Novel synthesis of P-chiral hydroxymethylphosphine-boranes through lipase-catalyzed optical resolution

Article abstract of DOI:10.1246/bcsj.76.833

A novel approach toward the lipase-catalyzed acylation of alkyl(1-hydroxymethyl)phosphine-boranes was achieved. Up to 99% of enantiomerically enriched phosphine-boranes was obtained by using lipase AK and CAL.

Full text of DOI:10.1246/bcsj.76.833

Short Articles
Bull. Chem. Soc. Jpn., 76, 833–834 (2003)
833
Novel Synthesis of P-Chiral Hy-
droxymethylphosphine–Boranes
through Lipase-Catalyzed Optical
Resolution
ꢀ
and Kentaro Okuma
Scheme 1.
Scheme 2.
Kosei Shioji, Yoshimitsu Kurauchi,
Department of Chemistry, Faculty of Science,
Fukuoka University, Jonan-ku, Fukuoka 814-0180
Table 1. CAL-Catalyzed Optical Resolution of Phosphine–
Boranes 1 in IPE or c-C6H12
(Received October 16, 2002)
1
Solv. Time/h Conv./% Ee% of 1 Ee% of 4 Ea)
A novel approach towardthe lipase-catalyzed acylation of
alkyl(1-hydroxymethyl)phosphine–boranes was achieved. Up
to 99% of enantiomerically enriched phosphine–boranes was
obtained by using lipase AK and CAL.
a
b
c,d
a
b
c
IPE
IPE
0.2
0.2
56
45
46
56
37
68
3
9
IPE
48
no reaction
c-C6H12
c-C6H12
c-C6H12
c-C6H12
0.5
0.8
3
73
63
47
49b)
92
91
29
53
34
55
33
n.d.c)
6
10
3
Optically active phosphines possessing a chiral center at the
phosphorus are precursors for a variety of C2-symmetrical
bisphosphine ligands.1 It has been reported that phosphine–
boranes are converted into phosphines under mild conditions
without racemization.2 In order to synthesize functional terti-
ary phosphine–boranes, secondary phosphine–boranes are key
intermediates.3 Imamoto et al. reported that secondary phosphi-
ne–boranes were obtained through stereoselective decarboxyl-
ation, followed by the oxidation of dialkyl(1-hydroxymeth-
yl)phosphine–boranes.4 Recently, we reported on the lipase-
catalyzed optical resolution of 1-hydroxymethylphosphine ox-
ides.5 These results prompted us to investigate the possibility of
the lipase-catalyzed optical resolution ofphosphine–boranes. We
would like to propose a new synthesis of P-chiral alkyl(1-
hydroxymethyl)phenylphosphine–boranes (1) via the lipase-
catalyzed optical resolution.
d
9
6
a) The E value was calculated according to Ref. 7. b) Isolated
yield. c) n.d., not determined by HPLC.
boranes 1a,b took place with poor stereoselectivity. On the other
hand, in c-C6H12, the stereoselectivity on the acylation increased
moderately. When the reactions of1a,b stoppedat conversions of
73and63%, enantiomericallyenrichedcompounds(R)-1a,bwere
obtained with 92 and 93% ee. The CD spectrum of (R)-
Hydroxymethyl(methyl)phenylphosphine–borane indicates a ne-
gative Cotton effect.4 Similarly, phosphine–borane 1 showed a
negative Cottoneffect atthe samewavelength, suggesting thatthe
absolute configurations of recovered 1a–c were (R). The
acylation of 1a–c favored the (S)-enantiomer. In contrast, (R)-
1d was acylated faster than its counterpart. This kind of reversed
enantioselectivity on the acylation of tert-butyl derivative 1d was
observed inthat of a phosphineoxide analogue. Thus, the binding
mode6 between phosphine–borane 1 and CAL is similar to that of
phosphine oxide.5 The lower steroselectivity in the acylation of
phosphine–boranes than that of phosphine oxides was affected by
the bulkiness of the BH3 group compared to the oxygen atom.
Actually, the acylation of phosphine–borane 1a proceeded with
lower enantioselectivity than that of 1b because of a small
difference in the bulkiness between BH3 and Et. Since the
isopropyl group of 1c is too large to bind for the medium
hydrophobic pocket of CAL, the acylation of 1c proceeded with
low stereoselectivity compared to the acylation of 1a,b. The E
values of CAL-catalyzed acylation were not influenced by
lengthening in the alkyl chain of vinyl esters.
Results and Discussion
Substituted 1-hydroxymethylphosphine–boranes 1 were pre-
pared by following two methods (Scheme 1). Racemic phos-
phine–boranes bearing a primary alkyl group, such as ethyl or
butyl-derivatives 1a,b, were synthesized in one-pot from phos-
phine oxides (2a,b) (Method A). Other racemic 1-hydroxyme-
thylphosphine–boranes 1c–d were prepared from secondary
phosphine–boranes (3c–d) (Method B).
The acylation of phosphine–boranes 1 in diisopropyl ether
(IPE) or cyclohexane (c-C6H12) was carried out using CAL from
Candida antarctica or Lipase AK from Pseudomonas fluo-
recence. In preliminary work, other lipases, such as lipase PS
from Pseudomonas cepacia and CRL from Candida rugosa,
showed no reaction or poor stereoselectivity. Acylation using
CAL with vinyl acetate as an acyl donor afforded the correspond-
ingacetylatedphosphine–boranes(4)(Scheme 2). Theresultsare
summarized in Table 1.
We then tried the acylation of phosphine–borane 1 using lipase
AK in the presence of vinyl butyrate. The lipase AK-catalyzed
acylation of phosphine–borane 1 was more sensitive toward an
organic solvent than CAL. It is well-known that the solvent
variation of lipase-catalyzed optical resolutions can influence
both the enantioselectivity and the reaction rate.8 In IPE
When IPE was used as a solvent, the acylation of phosphine–

834 Bull. Chem. Soc. Jpn., 76, No. 4 (2003)
Ó 2003 The Chemical Society of Japan
Table 2. Lipase AK-Catalyzed Optical Resolution of Phos-
phine–Boranes 1 Using Vinyl Butyrate in c-C6H12
CDCl3) ꢀ 0.59 (bq, 3H, J ¼ 85:2 Hz), 1.05 (q, 3H, J ¼ 7:2 Hz), 1.28
(q, 3H, J ¼ 7:2 Hz), 2.10 (bs, 1H), 2.43–2.50 (m, 1H), 4.22 (d, 2H,
J ¼ 2:4 Hz), 7.46–7.80 (m, 5H). 13C NMR (100 MHz, CDCl3) ꢀ
16.5, 16.7, 21.3, 21.7, 58.0, 58.4, 128.2, 128.4, 128.5, 131.8, 132.3.
Anal. Calcd for C10H18BOP: C, 61.27; H, 9.25%. Found: C, 60.82;
H, 9.12%. Cyclohexyl(1-hydroxymethyl)phenylphosphine–borane
(1d). t-Butyl(1-hydroxymethyl)phenylphosphine–borane (1e). Col-
orlesscrystals, mp58–59 ꢄC, 68%yield, 1H NMR(400 MHz, CDCl3)
ꢀ 0.69 (bq, 3H, J ¼ 97:2 Hz), 1.18 (d, 9H, J ¼ 13:6 Hz), 2.10 (bs,
1H), 4.34–4.48 (m, 2H), 7.45–7.74 (m, 5H). 13C NMR (100 MHz,
CDCl3) ꢀ 25.7, 28.9, 29.2, 55.7, 56.1, 124.6, 125.1, 128.3, 128.4,
131.4, 133.4. Anal. Calcd for C11H20BOP: C, 62.90; H, 9.60%.
Found: C, 62.64; H, 9.44%.
Entry
1
Time/h Conv./%
Ee% of 1
Ee% of 4
E
1
a
b
b
c
c
c
e
3
48
48
18
16
120
72
50
32
53
31
19
70
9
20
56
31
20
9
42
49
70
84
43
1
3
5
8
15
12
2a)
3
4a)
5
6
7
> 99
no reaction
a) Using vinyl acetate as an acyl donor.
(log P ¼ 1:9), a relatively non-polar solvent, as judged from
log P9 and other polar solvents, such as tetrahydrofuran (THF)
and acetonitrile (log P ¼ 0:49, ꢁ0:33), acylation did not take
place after 5 days. In the case of c-C6H12 (log P ¼ 3:2), which is
less polar than IPE, acylation afforded the corresponding esters 4.
The enantioselectivity in the lipase AK-catalyzed acylation of 1b
using vinyl butyrate was moderately higher than that of using
vinyl acetate (Table 2, Entry 2–5). The E-value for the acylation
of isopropyl derivative 1c was increased to 15.
Lipase-Catalyzed Optical Resolution of Phosphine–Boranes
1. A mixture of a racemic phosphine–borane 1 (5 mg), the enzyme
(40mg),molecularsieves3A(20mg),and vinylacetate(7equiv)was
ꢄ
stirred in an organic solvent (2 mL) at 36 C. The reaction was
analyzed by HPLC (Daicel Chiralpak AD-H). Optical resolutions on
a preparative scale were carried out with 40-times the quantity of the
analysis scale. The reactions were monitored by HPLC (Daicel
Chiralpak AD-H) and stopped at appropriate conversion. The
reaction mixture was separated by column chromatography. 1a: 31%
25
yield, 92%ee ((R)-form rich), ½ꢁꢅD ꢁ 23:3 (c ¼ 1:7, CHCl3). 4a:
58% yield, 34%ee ((S)-form rich). 1b: 23% yield, 91% ee ((R)-form
25
Experimental
rich), ½ꢁꢅD ꢁ 15:3 (c 1.2, CHCl3). 4b: 62% yield, 55% ee ((S)-form
25
rich). 1c: 32% yield, 79%ee ((R)-form rich), ½ꢁꢅD ꢁ 10:1 (c 1.1,
All solvents were distilled prior to use. Lipase PS (Amano PS) and
lipaseAK(AmanoAK)werepurchasedfromAmanoPharmaceutical
Co., Ltd. CAL (Chirazyme1 L-2,c.-f.,C2, lyo.) was purchased from
Roche Molecular Biochemicals. CRL was purchased from Sigma.
Secondary phosphine–boranes 3c–e were prepared according to a
procedure described in the literature.10
Preparation of Alkyl(1-hydroxymethyl)phosphine–Boranes
(1). Method A: To a solution of 1-hydroxymethylalkylphosphine
oxide 2a,b5 (0.5 mmol) in benzene (5.0 mL) was added HSiCl3 (2.2
mmol) at room temperature. After stirring for 3 h, to the reaction
CHCl3). 4c: 61% yield, 67%ee ((S)-form rich). 1d: 44% yield,
25
53% ee ((R)-form rich), ½ꢁꢅD þ 2:9 (c 1.0, CHCl3). 4d: 42% yield.
Enantiomeric excesses were determined by HPLC (Daicel Chiralpak
AD-H). 1a: tR ¼ 40:8 (R) and 76.0 (S) min, 4a: tR ¼ 14:7 (R) and
19.8 (S) min (hexane/2-propanol (97/3) including 0.1% TFA). 1b:
tR ¼ 37:4 (R) and 46.3 (S) min, 4b: tR ¼ 8:4 (R) and 11.3 (S) min
(hexane/2-propanol (99/1) including 0.1% TFA). 1c: tR ¼ 34:4 (R)
and 54.8(S)min, 4c:tR ¼ 6:8 (R)and 8.0(S)min (hexane/2-propanol
(98/2) including 0.1% TFA). 1d: tR ¼ 32:4 (R) and 44.6 (S) min
(hexane/2-propanol (99/1) including 0.1% TFA).
mixture was added THF BH3 (2.5 mmol); the resulting solution was
ꢂ
stirred for 2 h. To the reaction mixture was added 5 mL of benzene;
the resulting solution was quenched by 6 M HCl and extracted with
CH2Cl2 (5 mL ꢃ 3). The extract was dried over MgSO4 and
evaporated. The residue was purified by column chromatography on
silica gel (CH2Cl2) togive 1a,b. Ethyl(1-hydroxymethyl)phenylpho-
sphine–borane (1a). colorless oil, 70% yield, 1H NMR (400 MHz,
CDCl3) ꢀ 0.65 (bq, 3H, J ¼ 91:6 Hz), 1.10–1.19 (m, 3H), 2.01–2.09
(m, 2H), 2.03 (bs, 1H), 4.13 (s, 2H), 7.46–7.78 (m, 5H). 13C NMR
(100 MHz, CDCl3) ꢀ 6.6, 14.3, 14.7, 58.9, 59.3, 125.5, 126.0, 128.6,
128.7, 131.5, 132.2, 132.3. Anal. Calcd for C9H16BOP: C, 59.39; H,
8.86%. Found: C, 59.79; H, 8.77%. Butyl(1-hydroxymethyl)phe-
nylphosphine–borane (1b). Colorless oil, 48% yield, 1H NMR (400
MHz, CDCl3) ꢀ 0.67 (bq, 3H, J ¼ 94:8 Hz), 0.89 (t, 3H, J ¼ 7:2 Hz),
1.37–1.46 (m, 3H), 1.54–1.58 (m, 1H), 1.86 (bs, 1H), 1.97–2.04 (m,
2H), 4.12 (s, 2H), 7.46–7.79 (m, 5H). 13C NMR (100 MHz, CDCl3) ꢀ
13.4, 21.0, 21.4, 24.0, 24.1, 24.6, 59.5, 59.9, 125.9, 126.5, 128.7,
128.8, 131.6, 132.3. Anal. CalcdforC11H20BOP:C, 62.90;H, 9.90%.
Found: C, 63.14; H, 9.60%.
References
1
a) T. Imamoto, Pure Appl. Chem., 73, 373 (2001). b) H.
Tsuruta and T. Imamoto, Tetrahedron: Asymmetry, 10, 877 (1999).
c) K. M. Pietrusiewicz and M. Zablocka, Chem. Rev., 94, 1375
(1994), and references therein.
2
T. Imamoto, J. Watanabe, Y. Wada, H. Masuda, H. Yamada,
H. Tsuruta, S. Matsukawa, and K. Yamaguchi, J. Am. Chem. Soc.,
120, 1635 (1998).
3
M. Al-Masum, G, Kumaraswamy, and T. Livinghouse, J.
Org. Chem., 65, 4776 (2000).
K. Nagata, S. Matsukawa, and T. Imamoto, J. Org. Chem.,
65, 4185 (2000).
4
5
K. Shioji, Y. Ueno, Y. Kurauchi, and K. Okuma, Tetra-
hedron Lett., 42, 6569 (2001).
6
(1999).
7
W. V. Tuomi and R. J. Kazlauskas, J. Org. Chem., 64, 2638
Method B: To a solution of secondary phosphine–borane (3c–d)
(3.3 g, 20 mmol) in THF (40 mL) was added ButLi at ꢁ78 ꢄC. After
stirring for 2 h, gaseous formaldeyde was bubbled into the mixture.
After stirring for 1 h, the solution was quenched by aq ammonium
chloride and extracted with CH2Cl2 (40 mL ꢃ 3). The extract was
dried over anhydrous magnesium sulfate and evaporated to give
phosphine–borane (1c–d). 1-Hydroxymethyl(isopropyl)phenylpho-
sphine–borane (1c). Colorless oil, 60% yield, 1H NMR (400 MHz,
C. S. Chen, Y. Fujimoto, G. Girdauskas, and C. J. Sih, J. Am.
Chem. Soc., 104, 7294 (1982).
K. Faber, ‘‘Biotransformations in Organic Media,’’ 3rd ed,
Springer, Granz, Austria (1996).
C. Laane, S. Boeren, and C. Veeger, Biotechnol. Bioeng., 30,
8
9
81 (1987).
10 T. Imamoto, T, Oshiki, T. Onozawa, T. Kusumoto, and K.
Sato, J. Am. Chem. Soc., 112, 5244 (1990).