Enzymatic Desymmetrization of Diol

In the improved synthesis (20), shown in Fig. 6, a diol olefin 13 could be quickly assembled from bromodifluorobenzene. This underwent an iodocyclization reaction to form a 2,2,4-trisubstituted tetrahydrofuran in which the desired ds-isomer

Fig. 6. Improved synthesis.

Fig. 6. Improved synthesis.

14 predominated (85:15 cis-.trans) (21,22). Introduction of the triazole followed by formation of the chlorosulfonate provided the key intermediate 5, which could be condensed with the side chain to form SCH56592.

Like the (KJ-triol 8 in the previous route, diol 13 is also a 2-substituted-1,3-propanediol, providing a similar opportunity for biocatalysis. If a suitable enzyme could be found to desymmetrize diol 13, then the stereoselectivity of the iodocyclization would transfer the chirality generated by the enzyme to the newly formed benzylic center. Acylation of the pro-S hydroxyl of 13 would result in a five-step sequence to (K,S)-sulfonate 5 (acylation, iodocyclization, triazole introduction, deacylation, and sulfonation), whereas pro-K acylation would require two extra steps to invert the (2K) chiral center. Pro-K hydrolysis, requiring one extra step to generate diester 15, would result in a six-step sequence to 5 (Fig. 7).

Approximately 205 commercially available enzyme preparations were screened for selective acetylation of diol 13 in toluene using vinyl acetate as the acyl donor. The best results are collected in Table 2; C. antarctica lipase B showed the best pro-S selectivity, whereas a lipase from Humicola lanuginosa displayed excellent pro-K selectivity. As earlier, the major effort was to obtain not only the highest possible enantiomeric excess (ee) of the monoester product, but to maximize the yield of monoester 16 and minimize the yield of both unreacted diol 13 and the diester 15 formed by overreaction.

The availability of two enzymes with opposite prochiral selectivity allowed three approaches to the preparation of the desired (2S)-monoester precursor; pro-K acylation of 13 or pro-K hydrolysis of 15 using Amano lipase CE, or proS acylation of 13 using Novo SP435 (Fig. 8).

2.2.1. Pro-R Acylation

Pro-K acetylation of diol 13 using lipase CE gave the monoacetate (K)-16a with high chemical and optical yield. As the reaction profile illustrates (Fig. 9), the prochiral selectivity was excellent with monoacetate being formed in high enantiomeric excess and very little diacetate being formed; vK1 >> vS1 or vS2 (Fig. 8). However, protection/deacylation was required to invert the chiral center, resulting in a seven-step sequence from diol 13 to sulfonate 5. Because attempts to introduce

17 5

Fig. 7. Six-step sequence.

17 5

Fig. 7. Six-step sequence.

Table 2

Acetylation of Diol 13: Initial Screen Results

Conditions: Diol 13, 50 mg; enzyme, 10-200 mg; VinylOAc, 5-10 equiv; toluene, 1.0 mL; room temperature.

a tetrahydropyranyl protecting group resulted in significant racemization, presumably via 1,3 acyl migration, this route was not pursued.

Fig. 8. Approach to the preparation of the desired 2S-monoester precursor.

Fig. 8. Approach to the preparation of the desired 2S-monoester precursor.

2.2.1.1. Synthesis of (2R)-2[2'-(2",4"-Difluorophenyl)]-2'-Propenyl-1-Acetoxy-3-Hydroxypropane (R)-16b

Lipase CE (2.0 g) was added to a solution of diol 13 (10.0 g, 43.8 mmol) and vinyl acetate (8.0 mL, 86.8 mmol) in HPLC-grade toluene (200 mL). The mixture was stirred at room temperature with periodic HPLC monitoring. After 26 h the reaction mixture was filtered through a Celite pad that was washed with tert-butyl metyl ether (TBME) (20 mL). The filtrate was evaporated (<30°C) and purified by column chromatography, eluting with 10-35% EtOAc/ hexane. Pooling the relevant fractions yielded the diacetate 15b (0.66 g), a mixture of monoacetate and diacetate (0.89 g), and pure monoacetate (K)-14b as a viscous oil (10.37 g, 87.6%; 97.0% ee): [a]2D4 = +13.9 (c 1.085, EtOH).

2.2.2. Pro-R Hydrolysis

Lipase CE-catalyzed pro-K hydrolysis of the diester 15 would yield the (S)-monoester, for a six-step total sequence from diol 13 to the key sulfonate 5. Although hydrolysis of the dibutyrate 15b produced the monobutyrate (S)-16b with high enantiomeric excess (Fig. 10), the chemoselectivity was poor and the reaction mixture consisted of a mixture of diol 13, dibutyrate 15b, and the desired monoester (S)-16c.

Fig. 9. Acylation of 13 with Amano lipase CE: Diol 13, 5 g; lipase CE, 5 g; vinyl acetate, 5 equivalents.; toluene, 100 mL; room temperature.
Fig. 10. Amano lipase CE hydrolysis: DiOBu 15b, 7 g; Amano CE, 5 g; 50 mM KCl, 63 mL; pH 7.5; 22°C.

Time min

Fig. 11. Acylation of diol 13 with Novo SP435: Diol 13, 50 mg; Novo SP435, 10 mg; VinylOAc, 10 equivs; Toluene, 1.0 mL; room temperature.

Time min

Fig. 11. Acylation of diol 13 with Novo SP435: Diol 13, 50 mg; Novo SP435, 10 mg; VinylOAc, 10 equivs; Toluene, 1.0 mL; room temperature.

2.2.3. Pro-S Acylation

Novo SP435-catalyzed pro-S acetylation of diol 13 provided (S)-monoacetate 16a directly, for an overall five-step route to the key sulfonate 5. Because the prochiral selectivity is lower than for Lipase CE (vSl/vRl ~20), optical purity was purchased at the cost of chemical yield, and a significant amount of the diacetate 15a had to be formed before the ee of the remaining (SJ-monoacetate 16a reached useful levels (Figs. 8 and 11). In other words, the moderate prochiral selectivity of the initial desymmetrization reaction is compensated for in a subsequent kinetic resolution, in which the initially formed monoester 16 reacts further to form the diester 15 (23). Because the enzyme displays the same preference in both the desymmetrization and the kinetic resolution steps (vS2 > vR2), the undes-ired (R)-monoester is acylated faster than the desired (Sj-isomer and the ee of the monoester increases with conversion. However, this comes at the cost of reduced yield of monoester 16 and increased levels of the diester byproduct 15.

Because it provided the desired (S)-monoester directly under operationally simple conditions, the Novo SP435-catalyzed acetylation of diol 13 was selected for optimization. Because the prochiral selectivity of the enzyme was not absolute, the enantiomeric excess of the (S)-monoacetate 16a increased over the course of the reaction at the expense of chemical yield. The question

13 R,R' = H 15a R,R' = COCH3 (S)-16a R = COCH3, R' = H

13 R,R' = H 15a R,R' = COCH3 (S)-16a R = COCH3, R' = H

No Reaction

Fig. 12. Iodocyclization of unreacted diol 13, yielding ravemic material 14.

Fig. 12. Iodocyclization of unreacted diol 13, yielding ravemic material 14.

was: What compromise should be struck between optical and chemical yield? Like the (S)-monoacetate 16a, unreacted diol 13 also undergoes iodocyclization to yield racemic material 14 (Fig. 12), which, if carried forward, would ultimately degrade the optical purity of the key intermediate sulfonate 5. The diester 15a, lacking a hydroxyl function, does not participate in the iodocyclization reaction and, in fact, may be recovered as the starting diol 13 at a subsequent stage.

Although minimization of both diol and diester would be desirable and would give the maximum yield, it was more important that the amount of unreacted diol be minimized. Thus, it was preferable to run the enzymatic acetylation reaction until no diol 13 remained and to accept the yield loss as a result of the formation of the diacetate 15a. This ensured that the enantiomeric excess of the (5)-monoacetate 16a was sufficiently high. In practice, the reaction was carried out until <1-2% of diol 13 remained, at which point the optical purity of the monoacetate was typically >95% ee.

The optimization of this process has been described in detail elsewhere (19). Examination of a number of solvents, acetylating agents, and temperatures indicated that the reaction worked best using vinyl acetate in either toluene or acetonitrile at 0-15°C. As the iodocyclization reaction that followed this enzyme step was expected to be run in acetonitrile, this was chosen to avoid changing solvents between steps. Efforts to reduce the quantity of vinyl acetate showed that the reaction could be run with as little as 1.3 equivalents.

It was hoped that the use of a bulky acylating agent would decrease diester formation and increase the selectivity of the initial desymmetrization reaction. Using trifluoroethyl isobutyrate in toluene did suppress diester formation, but the monoester was formed with low ee; even after extended reaction and formation of diisobutyrate, the ee of the monoisobutyrate remained low. With trifluoroethyl 2-methylbutyrate, diester formation was completely suppressed, but the monoester was again formed in very low ee. Thus, vinyl acetate remained the acylating agent of choice.

Good volumetric productivity was attained by increasing the concentration of substrate to 20% and reducing the loading of the enzyme to 5% by weight of diol. To demonstrate enzyme recovery and reuse, a 50-mg sample of Novo SP435 was carried through multiple reactions over a period of 14 d. After 10 cycles, the enzyme sample still produced monoacetate of acceptable quality, but the rate of the reaction had slowed down considerably. This decrease in rate was attributed more to mechanical losses of enzyme during recycling than to inactivation of the enzyme. The conditions initially transferred to the pilot plant were the following:

• 2.0 Equivalents vinyl acetate

• Industrial-grade MeCN

• 0°C Reaction temperature

• Monoacetate specification: (S)-16a > 95% ee; <2% diol 13 remaining

A representative result is shown in Table 3 (run 1). The specifications for the enzymatic acylation were subsequently tightened to produce monoacetate (£)-16a of 98-99% ee with <1% unreacted diol remaining. This could be achieved using the above conditions but running the acetylation out to approx 30% diacetate formation (Table 3, run 2). The latter result also showed that there was little or no loss of optical purity during the iodocyclization reaction.

2.2.3.1. Synthesis of (2S)-2[2'-(2",4"-Difluorophenyl)]-2'-Propenyl-1-Acetoxy-3-Hydroxypropane (S)-16a

A mixture of diol 13 (5.01 g, 21.95 mmol) and Novo SP435 (0.26 g) in industrial-grade MeCN (25 mL) was cooled in an ice bath for 15 min. Vinyl acetate (4.0 mL, 43.4 mmol) was added, and the mixture stirred at 0°C. After 6 h, the reaction mixture was filtered, the beads washed with EtOAc (30 mL), and the combined organics evaporated (< 30°C). The residue was purified by column chromatography, eluting with 10-50% EtOAc/hexanes. Pooling the relevant fractions yielded the monoacetate (£)-16a (4.23 g, 71.3%; 98.2% ee): [a] D = -13.9 (c 1.688, EtOH).

2.2.4. Diol Desymmetrization: The Role of Solvent and Acylating Agent

The Novo SP435-catalyzed acetylation of diol 13 was used routinely for the preparation of hundreds of kilograms of the monoacetate (£)-16a. However,

Table 3

Scale-Up of Enzymatic Acetylation

aee of S-16a. bee of R,S-17.

the continued loss of up to 30% of product as the diacetate, even though it was potentially recoverable and reusable, was a major weakness in the process. The eventual solution of this problem illustrates the effect that a subtle interplay of solvent and acylating agent can have on the development of a commercial enzymatic process and the hazards of making assumptions while screening in an empirical field.

The enzymatic acetylation of diol 13 had been examined in a series of common solvents (Table 4), and the best results, in terms of product distribution and selectivity, were observed in MeCN and toluene, the reaction being faster and the diol more soluble in MeCN.

The enzymatic acylation was also investigated using 11 different acylating agents in either MeCN or toluene. Vinyl acetate and acetic anhydride showed similar selectivities when compared in MeCN (Table 5), as would be expected if the prochiral discrimination was dictated solely by approach of the diol substrate to the covalent acyl-enzyme intermediate at the active site of the enzyme. It was hoped that the use of bulky acylating agents would improve the product distribution of the reaction by (1) increasing the prochiral selectivity of the initial desymmetrization and (2) decreasing diester formation. Although the use of trifluoroethyl isobutyrate (TFEOiBu) in toluene did decrease formation of the diester, the enantiomeric excess of the monoester remained low; extending the reaction time did not improve the ee of the monoester even though substantial amounts of the diester were formed. Because MeCN and toluene performed equally well as media in the enzymatic reaction with vinyl acetate (Table 4, runs 7 and 8), and because the selectivity of the desymmetrization was presumed to depend only on the approach of the diol to

Table 4

Product Distribution and Enantiomeric Excess at >94% Diol Conversion in Various Solvents

Product distribution (%)

Table 4

Product Distribution and Enantiomeric Excess at >94% Diol Conversion in Various Solvents

Product distribution (%)

Vinyl acetate

Diol/enzyme ratio

Time

Diol

Monoacetate

Diacetate

% ee

Run

Solvent

equiv.

(w/w)

(min)

13

(S)-16a

15a

(S)

1

i-P^O

10.0

4.0

90

6

84

10

91

2

THF

10.0

4.0

120

2

77

22

90

3

Dioxane

10.0

4.0

90

1

75

24

93

4

Acetone

10.0

4.0

90

1

83

16

94

5

i-PrOAc

2.0

9.0

260

1

78

21

97

6

TBME

2.0

9.0

226

3

75

22

94

7

MeCN

5.0

4.0

60

1

85

14

96

8

Toluene

5.0

4.0

120

1

85

14

96

Conditions: Diol 13, 50 mg; solvent, 1.0 mL; 0°C (except run 3, 20°C).

Conditions: Diol 13, 50 mg; solvent, 1.0 mL; 0°C (except run 3, 20°C).

Table S

influence of Acylating Agent

Product distribution (%)

Diol

Diol/enzyme

Temp

Time

Diol

Monoester

Diester

%ee

Acyl agent

Equiv

Solvent

13 (M)

ratio (w/w)

(°C)

(min)

13

(S)-16

1S

(S)

1

VinylOAc

1.4

MeCN

0.93

5.0

0

115

1

l8

21

96

2

VinylOAc

Neat

Neat

0.90

10.0

0

165

1

84

15

96

3

VinylOAc

2.0

MeCN

0.88

19.0

0

2l0

0

84

16

95

4

TFEOAc

2.0

MeCN

0.85

15.0

0

385

12

86

2

91

5

TFEOiBu

10.0

Toluene

0.22

5.0

RT

420

0

96

4

89

1250

0

8l

13

91

6

Ac2O

2.0

MeCN

0.85

4.6

RT

120

1

81

18

9l

l

VinylOAc

2.0

MeCN

0.85

20.0

0

200

2

80

19

9l

8

VinylOAc

2.0

MeCN

0.85

94.0

0

1150

1

82

1l

9l

the acyl-enzyme intermediate, no other combinations of solvent and isobutyrate reagents were examined.

However, when the acylation of diol 13 was later examined in MeCN with isobutyric anhydride as the acylating agent, the reaction was found to be more selective than with vinyl acetate (24). At first, this result seemed strange, given that TFEOiBu had been a poor acylating agent and that the identical acyl-enzyme intermediate should be formed from either TFEOiBu or isobutyric anhydride. Subsequent investigation showed that the role of the solvent was crucial. When run under identical conditions, the Novo SP435-catalyzed acylation with TFEOiBu was found to be faster and significantly more selective in MeCN than in toluene (compare runs in Table 6 and Fig. 13).

When isobutyric anhydride was used as the acylating agent, a similar solvent difference was also observed. In this case, although the reaction was faster in toluene, it was more selective in MeCN. After 2 h in toluene, the monoisobutyrate 16c was formed in only 95% ee even though 22% of the diester had been formed (Table 7, run 1). In the same period in MeCN (run 2), only 9% diester had been formed, but the monoester 16c had already reached 99% ee. This difference in reactivity and selectivity is illustrated in Fig. 14.

Two other changes were made to the reaction conditions. The temperature was lowered to avoid any possible nonenzymatic acylation by the anhydride, resulting in a small improvement in selectivity. Solid NaHCO3 was also added to the reaction mixture to prevent any racemization by acid-catalyzed 1,3-acyl migration. These conditions have been used to prepare monoester 16c consistently with >98% ee. Based on these results, the plant process was modified as follows:

20% Diol 13 solution 5% Novo SP435 (w/w) 1.2 Equivalents isobutyric anhydride Industrial-grade MeCN -10 to -15°C Reaction temperature 0.7 Weight equiv. NaHCO3

Monoacetate specification: (S)-16c < 98% ee; <1% diol 13 remaining

In developing the desymmetrization of diol 13, 205 enzymes, 8 solvents and 11 acylating agents were examined. However, every one of the 18,040 possible combinations could not be individually examined. The linear approach to identifying the best enzyme followed by the optimum solvent followed by the optimum acylating agent did lead to a local maximum on the reaction yield surface: Novo SP435-catalyzed acetylation with vinyl acetate in MeCN. This local maximum was achieved after making some reasonable assumptions:

1. Because the reaction showed the same selectivity with vinyl acetate in toluene and MeCN, a similar lack of solvent effect would be seen with isobutyrate acylating agents (25).

Table 6

Acylation of Diol 13 with Trifluoroethyl Isobutyrate

J^-OH

Novo SP435

F Ii flT

JDCOiPr ^H 11

JDCOiPr

13

TFEOiBu MeCN or Toluene

CSJ-16C

15c

Run

Solvent

Time min

% Diol 13

% MonoOiBu (S)-16c

% DiOiBu 15c

% ee

1

Toluene

435

0.4

97.2

2.4

89

2

Toluene

420

0.3

96.0

3.7

89

1250

0

87.2

12.8

91

3

MeCN

114

0.2

93.0

6.8

97

4

MeCN

120

0

92.0

8.0

98

Conditions: Runs 1, 2, and 4: Diol 13, 60 mg; Novo SP435, 13-14 mg; TFEOiBu, 10 equiv.; solvent, 1.0 mL; RT. Run 3: Diol 13, 1.0 g, Novo SP435, 0.1 g; TFEOiBu 5 equiv.; solvent, 5.0 mL; RT.

Conditions: Runs 1, 2, and 4: Diol 13, 60 mg; Novo SP435, 13-14 mg; TFEOiBu, 10 equiv.; solvent, 1.0 mL; RT. Run 3: Diol 13, 1.0 g, Novo SP435, 0.1 g; TFEOiBu 5 equiv.; solvent, 5.0 mL; RT.

Fig. 13. The enzymatic isobutyrylation of diol 13 with TFEOiBu in Toluene (Graph A, Run 2, Table 6), or in MeCN (Graph B, Run 4, Table 6).

Table 7

Acylation of Diol 13 with Isobutyric Anhydride

Table 7

Acylation of Diol 13 with Isobutyric Anhydride

Time

% Diol

% MonoOiBu

%DiOiBu

%

Run

Solvent

min

13

(S)-16c

15c

ee

1

Toluene

120

0.0

77.5

22.4

95

2

MeCN

120

0.0

90.5

9.5

99

Conditions: Diol 13, 1.0 g; solvent, 5.0 mL; Novo SP435, 50 mg; (i-PrCO)2O, 1.3 equiv.; RT.

Conditions: Diol 13, 1.0 g; solvent, 5.0 mL; Novo SP435, 50 mg; (i-PrCO)2O, 1.3 equiv.; RT.

Fig. 14. The enzymatic isobutyrylation of diol 13 in Toluene (Graph A, Run 1, Table 7), or in MeCN (Graph B, Run 2, Table 7).

2. The reaction mechanism suggested that prochiral selectivity would be determined solely by approach of the substrate diol to the acyl-enzyme intermediate. Reactions in MeCN using vinyl acetate and acetic anhydride as acylating agents and proceeding through the same acyl-enzyme intermediate showed similar selectivities. Thus, it was assumed that because trifluoroethyl isobutyrate showed poor selectivity in toluene, isobutyric anhydride would also show poor selectivity in MeCN.

These assumptions proved to be incorrect. By failing to recognize the interplay of enzyme/solvent/acylating agent in determining selectivity, the initial process subjected to scale-up resulted in significant loss of yield. However, the proper choice of solvent and acylating agent has improved the yield of the desired product by up to 20%. Although seductive in their rational appearance and in their ability to simplify the task of reaction optimization, even simple assumptions concerning interrelated variables can be misleading. An automated approach, testing every possible variable, or a multivariate investigation of reaction parameters (26,27), might have led to a more global maximum sooner. This reaction is a specific example of the need for process chemists to utilize these tools.

2.2.4.1. Synthesis of (2S)-2[2'-(2",4"-Difluorophenyl)]-2'-Propenyl-1-Isobutyroxy-3-Hydroxypropane (S)-16c

A mixture of diol 13 (1.0 g, 4.4 mmol) and Novo SP435 (52 mg) was stirred in industrial-grade MeCN (5 mL) under N2 at 25°C. Isobutyric anhydride

(0.84 mL, 5.1 mM) was added and the reaction monitored by chiral HPLC every 15 min. After 100 min, the reaction mixture was filtered through a Celite pad that was washed with TBME (25 mL). The filtrate was washed with water (20 mL), saturated NaHCO3 (20 mL) water (20 mL), saturated NaCl (10 mL), and dried (MgSO4), filtered and evaporated. The residue was purified by column chromatography, eluting with 20-30% EtOAc/hexanes. Pooling the relevant fractions yielded the monoisobutyrate (S)-16c (1.2 g, 90.2%; 98.2% ee): [a] D = -13.95 (c 1.584, EtOH).

Was this article helpful?

0 0

Post a comment