shifts (d) are downfield from TMS. EI-mass spectra
were measured at 70 eV on a VG 70/70 spectrometer. ESI-mass
spectra were recorded on a Perkin–Elmer API1 spectrometer.
Elemental analyses were obtained with a Carlo Erba EA1110
elemental analyzer. UV spectra were obtained on a Perkin–Elmer
Lambda 2 or on a Unicam Helios b spectrophotometer. CD
spectra were recorded on a Jasco J-710 spectropolarimeter. HPLC
analyses were run on a Hewlett–Packard 1100 HPLC system. THF
was freshly redistilled from sodium/benzophenone. Flash column
chromatography was performed on silica gel 60H (230–400 mesh)
Merck 9385; thin layer chromatography was performed on silica-
coated Merck kieselgel 60F254 0.25 mm plates and visualized by UV
irradiation at 254 nm. Diketones 3a and 3j were purchased from
Sigma–Aldrich. Diketone 3h41 was synthesized according to the
literature. The hydroxy ketone 2b was obtained from diketone 3b
by selective hydrogenation of the aromatic carbonyl, as described
by Scho¨pf.42
General procedure for the synthesis of hydroxyketones 1a–e
Racemic hydroxyketones 1a–e (R = CH3) were obtained from
the lithium enolate of the appropriate aromatic ketone and
acetaldehyde.43 A solution of the corresponding ketone (15 mmol)
in 15 mL dry THF was added dropwise to a freshly prepared
solution of LDA (16.5 mmol) in 15 mL of dry THF at -78 ◦C.
After 30 min the aldehyde (15 mmol) in dry THF (15 mL) was
added dropwise and the reaction mixture was stirred for 15 min
at -78 ◦C. The reaction was quenched by addition of ice cold, sat.
aqueous NaHCO3 and extracted with CHCl3 (3 ¥ 25 mL). The
combined extracts were dried over an. Na2SO4 and evaporated
under reduced pressure; silica gel chromatography of the residue
(CH2Cl2) gave the pure products. Aldols 1a,43a 1b,43b and 1d44 are
1-(4-Nitrophenyl)butane-1,3-dione 3e
Pyridinium chlorochromate (160 mg, 0.740 mmol) was added in
small portions, at 25 ◦C, to aldol 1e (140 mg, 0.670 mmol) in dry
dichloromethane (20 mL) and the mixture was stirred for 6 h. Light
petroleum was added (100 mL) and the mixture was filtered. Upon
concentrating the resulting solution to approximately 10 mL, pure
3e separated as yellow crystals (105 mg, 75%); m.p. 110 ◦C (lit.45
111 ◦C).
This journal is
The Royal Society of Chemistry 2011
Org. Biomol. Chem., 2011, 9, 1987–1999 | 1995
©

View Online
Bakers’ yeast reduction of diketones
layers were extracted with water and sat. brine (30 mL each) and
dried over Na2SO4. Evaporation of the solvent under reduced
pressure gave the diols as mixtures of anti (4) and syn (5) isomers
which were separated by flash chromatography on silica gel, eluting
with 1 : 1 ethyl acetate/petroleum ether. Diols 4a and 5a,50 4c and
5c,51 4d and 5d,44 4g and 5g,52 4h and 5h,53 4j and 5j54 are known
compounds and were identified from their published spectral data.
Spectroscopic data for diols 4b,e,f,i and 5b,e,f,i and for their
acetonides are given below.
A suspension of fresh bakers’ yeast (Saccharomyces cerevisiae,
50 g) in water (1 L) containing sucrose (30 g) was kept at 37 C
◦
for 1 h. The diketone was added (1 g) and the mixture was
stirred at 37 ◦C for 3 days, adding sucrose (12 g) every 24 h.
The mixture was centrifuged and the aqueous phase was extracted
with dichloromethane (8 ¥ 100 mL); the combined organic phases
were dried over anhydrous sodium sulfate and evaporated under
reduced pressure. The crude products were purified as described
for the racemic hydroxyketones and the enantiomeric excess was
established from the NMR analysis of the corresponding Mosher’s
1
anti-1-p-Tolylbutane-1,3-diol (4b). Oil. H NMR d = 1.11 (d,
J = 6.3 Hz, 3H), 1.65–1.80 (m, 2H), 2.24 (s, 3H), 3.26 (2H, OH),
3.93 (m, 1H), 4.87 (dd, J = 7.9, 3.7 Hz, 1H), 7.04 (d, 2H), 7.13 (d,
2H). 13C NMR d = 21.1, 23.5, 46.4, 65.3, 71.5, 125.8, 129.3, 137.1,
141.8. EI-MS m/z (%): 180 (10) [M+], 162 (10), 147 (20), 121 (100),
120 (36), 119 (54), 91 (33), 77 (18). Acetonide: 1H NMR d = 1.27
(d, J = 6.2 Hz, 3H), 1.41 (s, 3H), 1.56 (s, 3H), 1.5–1.7 (m, 2H),
2.33 (s, 3H), 3.88 (m, 1H), 4.62 (dd, J = 11.3, 2.5 Hz, 1H), 7.15
(d, 2H), 7.27 (d, 2H). 13C NMR d = 20.1 (acetonide-CH3), 21.4
(acetonide-CH3), 21.5, 21.8, 40.6, 72.8, 78.5, 99.2, 126.2, 129.3,
137.4, 139.7.
esters. (S)-1a: e.e. 96%; [a]2D0 = +64.7 (c = 4.4 in CHCl3) (lit.49 [a]2D0
=
+68.7 (c = 1.01 in CHCl3)); (S)-1b: e.e.>99%; [a]2D0 = +60.4 (c = 4.1
in CHCl3); (S)-1c: e.e. 95%; [a]2D0 = +37.6 (c = 2.6 in CHCl3).
Mosher’s esters of hydroxyketones
The hydroxyketone (racemic or single enantiomer, 10 mg) and
(R)-(-)-a-methoxy-a-trifluoromethyl-phenylacetylchloride (1.3
equiv) in dry pyridine (0.2 mL) were stirred at room temperature
for 18 h. Dichloromethane was added (2 mL) and the solution
was extracted with 10% aqueous citric acid (1 mL). The aqueous
phase was extracted with dichloromethane (1 mL) and the
combined organic layers were passed through a short silica
column; the column was rinsed with dichloromethane (5 mL) and
the flow-through was evaporated under reduced pressure to give
the esters; 1H NMR data are listed below.
1
syn-1-p-Tolylbutane-1,3-diol (5b). Oil. H NMR d = 1.10 (d,
J = 6.2 Hz, 3H), 1.61 (ddd, J = 14.5, 2.8, 2.9 Hz, 1H), 1.73 (dt, J =
14.5, 10.0 Hz, 1H), 2.25 (s, 3H), 3.60 (1H, OH), 3.69 (1H, OH),
4.02 (m, 1H), 4.77 (dd, J = 10.0, 2.9 Hz, 1H), 7.06 (d, 2H), 7.15
(d, 2H). 13C NMR d = 21.2, 24.0, 47.1, 68.8, 75.2, 125.9, 129.4,
137.4, 141.8. EI-MS m/z (%): 180 (11) [M+], 162 (10), 147 (20),
121 (100), 120 (37), 119 (48), 91 (38), 77 (23). Acetonide: 1H NMR
d = 1.21 (d, J = 6.2 Hz, 3H), 1.4–1.5 (m, 1H), 1.50 (s, 3H), 1.56
(s, 3H), 1.68–1.73 (m, 1H), 2.33 (s, 3H), 4.13 (m, 1H), 4.86 (dd,
J = 11.9, 2.7 Hz, 1H), 7.15 (d, 2H), 7.26 (d, 2H). 13C NMR d =
20.1 (acetonide-CH3), 21.4, 22.4, 30.6 (acetonide-CH3), 41.2, 65.6,
71.6, 99.2, 126.2, 129.3, 137.4, 139.7.
Mosher’s ester of (S)-(+)-1a. 1H NMR (C6D6) d = 1.39 (d, J =
4.8 Hz, 3H), 3.05 (dd, J = 17.7, 8.6 Hz, 1H), 3.16 (dd, J = 17.7,
2.9 Hz, 1H), 3.47 (s, 3H), 5.81 (m, 1H), 6.86–7.73 (m, 10H).
Mosher’s ester of (R)-1a (from racemic 1a). 1H NMR (C6D6)
d = 1.47 (d, 3H), 2.80 (m, 1H), 3.12 (m, 1H), 3,41 (s, 3H), 5.81 (m,
1H), 6.86–7.73 (m, 10H).
1
anti-1-(4-Nitrophenyl)butane-1,3-diol (4e). Oil. H NMR d =
1.28 (d, J = 6.3 Hz, 3H), 1.89 (m, 2H), 4.09 (m, 1H), 5.17 (m, 1H),
6.25–6.75 (broad, 2H), 7.54 (d, 2H), 8.19 d (2H). 13C NMR d =
23.7, 45.7, 65.8, 71.1, 123.9, 126.6, 147.4, 152.4.
Mosher’s ester of (S)-(+)-1b. 1H NMR (C6D6) d = 1.07 (d, J =
6.23 Hz, 3H), 2.00 (s, 3H), 2.29 (dd, J = 17.2, 4.2 Hz, 1H), 2.95
(dd, J = 17.2, 8.3 Hz, 1H), 3.47 (s, 3H), 5.81 (m, 1H), 6.86–7.73
(m, 9H).
1
syn-1-(4-Nitrophenyl)butane-1,3-diol (5e). Oil. H NMR d =
1.27 (d, J = 6.3 Hz, 3H), 1.79 (m, 2H), 2.0–2.5 (broad, 2H), 4.23
(m, 1H), 5.06 (dd, J = 4.5, 2.2 Hz, 2H), 7.55 (d, 2H), 8.20 (d, 2H).
13C NMR d = 24.8, 47.1, 69.5, 74.5, 123.9, 126.6, 137.7, 152.0.
Mosher’s ester of (R)-1b (from racemic 1b). 1H NMR (C6D6)
d = 1.14 (d, J = 6.4 Hz, 3H), 1.99 (s, 3H), 2.39 (dd, J = 17.0, 5.5 Hz,
1H), 2.98 (dd, J = 17.0, 7.14 Hz, 1H), 3.42 (s, 3H), 5.78–5.90 (m,
1H), 6.83–7.77 (m, 9H).
1
Acetonide: H NMR d = 1.23 (d, J = 6.1 Hz, 3H), 1.36 (m, 1H),
1.53 (s, 3H), 1.57 (s, 3H), 1.79 (dt, J = 13.0, 2.6 Hz, 1H), 4.17 (m,
Mosher’s ester of (S)-(+)-1c. 1H NMR (C6D6) d = 1.35 (d, J =
6.22 Hz, 3H), 3.19 (dd, J = 17.6, 8.8 Hz, 1H), 3.32 (dd, J = 17.6,
2.7 Hz, 1H), 3,66 (s, 3H), 4.47 (m, 1H), 7.36–8.05 (m, 12H).
1H), 5.02 (dd, J = 11.8, 2.7 Hz, 1H) 7.54 (d, 2H), 8.20 (d, 2H). 13
C
NMR d = 20.0 (acetonide-CH3), 22.3, 30.4 (acetonide-CH3), 40.9,
65.3, 70.8, 99.4, 123.9, 126.7, 147.4, 150.0.
Mosher’s ester of (R)-1c (from racemic 1c). 1H NMR (C6D6)
d = 1.44 (d, J = 6.4 Hz, 3H), 2.91 (dd, J = 17.4, 3.3 Hz, 1H), 2.96
(dd, J = 17.6, 8.9 Hz, 1H), 3.75 (s, 3H), 4.47 (m, 1H), 7.36–8.05
(m, 12H).
1
anti-1-(4-Iodophenyl)butane-1,3-diol (4f). Oil. H NMR d =
1.14 (d, J = 6.1 Hz, 3H), 1.75 (m, 2H), 2.48 (1H, OH), 3.42 (1H,
OH), 3.96 (m, 1H), 4.91 (dd, J = 6.08, 4.86 Hz, 1H), 7.02 (d, 2H),
7.59 (d, 2H). 13C NMR d = 23.8, 46.0, 65.7, 71.4, 92.8, 127.9, 137.6,
144.3. Acetonide: H NMR d = 1.22 (d, J = 6.5 Hz, 3H), 1.23 (s,
6H), 1.40 (m, 1H), 1.83 (m, 1H), 4.04 (m, 1H), 4.99 (dd, J = 6.7,
1
NaBH4 reductions
4.4 Hz, 1H), 7.09 (d, 2H), 7.65 (d, 2H).
Sodium borohydride (1.14 g, 30 mmol) was added portionwise
to a solution of hydroxyketones 1, 2 or diketones 3 (5 mmol) in
10% aqueous acetonitrile (30 mL) and the mixture was stirred
at room temperature for 1–4 h, diluted with water (30 mL) and
extracted with ethyl acetate (3 ¥ 30 mL). The combined organic
syn-1-(4-Iodophenyl)butane-1,3-diol (5f). Oil. 1H NMR d =
1.15 (d, J = 6.2 Hz, 3H), 1.60–1.78 (m, 2H), 3.04 (1H, OH), 3.74
(1H, OH), 4.05 (m, 1H), 4.79 (dd, J = 9.91, 3.06 Hz, 1H) 7.02
(d, 2H), 7.59 (d, 2H). 13C NMR d = 24.4, 47.1, 69.2, 74.9, 93.1,
1996 | Org. Biomol. Chem., 2011, 9, 1987–1999
This journal is
The Royal Society of Chemistry 2011
©

View Online
127.9, 137.7, 144.4. Acetonide: 1H NMR d = 1.21 (d, J = 6.1 Hz,
3H), 1.39 (m, 1H), 1.50 (s, 3H), 1.55 (s, 3H), 1.68 (dt, J = 13.1,
2.5 Hz 1H), 4.12 (m, 1H), 4.84 (dd, J = 11.7, 2.5 Hz, 1H), 7.12 (d,
2H), 7.64 (d, 2H). 13C NMR d = 19.8 (acetonide-CH3), 22.1, 30.3
(acetonide-CH3), 40.8, 65.24, 71.0, 92.8, 99.1, 127.9, 137.5, 142.2.
HPLC analysis of the NaBH4 and NaBH4/albumin reductions
For the analysis, 200 mL aliquots of the reaction mixtures were
taken and acidified with TFA (5 mL). Samples containing albumin
were centrifuged at 4200 r.p.m. at 4 ◦C for 20 min to separate
the denaturated protein. The resulting solutions were directly
analyzed by HPLC with an Alltech Alltima C18 5 m column (250 ¥
4.6 mm). Elution with water–acetonitrile 68 : 32 gave in all cases
base-line separated peaks for the diketones, the corresponding
hydroxyketones and anti and syn 1,3-diols.
anti-1-p-Tolylhexane-1,3,6-triol (4i). Oil.1H NMR d = 1.50 (m,
2H), 1.60 (m, 2H), 1.60–1.85 (m, 2H), 2.30 (s, 3H), 3.53 (m, 2H),
3.83 (m, 1H), 4.0–5.0 (broad, 3H, OH), 4.90 (dd, J = 8.5, 3.1 Hz,
1H), 7.08 (d, 2H), 7.17 (d, 2H). 13C NMR d = 21.22, 29.0, 34.5,
45.4, 68.6, 71.0, 74.6, 125.7, 129.2, 136.8, 141.9.
syn-1-p-Tolylhexane-1,3,6-triol (5i). Oil.1H NMR d = 1.50 (m,
2H), 1.59 (m, 2H), 1.60–1.85 (m, 2H), 2.30 (s, 3H), 3.53 (m, 2H),
3.83 (m, 1H), 4.0–5.0 (broad, 3H, OH), 4.80 (dd, J = 9.7, 2.9 Hz,
1H), 7.09 (d, 2H), 7.18 (d, 2H). 13C NMR d = 21.24, 28.7, 35.0,
Fluoresceinated BSA
Fluorescein isothiocyanate (1.3 mg, 3.3 mmol) in DMF (30 ml) was
added to a solution of BSA (200 mg, 3 mmol) in 0.02 M phosphate
buffer, pH 8.2 (75 mL) and the mixture was stirred at 20 C for
◦
1
45.4, 62.6, 72.1, 74.6, 125.8, 129.2, 137.2, 141.6. Acetonide: H
NMR d = 1.49 (s, 3H), 1.55 (s, 3H), 1.57–1.71 (m, 6H), 2.32 (s,
3H), 3.64 (m, 2H), 4.00 (m, 1H), 4.86 (dd, J = 11.6, 2.7 Hz), 7.14
(d, 2H), 7.25 (d, 2H). 13C NMR d = 20.0 (acetonide-CH3), 21.3,
29.0, 30.4 (acetonide-CH3), 33.4, 39.4, 62.9, 69.6, 71.6, 99.4, 126.0,
129.3, 137.5, 139.4.
3 h. The solution was dialyzed against 0.01 M citrate buffer, pH
3, at 4 ◦C for 3 days and then against water, at 4 ◦C for 1 day.
A modification index of 0.93 was calculated for the conjugate,
assuming an extinction coefficient of 15.900 for fluoresceinated
BSA.9c,10b
Preparation of anti-diols 4b, 4e by reduction of aldols 2 with
Evans’ reagent
Calculations
A solution of aldol 2b42 or 2e55 (2 mmol) in dry acetonitrile (3.5 mL)
was added by cannula, at -40 C, to a freshly prepared solution
DFT calculations were carried out with the parallel version of the
Gaussian03 suite56 running on an Intel CoreTM2 cpu 6600@2.40
GHz, 3.50 GB RAM. All the makeup of the HSA pdb structure,
molecular mechanics optimizations, conformational searches and
dynamics were carried out with the Sybyl 7.3 suite running on
SGI Octane2 MIPS R12000 workstations, or on a RedHat Linux
Intel Pentium 4 2.53 GHz machine. Docking calculations were
performed with the Grid22b suite on a Kubuntu Linux Intel
CoreTM2 cpu 6600@2.40 GHz, 3.50 GB RAM.
◦
of tetramethylammonium triacetoxyborohydride (15 mmol) in
15 mL of a 1 : 1 mixture of dry acetonitrile and dry acetic acid.
The reaction mixture was stirred under argon, at -40 ◦C for
18 h and then quenched with 20 mL of 0.5 N aqueous sodium
potassium tartrate and allowed to reach room temperature. The
mixture was partitioned between ethyl acetate (40 mL) and sat.
aqueous NaHCO3 and the aqueous phase was extracted with ethyl
acetate (4 ¥ 30 mL). The combined organic phases were extracted
with brine and dried over anhydrous Na2SO4. Evaporation of the
solvent under reduced pressure gave the crude diols (87% and
85% from 2b and 2e, respectively) as 4 : 1 mixtures of anti/syn
isomers from which pure anti diols 4b and 4e were separated by
FC (CH2Cl2–EtOAc = 6/4 for 4b; CH2Cl2–CH3OH = 9/1 for 4e).
Homology model of BSA
The structure of the warfarin-HSA complex was downloaded from
the Protein Data Bank57 and all the crystallization water molecules
were removed before the optimization, with the exception of water
13 and 28. All the hydrogen atoms were then added, and the
structure was submitted to the multistage relaxation protocol of
Levit and Lifson.58 The Amber 4.1 force field was used for all
the optimizations;59 Gasteiger–Marsili charges were applied to
the protein atoms, while the charges on the ligand molecule were
obtained from a B3LYP/6.31G(d,p) calculation. After relaxing the
complex, the structure was allowed to reach thermal equilibrium at
300 K by a multistep molecular dynamic run in the NTV ensemble
(5000 fs at 100 K and at 200 K and then 405000 fs at 300 K).
Finally, the structure was globally reoptimized to give a model
of the HSA-warfarin complex. This geometry was used to build
the BSA homology model. The protein sequence was mutated at
the domain IIa level as follows: K195R, K199R, F211L, A215S,
R222K, S232T, T243K, E265D, S270T, E280D, N295K, E297A,
M298I, A300E. The mutated model was then submitted again to
the relaxation/dynamics/optimization scheme to give a model of
the IIa site of BSA. The model was conformationally consistent
and passed a Ramachandran analysis.
Reductions with NaBH4/albumin
The reduction of diketone 1b was carried out on the mg scale and
is described as a representative example. The diketone (26.4 mg,
150 mmol) in acetonitrile (15 mL) was added to a solution of
albumin (10 g, 150 mmol) in water (130 mL) and the mixture
was kept at room temperature for 30 min. A solution of NaB◦H4
(34 mg, 90 mmol) in water (5 mL) was added dropwise, at 20 C
and the resulting mixture was stirred at the same temperature for
2 h. The mixture was acidified with trifluoroacetic acid (5 mL)
and ethanol was added (150 mL). The denatured protein was
separated by centrifugation at 4200 r.p.m. at 4 ◦C for 20 min
and the solution was concentrated to 5 mL and extracted with
dichloromethane (3 ¥ 5 mL). The combined extracts were dried
over Na2SO4 and evaporated giving a 93 : 7 mixture of diols 4a and
5a (20.5 mg, 76%). The other reactions were carried out similarly
on the 1.5 mmol scale and analyzed as described in the following
section.
This journal is
The Royal Society of Chemistry 2011
Org. Biomol. Chem., 2011, 9, 1987–1999 | 1997
©

View Online
Docking
13 (a) S. Colonna, S. Banfi, F. Fontana and M. Sommaruga, J. Org. Chem.,
1985, 50, 769; (b) S. Colonna, N. Gaggero and M. Leone, Tetrahedron,
1991, 47, 8385; (c) S. Colonna, V. Pironti, F. Zambianchi, G. Ottolina,
N. Gaggero and G. Celentano, Eur. J. Org. Chem., 2007, 363.
14 S. V. Dzyuba and A. M. Klibanov, Biotechnology Letters, 2003, 25,
1961.
15 K. Ogura, M. Fujita and H. Iida, Tetrahedron Lett., 1980, 21, 2233.
16 S. Colonna, N. Gaggero, J. Drabowicz, P. Lyzwa and M. Mikolajczyk,
Chem. Commun., 1999, 1787.
17 T. Sugimoto, Y. Matsumura, S. Tanimoto and M. Okano, Chem.
Commun., 1978, 926.
18 (a) S. Colonna, A. Manfredi and R. Annunziata, Tetrahedron Lett.,
1988, 29, 3347; (b) M. T. Reetz and N. Jiao, Angew. Chem., Int. Ed.,
2006, 45, 2416.
The docking analysis was performed with the GRID 22b
package.37 To build the molecular interaction maps, all the residues
within 15 A from the Ne2 atom of His 242 were considered; the
˚
˚
residues were included into a 20 A cubic grid made of 8000 knots.
His 242 was considered to be in its neutral state, while Lys 222
was considered in the protonated cationic state. The grid was
mapped with eight different GRID molecular probes (H, OH2,
OH, O-, O , C3, C1 , DRY), and after having obtained the
maps, docking of diketone 3a and hydroxyketones 1a and 2a were
performed considering all the tautomeric forms of the diketone
and all the conformations of the hydroxyketones.
19 (a) S. Colonna, A. Manfredi, R. Annunziata and M. Spadoni,
Tetrahedron, 1987, 43, 2157; (b) S. Colonna, N. Gaggero, A. Manfredi,
M. Spadoni, L. Casella, G. Carrea and P. Pasta, Tetrahedron, 1988, 44,
5169.
20 T. Kokubo, T. Sugimoto, T. Uchida, S. Tanimoto and M. Okano, Chem.
Acknowledgements
Commun., 1983, 769.
21 T. Kokubo, T. Uchida, S. Tanimoto, M. Okano and T. Sugimoto,
Tetrahedron Lett., 1982, 23, 1593.
22 F. Benedetti, F. Berti, I. Donati and M. Fregonese, Chem. Commun.,
2002, 828.
This work was supported by the Ministry of University and
Scientific Research and by Regione Friuli Venezia Giulia. G. F.
is grateful to the University of Trieste for a research fellowship.
23 (a) R. Cheneˆvert and S. Thiboutot, Can. J. Chem., 1986, 64, 1599; (b) A.
Fauve and H. Veschambre, J. Org. Chem., 1988, 53, 5215.
24 (a) V. Prelog, Pure Appl. Chem., 1968, 9, 119; (b) K. Faber, Biotrans-
formations in Organic Chemistry 5th ed., Springer, Berlin, 2004, pp.
177–219.
25 For a recent review on the stereoselective reduction of b-hydroxyketones
and 1,3-diketones see: S. E. Bode, M. Wolberg and M. Mu¨ller,
Synthesis, 2006, 557; see also: O. Labeeuw, C. Roche, P. Phansavath
and J.-P. Genet, Org. Lett., 2007, 9, 105.
Notes and references
1 (a) T. Peters, Jr., All About Albumin; Biochemistry, Genetics and Medical
Applications, Academic Press, New York, 1996; (b) J. Figge, T. H.
Rossing and V. Fencl, J. Lab. Clin. Med., 1991, 117, 453.
2 (a) D. C. Carter and J. X. Ho, Adv. Protein Chem., 1994, 45, 153;
(b) U. Kragh-Hansen, V. T. G. Chuang and M. Otagiri, Biol. Pharm.
Bull., 2002, 25, 695; (c) J. Ghuman, P. A. Zunszain, I. Petitpas, A. A.
Bhattacharya, M. Otagiri and S. Curry, J. Mol. Biol., 2005, 353, 38.
3 (a) J. S. Stamler, D. J. Singel and J. Loscalzo, Science, 1992, 258, 1898;
(b) O. Rafikova, R. Rafikov and E. Nudler, Proc. Natl. Acad. Sci. U. S.
A., 2002, 99, 5913.
4 (a) M. Dockal, D. C. Carter and F. Ru¨ker, J. Biol. Chem., 1999, 274,
29303; (b) M. Dockal, M. Chang, D. C. Carter and F. Ru¨ker, Protein
Sci., 2000, 9, 1455.
5 (a) X. M. He and D. C. Carter, Nature, 1992, 358, 209; (b) D. C. Carter,
B. Chang, J. X. Ho, K. Keeling and Z. Krishnasami, Eur. J. Biochem.,
1994, 226, 1049; (c) S. Curry, H. Mandelkow, P. Brick and N. Franks,
Nat. Struct. Biol., 1998, 5, 827; (d) S. Sugio, A. Kashima, S. Mochizuki,
M. Noda and K. Kobayashi, Protein Eng., Des. Sel., 1999, 12, 439; (e) I.
Petitpas, A. A. Bhattacharya, S. Twine, M. East and S. Curry, J. Biol.
Chem., 2001, 276, 22804.
6 (a) R. D. Appel, A. Bairoch and D. F. Hochstrasser, TIBS, 1994,
19, 258; (b) K. Hirayama, S. Akashi, M. Furuya and K. Fukuhara,
Biochem. Biophys. Res. Commun., 1990, 173, 639.
7 (a) M. Yvon, P. Anglade and J. M. Val, FEBS Lett., 1988, 239, 237;
(b) N. Diaz, D. Sua´rez, T. L. Sordo and K. M. Merz Jr., J. Med. Chem.,
2001, 44, 250; (c) N. Diaz, D. Sua´rez, T. L. Sordo and K. M. Merz, Jr.,
J. Am. Chem. Soc., 2001, 123, 7574.
8 (a) G. Klein and J.-L. Reymond, Bioorg. Med. Chem. Lett., 1998, 8,
1113; (b) F. Badalassi, D. Wahler, G. Klein, P. Crotti and J.-L. Reymond,
Angew. Chem. Int. Ed., 2000, 39, 4067.
9 (a) R. P. Taylor, V. Chau, C. Bryner and S. Berga, J. Am. Chem. Soc.,
1975, 97, 1934; (b) R. P. Taylor, V. Chau, C. Bryner and S. Berga, J.
Am. Chem. Soc., 1975, 97, 1943; (c) R. P. Taylor, J. Am. Chem. Soc.,
1976, 98, 2684; (d) R. P. Taylor and A. Silver, J. Am. Chem. Soc., 1976,
98, 4650.
26 S. D. Rychnovsky and D. J. Skalitzky, Tetrahedron Lett., 1990, 31, 945;
D. A. Evans, D. L. Rieger and J. R. Gage, Tetrahedron Lett., 1990, 31,
7099.
27 D. A. Evans, K. T. Chapman and E. M. Carreira, J. Am. Chem. Soc.,
1988, 110, 3560.
28 P. E. Stein, A. G. Leslie, J. T. Finch and R. W. Carrell, J. Mol. Biol.,
1991, 221, 941.
29 J. U. Lowe, Jr. and L. N. Ferguson, J. Org. Chem., 1965, 30, 3000.
30 See, for example: (a) E. Karnaukhova, Biochem. Pharmacol., 2007, 73,
901; (b) F. Zsila, Z. Bika´di and M. Simonyi, Tetrahedron: Asymmetry,
2002, 13, 273.
31 B. Loun and D. S. Hage, Anal. Chem., 1994, 66, 3814.
32 H. Li, A. D. Robertson and J. H. Jensen, Proteins: Struct., Funct.,
Bioinf., 2005, 61, 704.
33 T. Svedberg and B. Sjo¨gren, J. Am. Chem. Soc., 1930, 52, 2855.
34 (a) A. D. Becke, Phys. Rev. A: At., Mol., Opt. Phys., 1988, 38, 3098;
(b) C. Lee, W. Yang and R. G. Parr, Phys. Rev. B:, 1988, 37, 785.
35 M. T. Cance`s, B. Mennucci and J. Tomasi, J. Chem. Phys., 1997, 107,
3032.
36 SYBYL 7.3, Tripos Inc., 1699 South Hanley Rd., St. Louis, Missouri,
63144, USA.
37 (a) P. J. Goodford, J. Med. Chem., 1985, 28, 849;
(b) GRID22b’Molecular Discovery Ltd. www.moldiscovery.com.
38 (a) Tripos Bookshelf 7.3, Tripos International, 1699 South Hanley Rd.,
St. Louis, Missouri, 63144, USA; (b) M. Saunders, K. N. Houk, W.
Yun-Dong, W. C. Still, M. Lipton, G. Chang and W. C. Guida, J. Am.
Chem. Soc., 1990, 112, 1419.
39 (a) M. Woberg and M. Mu˝ller, Synthesis, 2006, 557; (b) A. H. Hoveyda,
D. A. Evans and G. C. Fu, Chem. Rev., 1993, 93, 1307.
40 (a) L. B. Davin and N. G. Lewis, Curr. Opin. Biotechnol., 2005, 16, 398;
(b) B. Pickel, M.-A. Constantin, J. Pfannistiel, J. Conrad, U. Beifuss
and A. Schaller, Angew. Chem. Int. Ed., 2010, 49, 202.
41 B. R. Davis, M. G. Hinds and P. C. Ting, Aust. J. Chem., 1992, 45, 865.
42 C. Scho¨pf and K. Thierfelder, Justus Liebigs Ann. Chem., 1935, 518,
127.
43 (a) D. A. Oare, M. A. Henderson, M. A. Sanner and C. H. Heathcock,
J. Org. Chem., 1990, 55, 132; (b) V. A. Martin, D. H. Murray, N.
E. Pratt, Y. Zhao and K. F. Albizati, J. Am. Chem. Soc., 1990, 112,
6965.
44 A. Patti and S. Pedotti, Tetrahedron: Asymmetry, 2006, 17, 778.
45 T. K. Khatab, S. S. Elmorsy and D. S. Badawy, Phosphorus, Sulfur
Silicon Relat. Elem., 2005, 180, 109.
10 (a) F. Hollfelder, A. J. Kirby and D. S. Tawfik, Nature, 1996, 383, 60;
(b) K. Kikuchi, S. N. Thorn and D. Hilvert, J. Am. Chem. Soc., 1996,
118, 8184; (c) F. Hollfelder, A. J. Kirby, D. S. Tawfik, K. Kikuchi and
D. Hilvert, J. Am. Chem. Soc., 2000, 122, 1022; (d) L. C. James and
D. S. Tawfik, Protein Sci., 2001, 10, 2600; (e) Y. Hu, K. N. Houk, K.
Kikuchi, K. Hotta and D. Hilvert, J. Am. Chem. Soc., 2004, 126, 8197;
(f) G. Boucher, S. Robin, V. Fargeas, T. Dintinger, M. Mathe´-Allainmat,
J. Lebreton and C. Tellier, ChemBioChem, 2005, 6, 807.
11 See for example: S. S. Singh and J. Mehta, J. Chromatography B, 2006,
834, 108 and references therein.
12 S. V. Dzyuba and A. M. Klibanov, Tetrahedron: Asymmetry, 2004, 15,
2771.
1998 | Org. Biomol. Chem., 2011, 9, 1987–1999
This journal is
The Royal Society of Chemistry 2011
©

View Online
46 W. Bell, J. A. Crayston, C. Glidewell, M. A. Mazid and M. B.
Hursthouse, J. Organomet. Chem., 1992, 434, 115.
47 W. Hao, Y. Zhang and T. Ying, Synth. Commun., 1996, 26, 2421.
48 M. R. Detty, J. Org. Chem., 1979, 44, 2073.
49 A. Abate, E. Brenna, G. Fronza, C. Fuganti, S. Ronzani and S. Serra,
Helv. Chim. Acta, 2003, 86, 592.
50 K. Ahmad, S. Koul, S. C. Taneja, A. P. Singh, M. Kapoor, R. ul-Hassan,
V. Verma and G. N. Qazi, Tetrahedron: Asymmetry, 2004, 15, 1685.
51 M. Edin, J. Steinreiber and J.-E. Ba¨ckvall, Proc. Natl. Acad. Sci. U. S.
A., 2004, 101, 5761.
52 (a) Y. Kobayashi and M. Matsuumi, J. Org. Chem., 2000, 65, 7221; (b) T.
Cohen, I.-H. Jeong, B. Mudryk, M. Bhupaty and M. M. A. Awed, J.
Org. Chem., 1990, 55, 1528.
53 (a) C. Schneider, M. Hansch and T. Weide, Chem.–Eur. J., 2005, 11,
3010; (b) G. E. Keck, C. A. Wager, T. Sell and T. T. Wager, J. Org.
Chem., 1999, 64, 2172; (c) E. Vedejs, S. M. Duncan and A. R. Haight,
J. Org. Chem., 1993, 58, 3046; (d) T. H. Chan and K. T. New, J. Org.
Chem., 1992, 57, 6107.
54 R. Todesco, D. Van Bockstaele, J. Gelan, H. Martens and J. Put, J. Org.
Chem., 1983, 48, 4963.
55 T. Hoffmann, G. Zhong, B. List, D. Shabat, J. Anderson, S. Gra-
matikova, R. A. Lerner and C. F. Barbas III, J. Am. Chem. Soc., 1998,
120, 2768.
56 Gaussian 03, Revision C.02, M. J. Frisch, G. W. Trucks, H. B. Schlegel,
G. E. Scuseria, M. A. Robb, J. R. Cheeseman, J. A. Montgomery, Jr.,
T. Vreven, K. N. Kudin, J. C. Burant, J. M. Millam, S. S. Iyengar, J.
Tomasi, V. Barone, B. Mennucci, M. Cossi, G. Scalmani, N. Rega,
G. A. Petersson, H. Nakatsuji, M. Hada, M. Ehara, K. Toyota,
R. Fukuda, J. Hasegawa, M. Ishida, T. Nakajima, Y. Honda, O.
Kitao, H. Nakai, M. Klene, X. Li, J. E. Knox, H. P. Hratchian, J.
B. Cross, V. Bakken, C. Adamo, J. Jaramillo, R. Gomperts, R. E.
Stratmann, O. Yazyev, A. J. Austin, R. Cammi, C. Pomelli, J. W.
Ochterski, P. Y. Ayala, K. Morokuma, G. A. Voth, P. Salvador, J. J.
Dannenberg, V. G. Zakrzewski, S. Dapprich, A. D. Daniels, M. C.
Strain, O. Farkas, D. K. Malick, A. D. Rabuck, K. Raghavachari, J. B.
Foresman, J. V. Ortiz, Q. Cui, A. G. Baboul, S. Clifford, J. Cioslowski,
B. B. Stefanov, G. Liu, A. Liashenko, P. Piskorz, I. Komaromi, R.
L. Martin, D. J. Fox, T. Keith, M. A. Al-Laham, C. Y. Peng, A.
Nanayakkara, M. Challacombe, P. M. W. Gill, B. Johnson, W. Chen, M.
W. Wong, C. Gonzalez and J. A. Pople, Gaussian, Inc., Wallingford CT,
2004.
57 (a) H. M. Berman, K. Henrick and H. Nakamura, Nat. Struct. Biol.,
2003, 10, 980; (b) http://www.pdb.org/.
58 M. Levit and S. Lifson, J. Mol. Biol., 1969, 46, 269.
59 M. Clark, R. D. Cramer III and N. Van Optenbosch III, J. Comput.
Chem., 1989, 10, 982.
This journal is
The Royal Society of Chemistry 2011
Org. Biomol. Chem., 2011, 9, 1987–1999 | 1999
©