Calcium pantothenate. Part 2. Optimisation of oxynitrilase-catalysed asymmetric hydrocyanation of 3-hydroxy-2,2-dimethylaldehyde: Synthesis of (R)-pantolactone

Article abstract of DOI:10.1021/op050186s

The synthesis of (R)-pantolactone via oxynitrilase-catalysed asymmetric hydrocyanation of 3-hydroxy-2,2-dimethylaldehyde has been investigated. (R)-Oxynitrilases from almonds as well as from apple and plum kernels were employed as catalysts in the form of defatted meal. A number of factors influencing the hydrocyanation process have been studied and the conditions optimised using statistical methods. (R)-Pantolactone with 74% yield and 30% ee has been synthesised.

Full text of DOI:10.1021/op050186s

Organic Process Research & Development 2006, 10, 103−108
1
Calcium Pantothenate. Part 2. Optimisation of Oxynitrilase-Catalysed
Asymmetric Hydrocyanation of 3-Hydroxy-2,2-dimethylaldehyde: Synthesis of
(
R)-Pantolactone
Ludwik Synoradzki,* Tomasz Rowicki, and Marek Włostowski
Laboratory of Technological Processes, Faculty of Chemistry, Warsaw UniVersity of Technology, ul. Noakowskiego 3,
00-664 Warszawa, Poland
Abstract:
The asymmetric hydrocyanation of aldehydes is a well-
known reaction, catalysed among others by oxynitrilases, the
enzymes present in several plants. A number of reviews over
asymmetric hydrocyanation are available. In the case of (R)-
pantolactone synthesis, the simplest substrate is 3-hydroxy-
The synthesis of (R)-pantolactone via oxynitrilase-catalysed
asymmetric hydrocyanation of 3-hydroxy-2,2-dimethylaldehyde
has been investigated. (R)-Oxynitrilases from almonds as well
as from apple and plum kernels were employed as catalysts in
the form of defatted meal. A number of factors influencing the
hydrocyanation process have been studied and the conditions
optimised using statistical methods. (R)-Pantolactone with 74%
yield and 30% ee has been synthesised.
6
2,2-dimethylaldehyde (2), and a few efforts of its asymmetric
7
hydrocyanation have been reported. In all cases an (R)-
oxynitrilase from almonds was employed as a catalyst. The
isolated and purified enzyme was immobilised on a support
or used as a buffered solution, and the syntheses afforded
(R)-pantolactone in 62-96% yield and 20-89% enantio-
meric excess. However, the use of the isolated and purified
enzyme is not very practical since an additional stage of
enzyme preparation is necessary which obviously prolongs
and complicates the whole process and raises its cost. Thus,
we attempted to utilise the enzyme without isolating it from
the original material by application of defatted meal from
almonds as a catalyst in the hydrocyanation of aldehyde 2.
The aim of our work was to maximise both the yield and
enantiomeric excess of cyanohydrin (R)-3 and, conse-
equently, those of (R)-pantolactone. Since similar application
of the defatted almond meal instead of the purified enzyme
Introduction
The methods of (R)-pantolactone ((R)-dihydro-3-hydroxy-
4,4-dimethyl-2(3H)-furanone, (R)-4) manufacture are still the
object of research, despite over 60 years elapsing from its
first total synthesis. The reason for that is the importance
of (R)-pantolactone in calcium (R)-pantothenate and (R)-
5
panthenol synthesis. Both compounds, as vitamin B precur-
sors, find wide application in food, cosmetics and pharma-
ceuticals. The existing methods of (R)-pantolactone manu-
facturing are based on the resolution of the racemate, or
asymmetric hydrogenation of dihydro-4,4-dimethyl-2,3-
furandione (5). The drawback of both methods are the long
2
3
,4
5
8
have been reported in the case of different aldehydes, as
and complicated technologies starting from racemic 4
synthesis, which is subsequently resolved into optical isomers
or oxidised to ketolactone 5 and reduced back to (R)-4.
However, the methodology of (R)-4 synthesis can be radically
shortened and simplified by carrying out an asymmetric
hydrocyanation of aldehyde 2 that allows us to obtain (R)-
pantolactone directly. With such an approach the racemate
resolution or the oxidation/reduction processes are no longer
necessary (Scheme 1).
well as the possibility of using meals from other sources
mentioned, we also tested meals from apple and plum
kernels.
9
(6) (a) North, M. Synlett 1993, 807. (b) Effenberger, F. Angew. Chem., Int. Ed.
Engl. 1994, 33, 1555. (c) Griengl, H.; Hickel, A.; Johnson, D. V.; Kratky,
C.; Schmidt, M.; Schwab, H. Chem. Commun. 1997, 1933. (d) Gregory, R.
J. H. Chem. ReV. 1999, 99, 3649. (e) Effenberger, F. Chimia 1999, 53, 3.
(f) Schmidt, M.; Griengl, H. Top. Curr. Chem. 1999, 200, 193. (g) Seoane,
G. Curr. Org. Chem. 2000, 4, 283. (h) North, M. Tetrahedron: Asymmetry
2003, 14, 147. (i) Brunel, J.-M.; Holmes, I. P. Angew. Chem., Int. Ed. 2004,
43, 2752.
*
To whom correspondence should be addressed. E-mail: ludwik.synoradzki@
(7) (a) Ognyanov, V. I.; Datcheva, V. K.; Kyler, K. S. J. Am. Chem. Soc. 1991,
113, 6992. (b) Beisswenger, T.; Huthmacher, K.; Klenk, H. DE Patent
4,126,580, 1993; Chem. Abstr. 1993, 118, 233500h. (c) Effenberger, F.;
Eichhorn, J.; Roos, J. Tetrahedron: Asymmetry 1995, 6, 271.
ch.pw.edu.pl. Fax: +48-22-6255317.
1) Part 1: Rowicki, T.; Synoradzki, L.; Włostowski, M. Ind. Eng. Chem. Res.
(
(
(
2
006. In press.
2) Stiller, E. T.; Harris, S. A.; Finkelstein, J.; Keresztesy, J. C.; Folkers, K. J.
Am. Chem. Soc. 1940, 62, 1785.
(8) (a) Zandbergen, P.; van der Linden, J.; Brussee, J.; van der Gen, A. Synth.
Commun. 1991, 21, 1387. (b) Huuhtanen, T. T.; Kanerva, L. T. Tetrahe-
dron: Asymmetry 1992, 3, 1223. (c) van den Nieuwendijk, A. M. C. H.;
Warmerdam, E. G. J. C.; Brussee, J.; van der Gen, A. Tetrahedron:
Asymmetry 1995, 6, 801. (d) Kiljunen, E.; Kanerva, L. T. Tetrahedron:
Asymmetry 1996, 7, 1105. (e) Han, S.; Lin, G.; Li, Z. Tetrahedron:
Asymmetry 1998, 9, 1835. (f) Lin, G.; Han, S.; Li, Z. Tetrahedron 1999,
55, 3531. (g) Roos, J.; Effenberger, F. Tetrahedron: Asymmetry 1999, 10,
2817. (h) Han, S.; Chen, P.; Lin, G.; Huang, H.; Li, Z. Tetrahedron:
Asymmetry 2001, 12, 843. (i) Tromp, R. A.; van der Hoeven, M.; Amore,
A.; Brussee, J.; Overhand, M.; van der Marel, G. A.; van der Gen, A.
Tetrahedron: Asymmetry 2001, 12, 1109. (j) Chen, P.; Han, S.; Lin, G.;
Li, Z. J. Org. Chem. 2002, 67, 8251.
3) (a) Pfohl, S.; Paust, J.; Siegel, H.; Himmele, W.; Hoffmann, W.; Fraunberg,
K. DE Patent 2,404,305, 1975; Chem. Abstr. 1975, 83, 205780q. (b) Paust,
J.; Pfohl, S.; Reif, W.; Schmidt, W. Liebigs Ann. Chem. 1978, 1024.
4) (a) Sakamoto, K.; Shimizu, S.; Yamada, H. WO Patent 91/02081, 1991;
Chem. Abstr. 1991, 115, 27689m. (b) Sakamoto, K.; Shimizu, S.; Yamada,
H. WO Patent 92/06182, 1992; Chem. Abstr. 1992, 117, 110119n. (c)
Shimizu, S.; Kataoka, M.; Shimizu, K.; Hirakata, M.; Sakamoto, K.;
Yamada, H. Eur. J. Biochem. 1992, 209, 383.
(
(
5) (a) Broger, E. A.; Crameri, Y. EP Patent Appl. 158,875, 1985; Chem. Abstr.
1987, 106, 67506. (b) Broger, E. A.; Crameri, Y. EP Patent Appl. EP
218,970, 1986; Chem. Abstr. 1987, 107, 7394. (c) Schmidt, R. Chimia 1996,
50, 110.
(9) Kiljunen, E.; Kanerva, L. T. Tetrahedron: Asymmetry 1997, 8, 1225.
1
0.1021/op050186s CCC: $33.50 © 2006 American Chemical Society
Vol. 10, No. 1, 2006 / Organic Process Research & Development
•
103
Published on Web 12/14/2005

Scheme 1. (R)-Pantolactone manufacturing strategies
Scheme 2. Synthesis of (R)-pantolactone via asymmetric hydrocyanation of 3-hydroxy-2,2-dimethylaldehyde
Table 1. Factorial design I: variables at maximum and
minimum levelsa
Results and Discussion
R)-Pantolactone synthesis utilizing the asymmetric hy-
drocyanation consists of three steps: an aldol reaction of
-methylpropanal (1) with formaldehyde that leads to alde-
hyde (2) that undergoes asymmetric hydrocyanation to afford
R)-2,4-dihydroxy-3,3-dimethylbutylnitrile (3), which is sub-
sequently hydrolysed in acidic conditions with direct forma-
tion of (R)-4 (Scheme 2).
The first and the last reaction proceeds without complica-
tions with high yields; moreover, acidic conditions during
hydrolysis ensures that no racemisation of either the nitrile
or the lactone stereogenic centre takes place.7c Therefore,
(
z
z
z
z
i
natural variable
buffer concentration (% vol)
reaction time (h)
(-1)
6
6
(0)
9
12
20
(+1)
12
1
2
3
2
18
28
temperature (°C)
12
(
a
Enzymatic reaction conditions: 2 mmol of acetone cyanohydrin, 0.12 g of
almond meal and 2 mL of diisopropyl ether per mmol of aldehyde 2 respectively.
The experiments were performed at a scale of 2 mmol of 2, variables of the
5 6
factorial design II were constant (x ) -0.8, x ) -2.2).
2
and enantiomeric excess, y , of (R)-pantolactone have been
determined by capillary gas chromatography. The results
along with the experimental matrix are shown in Table 2.
The factorial design I statistical analysis and the signifi-
cant coefficients are summarised in Table 3. The coefficients
have been calculated using the polynomial function of the
experimental coded variables as given in eq 1.
the main object of our study was an asymmetric hydrocya-
nation of aldehyde 2, as the key step of the whole synthesis,
responsible for the optical purity of (R)-pantolactone. Among
the described syntheses of (R)-3 that employed the isolated
enzyme, the best results have been obtained using hydrogen
7
b,c
cyanide (HCN). The application of acetone cyanohydrin
ACH) as a HCN source has been reported to give a much
y ) b + b x +
b x x + e
(1)
0
∑
i
i
∑∑ ij
i j
(
7a
lower ee of nitrile (R)-3; however, considering safety, we
Mathematical models for the yield, yˆ
excess, yˆ , are outlined in eqs 2 and 3.
1
, and the enantiomeric
have decided to start with ACH in our work.
2
We have tested the influence of buffer concentration, z
the reaction time, z , and temperature, z , on the yield, y
1
,
,
2
3
1
3
yˆ ) 5.925 + 1.200‚x + 1.200‚x + 3.825‚x
(2)
(3)
1
1
2
3
2
and enantiomeric excess, y , of (R)-pantolactone using 2
factorial design.10 The three additional centre point experi-
yˆ
2
) 5.475 + 1.975‚x
0.775‚x ‚x
2
- 3.075‚x
3
-
ments, where the levels for all variables were chosen at an
intermediate value, have been set up for checking the
experiments’ repeatability (0 in Table 1).
Since the nitrile (R)-3 hydrolysis to the lactone (R)-4
proceeds in acidic conditions quantitatively and without any
racemisation, we have not measured the respective values
for the former but for the latter only. The experiments have
1
2
+ 0.875‚x
1
‚x
3
- 2.175‚x ‚x
2 3
The results obtained, both the yield and the ee of (R)-
pantolactone were rather disappointing. However, the value
measured for (R)-4 enantiomeric excess (16.6%) was similar
to that reported by Ognyanov et al. (ee ) 20% for nitrile
7
a
(
R)-3). This unfortunately meant that probably no signifi-
cant success could be achieved with the acetone cyanohydrin
methodology used. Comparison of the reaction conditions
applied here and reported by Ognyanov et al. with those
1
been performed at random, and for each one the yield, y ,
(
10) (a) Adler, Yu. P.; Markowa, E. V.; Granovsky, Yu. V. The Design of
Experiments To Find Optimal Conditions; Mir Publishers: Moscow, 1975.
b) Achnazarowa, S. Ł.; Kafarow, W. W. Optymalizacja eksperymentu w
7c
reported by Effenberger to give much better yield and ee
(
of (R)-4 clearly showed that employment of acetone cyano-
hydrin instead of the free hydrogen cyanide might be the
main reason for the poor results of the enzyme-catalysed
asymmetric hydrocyanation. Therefore, considering addition-
chemii i technologii chemicznej (Experiment optimization in chemistry and
chemical engineering); WNT: Warszawa, 1982. (c) Morgan, E. Chemo-
metrics: Experimental Design; J. Wiley & Sons: Chichester, 1991. (d)
Ja n˜ czewski, D.; R o´ z˘ ycki, C.; Synoradzki, L. Projektowanie Proces o´ w
Technologicznych cz. 2 (Process Design, Part 2); WPW: Warszawa, 2001.
104
•
Vol. 10, No. 1, 2006 / Organic Process Research & Development

Table 2. Factorial design I: experimental matrix and the results
coded variables yield of (R)-4 (%)
ee of (R)-4 (%)
response y calculated yˆ
trial no.
x
1
x
2
x
3
response y
1
calculated yˆ
1
2
2
1
2
3
4
5
6
7
8
9
-
+
-
+
-
+
-
+
0
-
-
+
+
-
-
+
+
0
-
-
-
-
+
+
+
+
0
0.9
1.3
1.5
4.7
6.4
10.3
10.1
12.2
4.5
5.8
5.1
-0.3
2.1
2.1
4.5
7.4
9.8
9.7
12.1
5.9
5.9
5.9
3.0
5.8
16.6
8.8
3.2
2.0
0.6
3.8
5.0
5.8
5.2
4.5
4.3
14.4
11.1
1.0
4.3
2.1
2.3
5.5
5.5
5.5
1
1
0
1
0
0
0
0
0
0
Table 3. Factorial design I: influence of variables and
statistical analysis
Table 5. Factorial design II: variables maximum and
minimum levels
values for yfield (yˆ
1
)
values for ee (yˆ
2
)
z
z
z
z
z
i
natural variable
reaction time (h)
buffer amount (ml/mmol of 2)
almond meal amount (g/mmol of 2)
solvent amount (ml/mmol of 2)
(-)
(0)
4
0.2
0.2
7.5
(+)
2
4
5
6
2
6
significant
0.1
0.1
5
0.3
0.3
10
coefficients b
i
b
i
t
bi
b
i
t
bi
b
b
b
b
b
b
b
0
1
2
3
5.925
1.200
1.200
3.825
25.757
5.217
5.217
16.628
-
5.475 37.195
-
-
1.975 13.417
-3.075 20.891
Therefore, we have performed the new series of asym-
metric hydrocyanation experiments using almond meal as
catalyst and free hydrogen cyanide as CN source. We have
decided to use 2.5 molar excess of HCN calculated with
respect to aldehyde 2 and, because of the very poor yields
obtained at lower temperature in factorial design I, to carry
out all the reactions at 28 °C. We have tested the influence
12
13
23
-
-
-
-0.775
0.875
5.265
5.944
-
-
-
-2.175 14.776
4.303
0.95
tbi critical
4.303
level of confidence
degrees of freedom
residual variance
0.95
2
0.956
2
14.625
of the amounts of buffer, z
4
, almonds meal, z
5
, and solvent,
, and
2
, of (R)-pantolactone using 2 factorial
z
6
, as well as the reaction time, z
2
, on the yield, y
1
Table 4. Asymmetric hydrocyanation of aldehyde 2
catalysed by isolated (R)-oxynitrilase from almonds
4
enantiomeric excess, y
(R)-oxynitrilase
(R)-3a
(R)-4
design (Table 5).
Again, the experiments have been performed at random,
and for each one the yield and enantiomeric excess of (R)-
pantolactone have been determined by capillary gas chro-
matography. The polynomial function calculated for the (R)-4
yield from experiments 12-27 using eq 1 did not, however,
fit the experimental results. Therefore, we decided to expand
our factorial design II to central composite design10 by
carrying out a number of additional experiments (trials 28-
38). Complete results along with the experimental matrix
are shown in Table 6.
amount activity
(
U/mmol of 2) (U/mg) yield (%) ee (%) yield (%) ee (%)
1
2
1
1
4
25b
50
50
75
25
160b
160b
38
74
136
-
-
55
84
77
-
-
83
89
89
58
85
-
62
80
-
b
-
-
-
-
a
Reference 7c. b According to the supplier: Sigma-Aldrich Chemie GmbH.
Similarly to the previous design, the experimental coded
variables coefficients have been calculated using a polyno-
mial function as given in eq 1, and along with statistical
analysis have been summarised in Table 7. An analysis of
the variable coefficients indicates a considerable complication
level of the reaction model studied. Among four variables,
ally the statistical analysis of the results obtained (Table 3),
we substantially revised our methodology.
We have performed the (R)-pantolactone synthesis via
aldehyde 2 asymmetric hydrocyanation catalysed by free (R)-
7
c
oxynitrilase from Sigma as described by Effenberger,
additionally followed by direct hydrolysis of nitrile (R)-3 to
R)-pantolactone. The yield and enantiomeric excess deter-
only the reaction time, z
expected, we observed relevant interaction coefficient be-
tween buffer, x , and almonds meal, x , amounts (b45).
4
, has been found insignificant. As
(
4
5
mined for (R)-4 were substantially similar to that reported
for nitrile (R)-3. Besides proving the positive influence of
the free hydrogen cyanide application, this also confirmed
our assumption about the nature of the hydrolysis, proceeding
in a quantitative manner and without racemisation of either
the nitrile or the lactone stereogenic centre (Table 4).
yˆ ) 31.893 + 6.280‚x - 4.481‚x ‚x - 2.606‚x ‚x +
1
5
2
5
4
5
2
6
.139‚x6 (4)
2
yˆ ) 20.747 - 2.766‚x + 4.529‚x + 3.640‚x - 1.797‚x
6
2
4
5
6
(5)
Vol. 10, No. 1, 2006 / Organic Process Research & Development
•
105

Table 6. Factorial design II: experimental matrix and the results
coded variables
yield of (R)-4 (%)
response y calculated yˆ
ee of (R)-4 (%)
response y calculated yˆ
2
trial no.
x
2
x
4
x
5
x
6
1
1
2
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
-
-
-
-
20.7
19.2
56.9
43.0
26.5
30.3
49.6
29.6
38.8
47.3
27.7
28.8
31.5
29.6
50.7
50.7
28.8
33.4
22.3
41.9
20.4
44.6
61.4
43.3
35.3
39.9
36.5
27.5
27.8
50.2
40.0
23.4
23.7
54.2
44.1
31.5
41.7
36.3
37.0
27.5
37.6
40.3
40.0
30.9
30.9
31.9
31.9
22.6
42.0
52.1
52.1
37.5
37.5
37.5
13.2
7.0
13.8
6.8
-
-
-
-
-
-
-
+
+
+
+
+
+
+
+
+
-
+
-
+
-
+
-
+
-
+
-
+
-
+
0
-
+
+
-
-
+
+
-
-
+
+
-
-
+
+
0
-
-
-
+
+
+
+
-
-
-
-
+
+
+
+
0
22.0
18.4
23.0
14.0
29.6
25.6
15.0
10.4
21.2
17.0
22.6
14.4
28.4
24.0
22.0
20.6
25.2
16.6
7.0
22.7
18.6
21.0
14.0
29.8
25.8
15.1
8.1
21.4
17.3
22.2
15.2
28.6
24.5
21.5
21.5
24.5
15.9
9.1
-1.547
+1.547
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
-1.547
+1.547
0
0
0
0
0
0
0
-1.547
+1.547
0
0
0
0
0
0
24.8
14.7
26.5
19.8
20.8
19.0
23.1
15.1
26.4
20.2
20.2
20.2
-1.547
+1.547
0
0
0
Table 7. Factorial design II: influence of variables and
statistical analysis, test F
investigated range. We have employed a steepest-ascent
procedure, as a useful tool for this purpose. Since the
equation without the interaction coefficients did not fit the
experimental results for the yield (according to the test F),
we have used the values of the enantiomeric excess to
estimate the new reaction conditions (Table 8). The steepest-
ascent line has been calculated from the polynomial function
values for the yield ( yˆ 1) values for the ee ( yˆ 2)
significant
coefficients
bi
tbi
bi
tbi
b0
b4
b5
b6
b55
b66
b25
b45
31.893
-
44.856
-
20.747
-2.766
4.529
3.640
-1.797
-
61.410
11.125
18.211
14.636
5.364
-
2
for the enantiomeric excess, yˆ , without the interaction
coefficients.
6.280
-
11.999
-
-
-
Only the first experiment carried out along the path of
steepest ascent brought some improvement of the yield, y
and the determined ee, y , has been in compliance with the
6.139
-4.481
-2.606
8.706
7.512
4.369
1
,
-
-
2
-
-
expected value. For the next steps (trials 40, 41) the
enantiomeric excess of (R)-pantolactone rapidly decreased
and similarly did the yield. An analysis of the reaction
conditions showed that this was not as surprising as it seemed
at the first glance. The buffer amount coefficient for the
tbi critical
4.303
0.95
2
4.303
0.95
2
level of confidence
degrees of freedom
residual variance
101.272
1.285
Mathematical models for the yield, yˆ
excess, yˆ , are outlined in eqs 4 and 5. Generally, the results
1
, and enantiomeric
1
enantiomeric excess had a negative sign (Table 7, b ), thus
indicating that inside the factorial design II a decrease in
2
achieved with the new methodology utilizing free hydrogen
cyanide were much better when compared with those
obtained with acetone cyanohydrin. The enantiomeric excess
and yield for (R)-pantolactone reached 29.6% and 49.6%
respectively (trial 18). The optimal conditions inside the
design core, calculated with MS Excel Solver were found at
the buffer amount had a positive influence on the ee of (R)-
4. However, the steepest-ascent line going beyond the range
of factorial design II brought this value almost to zero, that
made a qualitative change in the reaction conditions. It is
well-known that the enzyme requires small amounts of water
11
to retain its catalytic activity; therefore, the insufficient
the border of the area: x
1 for the yield, yˆ and x
enantiomeric excess, yˆ (reaction time, x
2
) -1, x
4
) -1, x
) +1, x
, was not present
5
) +1, x
) +1 for the
6
)
amount of buffer had certainly suppressed the enzymatic
reaction in steps 2 and 3 of the steepest-ascent procedure.
This not only caused rapid ee decrease but also significantly
+
1
4
) -1, x
5
6
2
2
in eq 5). This corresponded to the best result achieved;
however, such values were still not satisfactory, and thus,
we have sought the better reaction conditions beyond the
(11) Drauz, K.; Waldmann, H. Enzyme Catalysis in Organic Synthesis; VCH
Vergagsgesellschaft; Weinheim, 1995; pp 26-29.
106
•
Vol. 10, No. 1, 2006 / Organic Process Research & Development

a
Table 8. Reaction conditions and the results of the steepest-ascent procedure
step 1
step 2
step 3
natural variable/result
z
z
z
z
y
y
y
i
or y
i
trial 39
0.134
0.305
9.71
74.1
29.4
trial 40
0.068
0.411
11.93
47.3
12.1
37.2
trial 41
0.002
0.511
14.14
20.0
3.4
46.2
buffer amount (ml/mmol of 2)
almond meal amount (g/mmol of 2)
solvent amount (ml/mmol of 2)
yield (%)
4
5
6
response
response
calculated
1
2
2
enantiomeric excess (%)
28.2
a
2
Reaction time, z , of all experiments was 4 h.
Table 9. Comparison of the catalytic properties of meals
from almonds and kernels of mature garden plums and apples
by triple extraction with ethyl acetate and drying under
reduced pressure (300 hPa). Meals were refrigerated and
stored under argon. 3-Hydroxy-2,2-dimethylaldehyde (2) was
distilled under reduced pressure and was melted under argon
atmosphere prior to reaction. Pantolactone yield was deter-
mined by capillary GC chromatography: HP 6890 apparatus,
HP-1 column 30 m × 0.32 mm, temperature program 100-
from different sourcesa
no.
meal source
yield of (R)-4 (%)
ee of (R)-4 (%)
1
4
4
8
3
4
almonds
apple kernels
plum kernels
49.6
32.7
36.6
29.6
11.5
19.6
a
Reactions carried out at optimal conditions from factorial design II (x
1, x ) -1, x ) +1, x ) +1).
2
)
-
4
5
6
1
60 °C at 10 °C/min, injector 250 °C, detector 250 °C, carrier
gas He 2.0 mL/min. GC analysis of pantolactone enanti-
omers: HP 6890 chromatograph, capillary column with chiral
stationary phase Supelco 2-4304 (beta dex 120) 30 m × 25
mm, temperature program 140-160 °C at 4 °C/min, 160
°C 2 min, injector 235 °C, FID detector 235 °C, carrier gas
He 1.0 mL/min. Retention times: (S)-4, 7.0 min and (R)-4,
7.1 min.
lowered the yield, and the residual yield was the result of
the noncatalysed reaction.
Taking into account the literature data reporting good
results of asymmetric hydrocyanation of a number of
9
aliphatic aldehydes with meals from various plants, we have
decided to investigate the reaction catalysed by defatted
meals from kernels of plums and apples. The experiments
were performed at the optimum conditions calculated from
Immobilised (R)-Oxynitrilase-Catalysed Hydrocyana-
tion of 3-Hydroxy-2,2-dimethylaldehyde (2). (R)-Oxynit-
rilase from Sigma-Aldrich Chemie GmbH was immobilised
on Celite (250 U/300 mg or 500 U/600 mg) and inoculated
for 15 min. with 0.3 mL of 0.02 M acetate buffer pH ) 3.3.
Then 0.21 g (2 mmol) of freshly melted aldehyde 2 in
diisopropyl ether was added, followed by 4 mmol of
hydrogen cyanide (2.94 M solution in diisopropyl ether;
CAUTION: TOXIC, all operations with HCN and its
solutions should be carried out under a ventilation hood,
using appropriate protective clothes). The reaction was
maintained for 2 h at 20 °C; the immobilised enzyme was
filtered off and washed with diethyl ether. Combined ether
fractions were evaporated, affording nitrile (R)-3 as a
colourless oil, which was immediately hydrolysed to (R)-4.
Almond Meal-Catalysed Hydrocyanation of 3-Hy-
droxy-2,2-dimethylaldehyde (2) with Acetone Cyanohy-
drin. Defatted almond meal (0.6 g) was inoculated for 15
min with 0.02 M citrate buffer (pH 5.5, for amounts see Table
1). Then 0.21 g (2 mmol) of freshly melted aldehyde 2 in
diisopropyl ether was added, followed by 2 equiv of acetone
cyanohydrin. The reaction was maintained for given times
and conditions (Table 1, Table 2); the almond meal was then
filtered off and washed with diethyl ether. Combined ether
fractions were evaporated, affording nitrile (R)-3 as a
colourless oil, which was immediately hydrolysed to (R)-4.
Defatted Plant Meal-Catalysed Hydrocyanation of
the factorial design II (x
1), and the results comparing catalytic properties of kernel
meal from almonds, apples, and plums are shown in Table
2
) -1, x
4
) -1, x
5
) +1, x
6
)
+
9.
Application of meals from apple and plum kernels brought
no improvement either of the yield or the enentiomeric excess
of synthesised (R)-4. The best results remained the ee’s of
29.4% and 29.6% and yields 74.1% and 49.6% obtained in
trials 39 and 18, respectively.
Conclusions
In summary, we have completed the synthesis of (R)-
pantolactone (R)-4 with the key step of asymmetric hydro-
cyanation of 3-hydroxy-2,2-dimethylaldehyde (2) catalysed
by defatted meal from almonds as well as from kernels of
apples and plums. We showed that acetone cyanohydrin is
not a satisfactory cyanide source for a meal-catalysed
aldehyde 2 hydrocyanation. By employing free hydrogen
-
cyanide as the CN source we have studied the influence of
several factors on the yield and enantiomeric excess of the
product with the aid of DOE, finding the buffer, solvent,
and meal amounts to be most important. We have synthesised
(R)-4 with 74.1% yield and 29.4% enantiomeric excess using
the defatted meal from almonds. An attempt to apply meals
from plum and apple kernels brought no further improvment.
3-Hydroxy-2,2-dimethylaldehyde (2) with Hydrogen Cya-
Experimental Section
Commercially available solvents and reagents were used
without further purification. Defatted meals were prepared
nide: General Procedure. Defatted meal from almonds as
well as apple and plum kernels was inoculated for 15 min.
with 0.02 M citrate buffer (pH 5.5, for buffer amount see
Vol. 10, No. 1, 2006 / Organic Process Research & Development
•
107

Tables 5, 8-10). Then, freshly melted aldehyde 2 in
diisopropyl ether was added, followed by 2.5 equiv. of
hydrogen cyanide (2.94 M solution in diisopropyl ether;
CAUTION: TOXIC, all operations with HCN and its
solutions should be carried out under a ventilation hood,
using appropriate protective clothes). The reaction was
maintained for given times and conditions (Tables 5, 8-10);
the meal was then filtered off and washed four times with
diethyl ether. Combined ether fractions were evaporated,
affording nitrile (R)-3 as a colourless to pale-yellow oil,
which was immediately hydrolysed to (R)-4.
Nitrile (R)-3 Hydrolysis to (R)-Pantolactone ((R)-4):
General Procedure. Concentrated hydrochloric acid (1.5
equiv) was immediately added to the crude nitrile (R)-3
followed by 5-10 mL of methanol. The reaction mixture
was then maintained for 12-16 h at room temperature and
4.6 with sodium carbonate, and methanol was evaporated
off. The residue was dissolved in diethyl ether and dried with
magnesium sulphate. The solids were filtered off and washed
with diethyl ether, and the combined filtrates were evaporated
to give a colourless to pale-yellow crystallising syrup of (R)-
pantolactone. The yield and enantiomeric excess of (R)-4
was then determined by capillary gas chromatography.
Acknowledgment
We thank Professor Cezary R o´ z˘ ycki for valuable discus-
sion concerning statistical methods of design of experiments.
Financial support from Warsaw University of Technology,
Faculty of Chemistry is gratefully acknowledged.
Received for review October 4, 2005.
OP050186S
3-4 h at reflux. After cooling, the pH was adjusted to 4.3-
108
•
Vol. 10, No. 1, 2006 / Organic Process Research & Development