Quick E. A Fast Spectrophotometric Method to Measure the Enantioselectivity of Hydrolases

Full text of DOI:10.1021/jo9707803

4
560
J . Org. Chem. 1997, 62, 4560-4561
Qu ick E. A F a st Sp ectr op h otom etr ic
Meth od To Mea su r e th e En a n tioselectivity
of Hyd r ola ses
Lana E. J anes and Romas J . Kazlauskas*
Department of Chemistry, McGill University, 801
Sherbrooke Street West, Montr e´ al, Qu e´ bec H3A 2K6,
Canada
Received May 27, 1997
Hydrolase-catalyzed resolutions of racemates are often
1
the best route to enantiomerically pure compounds.
These reactions are often more selective and cheaper
than chemical methods. To find an enantioselective
hydrolase for a target compound, researchers first screen
commercial enzymes and cultures of microorganisms and
then optimize the reaction conditions. Both screening
and optimization require measuring the enantioselectiv-
ity of the reaction. The enantioselectivity of an enzyme
F igu r e 1. First step of the quick E measure of enantioselec-
tivity of lipases toward 4-nitrophenyl 2-phenylpropanoate, 1.
Lipase-catalyzed hydrolysis of (S)-1 and the reference com-
pound, resorufin tetradecanoate, yields yellow and pink chro-
mophores, respectively. The solution turns deep orange if both
substrates are hydrolyzed, pink if only the reference compound
is hydrolyzed. The second step of the quick E is the same,
except that it uses the (R)-enantiomer of the chiral ester.
Equations 2 and 3 yield the enantioselectivity.
M
is the ratio of the specificity constants, kcat/K , for the
enantiomers, eq 1.2
,3
(
kcat/K )
M fast enant
enantiomeric ratio ) E )
(1)
(kcat/K )
M slow enant
Currently, the best method to measure E is the
3
endpoint method developed by Sih’s group, but screening
ion.7 The increase in absorbance at 404 nm revealed the
initial rates of hydrolysis of each enantiomer, but the
ratio of these rates did not give the enantiomeric ratio,
Table 1. The ratio of rates over- and underestimated E
by as much as 70% because it ignored competitive binding
of the two enantiomers to the enzyme. In other cases,
researchers found that differences in KM for the enanti-
hundreds of commercial enzymes or cultures of microor-
ganisms by this method is time-consuming. To measure
E, researchers run a test resolution, work up the reaction,
and measure conversion and enantiomeric purity of the
starting material or product. A typical example, measur-
ing the enantioselectivity of a hydrolase toward 1,
required approximately 4.5 h.
Recognizing this difficulty, researchers have reported
alternative methods to measure E by measuring initial
omers contributed a factor of 3-4 to the enantioselectiv-
ity.8
4
rates of samples with varying ratios of enantiomers or
To reintroduce competition, we added resorufin tet-
radecanoate as a reference compound.9 We monitored
by analyzing reaction progression curves.5 Unfortu-
nately, they are not significantly faster and can be less
accurate than the endpoint method. In this paper, we
report a method to measure E in 1 min from relative
initial rates of hydrolysis of pure enantiomers and a
reference compound. Researchers previously used mix-
tures of substrates to measure enzyme selectivity.6 We
extend these techniques to enantioselectivity and to rapid
spectrophotometric measurements.
the initial rates of hydrolysis of (S)-1 at 404 nm and the
reference compound at 572 nm in the same solution,
Figure 1. After taking into account the initial concentra-
tions of both substrates, the ratio of these rates yielded
the selectivity of the hydrolase for (S)-1 over the reference
compound, eq 2.
(
kcat./K )
Hydrolyses of pure enantiomers of 4-nitrophenyl 2-phe-
nylpropanoate, (S)-1 and (R)-1, and 4-nitrophenyl 2-(4-
isobutylphenyl)propanoate (ibuprofen 4-nitrophenyl es-
ter), (S)-2 and (R)-2, liberates the yellow p-nitrophenoxide
(S)-1
reference
M (S)-1
selectivity )
)
(kcat./K )
M reference
ν(S)-1
[reference]
(2)
νreference [(S)-1]
(
1) Roberts, S. M., Ed. Preparative Biotransformations; Wiley: New
York, 1992-1997. Faber, K. Biotransformations in Organic Chemistry,
nd ed.; Springer: Berlin, 1995.
2) Fersht, A. Enzyme Structure and Mechanism, 2nd ed.; Free-
man: New York, 1985; pp 103-106.
3) Chen, C. S.; Fujimoto, Y.; Girdaukas, G.; Sih, C. J . J . Am. Chem.
Soc. 1982, 104, 7294-7299.
4) (a) J ongejan, J . A.; van Tol, J . B. A.; Geerlof, A.; Duine, J . A.
A second experiment using (R)-1 and the reference
compound yielded the selectivity of (R)-1 over the refer-
2
(
(
(7) Enantiomerically pure acid and 4-nitrophenol were coupled using
1 equiv of N-ethyl-N′-[3-(dimethylamino)propyl]carbodiimide and 1
equiv of 1-hydroxybenzotriazole in anhydrous dichloromethane at 0
°C for 15 min and then stirred at room temperature for 48 h. Esters
were purified by column chromatography on silica gel eluted with ethyl
acetate and recrystallized from hexanes/ethyl acetate, 44-60% yield.
(R)-1, 99.7% ee; (S)-1, 99.4% ee; (R)-3, 98.2% ee; (S)-3, 99.6% ee.
Starting acids showed the same enantiomeric purity. All enantiomeric
purities were measured by HPLC using a Chiracel OD-H column
(Daicel) at 25 °C, eluted at 1 mL/min. 2-Phenylpropanoic acid (98/2/1
(
Recl. Trav. Chim. Pays-Bas 1991, 110, 247-254. (b) van Tol, J . B. A.;
J ongejan, J . A.; Geerlof, A.; Duine, J . A. Recl. Trav. Chim. Pays-Bas
1
991, 110, 255-262.
5) (a) Lu, Y.; Zhao, X.; Chen, Z. N. Tetrahedron: Asymmetry 1995,
(
6
, 1093-1096. (b) Rakels, J . L. L.; Romein, B.; Straathof, A. J . A.;
Heijnen, J . J . Biotechnol. Bioeng. 1993, 43, 411-422. (c) Fourneron,
J . D.; Combemel, A.; Buc, J .; Pi e´ roni, G. Tetrahedron Lett. 1992, 33,
2
469-2472.
6) (a) Berman, J .; Green, M.; Sugg, E.; Anderegg, R.; Millington,
D. S.; Norwood, D. L.; McGeehan, J .; Wiseman, J . J . Biol. Chem. 1992,
(
hexanes/2-propanol/trifluoroacetic acid, k
) 2.13); 2-(4-isobutylphenyl)propanoic acid (100/1/0.1 hexanes/2-
propanol/trifluoroacetic acid, k ′ ) 3.31; k ′ ) 4.27; R ) 1.29; R
2.43); 3 (100/1/0.1 hexanes/2-propanol/trifluoroacetic acid, k ′ ) 2.41;
R S
′ ) 3.35; k ′ ) 4.05; R ) 1.2;
R
s
2
67, 1434-1437. (b) Petithorny, J . R.; Masiarz, F. R.; Kirsch, J . F.;
R
S
s
)
Santi, D. V. Proc. Natl. Acad. Sci. U.S.A. 1991, 88, 11510-11514. (c)
Schellenberger, V.; Siegel, R. A.; Rutter, W. J . Biochemistry 1993, 32,
R
S s
k ′ ) 2.93; R ) 1.21; R ) 1.66). For analysis, samples of 1 were
hydrolyzed to the acid in aqueous NaOH.
4
344-4348.
S0022-3263(97)00780-9 CCC: $14.00 © 1997 American Chemical Society

Communications
J . Org. Chem., Vol. 62, No. 14, 1997 4561
Ta ble 1. En a n tiom er ic Ra tios of Hyd r ola ses tow a r d (()-1 a n d (()-3 Mea su r ed Usin g th e En d p oin t Meth od a a n d th e
Ra tio of Sep a r a tely-Mea su r ed In itia l Ra tes of Hyd r olysis of th e En a n tiom er s
lipaseb
eesc (%)
eepc (%)
cd (%)
E endpoint
e
initial rate (S)
f
initial rate (R)
f
rate ratio S/Rg
substrate
1
PCL
nd
17
67
50
60
25
30
4
37
39
27
29
29 ( 3
15.32
2.81
0.771
2.38
20 ( 1
1
CRL
3.5 ( 0.2
>100
1 ( 0.2
1
IPA-CRL
PPL
CAL-A
PCL
CRL
43
98
3.56
0.092
0.757
0.355
0.579
0.274
0.0128
40 ( 2
1.4 ( 0.1
4 ( 0.2
1.1 ( 0.1 (R)
1.2 ( 0.1
55 ( 5
1
nd
20
28
5.3
40.1
2
23
43.2
14.5
>99
1.1 ( 0.1
1.9 ( 0.1
3.3 ( 0.1 (R)
1.4 ( 0.1
>100
1.11
1.40
0.535
0.330
0.702
1
3
3
3
IPA-CRL
a
Reaction conditions: room temperature, 10 mM Tris buffer, pH 8, 100 mg of crude lipase or 0.7 mg of protein for IPA-CRL, 1 mmol
b
of substrate dissolved in 1 mL of acetonitrile. For spectrophotometric measurements, the lipases were dissolved in Tris buffer (25 mM,
pH 7.5) and centrifuged to remove insoluble material. PCL (lipase from Pseudomonas cepacia, Amano lipase PS, 29.3 mg solid/mL), CRL
(
2
lipase from Candida rugosa, Sigma, 36.9 mg of solid/mL; 0.5 mg protein/mL by the Bio-Rad protein assay), IPA-CRL (CRL treated with
-propanol as in ref 12a; 0.7 mg/mL of protein by the Bio-Rad protein assay), PPL (porcine pancreatic lipase, Sigma, 42.3 mg of solid/mL),
c
CAL-A (lipase A from Candida antarctica, Boehringer Mannheim, 18.2 mg of solid/mL). ees ) enantiomeric excess of remaining substrate,
eep ) enantiomeric excess of product. Determined by HPLC as described in ref 7. d Extent of conversion. e Enantiomeric ratio calculated
using eq 2. Error limits for E were estimated assuming an error of (1% for enantiomeric purity. All hydrolyses favored the (S)-enantiomer
except the PCL-catalyzed hydrolysis of 3, which favored the (R)-enantiomer. f Substrates were emulsified in aqueous solution according
to: Vorderwulbecke, T.; Kieslich, K.; Erdmann, H. Enzyme Microb. Technol. 1992, 14, 631-639. An acetonitrile solution of 1 (7.8 mM,
0
(
.5 mL) or 3 (9.04 mM, 0.5 mL) was added dropwise to Tris buffer (9 mL, 50 mM, pH 8.0) containing 0.45-0.90 w/v % Triton X-100
Pierce Surfact-Amps) and vortexed until clear. This emulsion remained clear for at least 3 h. To measure initial rates of hydrolysis, the
lipase solution (100 µL) was added to the substrate emulsion (900 µL) at 25 °C, and the increase in absorbance at 404 nm was monitored
3
g
for 15 s. No spontaneous chemical hydrolysis was detected. Values are in absorbance per second × 10 . Average and standard deviation
for three experiments.
Ta ble 2. En a n tiom er ic Ra tios of Hyd r ola ses tow a r d (()-1 a n d (()-3 Usin g th e Qu ick E Meth od
(
S)-enantiomer + refb
(R)-enantiomer + refb
404 nm 572 nm
7.50
lipasea
404 nm
572 nm
E quick E
c
substrate
1
1
1
1
1
3
3
3
PCL
13.2
1.02
1.89
0.199
2.15
0.3341
3.408
2.154
9.79
9.48
7.58
0.186
1.53
6.78
49.3
8.78
0.379
0.28
29 ( 3
CRL
9.02
3.68
0.164
1.47
4.709
52.18
5.767
3.5 ( 0.3
210 ( 20
1.4 ( 0.2
2.3 ( 0.2
2.5 ( 0.3 (R)
3 ( 2
IPA-CRL
PPL
CAL-A
PCL
CRL
0.004
0.125
0.824
0.5750
1.342
d
IPA-CRL
<0.01
>140
a
See Table 1. b Substrate solutions as for Table 1, but an acetonitrile solution of resorufin tetradecanoate (0.5 mL, 1.6 mM) was also
added. The liberated 4-nitrophenoxide and resorufin were measured in two separate measurements at 404 and 572 nm, respectively, at
3
c
2
5 °C for 15 s. No spontaneous chemical hydrolysis was detected. Values are in absorbance per second × 10 . Enantiomeric ratio
d
calculated using eq 3. Average and standard deviation for three measurements. CRL (lipase from Candida rugosa, Sigma, 214 mg
solid/mL; 1.9 mg protein/mL by the Bio-Rad protein assay).
ence compound. The ratio of these two selectivities yields
the enantiomeric ratio, eq 3.
curacies of the endpoint method also apply quick E.11
Second, it requires much smaller amounts of hydrolase
because the entire reaction occurs in the spectrophotom-
eter. This advantage may be especially useful for screen-
ing biocatalyst libraries. Third, the quick E method can
measure high enantioselectivities more easily than the
endpoint method.
There are also several disadvantages to this screen.
First, it requires pure enantiomers, albeit in small
amounts (∼20 µg per measurement in a 200 µL cuvette).
Second, the current version of quick E measures the
enantioselectivity only for chromogenic substrates. Fu-
ture versions of quick E will extend the range to all
esters. Third, spectrophotometric measurements require
clear solutions. Dissolving hydrophobic substrates in
aqueous solution requires addition of surfactants.
(
S)-1
selectivity
selectivity
(
S)-1
reference
(R)-1
reference
E )
selectivity )
(3)
(R)-1
The quick E method agreed with the endpoint method
for both 1 and 2 using five different lipases, Table 2. We
measured low (E ) 1.4), average (E ) 27), and excellent
(E ) 210) enantiomeric ratios correctly by this technique.
Each hydrolysis experiment requires 30 s; thus, the
measurement time for E was only one minute.
There are three advantages to the quick E method.
First, it is many times faster than a typical endpoint
measurement, yet has equivalent or better accuracy.
Accuracy may be particularly important for screening in
directed evolution where the improvements of each
generation are small.10 Note that quick E is based on
the same equations as the endpoint methods, so inac-
Ack n ow led gm en t. Financial support of this work
by NSERC (Canada) gratefully acknowledged. Ac-
knowledgement is made to the donors of the Petroleum
Research Fund, administered by the American Chemi-
cal Society, for partial support of this research. We
thank Ms. Cecile Vaugelaude, from INSA Rouen, France,
for the synthesis of (()-1 and Dr. Michael Trani of the
NRC Biotechnology Research Institute, Montr e´ al, for
the gift of (R)- and (S)-ibuprofen.
(
8) Wu, S. H.; Guo, Z. W.; Sih, C. J . J . Am. Chem. Soc. 1990, 112,
1
990-1995; van der Lugt, J . P.; Elfrink, H.; Evenaar, J .; Doddema,
H. J . In Microbial Reagents in Organic Synthesis; Servi, S., Ed.; Kluwer
Academic: Dordrecht, 1992; pp 261-272.
J O9707803
(9) Resorufin acetate can also serve as a reference compound.
Previous use of resorufin acetate as a chromogenic lipase substrate:
Kramer, D. N.; Guilbault, G. G. Anal. Lett. 1964, 36, 1662-1663.
Herrmann, R. Chimia 1991, 45, 317-318.
(11) Due to assumptions made in deriving eq 1, both the endpoint
method and quick E will give inaccurate enantioselectivities in two
common situationssimpure biocatalyst and reactions inhibited by
product.
(
10) Moore, J . C.; Arnold, F. H. Nature Biotechnol. 1996, 14, 458-
4
68.