вход по аккаунту


Comprehensive Step-by-Step Engineering of an (R)-Hydroxynitrile Lyase for Large-Scale Asymmetric Synthesis.

код для вставкиСкачать
Tailor-Made Enzyme Catalyst
Comprehensive Step-by-Step Engineering of an
(R)-Hydroxynitrile Lyase for Large-Scale
Asymmetric Synthesis**
Anton Glieder,* Roland Weis, Wolfgang Skranc,
Peter Poechlauer, Ingrid Dreveny, Sandra Majer,
Marcel Wubbolts, Helmut Schwab, and Karl Gruber
Optically active a-hydroxyacids, such as mandelic acid, and
their derivatives are important building blocks for the
production of pharmaceuticals.[1] Attractive synthetic routes
are made available by hydroxynitrile lyases (HNLs). These
enzymes catalyze the asymmetric addition of HCN to
aldehydes to provide cyanohydrins in high yields and with
high selectivities (Scheme 1).[2, 3] The cyanohydrin products
Scheme 1. Enantioselective synthesis of (R)-a-hydroxycarboxylic acids
(R)-3 from aldehydes 1 by the nitrilase and the hydroxynitrile lyase
(HNL) routes. Other methods, such as the hydrolytic resolution of
cyanohydrins in the absence of an additional catalyst, are limited to a
maximum yield of 50 %.
undergo chemical hydrolysis under acidic conditions at
elevated temperature to give the corresponding a-hydroxyacids without racemization.[1] An alternative route, which
involves enantioselective nitrilases,[4, 5] should also theoretically lead to quantitative yields through substrate racemization in situ at high pH values. However, many nitrilases suffer
[*] Dr. A. Glieder, S. Majer, Prof. Dr. H. Schwab
Institut f-r Biotechnologie, Technische Universit2t Graz
Petersgasse 12, 8010 Graz (Austria)
Fax: (+ 43) 316-873-8434
R. Weis, Prof. Dr. K. Gruber
Research Centre Applied Biocatalysis
Steyrergasse 17, 8010 Graz (Austria)
Dr. I. Dreveny, Prof. Dr. K. Gruber
Institut f-r Chemie, Karl-Franzens-Universit2t Graz
Heinrichstrasse 28, 8010 Graz (Austria)
Dr. W. Skranc, Dr. P. Poechlauer, Dr. M. Wubbolts
R & D Center Linz
DSM Fine Chemicals Austria Nfg. GmbH & Co KG
St.-Peter-Strasse 25, 4021 Linz (Austria)
[**] We thank the SFB Biokatalyse (F0101), the TIG, the province of
Styria, SFG, and the city of Graz for financial support.
Supporting information for this article is available on the WWW
under or from the author.
Angew. Chem. Int. Ed. 2003, 42, 4815 –4818
from low turnover frequencies (TOF), low stereoselectivity,
and rapid deactivation at high substrate concentrations.
(R)-2-Chloromandelic acid (3 b) is a key intermediate for
the production of a widely administered anticoagulant that
reduces the risk of cardiovascular events in patients with
acute coronary syndromes. However, ortho-substituted substrates perform poorly in all known syntheses of chiral
cyanohydrins, and their use results in low yields, low
TOF values, and/or low ee values.[4, 6–9] Furthermore, the
conversion of aromatic aldehydes into cyanohydrins with
the only reported recombinant (R)-hydroxynitrile lyase
(LuHNL, cloned from Linum usitatissimum) does not reach
completion.[10] (R)-HNL from Prunus amygdalus (PaHNL),
which shows a broad substrate specificity, is only available
from very limited natural sources, and the recombinant
production of any active FAD (flavin adenine dinucleotide)containing HNL was previously unsuccessful.[11] Herein we
report the first recombinant PaHNL that retains its activity
under process conditions and can be used in emulsion systems
at low pH values. We also demonstrate the success of a
generally applicable engineering concept, which could be
used to improve the productivity of this enzyme during the
fermentation process, and which led to a high TOF value for
the challenging substrate 1 b. Our goal was to develop a
commercially viable biocatalyst for the production of (R)-3 b
by the HNL route, or in other words to make the production
of optically pure (R)-2 b possible in high yield in the presence
of a small quantity of a robust enzyme.
A considerable amount of data from preparative reactions
and from biochemical and structural studies[2, 3, 12–17] are
available for PaHNL. We cloned the Pa_hnl5 gene from
Prunus amygdalus to have unlimited access to PaHNL. This
gene was very similar to hnl genes from Prunus dulcis (mdl1,
99 % identity) and Prunus serotina (mdl5, 94 % identity),
which are expressed specifically in floral tissues. By overexpressing the Pa_hnl5 gene, including its native plant
secretory-signal sequence, in the methylotrophic yeast
Pichia pastoris, we obtained 250 mg of the active secreted
enzyme per liter of culture supernatant. The highly glycosylated enzyme showed a specific activity of 295 30 mmol min 1 mg for the cleavage of 2 a, which is twice as
high as the specific activity of PaHNL isolated from almond
seeds (Sigma M-6782, Lot41H4016; 160 30 mmol min 1 mg 1). The activity of the recombinant isoenzyme
PaHNL5 did not change significantly over the range
pH 2.5–6.5. PaHNL5 was surprisingly stable under acidic
conditions (Figure 1), under which the competitive chemical
reaction is suppressed[18, 19] and the enantiomerically pure
cyanohydrin product is also stable. This extremely important
technical feature of HNL reactions in emulsion systems
makes the production of pure enantiomers possible even from
substrates that react slowly and facilitates effective enzyme
recycling. No other HNL is known to be so stable at low
pH values. Although the molecular foundation for this
property is still unclear, overglycosylation by Pichia can be
excluded as the only cause, as deglycosylation of recombinant
PaHNL5 by endoglycosidase H in a nondenaturing buffer at
37 8C to leave just one N-acetylglucosamine substituent on
each of the modified Asn residues did not lead to instability at
DOI: 10.1002/anie.200352141
2003 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Table 1: Enantioselective synthesis and cleavage of chloro-substituted
mandelonitrile derivatives.[a]
Figure 1. Plot of the relative activities (rA) of a series of enzymes for
the cleavage of 2 a at pH 5.0 after incubation at pH 2.6. PaHNL5 (H21,
*), which was expressed in Pichia pastoris with the signal sequence of
the a mating factor of S. cerevisiae, and its A111G variant ( ! ) showed
remarkable stability at low pH values. The deglycosylated enzyme
(H21_deglyco, ! ) was slightly less stable. The commercially available
native PaHNL from almond seeds (Sigma; *) is shown for
pH 2.6, in contrast to the significant denaturation observed
for a PaHNL preparation from almond seeds (Sigma) under
the same conditions (Figure 1). The deglycosylated enzyme
showed a slight decrease in stability after treatment at 37 8C.
The estimated deactivation half-life values of the recombinant protein PaHNL5_L1Q (H21), the enzyme with the
additional mutation (A111G), the deglycosylated enzyme
(H21_deglyco), and the native enzyme from almond seeds
(Sigma) under these conditions were determined from the
deactivation curves as 530, 530, 250, and 20 min, respectively (Figure 1).
In the next step, the enzyme productivity was increased 4–
4.5-fold by changing the first amino acid of the mature protein
(leucine to glutamine; L1Q) and replacing the native plant
secretion-signal peptide with the a-mating-factor prepro
leader from Saccharomyces cerevisiae. The secretion of
near-pure enzyme into the culture supernatant and the
production of 1 g of enzyme per liter of supernatant ensures
almost unlimited availability of this novel biocatalyst.
High enantioselectivity was observed for the synthesis of
the commercially interesting intermediate (R)-2 b catalyzed
by the pH-stable recombinant PaHNL isoenzyme 5. The
TOF values were more than 10–15 times lower than those
observed in the reaction with the natural unsubstituted
substrate 1 a (Table 1). No effective method for the stereoselective conversion of ortho-substituted mandelonitrile or
benzaldehyde derivatives into cyanohydrins has yet been
reported. The observations made for the reaction of native
PaHNL from natural sources,[9] for the conversion of 2 b by
nitrilases,[4] and for the nonbiocatalytic route with titanium or
vanadium catalysts[6, 7, 20] were similar to those for the reaction
of PaHNL isoenzyme 5. To minimize unfavorable steric
interactions within the active site, alanine 111 and valine 317
were replaced by smaller glycine residues through sitedirected mutagenesis. Computer substrate-docking simulations of (R)-2 b, (R)-2 c, and (R)-2 d in the active site of a
homology model of PaHNL5 (based on the known structure
2003 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Substrate Product
(R)-2 a
(R)-2 a
(R)-2 a
(R)-2 b
(R)-2 b
(R)-2 b
(R)-2 c
(R)-2 c
(R)-2 c
(R)-2 d
(R)-2 d
(R)-2 d
(S)-2 a + 1 a
(S)-2 a + 1 a
(S)-2 a + 1 a
Mutation Specific activity[b]
[mmol min 1 mg]
1450 400
A111G 1150 400
V317G 295 50
67 25
A111G 409 60
155 30
A111G 390 100
V317G 290 50
298 50
A111G 566 40
26 13
325 30
6 1.5
[a] All synthetic reactions were performed with 15 mm substrate and
0.5 mg of the purified enzyme in an emulsion of citrate buffer (50 mm,
3.8 mL), tert-butyl methyl ether (2.1 mL), and HCN (1.2 mL) at 10 8C and
pH 3.4. All PaHNL variants were based on PaHNL5 with the mutation
L1Q. The cleavage activity for 2 a was determined photometrically within
the first 3 min of a linear increase in product concentration (l = 280 nm;
T = 25 8C; substrate concentration: 12 mm; phosphate–citrate buffer
(0.1 m; pH 5). [b] The specific activity is based on protein concentrations
determined by a Lowry assay with PaHNL from Sigma as a standard and
on product-formation rates within the first 15 min (1 b) or 5–10 min (1 a–
d). In these time periods 5–20 % of the product was formed.
of the homologous isoenzyme PaHNL1[15] from almond
seeds) allowed the identification of amino acid residues
situated near the chloro substituents of the substrate
(Figure 2). In the case of (R)-2 b, the side chains of A111
and V317 are in close contact with the chlorine atom at the
ortho position of the aromatic ring. Steric interactions with
these two residues, particularly between A111 and (R)-2 b, are
probably responsible for the decreased enzymatic performance observed with this substrate. A chloro substituent in the
meta position did not cause any noticeable steric problems
(Figure 2).
For both mutations, A111G and V317G, a dramatic
decrease in activity to just 2–5 % was observed in the cleavage
of 2 a (Table 1). This nearly total loss of activity was not
surprising, as there is a very high risk of destroying enzymatic
activity when highly conserved amino acid residues in the
hydrophobic cavity of the enzyme are changed. However, a
completely different situation arose in the synthetically more
important HCN addition reaction. Whereas the mutant
V317G also showed poor reactivity in the synthesis of (R)2 a, (R)-2 b, and (R)-2 d, the A111G variant retained high
hydrocyanating activity. Turnover rates of the A111G variant
with the natural substrate 1 a were only 20–30 % lower than
those of PaHNL5_L1Q. Moreover, the activity of the enzyme
in the synthesis of (R)-2 b was strongly enhanced through this
single mutation (specific activity of A111G: 409 mmol mg 1 min; TOF: 390 sec 1; six times faster than with the wildtype
enzyme). Both mutants catalyzed the conversion of compound 1 c at higher rates than the native enyzme and with
excellent enantioselectivity (> 99 % ee). The synthesis of (R)-
Angew. Chem. Int. Ed. 2003, 42, 4815 –4818
Figure 2. The active site of PaHNL5 with (R)-2 b (2-Cl-Man) and (R)-2 c
(3-Cl-Man). a) Steric interactions between the ortho chloro substituent
(green area) of the substrate 2-Cl-Man and A111 and V317 (gray
areas). b) Close contact of L343 and L331 (gray areas) with the
meta-chloro substituent (green area) of 3-Cl-Man.
An ee value as high as that observed with the A111G
variant (97 % ee) was described for the enzymatic step of the
nitrilase route,[22] albeit with low substrate concentrations.
Furthermore, the specific activity reported was more than 100
times lower, with only 3 mmol of product per minute per mg of
nitrilase I. Protein folding and protein stability are often
adversely affected by mutations of active-site residues, which
frequently impedes the generation of useful mutants. Our
A111G variant was also expressed well when the fermentation was scaled up and showed high stability at low pH values.
PaHNL was the first biocatalyst described for asymmetric
synthesis.[23] However, its instability at low pH values (conditions under which the competitive chemical reaction is
suppressed), limited availability from relatively expensive
natural sources, and low activity and selectivity in reactions
with ortho-substituted aryl aldehydes have contributed to the
fact that PaHNL has not yet found application in industrial
processes. We believe that variation in the fractions of the
individual isoenzymes in enzyme preparations from different
batches and suppliers are the reason for the varying enantioselectivities and conversion rates observed by Gotor and coworkers.[24] Through the cloning and expression engineering
of the gene that encodes HNL isoenzyme 5 from Prunus
amygdalus we have established the foundation for a cheap
and unlimited supply of a recombinant PaHNL with novel
technological features. Rational engineering of the active site
led to the adaptation of the enzyme for the difficult stereoselective conversion of the unnatural, industrially interesting
substrate 1 b into (R)-2 b with high enantioselectivity. Thus,
almost 100 years after the discovery of the first stereoselective
biocatalyst, an industrial application of this type of enzyme on
a multiton scale is now within reach, as a result of a
comprehensive molecular-engineering program.
Received: June 13, 2003 [Z52141]
Published Online: September 29, 2003
Keywords: asymmetric synthesis · C C coupling ·
enzyme catalysis · protein design · protein engineering
2 d was slightly improved through the mutation A111G. In a
reaction with a high concentration of aldehyde (3 m based on
the volume of the aqueous phase), 1 b (150 mmol) was
converted in the presence of the enzyme ( 5 mg) at pH 3.4
into (R)-2 b (21.1 g); 2.9 % of the starting aldehyde 1 b was
also recovered. This result corresponds to 95.9 % of the
theoretical yield of (R)-2 b. No by-products were observed
after a reaction time of 7 h and extraction with tert-butyl
methyl ether. The R enantiomer was obtained in 96.5 % ee.
This represents a tremendous improvement in the synthesis of
(R)-2 b by the HNL route. In the best case reported
previously,[9] an ee value of 83 % was observed at pH 3.3
when 25 times as much native enzyme was used (based on
calculations with the specific enzyme activity of the most
commonly used native PaHNL in the cleavage reaction of 2 a
(160 mmol min 1 mg)). In spite of significant advances, the use
of non-enzymatic catalysts in these asymmetric synthetic
steps led to only 80 % ee, low yields for ortho-substituted
substrates,[7, 20, 21] and the necessity for a recrystallization step
to obtain 3 b with high optical purity.
Angew. Chem. Int. Ed. 2003, 42, 4815 –4818
[1] H. Groger, Adv. Synth. Catal. 2001, 343, 547 – 558.
[2] F. Effenberger, S. Forster, H. Wajant, Curr. Opin. Biotechnol.
2000, 11, 532 – 539.
[3] D. V. Johnson, A. A. Zabelinskaja-Mackova, H. Griengl, Curr.
Opin. Chem. Biol. 2000, 4, 103 – 109.
[4] G. Desantis, Z. Zhu, W. A. Greenberg, K. Wong, J. Chaplin, S. R.
Hanson, B. Farwell, L. W. Nicholson, C. L. Rand, D. P. Weiner,
D. Robertson, M. J. Burk, J. Am. Chem. Soc. 2002, 124, 12 922.
[5] M. Ress-Loeschke, B. Hauer, R. Mattes, D. Engels (BASF, A.G.), DE 10010149, 2000 [Chem. Abstr. 2001, 135, 223444].
[6] C. Bensoussan, M. C. Bontoux, P. M. Gelo, F. Foray (Rhodia
Chimie), Fr 4014[2822460], 2001 [Chem. Abstr. 2002, 137,
[7] A. J. Blacker, I. N. Houson
(Avecia Limited), WO
GB667[0266410], 2002 [Chem. Abstr. 2002, 137, 201079].
[8] B. Kirschbaum, G. Wilbert, F. Effenberger (Clariant), US
4600[20020052523], 2001 [Chem. Abstr. 2002, 136, 355075].
[9] B. Kirschbaum, R. Fell, (Clariant), WO EP6681[0305010], 2002
[Chem. Abstr. 2003, 138, 106498].
[10] K. Trummler, J. Roos, U. Schwaneberg, F. Effenberger, S.
Forster, K. Pfizenmaier, H. Wajant, Plant Sci. 1998, 139, 19 – 27.
2003 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
[11] D. V. Johnson, A. A. Zabelinskaja-Mackova, H. Griengl, Curr.
Opin. Chem. Biol. 2000, 4, 103 – 109.
[12] K. R. Baerwald, L. Jaenicke, FEBS Lett. 1978, 90, 255 – 260.
[13] W. Becker, U. Benthin, E. Eschenhof, E. Pfeil, Angew. Chem.
1963, 75, 93; Angew. Chem. Int. Ed. Engl. 1963, 2, 44 – 45.
[14] J. Butenuth, E. Pfeil, Hoppe-Seyler's Z. Physiol. Chem. 1972,
353, 698 – 699.
[15] I. Dreveny, K. Gruber, A. Glieder, A. Thompson, C. Kratky,
Structure 2001, 9, 803 – 815.
[16] I. Dreveny, C. Kratky, K. Gruber, Protein Sci. 2002, 11, 292 – 300.
[17] M. Schmidt, H. Griengl, Top. Curr. Chem. 1999, 200, 193 – 226.
[18] M. Bauer, R. Geyer, M. Boy, H. Griengl, W. Steiner, J. Mol.
Catal. B 1998, 5, 343 – 347.
[19] U. Kragl, U. Niedermeyer, M. R. Kula, C. Wandrey, Ann. N. Y.
Acad. Sci. 1990, 613, 167 – 175.
[20] Y. N. Belokon', M. North, T. Parsons, Org. Lett. 2000, 2, 1617 –
[21] Y. N. Belokon', S. Caveda-Cepas, B. Green, N. S. Ikonnikov,
V. N. Khrustalev, V. S. Larichev, M. A. Moscalenko, M. North, C.
Orizu, V. I. Tararov, M. Tasinazzo, G. I. Timofeeva, L. V.
Yashkina, J. Am. Chem. Soc. 1999, 121, 3968 – 3973.
[22] G. Desantis, Z. Zhu, W. A. Greenberg, K. Wong, J. Chaplin, S. R.
Hanson, B. Farwell, L. W. Nicholson, C. L. Rand, D. P. Weiner,
D. E. Robertson, M. J. Burk, J. Am. Chem. Soc. 2002, 124, 9024 –
[23] L. Rosenthaler, Biochem. Z. 1909, 14, 238 – 253.
[24] G. de Gonzalo, R. Brieva, V. Gotor, J. Mol. Catal. B 2002, 8,
223 – 230.
2003 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2003, 42, 4815 –4818
Без категории
Размер файла
150 Кб
scala, engineering, asymmetric, step, synthesis, large, comprehension, hydroxynitrile, lyase
Пожаловаться на содержимое документа