Asymmetric Strecker reaction of N-benzhydrylimines utilising new tropos biphenyldiol-based ligands

Article abstract of DOI:10.1002/ejoc.200700763

The synthesis of a library of N-arenesulfonyl-1,3-oxazolidinyl-substituted biphenyldiols is presented. A set of two central intermediates together with easily accessible N-arenesulfonylamino alcohols furnish a broad variety of 1,3-oxazolidines. These are applied as chiral tropos ligands in a titanium-mediated Strecker-type reaction of N-benzhydrylimines. A correlation between the ee values in the product and the diastereomeric ratio concerning the chiral axis of the ligand is made. Those substituents in the ligand which proved to lead to a higher preference for one diastereomeric conformer of the chiral axis in related compounds now provide the most selective ligands. Two privileged ligands are identified that afford superior results in 8 of 13 screenings. Wiley-VCH Verlag GmbH & Co. KGaA, 2008.

Full text of DOI:10.1002/ejoc.200700763

FULL PAPER
DOI: 10.1002/ejoc.200700763
Asymmetric Strecker Reaction of N-Benzhydrylimines Utilising New Tropos
Biphenyldiol-Based Ligands
Stefan Wünnemann,[a] Roland Fröhlich,[a] and Dieter Hoppe*[a]
Dedicated to Professor Volker Jäger on the occasion of his 65th birthday
Keywords: Oxazolidines / Atropisomerism / Ligand design / Hydrocyanation / Titanium
The synthesis of a library of N-arenesulfonyl-1,3-oxazolid-
inyl-substituted biphenyldiols is presented. A set of two cen-
tral intermediates together with easily accessible N-arene-
made. Those substituents in the ligand which proved to lead
to a higher preference for one diastereomeric conformer of
the chiral axis in related compounds now provide the most
sulfonylamino alcohols furnish a broad variety of 1,3-oxazolid- selective ligands. Two privileged ligands are identified that
ines. These are applied as chiral tropos ligands in a titanium-
mediated Strecker-type reaction of N-benzhydrylimines. A
correlation between the ee values in the product and the dia-
stereomeric ratio concerning the chiral axis of the ligand is
afford superior results in 8 of 13 screenings.
(© Wiley-VCH Verlag GmbH & Co. KGaA, 69451 Weinheim,
Germany, 2008)
ated Strecker-type reaction of N-benzhydrylimines. We re-
cently disclosed the synthesis of a library of phosphite li-
gands 2 (Figure 1) based on chiral 1,3-oxazolidinyl-substi-
tuted biphenyldiols 1,[7] and we now wish to report an ex-
tension of the library of the intermediate diols. Further-
more, our earlier findings concerning the asymmetric induc-
tion of the 1,3-oxazolidines onto the chiral axis of the bi-
phenyl moiety will find application in this study.
Introduction
The asymmetric hydrocyanation of imines (Strecker reac-
tion)[1] has widely been used to furnish enantioenriched α-
amino acids.[2] A number of organocatalytic[3] approaches
have been developed, as well as several metal-catalysed[4]
processes. Therein, the addition of cyanide to benzhy-
drylimines plays an important role, as conversion into
amino acids is accomplished most simply in one single
step.[4b] Often, trimethylsilyl cyanide has been used as a
convenient source of hydrogen cyanide. Within the metal-
catalysed processes, the atropos binol derivatives play an
important role, but only one protocol was found utilising
tropos 2,2Ј-biphenyldiols (BIPOL).[5] Usually, chiral ligands
such as TADDOL and BINOL derivatives were required as
asymmetric activators. To the best of our knowledge, there
is no example of an asymmetric metal-catalysed Strecker
reaction with the use of chiral tropos ligands. One major
advantage of our concept, which uses tropos ligands, is the
fact that resolution of enantiomeric or diastereomeric forms
of the ligands is not necessary in contrast to atropos sys-
tems.[6]
Figure 1. Biphenyldiol 1 and related phosphite 2.
In an extensive study, we could demonstrate the influence
of each substituent in phosphite 2 on the diastereomeric
ratio of the conformers of 2.[7] Besides the influence of resi-
due R1, the bulkiness of the R2 and R3 substituents clearly
correlate with the preference for one diastereoisomer; par-
ticularly, a change in R2 from a para-methylphenyl group
to a 2,4,6-trimethylphenyl residue strongly enhanced the
population of the preferred diastereomer.
In this work, the impact of those substituents, which
clearly influence the diastereomeric ratio in phosphite 2, on
the selectivity induced by related biphenyldiols 1 will be in-
vestigated.
In our program on the application of tropos 3,3Ј-oxazol-
idinyl-substituted 2,2Ј-biphenyldiols 1, we investigated their
asymmetric induction in a titanium tetraisopropoxide medi-
[a] Organisch-Chemisches Institut, Westfälische Wilhelms-Uni-
versität Münster,
Corrensstr. 40, 48149 Münster, Germany
Fax: +49-251-8336531
E-mail: dhoppe@uni-muenster.de
Supporting information for this article is available on the
WWW under http://www.eurjoc.org or from the author.
684
© 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Org. Chem. 2008, 684–692

Asymmetric Strecker Reaction of N-Benzhydrylimines
Table 1. Synthesis of 1,3-oxazolidinyl-substituted ligands 1.
Results and Discussion
Our library of ligands 1 contains 27 members (Table 1)
and can be divided into different groups according to the
substitution pattern. The first distinction can be made re-
garding the biaryl backbone: two different R1 substituents
were introduced (these refer to the first letter added to the
compound numbers 1 and 10). A further division into sub-
groups can be made according to the arenesulfonyl group
R2. Four different substituents were introduced successfully,
and a fifth failed (the second letter added to the compound
numbers 1 and 10 divides these). Finally, ten different com-
binations of R3/R4 were introduced (distinguished by the
last letter added to compound numbers 1 and 10). Not all
combinations were synthesised and it appears helpful to di-
vide the library into three major groups (1aax, R1 = tBu,
R2 = Tol; 1abx, R1 = tBu, R2 = Mes; and 1bbx, R1 = H,
R2 = Mes;) and a few single members. The synthesis of
ligands 1 (Scheme 1) was improved in comparison to our
earlier report[7] and the number of variable sites was raised
by one, namely, the 5- and 5Ј-position on the biaryl back-
bone. The library contains two different substitution pat-
terns on the biaryl with R1 = tert-butyl and R1 = H; there-
fore, two different syntheses of central intermediate dialde-
hyde 7 were required.
The synthesis of tert-butyl derivative 7a commenced with
oxidative coupling of di-tert-butylphenol 2 in 88% yield[8]
and selective removal of the ortho-tert-butyl groups in excel-
lent 98% yield[9] after recrystallisation. Biphenol 3 was then
converted into carbamate 4 nearly quantitatively. In con-
trast to our previously reported synthesis,[7] we now applied
the Snieckus protocol[10] for double ortho lithiation of car-
bamate 4, as this proved more reliable on a multigram scale.
The slight drawback of the diethylcarbamoyl group,
namely, the more harsh conditions of its removal, does not
[a] The alphabetical keys refer to R1 (first letter), R2 (second letter)
and R3/R4 (third letter). [b] Tol = 4-methylphenyl. [c] Mes = 2,4,6-
trimethylphenyl. [d] Trip = 2,4,6-triisopropylphenyl. [e] Tbp = 4-
tert-butylphenyl. [f] As = 4-methoxyphenyl. [g] (S) configured, de-
rived from isoleucine.
Scheme 1. Synthesis of the 1,3-oxazolidinyl-substituted ligands 1 (see Table 1 for substituent key).
Eur. J. Org. Chem. 2008, 684–692
© 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
685

S. Wünnemann, R. Fröhlich, D. Hoppe
FULL PAPER
play any role here, as in this particular case the formyl
Finally, removal of the acetyl groups under mild condi-
group assists in the hydrolysis by a neighbouring group ef- tions in usually very high yields afforded desired diol li-
fect. Hence, the carbamate is cleaved under the basic condi- gands 10. The sometimes decreased yield in this deprotec-
tions of methanolic workup to furnish dialdehyde 7a in tion step might be due to partial hydrolysis of the oxazolid-
90% yield. The synthesis of unsubstituted dialdehyde 7b ines, as in some cases amino alcohol 9 was detected.
started from commercially available 2,2Ј-dihydroxybiphenyl
The N-arenesulfonylamino alcohols (Table 2) were pre-
(5) by a similar route. Nearly quantitative conversion into pared as shown in Scheme 3.[12] The synthesis of 9aj, 9bj
isopropyl carbamate 6, followed by the ortho lithiation se- and 9ej from norephedrine is a three-step sequence with
quence developed in our group[11] gave desired aldehyde 7b inversion of the stereocentre at C-1 by oxidation and dia-
in good yield.
stereoselective reduction.[13]
Acetylation of the hydroxy groups, in both cases in 99%
yield, afforded starting materials 8 for the synthesis of the
1,3-oxazolidines. It is noteworthy that not only the first
steps were conducted up to a 1-mol scale, but the lithiations
can also be performed on a 50-mmol scale for stockpiling
of the intermediates. The key step of our synthesis, the for-
mation of the chiral 2,4-cis-1,3-oxazolidines, proceeds in the
presence of 3 equiv. of N-sulfonylamino alcohol 9 and
5 equiv. of dichlorodimethylsilane[12] in refluxing toluene.
Depending on the substituents in amino alcohol 9, yields
up to 99% of product 10 could be isolated (Table 1). It
turned out that an excess amount of amino alcohol 9 and
the silane, as well as high concentrations and prolonged re-
action times are beneficial to the yield. In some cases, con-
version was nearly complete after 1 d, whereas in one case
(Table 1, Entry 18) the reaction was run over 5 d and still
afforded only 46% yield. In several cases, significant
amounts of intermediate monooxazolidine 11 were isolated,
which could be recycled (Scheme 2). The acetyl group in the
salicylic aldehyde moiety that was cleaved upon workup was
reinstalled to give 12, and its reaction with an excess
amount of amino alcohol 9 afforded additional quantities
of the product. In general, the reactivity decreases with in-
creasing steric demand of the arenesulfonyl group. With a
tosyl- or 4-tert-butylbenzenesulfonyl group excellent yields
were obtained (Table 1, Entries 1–9, 21–22), whereas the
more bulky mesitylene sulfonyl group (compounds 9xbx;
Table 2. Synthesis of amino alcohols 9.
Entry
R2
R3
R4
Product 9[a] Yield [%]
1
2
3
4
5
6
7
8
Tol
Tol
Tol
Tol
Tol
Tol
Tol
Tol
Me
Et
iPr
iBu
sBu
tBu
Ph
H
H
H
H
H
H
H
H
H
Ph
H
H
H
H
H
H
H
H
Ph
H
H
H
H
Ph
9aa (92)
ent-9ab[12]
9ac (95)
9ad (96)
9ae (90)
9af (99)
ent-9ag (95)
9ah (92)
9ai (90)
9aj (57, 3 steps)
9ba (97)
ent-9bb[14]
9bc (86)
9bd (87)
9be (89)
9bf (99)
Bn
9
Tol
Tol
(CH2)2SCH3
Me
Me
Et
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
Mes
Mes
Mes
Mes
Mes
Mes
Mes
Mes
Mes
Trip
Tbp
Tbp
As
iPr
iBu
sBu
tBu
Ph
Bn
Me
Et
Et
sBu
Ph
ent-9bg (99)
9bh (95)
9bj[15]
ent-9cb[16]
ent-9 db (97)
9de (87)
ent-9eg[15]
9ej[15]
As
Me
[a] The alphabetical keys refer to R2 (first letter) and R3/R4 (second
letter).
With a set of 27 ligands in hands, we then turned our
Table 1, Entries 11–18, 25–31) led to moderate-to-excellent focus on exploring their application in asymmetric catalysis.
yields, depending on the substituents R3 and R4; in one case Because titanium forms strong bonds to oxygen-containing
(Table 1, Entry 32), no product could be isolated. The most ligands and because titanium tetraisopropoxide is readily
bulky triisopropylbenzenesulfonyl group did not give any available as a precursor for the active complex, we sought
desired products 8 (Table 1, Entry 20). To our surprise, the titanium-mediated reactions. It was our hope that two of
reaction of norpseudoephedrine with a para-methoxyben- the isopropoxide residues might be easily replaced by our
zenesulfonyl group attached to it readily proceeded, but the diol ligands 1 for entropic reasons. Furthermore, the tita-
same sulfonyl substituent on phenyl glycine did not lead to nium atom provides additional ligation sites, which can be
any product (compare Table 1, Entries 10, 19, 23 and 24).
occupied by a substrate molecule. In our expectation, the
Scheme 2. Conversion of intermediates 11 into 1,3-oxazolidines 10.
686
www.eurjoc.org
© 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Org. Chem. 2008, 684–692

Asymmetric Strecker Reaction of N-Benzhydrylimines
without any ligand nearly no conversion was observed. Sec-
ond, we investigated the ideal reaction conditions: (1) the
solvent of choice is toluene; (2) the best enantiocontrol is
obtained at 0 °C, as higher and lower temperatures led to a
decrease in the ee values;[17] (3) the catalyst complex is
formed in situ at room temperature, and elevated tempera-
tures and the addition of amines[4e] are not required to as-
sist the complex formation Ϫ both led to poor selectivity
and reaction rate. With the optimum setup for the enantio-
selective Strecker-type reaction, we then turned our atten-
tion to the screening of ligands. The conversion in all reac-
tions was nearly complete after 18 h; only traces of starting
material were detected by HPLC. Only the naphthaldimines
required a prolonged reaction time of 72 h; the addition of
2-propanol was also carried out over the same duration.
For each substrate, Table 3 shows the best ligand and the
isolated yield after chromatography at full conversion.
Scheme 3. Synthesis of amino alcohols 9 (see Table 2 for substitu-
ent key).
Table 3. Enantioselective Strecker reaction of N-benzhydrylimines
17.
steric demand of the remaining isopropoxy ligands should
lead to enhanced asymmetric induction of the 1,3-oxazol-
idines onto the chiral axis. This would be in good agreement
with our previous observations on the diastereomeric ratio
in phosphites 2 derived from oxazolidines 1 (Scheme 4).[7]
Entry
R
Product 18
ee[b]
Ligand 1
Yield [%][a]
(abs. config.)[c]
1
2
3
4
5
6
7
8
Ph
18a (94)
ent-18b (92)
18c (83)
18d (67)
18e (87)
89 (R)
68 (S)
59 (R)
48 (S)
84 (R)
20(R)
62 (R)
80 (R)
72 (R)
85 (R)
55 (S)
72 (S)
63 (R)
1abf
ent-1abg
1bbd
1aaf
1abc
1abd
1abe
1bba
1abc
4-OMePh
3-OMePh
2-OMePh
4-ClPh
3-ClPh
2-ClPh
4-MePh
3-MePh
2-MePh
1-naphthyl[d]
2-naphthyl[d]
tBu
18f (87)
18g (93)
18h (92)
18i (83)
9
10
11
12
13
18j (91)
1abc
ent-18k (34)
ent-18l (99)
18m (86)
ent-1abb
ent-1adb
1abe
[a] Isolated yield after chromatography; only traces of starting ma-
terial were left in the crude reaction mixture by HPLC. [b] Deter-
mined by chiral HPLC. [c] Determined by comparison with litera-
ture values of the optical rotation.[4b,18,19] [d] Prolonged reaction
time and duration of the addition of the 2-propanol solution: 72 h.
Scheme 4. Titanium-mediated Strecker reactions of N-benzhy-
drylimines 17 (see Table 3 for substituent key).
The results from the first substrate, N-benzhydrylbenzald-
imine (17a), are shown in Figure 2. The best results were
obtained from ligands 1abc (R1 = tBu, R2 = Mes, R3 = iPr)
(87%ee) and 1abf (R1 = tBu, R2 = Mes, R3 = tBu)
First, we turned our attention to the development of
ideal reaction conditions. With trimethylsilyl cyanide as the
reagent it was reported by other researchers[4] that slow ad-
dition of protic additives is beneficial to the reaction rate
and enantioselectivity. Our own findings support these ob-
servations with a 2-propanol solution as the superior addi-
tive. The addition of alcohol should liberate hydrogen cya-
nide, which is the active cyanating agent. Additionally, the
remaining alkoxide might bind to the TMS cation, which
would therefore make it less prone to catalyse the reaction
competitively to the titanium. The optimum was found to
be an addition period of 18 h; slower additions did not im-
prove the enantioselectivity. However, the addition must not
be quicker than the Ti-catalysed reaction of the hydrogen
cyanide with the imine to suppress a nonselective reaction
mediated by free hydrocyanic acid. Another important fact
for this reaction is to prove ligand acceleration. Indeed, Figure 2. Enantiomeric excess of nitrile 18a.
Eur. J. Org. Chem. 2008, 684–692
© 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
687

S. Wünnemann, R. Fröhlich, D. Hoppe
FULL PAPER
(89%ee), other good ee values were obtained with ligands ligand 1abc (R1 = tBu, R2 = Mes, R3 = iPr) furnished the
1abe (R1 = tBu, R2 = Mes, R3 = sBu) (80%ee) and 1bbc product in good 84%ee, whereas 1abe (R1 = tBu, R2
(R1 = H, R2 = Mes, R3 = iPr) (74%ee). Mes, R3 = sBu) gave 75%ee. These results again follow the
=
As a general trend, those ligands bearing the more bulky general trend: the more bulky substituents provided better
mesitylene sulfonyl group in the 1,3-oxazolidine moiety selectivity; also, the unsubstituted biaryl backbone (ligands
(compounds 1xbx) seem to afford superior results. This is 1bxx) gave only moderate results (decrease in ee around
in good agreement with our expectation that the induction 20%). Once more, ligands 1abf and 1bbf did not follow this
onto the chiral axis of the ligand should be higher with rule. The set of reactions with meta-chloro substrate 17f
sterically more-demanding sulfonyl substituents. Addition- displayed a similar trend, but the best ee obtained was only
ally, the more branched side groups on C-4 of the oxazolid- 20% [compounds 1abd (R1 = tBu, R2 = Mes, R3 = iBu)
ine give better enantioselectivity. Secondly, the R1 group on and 1bbe (R1 = H, R2 = Mes, R3 = sBu)]; several ligands
the biaryl backbone of the ligand seems to have some influ- afforded just racemates. Better enantioselectivity was ob-
ence; those derivatives with a tert-butyl substituent (com- tained from the reactions of ortho-chlorobenzaldimine 17g
pounds 1axx) display a trend to higher ee values.
[ligand 1abe (R1 = tBu, R2 = Mes, R3 = sBu) 62%ee, li-
Next, we examined the substrate scope of the reaction. gands ent-1bbb (R1 = H, R2 = Mes, R3 = Et) 51%ee and
A number of ortho-, meta- and para-substituted N-benzhy- 1bbc (R1 = H, R2 = Mes, R3 = iPr) 56%ee]. The other
drylbenzaldimines were subjected to the Strecker-type reac- reactions furnished results within the general rule. Surpris-
tion. Further substrates were derived from 1- and 2-naphth- ingly, with substrate 17g, ligand 1aej (R1 = tBu, R2 = As,
aldehyde and pivaldehyde. The results (Supporting Infor- R3 = Me, R4 = Ph) afforded the product in 60%ee, and in
mation, Figures S1–S12) demonstrate: (1) a strong depen- all other reactions this oxazolidine performed much worse.
dence of the enantioselectivity from the substitution pattern
Next, we turned our attention to methyl-substituted sub-
in the substrate, (2) a preference for the mesitylene sulfonyl strates (Supporting Information, Figures S7–S9) by as-
group as R2 in the ligand (only two exceptions), (3) moder- suming that these should electronically be very close to un-
ate-to-good enantioselectivity at nearly complete conver- substituted N-benzhydrylbenzaldimine 17a, but behave dif-
sion for all substrates and (4) a superior lead structure ex- ferently for steric reasons. The reaction of para-methyl-sub-
ists, which still requires a ligand library.
stituted imine 17h was the most selective in the presence of
The first variation in the substrate was the introduction ligand 1bba (R1 = H, R2 = Mes, R3 = Me) (80%ee); ee
of a methoxy group, which provided a more electron-rich values above 70%ee were observed for ligands ent-1bbb (R1
imine (Supporting Information, Figures S1–S3). In the re- = H, R2 = Mes, R3 = Et), 1bbd (R1 = H, R2 = Mes, R3 =
action of para-substituted imine 17b, the best selectivity iBu), 1bbe (R1 = H, R2 = Mes, R3 = sBu) and ent-1bbg (R1
(68%ee) was achieved with three different ligands, [1abf (R1 = H, R2 = Mes, R3 = Ph). In good agreement with the
= tBu, R2 = Mes, R3 = tBu), 1abg (R1 = tBu, R2 = Mes, results from the para-methoxy substrate, the slightly less
R3 = Ph) and ent-1bbb (R1 = H, R2 = Mes, R3 = Et)]. bulky residues R3 furnished higher enantioselectivities, but
Again, the more bulky arenesulfonyl group at the oxazolid- this time the unsubstituted biaryl backbone performed bet-
ine moiety performed superiorly, but in contrast, the ter. meta-Methylbenzaldimine 17i displayed behaviour ac-
smaller R3 substituents furnish the better results; ligand cording to the general trend, and the best result was ob-
1abf is the exception to this trend and proved to behave tained for ligand 1abc (R1 = tBu, R2 = Mes, R3 = iPr)
differently with other substrates as well (vide infra). meta- (72%ee). The results from ortho-methyl substrate 17j are
Methoxybenzaldimine 17c gave moderate selectivity comparable to those of the same ligand that gave the best
(59%ee), but with this substrate, the same trends as those selectivity, and it afforded a good enantiomeric excess of
of the parent benzaldimine were observed: the more bulky 85%.
the residues, the better the ee. As indicated earlier, ligands
We then focused on 1- and 2-naphthaldimines (Support-
1abf and 1bbf (R3 = tBu) are an exception, as they provide ing Information, Figures S10 and S11). Moderate-to-good
very poor enantiomeric excesses. The next result, the reac- selectivities were observed for both substrates, and the li-
tion of ortho-methoxybenzaldimine 17d, was even more sur- gand performance was found to follow the general trend.
prising to us. The general trend for all other substrates that 1-Naphthaldehyde-derived substrate 17k gave only 55%ee
the mesitylene sulfonyl group as R2 gave the best selectivity [ligand ent-1abb (R1 = tBu, R2 = Mes, R3 = Et)], whereas
is reversed: the tosyl-substituted ligands provided superior isomeric imine 17l afforded the product in 72%ee. Surpris-
results, and ligand 1aaf (R1 = tBu, R2 = Tol, R3 = tBu) ingly, the latter result was obtained with ligand 1adb (R1 =
gave a moderate 48%ee as the highest selectivity. However, tBu, R2 = Tbp, R3 = Et), which bears a medium-sized 4-
the low enantiocontrol in this set might be due to a mis- tert-butylbenzenesulfonyl group. Finally, we were interested
matched correlation of the induction within the ligand in the pivaldimine as an entirely different substrate (Sup-
structure (chiral axis vs. oxazolidine moieties) for this sub- porting Information, Figure S12). We were pleased to find
strate; this is a plausible explanation for the unexpected re- that moderate-to-good enantioselectivities could be
versal of the general trend.
achieved in the reaction of imine 17m, as often those li-
The next set of reactions was carried out on the more gands affording good selectivity with aromatic imines per-
electron-poor chloro-substituted imines (Supporting Infor- form much worse with nonaromatic derivatives. Once again,
mation, Figures S4–S6). With para-chlorobenzaldimine 17e, ligand 1abe (R1 = tBu, R2 = Mes, R3 = sBu) furnished
688
www.eurjoc.org
© 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Org. Chem. 2008, 684–692

Asymmetric Strecker Reaction of N-Benzhydrylimines
product 18m in the best ee of this set (63%). Two further fore, the product-determining face discrimination becomes
ligands produced ee values greater than 50% [oxazolidines more unselective. On the other hand, given that the oxazol-
1abc (R1 = tBu, R2 = Mes, R3 = iPr) and 1abd (R1 = tBu, idine moiety shields one face of the substrate, the reaction
R2 = Mes, R3 = iBu)].
must be more efficient with more bulky substituents in the
To conclude the results from this screening, a general 1,3-oxazolidine.
trend concerning the relation between substitution pattern
in the ligand and ee in the product is observed to result
from two leading ligands: 1abc (R1 = tBu, R2 = Mes, R3 =
iPr) and 1abe (R1 = tBu, R2 = Mes, R3 = sBu). In 8 out of
13 screening sets, one or both oxazolidines afforded the best
or slightly lower second best ee in the product. By compar-
ing the substrates, it can be noted that substitution of the
phenyl ring in the meta position leads to lower enantio-
selectivities with a particularly strong effect of chlorine. The
reactions of naphthyl- and tert-butyl imines 17 proceed with
moderate ee values.
The absolute configuration in product 18 in all cases is
Figure 4. X-ray-crystallographic structure of ligand ent-1aab.
(R) when (2S,4S)-configured oxazolidines 1 were used as
ligands, and (S) with their optical antipodes ent-1; the ex-
ception is ortho-methoxybenzaldimine 17d. Figure 3 might
explain the observed selectivity where the Si face of the sub-
strate is more open to the approaching nucleophile; the Re
face is slightly shielded by the Ϫ in this view rear Ϫ 1,3-
Conclusions
In this report, the synthesis of an extended ligand library
oxazolidine moiety. A possible explanation for the behav- is disclosed. A set of 27 1,3-oxazolidine-substituted biphen-
iour of substrate 17d is a different binding mode of the sub- yldiol ligands was synthesised in good-to-excellent yields on
strate. Steric reasons seem not to play a role here, as ortho- a multigram scale. These chiral tropos ligands were used in
methylbenzaldimine 17j affords products with good ee val- a titanium-catalysed hydrocyanation of imines. A privileged
ues and the usually privileged ligands afford the best results. structure is identified with the bulky mesitylene sulfonyl
Similar observations are made with ortho-chloro substrate group and sterically more-demanding alkyl side chains in
17g. The oxygen atom of the methoxy group probably coor- the oxazolidine moiety. Both are expected, in agreement
dinates to the metal centre as well and, thus, the substrate with our earlier findings, to enhance the diastereomeric ra-
is a bidentate ligand to the titanium. The complex should tio concerning the chiral biaryl axis in the ligand. A broad
be significantly different from that shown in Figure 3 and, range of substrates was investigated, and moderate-to-good
hence, lead to the unexpected results. Such coordination is enantioselectivities were observed. Substituents in the meta
only possible for the oxygen atom of an ortho-methoxy position of N-benzhydrylbenzaldimine afford products with
group and therefore here is the sole exception.
lower ee values. Our results demonstrate that a ligand li-
brary may contain a privileged member, but the design of
fine-tunable ligands with several variable sites is still a
powerful tool in asymmetric catalysis to cover a broad sub-
strate range.
Experimental Section
General: Reactions were carried out in flame-dried glassware under
an atmosphere of argon, where appropriate. Toluene was distilled
from Na/benzophenone and THF from K/benzophenone; CH2Cl2
was dried with CaH2. TMEDA and NEt3 were distilled from CaH2
and stored under an atmosphere of argon in the dark. DMF was
stirred over CaH2, distilled under reduced pressure and stored over
4 Å molecular sieves. 2-Propanol was distilled from CaH2 and
stored over 4 Å molecular sieves. TIPT was distilled and stored
under an atmosphere of argon. All other reagents were used as
supplied by the commercial sources. Flash-column chromatography
(FCC) was carried out by using Merck 60 silica gel (40–60 µm,
230–400 mesh ASTM), and monitored by thin-layer chromatog-
raphy (TLC) on Merck 60 F254 TLC plates. NMR spectra were
recorded with Bruker ARX 300 and ARX 400 spectrometers.
Chemical shifts are relative to TMS (δ = 0.00 ppm). Calibration
Figure 3. Model for a substrate–catalyst complex.
From X-ray-crystallographic analysis (Figure 4) we know
that the chiral axis in free (2R,4R)-configured ligand ent-
1aab (R1 = tBu, R2 = Tol, R3 = Et) adopts the (S) configu-
ration. Together with our earlier findings in phosphites 2
that more bulky N-arensulfonyl groups in the oxazolidine
moiety strongly enhance the preference for the energetically
favoured diastereomer,[7] we can explain the general trend
that ligands 1xbx afford superior results. On the one hand,
in the unfavoured diastereomer of 1 the oxazolidine moie-
ties point towards each other and the metal centre. There- was performed by using the signals of TMS (δ = 0.00 ppm, 1H
Eur. J. Org. Chem. 2008, 684–692
© 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
689

S. Wünnemann, R. Fröhlich, D. Hoppe
FULL PAPER
NMR spectra) and CDCl3 (δ = 77.0 ppm, 13C NMR spectra). Peak
assignments were made by using routine 2D spectra; diastereotopic
methylene protons are named HA and HB. IR spectra were re-
corded with a Nicolet 5DCX, Bruker IFS 28 or Varian 3100 Excali-
bur Series with Specac Golden Gate Single Reflection ATR. Mass
spectra were recorded with a Bruker Micro-TOF and a Waters
Micromass Quatro-LC (both ESI) or a Finnigan MAT 8200 (EI)
spectrometer. Optical rotations were measured with a Perkin–El-
mer 341 apparatus. Melting points were measured with a Stuart
Melting Point Apparatus SMP3 (Bibby Sterlin Ltd) and are uncor-
rected. Combustion analyses were carried out with a Vario EL III
(Elementar-Analysensysteme GmbH). HPLC measurements were
performed with a Waters 600E Multisolvent Delivery System and
996 PDA detector or Knauer Smartline UV detector 2600, Pump
1000 and Manager 5000. As stationary phases were used: Chira
Grom 1 and Chira Grom 2 (2.0ϫ250 mm) by Grom, and Eurocel
(4.6ϫ250 mm) by Knauer.
tered off. The characterization of compound 7a was reported ear-
lier by us.[7]
2,2Ј-Bis(isopropylcarbamoyloxy)biphenyl (6): To a solution of 2,2Ј-
dihydroxybiphenyl (5; 55.80 g, 300.0 mmol, 1.0 equiv.) and DMAP
(1.098 g, 9.00 mmol, 0.03 equiv.) in THF (75 mL) was added iso-
propyl isocyanate (69 mL, 700.0 mmol, 2.3 equiv.), and the mixture
was stirred at 60 °C for 24 h. The solution was washed with aque-
ous HCl (2 , 145 mL), and the aqueous phase was extracted with
TBME (3ϫ50 mL). The combined organic layer was dried with
sodium sulfate, and the solvent was removed in vacuo. Crude car-
bamate 6 was purified by FCC (n-pentane/diethyl ether, 1:1) to
yield 6 (106 g, 99%) as a white solid, m.p. 123 °C, Rf = 0.29 (n-
pentane/diethyl ether, 1:1). 1H NMR (400 MHz, CDCl3): δ = 7.39–
3
7.34 (m, 2 H, aromatic), 7.30–7.22 (m, 6 H, aromatic), 5.23 (d, J
3
= 6.7 Hz, 2 H, NH), 3.67 (oct., J = 6.6 Hz, 2 H, CH), 1.05 (br. d,
3J = 5.7 Hz, 12 H, CH3) ppm. 13C NMR (100 MHz, CDCl3): δ =
154.1 (CO), 148.5 (C-2), 130.9 (CH), 130.7 (C-1), 128.6 (CH), 125.3
(CH), 122.8 (CH), 43.0 (CH), 22.5 (CH ) ppm. IR (ATR): ν = 3290
˜
3
5,5Ј-Di-tert-butyl-2,2Ј-bis(diethylcarbamoyloxy)biphenyl (4): To a
solution of 3 (18.49 g, 62.0 mmol, 1.0 equiv.) and DMAP (397 mg,
3.25 mmol, 0.05 equiv.) in THF (100 mL) was cautiously added a
suspension (60% wt.) of sodium hydride in mineral oil (9.10 g,
228.0 mmol, 3.5 equiv.) at 0 °C. After gas evolution had ceased, the
mixture was warmed to r.t. and diethylcarbamoyl chloride
(20.5 mL, 162.5 mmol, 2.5 equiv.) was added. The solution was
stirred at r.t. for 16 h, followed by the careful addition of aqueous
HCl (2 , 50 mL). The phases were separated, and the aqueous
phase was extracted with TBME (3ϫ30 mL). The combined or-
ganic layer was dried with sodium sulfate, and the solvent was evap-
orated. The remaining oil was filtered through a plug of silica, and
the solvent was removed in vacuo. After 2 d, white crystals had
formed, which were filtered and washed with cold n-pentane. The
mother liquor was subjected to FCC to give a combined amount
of a white solid (30.75 g, 62.0 mmol, 99%), m.p. 73 °C. Rf = 0.54
(n-pentane/diethyl ether, 1:1). 1H NMR (300 MHz, CDCl3): δ =
(br), 2975 (m), 1724 (s), 1698 (s), 1530 (s), 1238 (s), 1204 (s), 1172
(s), 1099 (s), 1029 (s), 762 (s) cm–1. MS (ESI): m/z (%) = 379.162
(100) [M + Na]+; calcd. 379.163. C20H24N2O4 (356.42): calcd. C
67.40, H 6.79, N 7.86; found C 67.39, H 6.55, N 7.90.
2,2Ј-Dihydroxybiphenyl-3,3Ј-dicarboxaldehyde (7b): To a stirred
solution of 5 (8.900 g, 25.00 mmol, 1.0 equiv.) in THF (200 mL)
was added TMEDA (8.9 mL, 60.0 mmol, 2.4 equiv.) and TMSOTf
(9.05 mL, 50.00 mmol, 2.00 equiv.) at 0 °C. The mixture was stirred
at r.t. for 30 min and then cooled to –78 °C. Again, TMEDA
(8.9 mL, 60.0 mmol, 2.4 equiv.) was added, followed by a solution
of tert-butyllithium (1.5  in pentane, 100 mL, 150.0 mmol,
6.0 equiv.). The mixture was stirred at –78 °C for 4 h and, sub-
sequently, DMF (11.7 mL, 150.0 mmol, 6.0 equiv.) was added
dropwise. The reaction mixture was warmed to r.t. before methanol
(30 mL) and a solution of NaOH (2  in water, 80 mL) were added.
The solution was stirred at r.t. for 3 h; then, aqueous HCl (2 , ca.
150 mL) was added until acidic pH (change of colour to bright
yellow). The phases were separated, and the aqueous phase was
extracted with dichloromethane (3ϫ80 mL). The combined or-
ganic layer was dried with sodium sulfate, and the solvent was re-
moved in vacuo. The remaining solid was suspended in TBME and
the residual yellow precipitate (4.590 g, 18.97 mmol, 76%) was fil-
tered off. M.p. 158 °C. Rf = 0.45 (CH2Cl2/MeOH, 500:1). 1H NMR
(300 MHz, CDCl3): δ = 11.43 (s, 2 H, OH), 9.96 (s, 2 H, CHO),
7.63 (dd, 3J = 8.0 Hz, 4J = 1.9 Hz, 2 H, 4-H), 7.61 (dd, 3J = 7.7 Hz,
4J = 1.9 Hz, 2 H, 6-H), 7.11 (t, 3J = 7.9 Hz, 2 H, 5-H) ppm. 13C
NMR (100 MHz, CDCl3): δ = 196.6 (CHO), 159.1 (C-2), 138.8 (C-
6), 133.7 (C-4), 125.1 (C-1), 120.9 (C-3), 119.5 (C-5) ppm. IR
4
4
7.34 (dd, 3J = 8.5 Hz, J = 2.5 Hz, 2 H, 4-H), 7.29 (d, J = 2.5 Hz,
3
2 H, 6-H), 7.16 (d, J = 8.4 Hz, 2 H, 3-H), 3.19 (br. s, 4 H, CH2),
3.08 (br. s, 4 H, CH2), 1.31 (s, 18 H, tBu), 1.01 (br. s, 6 H, CH3),
0.81 (br. s, 6 H, CH3) ppm. 13C NMR (75 MHz, CDCl3): δ = 154.3
(CO), 147.2 (C-2), 147.1 (C-5), 131.1 (C-1), 128.3 (C-4), 125.6 (C-
6), 122.5 (C-3), 41.6 (br., CH2), 41.35 (br., CH2), 34.2 (quart. tBu),
31.2 (tBu), 15.0 (br., CH ), 13.8 (br., CH ) ppm. IR (KBr): ν =
˜
3
3
2966 (s), 2873 (s), 1714 (s), 1472 (m), 1418 (m), 1211 (s), 1159 (s),
807 (m) cm–1. MS (EI, 70 eV): m/z (%) = 496 (1) [M]+, 100 (100).
C30H44N2O4 (496.68): calcd. C 72.55, H 8.93, N 5.64; found C
72.61, H 9.10, N 5.33.
5,5Ј-Di-tert-butyl-2,2Ј-dihydroxybiphenyl-3,3Ј-dicarbaldehyde (7a):
To a solution of TMEDA (20.8 mL, 139.5 mmol, 3.0 equiv.) in
THF (300 mL) at –78 °C was added a solution of tert-butyllithium
(1.5  in pentane, 100 mL, 150.0 mmol, 3.2 equiv.). The mixture
was stirred at this temperature for 15 min before a solution of 3
(23.064 g, 46.50 mmol, 1.0 equiv.) in THF (75 mL) was added over
20 min (syringe pump). The reaction mixture was held at –78 °C
for 90 min. Then, DMF (15.5 mL, 200 mmol, 4.3 equiv.) was added
dropwise into the solution over 20 min, and the reaction mixture
was warmed to r.t. over 16 h. Subsequently, methanol (30 mL) was
added and stirring was continued for 2 h. Then, aqueous HCl (2 ,
ca. 100 mL) was added until acidic pH was achieved (change of
colour to bright yellow). The phases were separated, and the aque-
ous phase was extracted with TBME (3ϫ100 mL). The combined
organic layer was dried with sodium sulfate, and the solvent was
evaporated. The remaining solid was suspended in n-pentane and
the residual yellow precipitate (18.860 g, 41.86 mmol, 90%) was fil-
(ATR): ν = 3535 (m), 3023 (br), 2871 (m), 1628 (s), 1612 (s), 1423
˜
(s), 1386 (s), 1293 (s), 1219 (s), 1064 (s), 727 (s) cm–1. MS (EI,
70 eV): m/z (%) = 242 (100) [M]+. C14H10O4 (242.23): calcd. C
69.42, H 4.14; found C 69.32, H 4.13.
2,2Ј-Diacetoxy-5,5Ј-di-tert-butylbiphenyl-3,3Ј-dicarboxaldehyde (8b):
To a stirred solution of 6b (4.590 g, 18.97 mmol, 1.0 equiv.) in
dichloromethane (80 mL) was added triethylamine (5.9 mL,
42 mmol, 2.2 equiv.) and acetyl chloride (3.0 mL, 42 mmol,
2.2 equiv.) at 0 °C. The mixture was stirred at r.t. for 2 h. Then,
TBME (80 mL) was added, and subsequently, the suspension was
filtered through a plug of silica. The filtrate was concentrated under
vacuum to give acetate 8b (6.182 g, 18.97 mmol, 99%) as a white
solid, m.p. 90 °C. Rf = 0.17 (n-pentane/diethyl ether, 1:1). 1H NMR
(400 MHz, CDCl3): δ = 10.07 (s, 2 H, CHO), 7.96 (dd, 3J = 8.0 Hz,
3
4
4J = 1.9 Hz, 2 H, 4-H), 7.56 (dd, J = 7.7 Hz, J = 1.9 Hz, 2 H, 6-
H), 7.50 (t, 3J = 7.6 Hz, 2 H, 5-H), 2.12 (s, 6 H, Ac-CH3) ppm.
690
www.eurjoc.org
© 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Org. Chem. 2008, 684–692

Asymmetric Strecker Reaction of N-Benzhydrylimines
13C NMR (100 MHz, CDCl3): δ = 188.2 (CHO), 168.7 (Ac-CO), 5,5Ј-Di-tert-butyl-3,3Ј-bis[(2S,4S)-4-methyl-3-(4-methylbenzene-
148.6 (C-2), 136.6 (C-6), 131.6 (C-4), 131.0 (C-1), 128.5 (C-3), 126.2 sulfonyl)-1,3-oxazolidinyl]biphenyl-2,2Ј-diol (1aaa): After FCC, the
(C-5), 20.3 (Ac-CH ) ppm. IR (ATR): ν = 2975 (m), 1670 (m), 1604 product was obtained in 78% yield, m.p. 106–109 °C. Rf = 0.08 (n-
˜
3
20
(m), 1455 (s), 1309 (s), 1227 (m), 1153 (s), 1081 (s), 952 (s), 669 (s) pentane/diethyl ether, 1:1). [α]
= –155.4 (c = 1.04, CHCl3). 1H
D
cm–1. MS (EI, 70 eV): m/z (%) = 326 (15) [M]+, 242 (100).
NMR (400 MHz, CDCl3): δ = 7.79 (d, J = 8.2 Hz, 4 H, 2ЈЈЈ-H),
7.60 (d, J = 2.6 Hz, 2 H, 4-H), 7.33 (d, J = 8.1 Hz, 4 H, 3ЈЈЈ-H),
7.28 (d, J = 2.4 Hz, 2 H, 6-H), 7.26 (s, 2 H, OH), 6.48 (s, 2 H, 2Ј-
H), 4.04 (sext., J = 6.3 Hz, 2 H, 4Ј-H), 3.84 (dd, J = 8.6 Hz, J
= 7.0 Hz, 2 H, 5Ј-HA), 3.56 (dd, J = 8.8 Hz, J = 5.9 Hz, 2 H, 5Ј-
HB), 2.43 (s, 6 H, Ts-CH3), 1.43 (d, J = 6.3 Hz, 6 H, 1ЈЈ-H), 1.34
3
4
3
C18H14O6 (326.30): calcd. C 66.26, H 4.32; found C 66.24, H 4.45.
4
General Procedure for the Synthesis of Oxazolidines 10: To a stirred
solution of 8 (1.0 equiv.) and N-sulfonyl-2-amino-1-alkanol 9
(3.0 equiv.) in toluene (10.0 mL per mmol) was added dichlorodi-
methylsilane (5.0 equiv.). The solution was heated at reflux under
an atmosphere of argon for 48 h and then cooled to r.t. Then, the
mixture was poured into saturated aqueous potassium carbonate
solution (10 mL per mmol) and washed. After phase separation,
the aqueous phase was extracted with TBME (3ϫ5 mL per mmol),
and the combined organic layers were dried with sodium sulfate.
The solvents were removed in vacuo, and the residue was purified
by FCC with mixtures of n-pentane and diethyl ether. The charac-
terisation of one representative example is given, for further com-
pounds, see Supporting Information.
3
2
3
2
3
3
(s, 18 H, tBu) ppm. 13C NMR (100 MHz, CDCl3): δ = 149.2 (C-
2), 144.2 (C-4ЈЈЈ), 143.3 (C-5), 134.5 (C-1ЈЈЈ), 129.9 (C-3ЈЈЈ), 129.7
(C-6), 128.0 (C-2ЈЈЈ), 126.0 (C-3), 125.5 (C-4), 123.4 (C-1), 90.0 (C-
2Ј), 71.8 (C-5Ј), 55.5 (C-4Ј), 34.3 (quart. tBu), 31.5 (tBu), 21.5 (Ts-
CH ), 20.2 (C-1ЈЈ) ppm. IR (ATR): ν = 2972 (m), 2875 (w), 1344
˜
3
(s), 1161 (s), 1095 (m), 815 (s), 667 (s) cm–1. MS (ESI): m/z (%) =
799.305 (100) [M + Na]+, calcd. 799.306. C42H52N2O8S2 (777.00):
calcd. C 64.92, H 6.75, N 3.61; found C 64.88, H 6.60, N 3.38.
General Procedure for the Titanium-Mediated Addition of TMSCN
to Imines: A 10-fold reactor was charged with ligand 1 (0.05 mmol,
0.1 equiv.) and flushed with argon. The ligand was dissolved in
toluene (3.0 mL) before a solution of Ti(OiPr)4 (0.1  in toluene,
0.50 mL, 0.05 mmol, 0.1 equiv.) was added. The catalyst mixture
was stirred at r.t. for 30 min and then the imine (0.50 mmol,
1.0 equiv.) was added. After cooling to 0 °C, TMSCN (0.13 mL,
1.00 mmol, 2.0 equiv.) was added. A solution of 2-propanol (1.0 
in toluene, 1.00 mL, 1.00 mmol, 2.0 equiv.) was added over 18 h at
0 °C by means of a syringe pump. The reaction mixture was passed
through a silica plug with mixtures of n-pentane/diethyl ether, con-
centrated and analysed by chiral HPLC (λ = 210 nm). For each
substrate, one representative example was purified by FCC with
mixtures of n-pentane/diethyl ether.
General Pattern for the Numbering of Atoms: The atoms are num-
bered for the assignment as shown. The substituents R3 are labelled
as 1ЈЈ etc. Assignments have been made by using DEPT, COSY,
HSQC, and HMBC spectra.
2,2Ј-Diacetoxy-5,5Ј-di-tert-butyl-3,3Ј-bis[(2S,4S)-4-methyl-3-(4-
methylbenzenesulfonyl)-1,3-oxazolidinyl]biphenyl (10aaa): After
Experimental Data for the Crystal Structure of ent-1aab:
X-ray crystal structure analysis for HOP2807:[20] formula
FCC, the product was obtained in 99% yield. M.p. 110–113 °C. Rf
C45H58Cl2N2O8S2,
0.35ϫ0.15ϫ0.05 mm, a = 14.064(1) Å, b = 12.017(1) Å, c =
14.757(1) Å, 109.35(1)°, =
2353.2(3) Å3, ρcalcd.
M
=
889.95,
colourless
crystal
20
= 0.17 (n-pentane/diethyl ether, 1:1). [α]
= –109.4 (c = 1.02,
D
CHCl3). 1H NMR (400 MHz, CDCl3): δ = 7.81–7.79 (m, 4 H, 2ЈЈЈ-
β
=
V
=
H), 7.70 (br. s, 2 H, 4-H), 7.36–7.34 (m, 4 H, 3ЈЈЈ-H), 7.32 (br. s, 2
1.256 gcm–3, µ = 2.489 mm–1, empirical absorption correction
(0.476ՅTՅ0.886), Z = 2, monoclinic, space group P21 (No. 4) λ
= 1.54178 Å, T = 223 K, ω/2θ scans, 5242 reflections collected,
(+h, –k, Ϯl), [(sinθ)/λ] = 0.62 Å–1, 5038 independent (Rint = 0.025)
and 3741 observed reflections [IՆ2σ (I)], 595 refined parameters,
R = 0.047, wR2 = 0.134, Flack parameter 0.00(2), max. residual
electron density 0.20 (–0.42) eÅ–3, hydrogen atoms calculated and
refined as riding atoms.
3
H, 6-H), 6.30 (s, 2 H, 2Ј-H), 3.90 (sext., J = 6.3 Hz, 2 H, 4Ј-H),
3.76 (dd, 2J = 8.8 Hz, 3J = 7.0 Hz, 2 H, 5Ј-HA), 3.39 (dd, 2J =
3
8.9 Hz, J = 7.0 Hz, 2 H, 5Ј-HB), 2.44 (s, 6 H, Ts-CH3), 1.91 (s, 6
3
H, Ac-CH3), 1.42 (d, J = 5.8 Hz, 6 H, 1ЈЈ-H), 1.33 (s, 18 H, tBu)
ppm. 13C NMR (100 MHz, CDCl3): δ = 169.3 (Ac-CO), 148.2 (C-
5), 144.6 (C), 144.1 (C-4ЈЈЈ), 134.7 (C-1ЈЈЈ), 131.2 (C), 129.8 (C-
3ЈЈЈ), 129.6 (C), 129.2 (C-6), 127.9 (C-2ЈЈЈ), 125.7 (C-4), 89.1 (C-
2Ј), 71.7 (C-5Ј), 55.5 (C-4Ј), 34.7 (quart. tBu), 31.3 (tBu), 21.5 (Ts-
Supporting Information (see footnote on the first page of this arti-
cle): Complete results of all asymmetric Strecker reactions and full
characterisation of all new compounds.
CH ), 20.5 (Ac-CH ), 19.6 (C-1ЈЈ) ppm. IR (ATR): ν = 2965 (m),
˜
3
3
2873 (w), 1765 (s), 1351 (s), 1192 (m), 1163 (s), 1094 (m), 903 (s),
815 (s), 666 (s) cm–1. MS (ESI): m/z (%) = 883.326 (100) [M +
Na]+; calcd. 883.327. C46H56N2O10S2 (861.07): calcd. C 64.16, H
6.56, N 3.25; found C 64.19, H 6.52, N 2.98.
Acknowledgments
General Procedure for the Synthesis of Ligands 1: To a stirred solu-
tion of oxazolidine 10 (1.0 equiv.) in MeOH (20 mL per mmol) was
added water (0.5 mL per mmol) and potassium hydroxide
(3.0 equiv.). The yellow solution was stirred at r.t. for 2 h. The same
volume of saturated aqueous ammonium chloride solution was
added, followed by the same amount of dichloromethane. After
phase separation, the aqueous layer was extracted with dichloro-
methane (3ϫ10 mL per mmol), and the combined organic layer
was dried with sodium sulfate. The solvent was removed in vacuo,
and the residue was purified by FCC on silica (n-pentane/diethyl
ether). The characterisation of one representative example is given,
for further compounds, see Supporting Information.
The authors wish to thank Mrs. Karin Gottschalk for excellent
conduction of countless reactions. We are also very grateful to the
Deutsche Forschungsgemeinschaft for generous funding (project
No. HO 577-15-1) and to the Fonds der Chemischen Industrie for
financial support.
[1] A. Strecker, Ann. Chem. Pharm. 1850, 75, 27.
[2] For reviews, see: a) H. Gröger, Chem. Rev. 2003, 103, 2795; b)
L. Yet, Angew. Chem. 2001, 113, 900; Angew. Chem. Int. Ed.
2001, 40, 875; c) C. Spino, Angew. Chem. 2004, 116, 1796; An-
gew. Chem. Int. Ed. 2004, 43, 1764.
Eur. J. Org. Chem. 2008, 684–692
© 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
691

S. Wünnemann, R. Fröhlich, D. Hoppe
FULL PAPER
[3] a) M. Rueping, E. Sugiono, C. Azap, Angew. Chem. 2006, 118,
2979; Angew. Chem. Int. Ed. 2006, 45, 2617; b) E. J. Corey,
[9] Yamato, K. Hasegawa, Y. Saruwatari, L. Doamekpor, Chem.
Ber. 1993, 126, 1435.
M. J. Grogan, Org. Lett. 1999, 1, 157; c) J. Huang, E. J. Corey, [10] A. S. Parsons, J. M. Garcia, V. A. Snieckus, Tetrahedron Lett.
Org. Lett. 2004, 6, 5027; d) M. S. Iyer, K. M. Gigstad, N. D.
Namdev, M. Lipton, J. Am. Chem. Soc. 1996, 118, 4910; e)
M. S. Sigman, E. N. Jacobsen, J. Am. Chem. Soc. 1998, 120,
4901; f) M. S. Sigman, P. Vachal, E. N. Jacobsen, Angew. Chem.
2000, 112, 1336; Angew. Chem. Int. Ed. 2000, 39, 1279; g) P.
Vachal, E. N. Jacobsen, Org. Lett. 2000, 2, 867; h) P. Vachal,
E. N. Jacobsen, J. Am. Chem. Soc. 2002, 124, 10012; i) J. Hu-
ang, X. Liu, Y. Wen, B. Quin, X. Feng, J. Org. Chem. 2007,
72, 204.
1994, 35, 7537.
[11] M. Kauch, V. Snieckus, D. Hoppe, J. Org. Chem. 2005, 70,
7149.
[12] I. Hoppe, H. Hoffmann, I. Gärtner, T. Krettek, D. Hoppe, Syn-
thesis 1991, 1157.
[13] M. Brüggemann, PhD Thesis 2001, University of Münster.
[14] D. Overrödder, PhD Thesis 1993, University of Kiel.
[15] D. Overrödder, Diploma Thesis 1991, University of Kiel.
[16] D. K. G. Heimbach, R. Fröhlich, B. Wibbeling, D. Hoppe,
Synlett 2000, 950.
[4] a) M. S. Sigman, E. N. Jacobsen, J. Am. Chem. Soc. 1998, 120,
5315; b) C. A. Krueger, K. W. Kuntz, C. W. Dzierba, W. G. [17] On first view, it is surprising that lower temperatures lead to
Wirschun, J. D. Gleason, M. L. Snapper, A. H. Hoveyda, J.
Am. Chem. Soc. 1999, 121, 4284; c) J. R. Porter, G. Wirschun,
K. W. Kuntz, M. L. Snapper, A. H. Hoveyda, J. Am. Chem.
Soc. 2000, 122, 2657; d) H. Ishitani, S. Komiyama, S. Kobaya-
decreased enantioselectivity. However one should take into ac-
count that titanium often forms oligomeric or bi- and trinu-
clear complexes; this should be temperature-dependent and
therefore may influence the selectivity of the reaction.
shi, Angew. Chem. 1998, 110, 3369; Angew. Chem. Int. Ed. Engl. [18] Z. Jiao, X. Feng, B. Liu, F. Chen, G. Zhang, Y. Jiang, Eur.
1998, 37, 3186; e) H. Ishitani, S. Komiyama, Y. Hasegawa, S.
Kobayashi, J. Am. Chem. Soc. 2000, 122, 762; f) M. Chavarot,
J. J. Byrne, P.-Y. Chavant, Y. Vallée, Tetrahedron: Asymmetry
2001, 12, 1147; g) S. Masumoto, H. Usuda, M. Suzuki, M.
Kanai, M. Shibasaki, J. Am. Chem. Soc. 2003, 125, 5634; h)
N. Kato, M. Suzuki, M. Kanai, M. Shibasaki, Tetrahedron
Lett. 2004, 45, 3147.
J. Org. Chem. 2003, 3818. In this publication, a wrong R/S
assignment was detected. The values of the optical rotation dif-
fer significantly (by a factor of 10–20) from those found by us
and those quoted in ref.[4b]
[19] V. Banphavichit, W. Mansawat, W. Bhanthumnavin, T. Vilai-
van, Tetrahedron 2004, 60, 10559.
[20] CCDC-637854 contains the supplementary crystallographic
data for this paper. These data can be obtained free of charge
from The Cambridge Crystallographic Data Centre via
www.ccdc.cam.ac.uk/data_request/cif.
[5] J. J. Byrne, M. Chavarot, P.-Y. Chavant, Y. Vallée, Tetrahedron
Lett. 2000, 41, 873.
[6] For a review on atropos ligands, see: K. Mikami, K. Aikawa,
Y. Yusa, J. Jodry, M. Yamanaka, Synlett 2002, 1561.
[7] S. Wünnemann, R. Fröhlich, D. Hoppe, Org. Lett. 2006, 8,
2455.
Received: August 16, 2007
Published Online: December 4, 2007
[8] G. J. H. Buisman, P. C. J. Kamer, P. W. N. M. van Leeuwen,
Tetrahedron: Asymmetry 1993, 4, 1625.
692
www.eurjoc.org
© 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Org. Chem. 2008, 684–692