syn/anti Diastereoselectivity in the aldol reaction of aldehydes with the C(3) carbanion of 1,3-dihydro-2H-1,4-benzodiazepin-2-one

Article abstract of DOI:10.1002/(SICI)1522-2675(20000315)83:3<603::AID-HLCA603>3.0.CO;2-1

The aldol reaction of the C(3) carbanion of 7-chloro-1,3-dihydro-1- methyl-5-phenyl-2H-1,4-benzodiazepin-2-one (2) with a series of aromatic and aliphatic aldehydes at -78°afforded threo/erythro diastereoisomers 3-16 of 7-chloro-1,3-dihydro-3-(hydroxymethyl)-1-methyl-5-phenyl-2H-1,4- benzodiazepinones, substituted at the C(3) side chain, in a ratio from 55:45 to 94:6 (Scheme 1). Lewis acids exhibited limited effect on the syn/anti diastereoselectivity of this reaction, and kinetic control of the reaction was confirmed. ¹H-NMR Data suggested the assignment of the threo relative configuration to the first-eluted diastereoisomers 3, 5 7, and 9 on reversed- phase HPLC, and the erythro configuration to the second-eluted counterparts 4, 6, 8, and 10, respectively. The structures and relative configurations threo and erythro of the diastereoisomers 5 and 6, respectively, were established by single-crystal X-ray analysis, confirming the assignment based on the ¹H-NMR data. A tentative mechanistic explanation of the diastereoselectivity invokes the enolate anion of 1,3-dihydro-2H-1,4- benzodiazepin-2-one as the reactive species (Scheme 2). Acid-catalyzed hydrolytic ring opening of 3 afforded threo-β-hydroxy-phenylalanine 17, whereas from 4, the N-(benzyloxy)carbonyl derivative 18 of erythro-β- hydroxy-phenylalanine was obtained (Scheme 3); in both cases, neither elimination of H₂O from the C(3)-CHOH moiety nor epimerization at C(3) were observed. This result opens a new pathway to various configurationally uniform α-amino-β-hydroxy carboxylic acids and their congeners of biological importance.

Full text of DOI:10.1002/(SICI)1522-2675(20000315)83:3<603::AID-HLCA603>3.0.CO;2-1

Helvetica Chimica Acta ± Vol. 83 (2000)
603
syn/anti Diastereoselectivity in the Aldol Reaction of Aldehydes with the
C(3) Carbanion of 1,3-Dihydro-2H-1,4-benzodiazepin-2-one
Ï
by Dean MarkovicÂ, Zdenko HamersÏak, Aleksandar VisÏnjevac, Biserka KojicÂ-ProdicÂ, and Vitomir SunjicÂ*
Ï
Â
Ruder Boskovic Institute, P.O.B. 1016, HR-10000 Zagreb, Croatia
Å
The aldol reaction of the C(3) carbanion of 7-chloro-1,3-dihydro-1-methyl-5-phenyl-2H-1,4-benzodiaze-
pin-2-one (2) with a series of aromatic and aliphatic aldehydes at 788 afforded threo/erythro diastereoisomers
3 ± 16 of 7-chloro-1,3-dihydro-3-(hydroxymethyl)-1-methyl-5-phenyl-2H-1,4-benzodiazepinones, substituted at
the C(3) side chain, in a ratio from 55 :45 to 94 :6 (Scheme 1). Lewis acids exhibited limited effect on the syn/
anti diastereoselectivity of this reaction, and kinetic control of the reaction was confirmed. 1H-NMR Data
suggested the assignment of the threo relative configuration to the first-eluted diastereoisomers 3, 5, 7, and 9 on
reversed-phase HPLC, and the erythro configuration to the second-eluted counterparts 4, 6, 8, and 10,
respectively. The structures and relative configurations threo and erythro of the diastereoisomers 5 and 6,
respectively, were established by single-crystal X-ray analysis, confirming the assignment based on the 1H-NMR
data. A tentative mechanistic explanation of the diastereoselectivity invokes the enolate anion of 1,3-dihydro-
2H-1,4-benzodiazepin-2-one as the reactive species (Scheme 2). Acid-catalyzed hydrolytic ring opening of 3
afforded threo-b-hydroxy-phenylalanine 17, whereas from 4, the N-(benzyloxy)carbonyl derivative 18 of
erythro-b-hydroxy-phenylalanine was obtained (Scheme 3); in both cases, neither elimination of H2O from the
C(3) CHOH moiety nor epimerization at C(3) were observed. This result opens a new pathway to various
configurationally uniform a-amino-b-hydroxy carboxylic acids and their congeners of biological importance.
1. Introduction. ± The aldol reaction is a useful method for preparing b-hydroxy-
substituted carbonyl compounds, optionally with an additional substituent at the a-
position, and has, therefore, attracted a great deal of attention from synthetic organic
chemists. The generality, versatility, and selectivity associated with this reaction has
been the subject of authoritative summaries [1]. The aldol reaction of nitriles [2],
glycine derivatives [3], and many other C H acidic compounds is a useful method for
the simultaneous construction of two stereogenic centers and the installation of latent
functionalities. In continuation of our project aiming at the use of 3-substituted 2H-1,4-
benzodiazepin-2-ones as templates for the preparation of optically pure a-amino acids
and their congeners [4], we studied the diastereoselectivity of the aldol reaction
between aldehydes and the C(3) carbanion of 1,3-dihydro-2H-1,4-benzodiazepin-2-one
2. Compound 2 comprises a glycine moiety in the conformationally rigid seven-
membered ring, and offers a specific stereoelectronic surrounding for carbanions
generated at a methylene C-atom. Exploration of such a carbanion in the aldol reaction
seems to be a synthetically worthwhile goal, and here, we report on the syn/anti
diastereoselectivity1) of this reaction, on the determination of the relative config-
uration of the separated diastereoisomeric racemates, and on an example of trans-
1
)
The descriptors syn/anti refer to the relative position of the aldehyde to the 1,3-dihydro-2H-1,4-
benzodiazepin-2-one anion in the transition state (see below, Scheme 2), and erythro/threo are used in the
traditional (carbohydrate) sense to describe the relative configuration at the two stereogenic (N- and OH-
substituted) centers (see below, Footnote 3).

604
Helvetica Chimica Acta ± Vol. 83 (2000)
formation of diastereoisomerically pure products into threo- and erythro-a-amino-b-
hydroxycarboxylic acid derivatives1).
2. Results and Discussion. ± 2.1. syn/anti Diastereoselectivity of the Aldol Reaction.
The carbanion at C(3) of 2 was generated in situ by treatment with BuLi and
diisopropylamine and then exposed to some representative aromatic and aliphatic
aldehydes (Scheme 1 and Table 1). In all cases, a mixture of racemic threo/erythro
diastereoisomers1) (see 3 ± 16) was formed. The yield of the condensation with
aromatic aldehydes varied between 75 and 83%, and the threo/erythro ratio varied from
30 :70 for 3/4 to 55 :45 for 9/10, as determined by HPLC. With three representative
aliphatic aldehydes, a notably higher diastereoselectivity was obtained, ranging from
11:89 for 13/4 to 6 :94 for 11/12.
Scheme 1
i) MeONa, MeOH, Me2SO4, ‡ 58. ii) Lithium diisopropylamide (LDA), 788. iii) R-CHO, 788.
Table 1. Aldol Reaction of 1,3-Dihydro-2H-1,4-benzodiazepin-2-one 2 with Aldehydes RCHO: Diastereoisomer
Ratio and Yields of the Products
R
HPLC Ratioa)
Yield [%]
3/4
5/6
7/8
9/10
11/12
13/14
15/16
Ph
30 : 70
37: 63
45 : 55
55 : 45
6 : 94
75
77
77
83
65
78
71
4-Cl C6H4
4-NO2 C6H4
4-MeO C6H4
Et
Pr
11 : 89
8 : 92
Me2CHCH2
a) First-eluted threo-isomer/second-eluted erythro-isomer.
As seen from Table 1, the second-eluted diastereoisomer on reversed-phase HPLC
was regularly formed as the prevailing product. All diastereoisomers were separated on
preparative scale by chromatography and then crystallized. Interestingly, the first-
eluted diastereoisomers regularly formed slightly yellow crystals, whereas crystals of
the second-eluted diastereoisomers were colorless.
With these results in hand, we attempted to improve the diastereoselectivity of the
aldol reaction, in particular, in the case of aromatic aldehydes. Thus, various auxiliary
agents were examined in the reaction of 2 with benzaldehyde (Table 2). Generally,

Helvetica Chimica Acta ± Vol. 83 (2000)
605
Table 2. Effect of the Base and Auxiliary Agents on the Diastereoselectivity of the Formation of 3/4. Reaction
temperature 708.
Basea)
Base/2 [mol/l]
Lewis
threo/erythro
acid
(HPLC)
LDAb)
LDA
0.60 : 0.14
0.03 : 0.03
0.25 : 0.05
0.60 : 0.14
0.60 : 0.14
0.60 : 0.14
0.60 : 0.14
0.60 : 0.14
0.60 : 0.14
0.60 : 0.14
±
±
±
±
30 : 70
no reaction
28 : 72
21 : 79
25 : 75
23 : 77
27: 73
21 : 79
no reaction
40 : 60
LDAc)
LDAc)
LDAc)
LDA
ZnCl2
ZnBr2
MgBr2
BF3Et2O
TiCl4
±
LDA
LDA
Ti(OPr)4
NaH
a) LDA ˆ Lithium diisopropylamide. b) Reference experiment.c) Reaction temperature 1008.
diastereoselectivity can be influenced by temperature, solvent, and dilution [5], and,
particularly, by Lewis acids [6]. Although Lewis acids were repeatedly reported to
improve the diastereoselectivity of the aldol reaction, no significant effect was observed
in the case of the carbanion of 1,3-dihydro-2H-1,4-benzodiazepin-2-one 2 (Table 2).
We assume that other binding sites within the heterocyclic enolate anion, in particular
the N(4) atom, compete with the carbonyl O-atom of the aldehyde for coordination to
the Lewis acid.
The notably higher diastereoselectivity with aliphatic aldehydes can tentatively be
explained by a generally higher steric demand of the alkyl group attached to the
carbonyl function; similarly, high diastereoisomeric excesses (de) were observed for
aliphatic aldehydes in the aldol reaction with other enolate anions [7].
1
2.2. H-NMR Study of the Relative Configuration and Conformation in Solution.
1
The H-NMR spectra of the first-eluted diastereoisomers regularly reveal smaller J
values (ca. 5 Hz), while for the second-eluted diastereoisomers, nearly doubled values
for J (ca. 9 Hz) are observed (Table 3). Besides, the H-atom of OH C(12)2) of the
second-eluted diastereoisomers 4, 6, and 8 shows coupling with H C(12) (J ꢀ 2.5 ±
3.0 Hz), which is not observed for the first-eluted diastereoisomers. Inspection of
Dreiding models for the diastereoisomeric compounds, combined with 1H-NMR data,
indicate that the second-eluted diastereoisomers should possess erythro-configuration,
since on H-bonding C(12) OH ´´´ OˆC(2), they could adopt a low-energy six-
membered chelate-ring conformation I with the larger groups at C(3) and C(12) in
pseudoequatorial position. threo-Diastereoisomers, instead, would possess larger
phenyl (or alkyl) groups in pseudoaxial position if a H-bond C(12) OH ´´´ OˆC(2)
was formed. This H-bonded chelate should, therefore, be energetically unfavorable for
threo-diastereoisomers; H-bonding to N(4) (see II), instead, is much weaker and,
according to our MM2 calculations [4c], its contribution to conformational stabilization
much lower. Consequently, threo-configuration is ascribed to the first-eluted diaster-
eoisomers [8a].
2
)
Arbitrary atom numbering according to the X-ray crystal-structure representations (see Figs. 1 and 2).

606
Helvetica Chimica Acta ± Vol. 83 (2000)
Table 3. 1H-NMR Data for Diastereoisomeric Aldol Products 3 ± 16 (Ha C(3) CHb(12) OHc2))
First-eluted
d(a)
d(b)
J(a,b)
J(b,c)
Second-eluted
d(a)
d(b)
J(a,b)
J(b,c)
(threo)
[ppm]
[ppm]
[Hz]
[Hz]
(erythro)
[ppm]
[ppm]
[Hz]
[Hz]
3
5
7
3.69
3.62
3.66
3.66
±
5.60
5.56
5.68
5.55
±
4.9
5.1
4.6
5.4
±
±
±
±
±
±
±
±
4
6
8
10
12
14
16
3.64
3.56
3.58
3.62
3.37
3.35
3.34
5.69
5.67
5.79
5.67
4.53
4.62
4.67
9.0
9.2
9.0
8.7
9.0
8.5
±
2.8
3.1
2.4
±
±
±
9
11a)
13
15
3.42
3.38
4.42
4.50
±
±
±
a) Formed in minor amounts, not isolated.
1H,1H-Coupling between H C(12) and HO C(12) would be expected for a
chelated erythro-diastereoisomer, not for the unchelated OH group in the threo-
1
diastereoisomers. The smaller J values (ca. 5 Hz) for the H,1H-coupling of H C(3)
and H C(12) in the first-eluted diastereoisomers are consistent with a ca. 1608 dihedral
angle in the chelated conformer of this diastereoisomer.
To unequivocally establish the threo/erythro assignment, however, single-crystal X-
ray structure analysis was performed for the diastereoisomeric racemates 5 and 63).
2.3. Crystal Structures of 5 and 6. The structures of 5 and 6 with the atom numbering
are shown in Figs. 1 and 2. Characteristic bond lengths and angles of 5 and 6 are listed
in Table 4, and the conformations of the seven-membered 1,4-benzodiazepine rings are
described by the selected torsional angles given in Table 5. H-Bonds are listed in
Table 6 and their patterns shown in the crystal packings (Figs. 3 and 4).
Both crystal structures are racemic mixtures: (R,S)- and (S,R)-enantiomers were
encountered in 5, whereas (R,R)- and (S,S)-enantiomers were found in 6. Two
crystallographically independent molecules A and B in the crystal of 6 are related by an
approximate inversion operation. These two conformers reveal significant conforma-
tional difference about the bond N(4) C(5) C(20) C(21)2); the corresponding
torsion angle is 30.0(6)8 in A and 147.9(5)8 in B (Table 5). In both molecules, the
seven-membered ring adopts a boat conformation with an approximate Cs symmetry.
The conformation about the bond N(4) C(3) C(12) O(19) of 5, comprising two
chiral centers of different absolute configuration ((3R,12S) and its enantiomer), is threo
(torsion angle 56.9(2)8). In 6, the bond C(3) C(12) connects two chiral centers of the
3
)
It is important to note that the ꢀzig-zagꢁ conformation of the C-chain C(2) C(3) C(12) C(13) should be
rotated to get a Fischer-type projection to assign threo/erythro relative configurations to the diastereoiso-
meric racemates [8a]. See also Footnote 1.

Helvetica Chimica Acta ± Vol. 83 (2000)
607
Fig. 1. ORTEP [9] View of 5. Arbitrary atom numbering; the thermal ellipsoids are scaled at the 30% level.
Fig. 2. ORTEP [9] View of 6 (A and B). Arbitrary atom numbering; the thermal ellipsoids are scaled at the
30% level.
same relative configuration ((3R,12R) and its enantiomer), and the conformation is
erythro (torsion angle 179.7(4)8 for A and 177.4(4)8 for B). Thus, the group
O(19) H in 5 is oriented towards N(4), whereas in 6, it is in the vicinity of the carbonyl
O(11)-atom. As a consequence, two chemically distinctive H-bonds are present in 5 and
6 (Table 6). In both crystal structures, the centrosymmetric dimers are formed; in 5,

608
Helvetica Chimica Acta ± Vol. 83 (2000)
Fig. 3. Crystal packing of 5 with dimers formed through O(19) H ´´´ N(4) H-bonds, connected into an infinite
chain via C(10) H ´´´ Cl(2) interactions
Fig. 4. Crystal packing of 6 with dimers formed of molecules related by inversion symmetry (A ´´´ A, B ´´´ B)
through O(19) H ´´´ O(11) H-bonds

Helvetica Chimica Acta ± Vol. 83 (2000)
609
Table 4. Geometry of the 1,4-Benzodiazepine Ring in 5 and 6: Bond lengths [Š] and Angles [8]. For numbering,
see Figs. 1 and 2.
5
6
Molecule A
Molecule B
N(1) C(2)
C(2) O(11)
C(2) C(3)
C(3) N(4)
N(4) C(5)
C(5) C(51)
C(51) C(91)
C(2) N(1) C(91)
C(3) C(2) N(1)
C(2) C(3) N(4)
C(3) N(4) C(5)
N(4) C(5) C(51)
C(5) C(51) C(91)
C(51) C(91) N(1)
1.374(3)
1.209(3)
1.519(4)
1.467(3)
1.286(4)
1.491(3)
1.395(4)
124.0(2)
115.5(2)
104.7(2)
116.5(2)
124.0(2)
122.3(2)
121.7(2)
1.37(1)
1.225(7)
1.526(7)
1.461(6)
1.279(6)
1.482(7)
1.392(7)
123.5(5)
116.9(5)
104.7(4)
117.7(4)
125.1(4)
122.0(5)
121.9(5)
1.36(1)
1.221(8)
1.514(8)
1.473(7)
1.275(6)
1.497(7)
1.399(8)
123.6(5)
115.0(5)
105.6(5)
117.9(4)
123.3(4)
122.5(5)
121.3(5)
Table 5. Selected Torsional Angles [8] of 5 and 6. For numbering, see Figs. 1 and 2.
5
6
Molecule A
Molecule B
C(3) C(2) N(1) C(91)
C(2) N(1) C(91) C(51)
N(1) C(91) C(51) C(5)
C(91) C(51) C(5) N(4)
C(51) C(5) N(4) C(3)
C(5) N(4) C(3) C(2)
N(4) C(3) C(2) N(1)
N(4) C(3) C(12) O(19)
N(4) C(3) C(12) C(13)
N(4) C(5) C(20) C(21)
3.8(4)
40.3(4)
0.3(5)
46.7(4)
2.6(4)
73.3(3)
78.2(2)
56.9(2)
178.3(2)
36.9(4)
0.4(7)
43.9(8)
3.1(8)
43.1(7)
0.1(6)
72.7(5)
74.5(5)
179.7(4)
57.5(6)
30.0(6)
1.5(8)
47.9(9)
6.2(9)
40.8(8)
0.2(8)
74.5(6)
75.2(6)
177.4(4)
55.4(6)
147.9(5)
Table 6. H-Bonds in the Crystals of 5 and 6. For numbering, see Figs. 1 and 2.
D-HA [Š] D-H [Š] H-A [Š] D-HA [8] Symmetry operations on A
H-Bond
5: O(19) H(19) ´´´ N(4)
6: O(19) H(19) ´´´ O(11)
O(19A) H(19A) ´´´ O(11A) 2.723(6)
2.934(2)
2.770(5)
0.801
0.820
0.820
2.134
1.951
1.916
175
177
168
1
1
x, 1 y,
x, y, 2
x, 1 y,
z
z
z
O(19) H ´´´ N(4) completes a R2 (10) ring (Fig. 3), whereas in 6 the R2 (12) motif is
2
2
generated via the O(19) H ´´´ O(11) H-bond (Fig. 4) [10]. In the crystal packing of 5, a
weak interaction C(10) H ´´´ Cl connects dimers to an infinite chain (Fig. 3).
The H-bonding patterns in the crystal and in solution are completely different.
Whereas in the crystal, both diastereoisomers 5 and 6 exhibit intermolecular H-
bonding, in solution, only the erythro-diastereoisomer is conformationally stabilized by
intramolecular H-bonding.

610
Helvetica Chimica Acta ± Vol. 83 (2000)
2.4. Mechanism of the Aldol Reaction. The aldol reaction is reversible, but may be
controlled, according to conditions and the nature of partners, either kinetically or
thermodynamically. The syn/anti stereochemistry of addition under kinetic control can
be successfully rationalized by considering the geometry of the transition state. In the
majority of cases, a cyclic transition state is the most probable [11][12]. We assumed
that the C(3) carbanion of 2 is delocalized, forming a cyclic (E)-enolate anion
(Scheme 2). According to the repeatedly proposed mechanism [13][8b], this enolate
anion coordinates the carbonyl electrophile via the lithium cation. It seems that specific
steric requirements around the C(3) carbanion of the 1,3-dihydro-2H-1,4-benzodiaze-
pine-2-one, in particular the effects of ring puckering and the proximity of the
annellated aromatic ring and 5-phenyl group, preclude an optimal approach of bulky
aldehydes and diminish the enthalpy difference of the two transition states due to
complex steric interactions within both transition states (Scheme 2).
Scheme 2
In a recent paper [14], we have reported that, under similar reaction conditions,
2 mol of formaldehyde can react with the C(3) carbanion of 2, i.e. formation of a
tertiary carbanion occurs. Such a carbanion derived from 3 ± 16, if not attacked by the
second mol of the sterically bulky aldehyde, can epimerize. To examine the influence of
epimerization at C(3) on the diastereoselectivity of the aldol reaction, pure 4 was
treated with an excess of base at 788. HPLC Control of the reaction revealed a very
slow epimerization; after 8 h, a 40 :60 ratio of 3/4 was reached. Since no traces of 2
were observed by HPLC, epimerization via retro-aldol reaction can be excluded.
However, HPLC monitoring of the aldol reaction, which is usually completed within
30 min, revealed that the product ratio does not vary with time; thus, epimerization at
C(3) can have only a minor influence on the product ratio. The most acceptable

Helvetica Chimica Acta ± Vol. 83 (2000)
611
explanation seems to be that the carbanion formed by the attack of the base is very
similar to the transition state shown in Scheme 2, leading thus to a constant ratio of
diastereoisomers.
2.5. Synthetic Utility of the Aldol Reaction. To establish the feasibility of the
investigated aldol reaction as a new approach to a-amino-b-hydroxycarboxylic acids,
diastereoisomer 3 was subjected to acid-catalyzed hydrolytic ring opening, affording
erythro-b-hydroxy-phenylalanine 18 (Scheme 3), which was isolated as ammonium salt.
Similarly, 4 afforded threo-b-hydroxy-phenylalanine, which was isolated in 45% yield as
the N-(benzyloxy)carbonyl derivative 18. In both cases, neither epimerization nor
elimination of H2O, to form a C(3)ˆC(12)2) bond, was observed.
Scheme 3
i) conc. HCl soln./H2O/MeOH 3:1:3. ii) CbzCl/aq. NaOH
In summary, we have described the first examples of a diastereoselective aldol
condensation of the C(3) carbanion of a 1,3-dihydro-2H-1,4-benzodiazepin-2-one. The
condensation proceeds with good yield and, in the case of aliphatic aldehydes, also with
high diastereoselectivity. Diastereoisomeric products can be separated and hydrolyzed
to rac-a-amino-b-hydroxycarboxylic acids. Our future work will be focused on the
understanding of the nuances of diastereoselectivity of this process and on its
enantioselective version.
Experimental Part
General. TLC: Precoated silica-gel thin-layer sheets 60 F 254 from Merck. Prep. column chromatography
(CC): silica gel (Baker, 63 ± 200 mm). Ion-exchange chromatography: Amberlite XAD-16 (Rohm & Haas, 20 ±
50 mesh). HPLC: Hewlett-Packard-1050 instrument, HP-3396A integrator; detection at 254 nm; RP 18 column
(Nucleosil 200 Â 4 mm, 7 mm); flow rate 0.8 ml/min; gradient elution from 100% A to 100% B within 20 min
(A ˆ 50% aq. MeOH, B ˆ MeOH). M.p.: Electrothermal-9100-MP aparatus; no correction. IR: Perkin-
Elmer 297; KBr pellets; in cm 1. NMR: Varian XL-Gem 300; CDCl3 solns., if not stated otherwise; d in ppm rel.
to SiMe4, J in Hz.
Aldol Reaction: General Procedure. A soln. of 2.5m BuLi in hexane (4.93 ml, 11.2 mmol) was added at 08 to
a soln. of (i-Pr)2NH (1.17 ml, 8.4 mmol) in THF (20 ml) under Ar. After 30 min, the stirred mixture was cooled
to 788 (dry ice/acetone), and a soln. of 2 (0.8 g, 2.8 mmol) in THF (12.5 ml) was added via syringe. After
30 min stirring, the soln. of the aldehyde in dry THF (5.6 mmol in 12.5 ml) was added dropwise. The reaction
was followed by TLC (CH2Cl2/AcOEt 9 :1) and the final diastereomer composition determined by HPLC. The
diastereoisomer mixtures were obtained in 65 ± 83% yield, and the diastereoisomers were separated by CC
(silica gel, CH2Cl2/AcOEt 9 :1).

612
Helvetica Chimica Acta ± Vol. 83 (2000)
(3RS)-7-Chloro-1,3-dihydro-3-[(1SR)-hydroxy(phenyl)methyl]-1-methyl-5-phenyl-2H-1,4-benzodiazepin-
2-one (3; threo): M.p. 215 ± 2178. IR: 3450, 1670, 1605, 1480, 1400, 1320, 1110, 700. 1H-NMR: 3.38 (s, 3 H); 3.69
(d, J ˆ 5.1, 1 H); 4.34 (s, 1 H); 5.60 (d, J ˆ 4.9, 1 H); 7.26 ± 7.59 (m, 13 H). 13C-NMR: 35.0; 68.1; 72.4; 123.0;
127.5; 127.6; 128.1; 128.4; 129.6; 129.9; 130.2; 130.9; 131.7; 137.8; 141.1; 141.8; 168.5; 169.7. Anal. calc. for
C23H19ClN2O2 (390.87): C 70.68, H 4.90, N 7.17; found: C 70.64, H 5.16, N 7.18.
(3RS)-7-Chloro-1,3-dihydro-3-[(1RS)-hydroxy(phenyl)methyl]-1-methyl-5-phenyl-2H-1,4-benzodiazepin-
2-one (4; erythro): M.p. 173 ± 1748. IR: 3450, 1660, 1605, 1480, 1400, 1320, 1110, 695. 1H-NMR: 3.44 (s, 3 H); 3.64
(d, J ˆ 9.0, 1 H); 3.90 (d, J ˆ 2.8, 1 H); 5.69 (d, J ˆ 9.0, 1 H); 7.12 ± 7.53 (m, 13 H). 13C-NMR: 35.0; 69.1; 74.0;
122.9; 127.4; 127.6; 128.0; 128.3; 129.5; 129.8; 130.1; 130.7; 131.6; 131.8; 140.6; 141.6; 167.0; 170.9. Anal. calc. for
C23H19ClN2O2 (390.87): C 70.68, H 4.90, N 7.17; found: C 70.97, H 4.99, N 7.06.
(3RS)-7-Chloro-3-[(1SR)-(4-chlorophenyl)hydroxymethyl]-1,3-dihydro-1-methyl-5-phenyl-2H-1,4-benzo-
diazepin-2-one (5; threo): M.p. 183 ± 1848. IR: 3450, 1670, 1605, 1480, 1400, 1320, 1110, 700. 1H-NMR: 3.38
(s, 3 H); 3.62 (d, J ˆ 5.1, 1 H); 4.36 (s, 1 H); 5.56 (d, J ˆ 4.9, 1 H); 7.12 ± 7.53 (m, 12 H). 13C-NMR: 35.0; 68.0;
71.9; 123.0; 128.2; 128.5; 128.9; 129.6; 129.7; 129.9; 130.1; 131.0; 133.8; 133.3; 137.7; 139.6; 141.7; 168.7; 169.6.
Anal. calc. for C23H18Cl2N2O2 (425.36): C 64.95, H 4.27, N 6.59; found: C 64.86, H 4.22, N 6.63.
(3RS)-7-Chloro-3-[(1RS)-(4-chlorophenyl)hydroxymethyl]-1,3-dihydro-1-methyl-5-phenyl-2H-1,4-benzodi-
azepin-2-one (6; erythro): M.p. 189 ± 1908. IR: 3500, 1660, 1605, 1480, 1400, 1320, 1120, 700. 1H-NMR: 3.45
(s, 3 H); 3.56 (d, J ˆ 9.2, 1 H); 3.90 (d, J ˆ 3.1, 1 H); 5.67 (dd, J ˆ 9.0, 3.1, 1 H); 7.19 ± 7.60 (m, 12 H). 13C-NMR:
35.0; 69.0; 73.4; 122.8; 128.1; 128.4; 128.7; 129.4; 129.8; 130.1; 130.8; 131.6; 133.3; 137.6; 139.1; 141.5; 155.1; 167.2;
170.8. Anal. calc. for C23H18Cl2N2O2 (425.36): C 64.95, H 4.27, N 6.59; found: C 64.77, H 4.51, N 6.53.
(3RS)-7-Chloro-1,3-dihydro-3-[(1SR)-hydroxy(4-nitrophenyl)methyl]-1-methyl-5-phenyl-2H-1,4-benzodi-
azepin-2-one (7; threo): M.p. 958 (dec.). IR: 3450, 1670, 1600, 1520, 1400, 1350, 1110, 700. 1H-NMR: 3.42
(s, 3 H); 3.66 (d, J ˆ 4.6, 1 H); 4.58 (s, 1 H); 5.68 (d, J ˆ 4.1, 1 H); 7.25 ± 8.20 (m, 12 H). 13C-NMR: 35.1; 67.6;
71.9; 123.0; 123.3; 128.3; 128.5; 129.5; 129.8; 130.0; 131.2; 132.0; 137.5; 141.5; 147.4; 148.4; 153.0; 169.0; 169.6.
Anal. calc. for C23H18ClN3O4 (435.87): C 63.38, H 4.16, N 9.64; found: C 63.56, H 4.44, N 9.61.
(3RS)-7-Chloro-1,3-dihydro-3-[(1RS)-hydroxy(4-nitrophenyl)methyl]-1-methyl-5-phenyl-2H-1,4-benzodi-
azepin-2-one (8; erythro): M.p. 224 ± 2258. IR: 3500, 1660, 1605, 1520, 1350, 840, 690. 1H-NMR: 3.47 (s, 3 H); 3.58
(d, J ˆ 9.0, 1 H); 4.02 (d, J ˆ 2.4, 1 H); 5.79 (d, J ˆ 9.0, 1 H); 7.17 ± 8.23 (m, 12 H). 13C-NMR: 35.1; 69.0; 73.2;
122.9; 123.1; 128.3; 128.5; 129.4; 129.9; 130.0; 131.0; 131.8; 137.5; 141.4; 147.5; 148.1; 153.2; 167.2; 170.5. Anal.
calc. for C23H18ClN3O4 (435.87): C 63.38, H 4.16; N 9.64; found: C63.52, H 4.42, N 9.56.
(3RS)-7-Chloro-1,3-dihydro-3-[(1SR)-hydroxy(4-methoxyphenyl)methyl]-1-methyl-5-phenyl-2H-1,4-benzo-
diazepin-2-one (9; threo): M.p. 188 ± 1898. IR: 3300, 1680, 1600, 1510, 1400, 1240, 1100, 830, 700. 1H-NMR: 3.37
(s, 3 H); 3.66 (d, J ˆ 5.4, 1 H); 3.79 (s, 3 H); 4.22 (s, 1 H); 5.55 (d, J ˆ 4.9, 1 H); 6.88 (d, J ˆ 8.7, 2 H); 7.25 ± 7.60
(m, 10 H). 13C-NMR: 34.9; 55.0; 68.2; 72.0; 113.4; 122.9; 128.4; 128.6; 129.6; 129.9; 130.2; 131.7; 133.2; 137.7;
139.5; 141.8; 159.0; 168.4; 169.6. Anal. calc. for C24H21ClN2O3 (420.90): C 68.49, H 5.03, N 6.66; found: C 68.26,
H 5.17, N 6.51.
(3RS)-7-Chloro-1,3-dihydro-3-[(1RS)-hydroxy(4-methoxyphenyl)methyl]-1-methyl-5-phenyl-2H-1,4-benzo-
diazepin-2-one (10; erythro): M.p. 171 ± 1728. IR: 3400, 1660, 1510, 1400, 1240, 1030, 820. 1H-NMR: 3.45 (s, 3 H);
3.62 (d, J ˆ 8.7, 1 H); 3.81 (s, 3 H); 5.67 (d, J ˆ 9.0, 1 H); 6.91 (m, 2 H); 7.30 ± 7.53 (m, 10 H). 13C-NMR: 35.0;
55.0; 69.2; 73.5; 113.4; 122.8; 128.3; 128.5; 129.5; 129.8; 130.1; 130.7; 131.6; 132.8; 137.8; 141.7; 159.0; 167.0; 171.0.
Anal. calc. for C24H21ClN2O3 (420.90): C 68.49, H 5.03, N 6.66; found: C 68.54, H 5.05, N 6.71.
(3RS)-7-Chloro-1,3-dihydro-3-[(1RS)-1-hydroxypropyl]-1-methyl-5-phenyl-2H-1,4-benzodiazepin-2-one
(12; erythro): M.p. 167 ± 1688. IR: 3500, 1670, 1605, 1480, 1400, 1110, 690. 1H-NMR: 1.09 (t, J ˆ 7.4, 3 H); 1.36 ±
1.46 (m, 1 H); 1.94 ± 2.01 (m, 1 H); 3.37 (d, J ˆ 9.0, 1 H); 3.42 (s, 3 H); 4.51 ± 4.57 (m, 1 H); 7.26 ± 7.59 (m, 8 H).
13C-NMR: 9.7; 25.4; 29.5; 34.9; 67.7; 72.8; 122.8; 128.4; 129.4; 129.7; 130.4; 130.8; 131.6; 137.8; 141.8; 167.4; 171.0.
Anal. calc. for C19H19ClN2O2 (342.83): C 66.57, H 5.59, N 8.17; found: C 66.49, H 5.62, N 8.16.
(3RS)-7-Chloro-1,3-dihydro-3-[(1SR)-1-hydroxybutyl]-1-methyl-5-phenyl-2H-1,4-benzodiazepin-2-one (13;
threo): M.p. 158 ± 1598. IR: 3450, 1670, 1620, 1480, 1400, 1325, 1110, 695. 1H-NMR: 0.95 (t, J ˆ 6.9, 3 H);
1.40 ± 1.72 (m, 4 H); 3.40 (s, 3 H); 3.42 (s, 1 H); 3.80 (s, 1 H); 4.42 (s, 1 H); 7.26 ± 7.67 (m, 8 H). 13C-NMR: 13.9;
18.7; 34.8; 35.5; 66.2; 70.1; 122.8; 128.4; 129.6; 129.8; 130.4; 130.9; 131.6; 137.8; 142.0; 168.4; 170.5.
(3RS)-7-Chloro-1,3-dihydro-3-[(1RS)-1-hydroxybutyl]-1-methyl-5-phenyl-2H-1,4-benzodiazepin-2-one (14;
erythro): M.p. 157 ± 1588. IR: 3500, 1675, 1610, 1480, 1400, 1120, 825, 690. 1H-NMR: 1.00 (t, J ˆ 7.4, 3 H);
1.31 ± 1.90 (m, 4 H); 3.35 (d, J ˆ 8.46, 1 H); 3.37 (s, 1 H); 3.41 (s, 3 H); 4.62 (m, 1 H); 7.26-7.56 (m, 8 H).
13C-NMR: 14.0; 18.5; 34.9; 34.9; 68.0; 71.4; 122.8; 128.4; 129.5; 129.7; 130.4; 130.8; 131.6; 137.8; 141.9; 167.4;
171.0. Anal. calc. for C20H21ClN2O2 (356.85): C 67.32, H 5.93, N 7.85; found: C 67.15, H 5.66, N 7.75.

Helvetica Chimica Acta ± Vol. 83 (2000)
613
Table 7. Crystallographic Data, Structure Solutions, and Refinement of 5 and 6
5
6
Formula
Solvent
Mr
C23H18Cl2N2O2
±
425.31
C23H18Cl2N2O2
±
425.31
Crystal system
Space group
a [Š]
b [Š]
c [Š]
a [8]
b [8]
g [8]
V[Š3]
Z
triclinic
P-1
triclinic
P-1
8.490(2)
10.410(2)
12.462(4)
70.39(2)
75.26(3)
87.31(3)
1002.5(5)
2
10.25(1)
12.32(1)
17.59(1)
81.05(7)
76.16(7)
87.8(1)
2130(3)
4
F(000)
Dx [Mgm
Radiation
440
1.4(4)
MoKa
880
1.34(2)
MoKa
3
]
Wavelength [Š]
0.71069
3.81 ± 19.91
0.35
0.71069
4.68 ± 10.70
0.33
Range [8]
1
m [mm
]
Temp. [K]
Crystal form
295(5)
plate
295(5)
plate
Crystal size [mm]
Crystal color
Data-collection method
Absorption correction
Total data collected
Unique data
0.12 Â 0.08 Â 0.13
Colorless
q scans
no correction
4311
0.16 Â 0.08 Â 0.22
Colorless
2q scans
no correction
9129
4075
8624
Observed data (criterion)
Rint
2456 (I > 2(I))
0.0332
2985 (I > 2(I))
0.0361
qmax [8]
Range of h, k, l
ꢁ 26.32
10 ꢁ h ꢂ 10
0 ꢁ k ꢂ 13
15 ꢁ l ꢂ 15
3
120
2.9
F2
ꢁ 26.32
12 ꢁ h ꢂ 0
15 ꢁ k ꢂ 15
21 ꢁ l ꢂ 21
3
180
0.6
F2
No. of standard reflections
Frequency of standard reflections [min]
Intensity decay [%]
Refinement on
R1 (Fo > s(Fo))
wR2 (F2), all data
S
0.041
0.1215
0.82
0.050
0.1021
0.77
Parameters
Weighting scheme
324
544
w ˆ 1/[2(Fo2) ‡ (0.0401P)2 ‡ 0.46P] where
P ˆ (Fo2 ‡ 2Fc2)/3
3
D1 (max; min) [eŠ
Data-reduction programme
Structure-solution programme
]
0.22, 0.25
0.23, 0.26
HELENA
SIR97
HELENA [15]
SIR97 [16]
Structure-refinement programme
Preparation of material for publication (programme)
SHELXL97 [17]
PLATON98 [18]
SHELXL97
PLATON98

614
Helvetica Chimica Acta ± Vol. 83 (2000)
(3RS)-7-Chloro-1,3-dihydro-3-[(1RS)-1-hydroxy-3-methylbutyl]-1-methyl-5-phenyl-2H-1,4-benzodiazepin-
2-one (16; erythro): M.p. 189 ± 1918. IR: 3520, 1670, 1610, 1480, 1400, 1120, 695. 1H-NMR: 0.99 (d, J ˆ 6.7, 3 H);
1.05 (d, J ˆ 6.4, 3 H); 1.25 ± 1.37 (m, 1 H); 1.56 ± 1.64 (m, 1 H); 1.96 ± 2.08 (m, 1 H); 3.31 ± 3.36 (m, 2 H); 3.41
(s, 3 H); 4.62 ± 4.69 (m, 1 H); 7.71 ± 7.64 (m, 8 H). 13C-NMR: 21.6; 23.9; 24.2; 34.8; 42.1; 68.5; 69.9; 122.8; 128.4;
129.5; 129.7; 130.4; 130.8; 131.5; 137.8; 141.9; 167.2; 171.1. Anal. calc. for C21H23ClN2O2 (370.87): C 68.01, H 6.25,
N 7.55; found: C 67.87, H 6.15, N 7.68.
(aRS,bSR)-b-Hydroxy-phenylalanine Ammonium Salt (17 ´ NH3; threo). A soln. of 3 (300 mg, 0.76 mmol)
in a mixture of conc. HCl soln. (3.0 ml), H2O (1.0 ml), and MeOH (3.0 ml) was refluxed for 18 h. The mixture
was neutralized to pH 10 with 10% Na2CO3 soln. and was washed with CH2Cl2 (3 Â 5 ml) to remove 5-chloro-2-
(methylamino)benzophenone. Upon adjustment of the pH to 2.5 with 10% H3PO4 soln., the aq. soln. was
purified by ion-exchange chromatography (Amberlite XAD-16 (5 g), H2O (50 ml), followed by 1% NH3 soln.
(50 ml)). Evaporation of the NH3 eluate afforded 73 mg (44%) of 17 ´ NH3, 99% pure by HPLC. M.p. 3508. IR:
3620, 3150, 1645, 1490, 1400, 1010, 705. 1H-NMR (D2O): 3.80 (d, J ˆ 4.3, 1 H); 5.19 (d, J ˆ 4.2, 1 H); 7.29 ± 7.38
(m, 5 H). 13C-NMR (D2O): 61.7; 72.1; 126.8; 129.5; 129.9; 140.1; 173.0.
(aRS,bRS)-N-[(Benzyloxy)carbonyl]-b-hydroxy-phenylalanine (18; erythro). As described for 17 ´ NH3,
with 4 (1.0 g, 2.5 mmol), conc. HCl (9.0 ml), H2O (3.0 ml), and MeOH (9.0 ml). After washing with CH2Cl2
(3 Â 10 ml) the aq. soln. was treated at r.t. with benzyl carbonochloridate (0.7 ml, 5.1 mmol), in portions within
6 h. The mixture was stirred for additional 10 h and then extracted with Et2O (2 Â 10 ml). Upon adjustment of
the pH to 3 with 5% HCl soln., the precipitated product was filtered and dried: 347 mg (45%) of 18; 99% pure
by HPLC. M.p. 3508. IR: 3400, 3330, 1695, 1540, 1250, 1020, 700. 1H-NMR ((D6)DMSO): 4.16 (t, J ˆ 8.7, 1 H);
4.76 (d, J ˆ 8.4, 1 H); 4.92 (s, 2 H); 7.13 ± 7.55 (m, 10 H). 13C-NMR ((D6)DMSO): 60.6; 65.4; 72.6; 127.2; 127.6;
127.7; 127.9; 128.0; 128.6; 137.2; 142.5; 155.9; 172.7.
X-Ray Structure Determination of 5 and 6. Compound 5 was crystallized from MeOH/CH2Cl2 3 :1, and 6
from i-PrOH with a few drops of CH2Cl2. Table 7 summarizes crystal data and experimental details of data
collection, refinement, and software used. Intensities were measured on an Enraf-Nonius-CAD4 diffractometer
with graphite-monochromated MoKa radiation using w/q (5) and w/2q (6) scans. During measurements, there
were no variations in intensities for three standard reflections. The data were corrected for Lorentz and
polarization effects [15]. The crystal-structure determination of 6 revealed two molecules (A and B) in the
asymmetric unit. The free rotation about the bonds C(12) C(13)1) resulted in two slightly different conformers
in the asymmetric unit. These two molecules are related by a pseudo inversion centre; a hypercentric
distribution of the normalized structure factors was observed.
H-Atom coordinates of 5 were determined from the subsequent difference Fourier maps with exception of
a Me group C(10); its H-atom positions were calculated on the basis of stereochemical considerations. For 6, all
H-atom positions were calculated on the basis of stereochemical considerations. Atomic scattering factors and
anamalous dispersion values for the Cl-atoms were those included in SHELXL97 [17].
Crystallographic data (excluding structure factors) for the structures reported in this paper have been
deposited with the Cambridge Crystallographic Data Centre as deposition No. CCDC-136204 and CCDC-
136242. Copies of the data can be obtained, free of charge, on application to the CCDC, 12 Union Road,
Cambridge CB21EZ, UK (fax: ‡44(1223)336033; e-mail: deposit@ccdc.cam.ac.uk).
REFERENCES
[1] D. H. Evans, J. V. Nelson, T. R. Taber, in ꢀTopics in Stereochemistryꢁ, Vol. 12, Ed. G. L. Geoffroy, John
Wiley & Sons, New York, 1981; M. T. Reetz, Angew. Chem., Int. Ed. 1984, 23, 556; M. Braun, in ꢀ Houben-
Weyl: Methods of Organic Chemistryꢁ, Eds. G. Helmchen, R. W. Hofmann, J. Mulzer, and E. Schaumann,
G. Thieme Verlag, Stuttgart, 1995, Vol. E 21, p. 1603; C. J. Cowden, I. Peterson, in ꢀOrganic Reactionsꢁ,
Vol. 51, Ed. L. A. Paquette, John Wiley & Sons, New York, 1997, p. 1.
[2] P. R. Carlier, K. M. Lo, M. M.-C. Lo, I. D. Williams, J. Org. Chem. 1995, 60, 7511; P. R. Carlier, K. M. Lo,
M. M.-C. Lo, P. C.-K. Lo, C. W.-S. Lo, J. Org. Chem. 1997, 62, 6316; P. R. Carlier, W. W.-F. Lam, N. C. Wan,
I. D. Williams, Angew. Chem., Int. Ed. 1998, 37, 2252.
[3] U. Schöllkopf, U. Groth, C. Deng, Angew. Chem., Int. Ed. 1981, 20, 798; Y. N. Belokon, A. G. Bulichev, S. V.
Vitt, Y. T. Struchkov, A. S. Batsonov, T. V. Timofeeva, V. A. Tsyryapkin, M. G. Rizhov, L. A. Lysova, V. I.
Bakhmutov, V. M. Belikov, J. Am. Chem. Soc. 1985, 107, 4252.
Ï
Ï
Â
Ï Â
Ï
Â
Â
[4] a) E. Decorte, R. Toso, A. Sega, V. Sunjic, Z. Ruzic-Toros, B. Kojic-Prodic, N. Bresciani-Pahor, G. Nardin,
Ï
Â
Â
L. Randaccio, Helv. Chim. Acta 1981, 64, 1145; b) A. Avdagic, V. Sunjic, Helv. Chim. Acta. 1998, 81, 85;
c) A. Avdagic, A. Lesac, Zs. Majer, M. Hollosi, V. Sunjic, Helv. Chim. Acta. 1998, 81, 1567.
Ï
Â
Â

Helvetica Chimica Acta ± Vol. 83 (2000)
615
[5] A. Houri, M. T. Didiuk, Z. Xu, N. R. Horan, A. H. Hoveyda, J. Am. Chem. Soc. 1993, 115, 6614.
[6] H. Marumoto, H. Kogen, S. Naruto, Chem. Commun. 1998, 2253.
[7] C. H. Heathcock, C. T. Buse, W. A. Kleschick, M. C. Pirrung, J. E. Sohn, J. Lampe, J. Org. Chem. 1980, 45,
1066.
[8] E. L. Eliel, S. H. Wilen, ꢀStereochemistry of Organic Compoundsꢁ, J. Wiley & Sons, New York, 1994,
a) p. 117; b) p. 913.
[9] C. K. Johnson, ꢀORTEPIIꢁ, Report ORNL-5138, Oak Ride Nat. Laboratory, Tennessee, USA, 1976.
[10] J. Bernstein, R. E. Davies, L. Shimoni, N.-L. Chang, Angew. Chem., Int. Ed. 1995, 34, 1555.
[11] H. E. Zimmerman, M. D. Traxler, J. Am. Chem. Soc. 1957, 79, 1920.
[12] M. Nogrady, ꢀStereoselective Synthesesꢁ, VCH, Weinheim, 1995, p. 178.
[13] C. H. Heathcock, ꢀThe Aldol Addition Reactionsꢁ, in ꢀAsymmetric Synthesisꢁ, Vol. 3, Ed. J. D. Morrison,
Academic Press, New York, 1984, p. 111.
Ï
Â
Â
[14] A. Avdagic, A. Lesac, V. Sunjic, Tetrahedron 1999, 55, 1407.
[15] L. Spek, ꢀ HELENA, Program for Data Reductionꢁ, University of Utrecht, The Netherlands, 1993.
[16] A. Altomare, G. Gascarano, C. Giacovazzo, A. Guagliardi, A. G. G. Moliterni, M.-C. Burla, G. Polidori, M.
Camalli, R. Spegna, ꢀSIR 97: A Package for Crystal Structure Solution by Direct Methods and
Refinementꢁ, Instituto di Ricerca per lo Sviluppo di Metodologie Cristalografiche, Bari, Dipartimento di
Scienze della Terra, Perugia, and Instituto di Strutturistica Chimica ꢀG. Giacomelloꢁ, Roma, Italy, 1997.
[17] G. M. Sheldrick, ꢀSHELXL97: Program for the Refinement of Crystal Structuresꢁ, Universität Göttingen,
Germany, 1997.
[18] L. Spek, ꢀPLATON 98: A Multipurpose Crystallographic Tool, 120398 Versionꢁ, University of Utrecht,
Utrecht, The Nederlands, 1998.
Received November 8, 1999