Comparative study of substrate- and stereospecificity of penicillin G amidases from different sources and hybrid isoenzymes

Article abstract of DOI:10.1007/s007060070090

Four natural pencillin G amidase variants from different sources and two genetically constructed hybrid enzymes were produced and purified to homogeneity. The specificity constants of one enzyme (E. coli) were found to differ six orders of magnitude for hydrolytic transformations within a wide range of substrates. The substrate specificity of the homologous penicillin amidases was found to differ less than one order of magnitude for hydrolysis of the most specific and up to two orders of magnitude for the less specific substrates. The S′₁-substrate specificity in hydrolytic and transfer reactions (studied mainly with the E. coli enzyme) varied more than three orders of magnitude for the different substrates. The penicillin amidases were found to be R-specific in the S₁-binding site and S-specific in the S′₁-binding site. The S₁-stereoselectivity differs less than one order of magnitude for the different variants. The S′₁-stereoselectivity is more pronounced, increases with nucleophile specificity, and was found to differ up to three orders of magnitude in transfer reactions for the enzyme from E. coli. The observed variation of enatioselectivity for different penicillin amidases and one substrate can also be achieved by changes in temperature. Comparison of substrate-and stereospecificity of penicillin amidases from different sources and hybrid isoenzymes suggests that similar changes can be expected for enzyme variants derived by rational protein design or directed evolution.

Full text of DOI:10.1007/s007060070090

Monatshefte f uÈ r Chemie 131, 623±632 (2000)
Comparative Study of Substrate- and
Stereospeci®city of Penicillin G Amidases
from Different Sources and Hybrid Isoenzymes
Boris Galunsky, Karsten Lummer, and Volker KascheÃ
Arbeitsbereich Biotechnologie II, Technische Universit aÈ t Hamburg-Harburg, D-21071 Hamburg,
Germany
Summary. Four natural pencillin G amidase variants from different sources and two genetically
constructed hybrid enzymes were produced and puri®ed to homogeneity. The speci®city constants of
one enzyme (E. coli) were found to differ six orders of magnitude for hydrolytic transformations
within a wide range of substrates. The substrate speci®city of the homologous penicillin amidases
was found to differ less than one order of magnitude for hydrolysis of the most speci®c and up to two
0
orders of magnitude for the less speci®c substrates. The S -substrate speci®city in hydrolytic and
1
transfer reactions (studied mainly with the E. coli enzyme) varied more than three orders of
magnitude for the different substrates. The penicillin amidases were found to be R-speci®c in the S1-
0
binding site and S-speci®c in the S -binding site. The S -stereoselectivity differs less than one order
1
1
0
of magnitude for the different variants. The S -stereoselectivity is more pronounced, increases with
1
nucleophile speci®city, and was found to differ up to three orders of magnitude in transfer reactions
for the enzyme from E. coli. The observed variation of enatioselectivity for different penicillin
amidases and one substrate can also be achieved by changes in temperature. Comparison of
substrate- and stereospeci®city of penicillin amidases from different sources and hybrid isoenzymes
suggests that similar changes can be expected for enzyme variants derived by rational protein design
or directed evolution.
Keywords. Homologous enzymes; Penicillin amidase (acylase); Stereospeci®city; Substrate spec-
i®city.
Introduction
During the evolution, natural enzyme variants with a common function have been
selected for favoured biological properties, e.g. high speci®city ꢀk =K † and
cat
m
stereoselectivity for natural substrates at conditions in the main habitat of the
microorganisms. Therefore, the enzyme properties differ for the homologous
enzymes depending on their source. The evolution optimizes the biocatalysts with
regard to their function in vivo, whereas their application as industrial biocatalysts
requires tolerance to reaction conditions outside the physiological range and
à Corresponding author

624
B. Galunsky et al.
transformation of substrates not encountered in nature [1]. During the last ®fteen
years much work has been done to optimize enzymes for industrial applications by
rational protein design [2±5] or directed evolution [6±10]. Despite the progress in
improving enzymes from a biotechnological point of view, fundamentally altered
properties have hardly been achieved. The question on the relation between
function and basic biocatalytic properties on one hand and structural parameters as
the amino acid sequence on the other one still remains actual. Comparison of data
on substrate- and stereoselectivity of homologous enzymes and mutation studies
may provide insight into the role of selection in optimizing the kinetic behaviour of
native enzymes and to what extent it could be modulated.
The family of pencillin G amidases (PA; EC 3.5.1.11) is suitable for such an
analysis. These enzymes catalyze the hydrolysis of benzylpenicillin (1) to
phenylacetic acid (2) and 6-aminopenicillanic acid (3) ± one of the key
intermediates for the production of semisynthetic ꢀ-lactam antibiotics [11]. PA
can also catalyze the synthesis of these antibiotics [12, 13]. Due to the
biotechnological interest, different natural and arti®cial variants of this enzyme
have been produced [14±20]. The genes encoding penicillin G amidase in E. coli,
K. citrophila, P. rettgeri, A. faecalis, and A. viscosus have been cloned and
sequenced. The nucleotide sequences encode precursors that are processed to two
nonidentical peptide chains (A and B) [11]. The homology in the primary structure
of all compared enzymes decreases with increasing genetic distances between the
different species which are supposed to be diverged from a common evolutionary
ancestor [21]. Despite the large differences in the homology between the most
distant variants (up to 70%), multiple sequence alignment data reveal in both A-
and B-chains segments of strongly conserved amino acid sequences essential for
the catalytic function of PA. The biological function of the penicillin amidase
family is still unknown. The same applies for whether it is controlled by the same
groups that control the biotechnical functions as a biocatalyst for benzylpenicillin
hydrolysis and production of the penem nucleus (3) [22, 23]. The penicillin
amidases have a wide range of substrate speci®city and differ in properties
important for existing and possible new industrial applications. The 3D-structure of
the E. coli enzyme reveals a single amino acid (Ser B1) catalytic site [24], and
supports its classi®cation as an Ntn-hydrolase [25]. The 3D-structure con®rms also
the acyl-enzyme reaction mechanism [26, 27]. The same mechanism has been
inferred for the other compared enzymes [11].
In this study we compare quantitative data on substrate- and stereoselectivity of
different natural and genetically constructed hybrid enzyme variants. The results
can yield information on how much these properties can be changed by directed or
random mutagenesis.
Results and Discussion
The molar concentrations of the penicillin amidases were determined by active site
titration [28]. Active site titration curves for all studied PA variants were almost
identical which implies that only their N-terminal Ser B1 are accessible for
acylation with phenylmethylsulfonyl ¯uoride (4).

Speci®ty of Penicillin G Amidases
625
Speci®city constants in hydrolytic reactions
1
The S -subsite speci®city of penicilin G amidase is mainly restricted to phenylacetic
1
acid and its derivatives [11], although other acyl moieties as acetyl, 2-benzoxazolon-
3-yl-acetyl, or mandelyl are also accepted [30±33]. Most of the Pa-catalyzed
transformations published so far have been done with the E. coli enzyme. For the
substrates we studied with E. coli PA and presented in Fig. 1, the speci®city
constants for equilibrium controlled hydrolysis vary by 6 orders of magnitude. The
observed differences between substrates varying in P were predominantly due
1
0
to larger differences in K . The S - and S -subsite speci®cities of puri®ed
m
(
1
1
homogeneous) natural and hybrid PA variants have been studied in detail for the
hydrolysis of 1, 6-nitro-3-phenylacetamido-benzoic acid (11) the less speci®c R-
phenylglycine amide (7), R-phenylglycine methyl ester (6), p-hydroxy-R-phenylgly-
cine amide (9), and 6-nitro-3-R-phenylglycylamidobenzoic acid (10). The results are
shown in Fig. 2. The k =K values for hydrolysis of 1 differ less than one order of
cat
m
magnitude for the different enzyme variants. The enzyme from A. faecalis has the
highest speci®city constant compared to the other natural enzymes. Similar data have
been reported by other authors [34]. For the hydrolysis of 11, 7, and 6 the differences
are larger ± up to 2 orders of magnitude. Here differences in both the turnover
number and the Michaelis-Menten constant account for the observed variation in the
Fig. 1. Substrate speci®city of af®nity puri®ed penicillin amidase from E. coli in equilibrium
ꢁ
controlled hydrolytic reactions. Reaction conditions: phosphate buffer, pHˆ 7.5 (I ˆ 0.2), 25 C. kcat
�
1
� 1
varies from 0.02 s for 5 to 200 s for 16. Km varies from 10 mM for 1 to 30mM for 6. Abbreviations
used: BP, benzylpenicillin; Amp, ampicillin; PAA, phenylacetic acid; Phg, phenylglycine; 6-APA,
0
6
-aminopenicillanic acid; NAB, 6-nitro-3-aminobenzoic acid; dAdo, 2 -deoxyadenosine; dGuo,
0
2
-deoxyguanosine; three-letter symbols for the amino acids
1
Using the convention for peptidases to designate the binding subsites [29]

626
B. Galunsky et al.
Fig. 2. Speci®city constants of penicillin amidases from different sources and two genetically
constructed hybrid enzymes (Hybrid I AKcBEc and Hybrid II AEcBKc for the hydrolysis of 1, 11, 7,
ꢁ
and 6, at pH ˆ 7.5 (I ˆ 0.2) 25 C. All data were obtained with (af®nity) puri®ed enzymes. Both kcat
and Km differ 3±4 fold for transformations of the same substrate catalyzed by different PA variants.
Abbreviations: for 1, 6, and 7, cf. Fig. 1; NIPAB: 6-nitro-3-phenylacetamido-benzoic acid
speci®city constants of the homologous enzymes. The speci®city constants differ
also for the different processed forms of one enzyme. For E. coli PA, the last
processed form with pI ˆ 6.6 ꢀPA † and an A-chain which is by 8 amino acids
6
:7
2
shorter than PA7 , has a speci®city constant for hydrolysis of 1 lower by 30% [35].
:0
The last processed form of A. faecalis PA with pI ˆ 5.3 ꢀPA †, with the A-chain
5
:3
shortened for 3 amino acids from the C-end and converted N-terminal Gly a 30%
larger speci®city constant for the same reaction as compared to PA5:5 [36].
3
We found that genetically constructed hybrids are active and have speci®city
constants for the studied hydrolytic reactions similar to those of the wild type
enzymes [18]. The similarity of the kinetic behaviour of the homologous PA is not
surprising. Besides the identity of Ser B1, Gln B23, Ala B69, Asn B241, and Arg
B263 forming the tetrahedral arrangements of ligands of O and the ꢂ-amino group
ꢁ
[
24], the amino acid sequences in their vicinity are strongly conserved as well: of 32
Ê
residues within 10 A of Ser B1 only 3 differ in the different variants [17]. Changes at
larger distances from the active site (the C-terminus of the A-chain) can also result in
small activity changes, as discussed for the different processed forms. The smaller
differences between the speci®city constants of the homologous enzymes for
hydrolysis of 1 compared to 11, 7, and 6 suggest that the transformation of one of the
most speci®c substrates is probably a reaction congruent to the biological function of
PA, which has been optimized during the evolution. Thus, there are hardly to be
expected large differences in the catalytic behaviour of mutant enzymes for the most
speci®c substrates. For catalyzing reactions of less speci®c substrates (more distant
2
The dominating processed form with C-terminal Ala209 of the A-chain and pI ˆ 7.0
3
The processed form with C-terminal Ala202 of the A-chain and pI ˆ 5.5

Speci®ty of Penicillin G Amidases
627
from the biological function), mutant enzymes with improved properties are more
probable. The few site directed or randomly created PA mutants have shown lower
speci®city constants for hydrolysis of 1 [39±40].
0
S -site speci®city
1
0
The S -speci®city re¯ects the ability of penicillin G amidase to catalyze acyl
1
transfer reactions [12]. It is broader and varies from penem and cephem nuclei to
ꢂ- and ꢀ-amino acids, their phosphonic and phosphonous analogs, peptides,
aromatic and aliphatic amines and alcohols, and nucleosides. These have been used
(also on an industrial scale) for the kinetically controlled synthesis of amoxicillin,
ampicillin, cephadroxil, and cephalexin, for deprotection in peptide and nucleotide
synthesis, and for pro-drug activation in cancer therapy [13, 40±46]. We studied the
nucleophile speci®city in both hydrolytic and synthetic directions mainly with the
E. coli enzyme. In Fig. 1, the speci®city constants for hydrolytic transformation of
substrates with phenylacetic acid in P -position and different moieties in positions
1
0
0
P and P are compared. The k =K values vary by more than 3 orders of
cat
1
magnitude for the different leaving groups. The few quantitative data on S -
2
m
0
1
speci®city published by other authors reveal speci®city constants of the same order
of magnitude [41].
0
Interactions in the S -binding subsite depend on electrostatic interactions
0
1
between the basic amino acid residue in S -position (most probably Arg A145) and
2
the unprotected carboxyl groups of amino acid and ꢀ-lactam structures [12, 41,
4
synthetic reactions where the enzyme catalyzes the transfer of an acyl moiety to a
7]. This has been observed also with A. faecalis PA [33]. In kinetically controlled
0
nucleophile, the measure for the S -speci®city is the transferase to hydrolase ratio
k =k [12]. We determined its valuewith E. coli PA for transfer of phenylacetyl or
1
T
H
phenylglycyl moieties from phenylacetyl-Gly 17, 7, and 6 to amino acids or
peptides. For the different substrates this ratios differ up to 3 orders of magnitude
1
4
(
1Â10 ±3Â10 ). For one nonspeci®c acyl moiety, k =k was found to have a value
T
H
5
typical for a transferase (2Â10 ), [48]. The E. coli enzyme was shown to be better
for the acyl transfer to cephem nuclei than the PA from A. faecalis. Here
comparative studies with other natural and hybrid penicillin G amidases are
necessary to analyze (together with 3D-structures of enzyme-substrate complexes)
a possible improvement of the transferase properties of the enzymes. This can be
relevant for the synthesis of ꢀ-lactam antibiotics and for new biotransformations
where C±N and C±O bonds are synthesized or hydrolyzed.
Stereoselectivity
4
The enantioselectivity of an enzyme E should be > 100 or < 0.01 to obtain pure
and R-enantiomers, respectively, at the end-point in biotransformations [49, 50].
PA stereospeci®city is also very important for its biotechnical application. E. coli
4
The quantitative measure for enzyme stereospeci®city is the enantioselectivity or enantiomeric
ratio E. It is de®ned as the ratio of the speci®city constants for hydrolysis of enantiomeric S- and R-
substrates or as the ratio of the transferase to hydrolase ratios for acyl transfer to enantiomeric S- and
R-nucleophiles.

628
B. Galunsky et al.
Fig. 3. S -subsite stereoselectivity of PA from different sources and two genetically constructed
1
hybrid enzymes (Hybrid I AKcBEc and Hybrid II AEcBKc), expressed as the enantiomeric ratio for
ꢁ
0
various substrates hydrolysed at pH ˆ 7.5 (I ˆ 0.2) and 25 C. For comparison, the S -subsite
1
stereoselectivity is given for 2 substrates hydrolyzed at the same conditions by E. coli and A. faecalis
enzymes. Abbreviations as in Fig. 1
PA has been used for the production of chiral synthons in peptide and nucleotide
synthesis [41±44] and for kinetically controlled enantioselective acylation during
loracarbef synthesis [51]. Considering the 3D-structure of E. coli PA and the
sequence homology of the enzymes from different sources, very similar 3D-
structures of all variants can be predicted [17]; thus, similar enantioselectivity
could be expected.
S1-enantioselectivity was studied with puri®ed homologous penicillin G
amidases only for a limited number of chiral substrates. The results are shown in
Fig. 3. With the same enzyme, E was found to differ about one order of magnitude
for the different substrates. A similar difference in enantioselectivity was observed
for the transformation of one substrate by different PA variants. Differences mainly
in k , i.e. the activation energy, account for the observed changes.
cat
0
S -enantioselectivity was studied in both hydrolytic and transfer reactions,
mainly with the E. coli enzyme. For comparison, in Fig. 3 the enantiomeric ratios
1
for the hydrolytic transformations of two substrates catalyzed by two PA variants
0
are given. S -enantioselectivity is more pronounced and increases with nucleophile
speci®city. As for the S -stereoselectivity, differences mainly in k account for
1
1
cat
differences in E. Literature data for the hydrolysis of phenylacetic acid derivatives
of ꢂ-amino acids, peptides, ꢀ-amino acids, and alcohols reveal differences in E up
to 3 orders of magnitude [41]. With the E. coli enzyme this was observed also in
transfer reactions where ꢀk =k † =ꢀk =k † differ up to 3 orders of magnitude
T
H
T
H
R
S
(0.5±500) for the different substrates.
The penicillin G amidase was found to be R-speci®c in the S - and S- speci®c in
1
0
the S -binding subsites. This observation demonstrates that an enzyme consisting
1

Speci®ty of Penicillin G Amidases
629
of homochiral amino acids can simultaneously catalyze reactions of two opposite
chiral centres. This applies for all studied PA variants. Since the most speci®c
substrates of PA [41] have no chiral moieties in P -position, it may imply that S -
1
1
0
subsite stereospeci®city is not important for the biological function. S -subsite
stereospeci®city is more pronounced, moreover many of the most speci®c
substrates [41] possess chiral centres in the P -position. For directed modulation
1
0
1
of E, however, more information on the structural determinants of enantiomer
discrimination for different enzyme variants is necessary. Directed evolution
studies have shown only moderate improvement of the enantiomeric ratio [52],
whereas temperature variation have resulted in larger (more than order of
magnitude) in¯uence [30, 53]. Thus, alternative modulation of E by the reaction
conditions should be considered as well.
Conclusions
1
. The speci®city constants of a homogeneous PA from E. coli were found to vary
0
0
by 6 orders of magnitude for different P � P � P -substrates. The differences
1
1
between substrates varying in P is predominantly due to larger differences in
2
1
K Á This indicates that the catalytic power of this enzyme is in¯uenced both by
m
substrate binding and lowering of the activation energy of the transition state.
0
2
3
4
. The enantioselectivity of the S - (R-speci®c) and S -(S-speci®c) binding sites of
1
1
0
this enzyme varied by up to 3 orders of magnitude for the hydrolysis of P � P
1
1
substrates. The differences are larger in the S -subsite and in both cases mainly
0
1
due to changes in k Á
cat
. Four natural and two hybrid variants of PA were puri®ed to homogeneity. They
differ up to 70% in primary structure. There is, however, a large homology of
amino acid residues that are located near the B-chain N-terminal Ser in the E.
coli enzyme, acylated in the reactions catalyzed by this enzyme.
0
0
. The speci®city constants of the PA variants for hydrolysis of one P � P � P
1
1
2
substrate or kinetically controlled synthesis with one nucleophile were found
to differ by maximal two orders of magnitude. The changes decrease with
the speci®city constants for the substrates and are due to changes both in k andK Á
cat
m
0
. The enantioselectivity of the S - and S -binding subsites of the enzyme variants
1
1
5
6
varied up to one order of magnitude. Such a change for one PA variant and one
substrate can also be achieved by changes in temperature.
. Based on this we conclude that changes in substrate- and stereospeci®city of new
active PA variants created by directed or random mutations will be of the order of
magnitude observed for the homologous enzymes, i.e. less than two orders of
magnitude for speci®city constants (the changes decreasing with the speci®city
constants) and less than one order of magnitude for stereoselectivities.
Experimental
Substrates
R- and S-enantiomers of 5, 6, and 7 were kindly provided by R oÈ hm (Dermstadt, Germany); R- and S-
enantiomers of 9 were kindly provided by DSM Research (Geleen, The Netherlands); S-enantiomers
of 15 and 16 were kindly provided by Dr. I. Stoineva, Institute of of Organic Chemistry, Bulgarian

630
B. Galunsky et al.
Academy of Sciences, So®a, Bulgaria); 17, the S-enantiomer of 18, and S-and R-enantiomers of 19
were synthesized as described in Ref. [53]; S,S-diastereomer of 14 was synthesized as described in
Ref. [44]; 12 and 13 were synthesized as described in Ref. [43]; 11 was synthesized as described in
Ref. [54]; 1, 4, and 8 were purchased from Sigma (Desenhofen, Germany); all other chemicals were
of analytical grade.
Preparation of PA variants
Penicillin amidase from E. coli and Hybrid I (A-chain from K. citrophila and B-chain from E. coli)
were produced in recombinant E. coli strains and isolated by af®nity chromatography using a
biospeci®c adsorbent with an immobilized monoclonal antibody directed against the B-chain of
E-coli enzyme as described in Refs. [18, 56]. The enzymers from K. citrophila (the gene in the
plasmid pYKH5 was kindly provided by Dr. Barbero, Antibioticos, Leon, Spain), A. feacalis (the
gene was ampli®ed by the PCR-technique from the total DNA prepared from strain NCTC 415 by
DSMZ, Braunschweig, Germany), and Hybrid II (A-chain from E. coli and B-chain from K.
citrophila) were produced in recombinant E. coli strains as described in Refs. [30, 36, 18]. The
Enzyme from A. viscosus was prepared using DNA from the wild strain ATCC 15294. These PA
variants were puri®ed from a concentrated homogenate by ion exchange chromatography using the
same procedure as described previously [35, 36]. The purity and homogeneity of the prepared
enzymes from different sources was analyzed by activity and protein stains of isoelectric focusing
gels [35]. E. coli and A. faecalis enzymes used here were the proteolytically processed forms with
isoelectric points of 7.0 and 5.3.
Active site titration
The molar concentrations of the PA variants were determined by active site titration with 4 [28] using
11 as a substrate.
Determination of kcat and KmÁ
The substrates were dissolved and diluted to the desired concentration in phosphate buffer (pH ˆ 7.5,
ꢁ
I ˆ 0.2 M). The reaction was performed in this buffer at 25 C, initiated by addition of enzyme. The
system was kept at constant temperature and pH value. The cleavage of 11 was followed
spectrophotometrically at 380 nm by 6-nitro-3-aminobenzoic acid release [57]. With the other
substrates, samples were withdrawn at different times and analysed by HPLC using a RP 8 column.
The elution conditions and wavelengths used to quantitate the products are described in Refs. [58,
30, 44]. Eadie-Hofstee plots were used to determine kcat and Km.
Determination of kT =kH
The substrates (6, 7, and 17) and the nucleophiles (amino acids and peptides with free carboxyl
function) were dissolved in carbonate buffer (pH ˆ 9, I ˆ 0.2 M). The ratio was determined in this
ꢁ
buffer at 25 C. Samples were withdrawn and analysed by HPLC using a RP 18 column. Gradient
elution was used as described in Ref. [30]. The peak areas at 225 nm were used to quantitate the
products. kT=kH was calculated from the initial rates of condensation and hydrolysis product
formation as described in Ref. [12].
Acknowledgements
This work has been supported by the Deutsche Forschungsgemeinschaft (Graduiertenkolleg
Biotechnologie, GK 95-3-98).

Speci®ty of Penicillin G Amidases
631
References
[
[
[
[
[
1] Rubingh DN (1997) Curr Opin Biotechnol 8: 417
2] Takagi H (1993) Int J Biochem 25: 307
3] Joyet P, Decklerk N, Gaillardin C (1992) Bio-Technology 10: 1579
4] You L, Arnold FH (1994) Protein Eng 9: 77
5] Rubingh DN (1996) Engineering Proteases with Improved Properties for Detergents. In: Savage
L, Southborough MA (eds) Enzyme Technology for Industrial Applications. IBC, p 98
6] Stemmer WPC (1994) Nature 370: 389
[
[
[
[
7] Stemmer WPC (1994) Proc Natl Acad Sci 91: 10747
8] Arnold FH (1996) Chem Eng Sci 51: 5091
9] Arnold FH, Moor JC (1997) Adv Biochem Eng 58: 1
[
[
[
[
[
[
10] Moore JC, Arnold FH (1996) Nat Biotechnol 14: 458
11] Virden R (1990) Biotechnol Gen Eng Rev 8: 189
12] Kasche V (1986) Enzyme Microb Technol 8: 4
13] Brugging A, Roos EC, de Vroom E (1998) Organic Process Research and Development 2: 128
14] Schumacher G, Sizmann D, Haug H, Buckel P, Bock A (1986) Nucleic Acids Res 14: 5713
15] Barbero JL, Buesa JM, Gonzales de Buitrago G, Mendez E, Perez Aranda A, Garcia JL (1986)
Gene 49: 69
[
16] Ljubijankic C, Konstantinovic M, Glisin V (1992) DANN Seq 3: 195
17] Verhaert RMD, Riemens AM, van der Laan J-M, van Duin J, Quax WJ (1997) Appl Environ
Microbiol 63: 3412
[
[18] Piotraschke E, Nurk A, Galunsky B, Kasche V (1994) Biotechnol Lett 16: 119
[19] Martin L, Prieto MA, Cortes E, Garcia JL (1995) FEMS Microbiol Lett 125: 287
[20] Ohashi H, Katsuna Y, Hashizume T, Abe SN, Kajiura H, Hattori H, Kamei T, Yano M (1988)
Appl Environ Microbiol 54: 2603
[21] Olsen GJ, Woese CR and Overbeek R (1994) J Bacteriol 176: 1
[22] Shewale JG, Sivaraman H (1989) Process Biochem 24: 146
[23] Shewale JG, Deshpande BS, Sudhakaran VK, Ambedkar SS (1990) Process Biochem 25: 97
[24] Duggleby H, Tolley SP, Hill CP, Dodson EJ, Dodson G, Moody PCE (1995) Nature 373: 264
[25] Brannigan JA, Dodson G, Duggleby H, Moody PCE, Smith JL, Tomchik DR, Murzin AG (1995)
Nature 378: 416
[26] Konecny J (1981) Biotechnol Lett 3: 112
[27] Hau¯er U (1987) PhD Thesis, University of Bremen, Germany
[28] Svedas VK, Margolin AL, Sherestyuk CF, Klesov AA, Berezin IV (1977) Bioorgan Khim (Russ)
3
: 546
[29] Schechter I, Berger A (1967) Biochem Biophys Res Commun 27: 157
[30] Kasche V, Galunsky B, Nurk A, Piotraschke E, Rieks A (1996) Biotechnol Lett 18: 455
[
31] Niersbach H, K uÈ hne A, Tischer W, Weber M, Wedekind F, Plapp R (1995) Appl Microbiol
Biotechnol 43: 679
[32] Stambolieva N, Mincheva Z, Galunsky B, Kalcheva V (1992) Enzyme Microb Technol 14: 496
[33] Vojtisek V, Slezak J (1975) Folia Microbiol (Praha) 20: 224
[34] Svedas, V, Guranda D, van Langen L, van Rantwijk F, Sheldon R (1997) FEBS Lett 417: 414
[35] Kasche V, Hau¯er U, Markowsky D, Ziech A, Galunsky B (1987) Ann NY Acad Sci 501: 97
[
36] Ignatova Z, Stoeva S, Galunsky B, H oÈ rnle C, Nurk A, Piotraschke E, Voelter W, Kasche V (1998)
Biotechnol Lett 20: 997
[
37] Prieto I, Martin J, Arche R, Fernandez P, Perez-Aranda A, Barbero JL (1990) Appl Microbiol
Biotechnol 33: 553
38] Williams JA, Zuzel TJ (1985) J Cell Biochem 9B: 99
[
[
39] Martin J, Prieto R, Barbero JL, Perez-Gil J, Mencheno JM, Arche R (1990) Biochim Biophys
Acta 1037: 133

632
B. Galunsky et al.: Speci®ty of Penicillin G Amidases
[
[
[
[
[
[
[
[
40] Forney LJ, Wong DCL, Ferber DM (1989) Appl Environ Microbiol 55: 2550
41] Waldmann H, Sebastian D (1994) Chem Rev 94: 911
42] Svedas VK, Savchenko MV, Beltser AI, Guranda DF (1996) Ann NY Acad Sci 799: 659
43] Waldmann H (1989) Tetrahedron Lett 30: 3057
44] Dineva M, Galunsky B, Kasche V, Petkov D (1993) Bioorg Med Chem Lett 3: 2781
45] Stoineva I, Galunsky B, Lozanov VS, Ivanov IP, Petkov DD (1992) Tetrahedron 48: 1115
46] Vrudhula VM, Senter PD, Fischer KJ, Wallace PM (1993) J Med Chem 36: 919
th
47] Alkema WBL, Prins AK, Janssen DB (1999) 4 International Symposium on Biocatalysis and
Biotransformations, Giardini Naxos-Taormina, Italy, Abstract P112
[
[
[
[
[
[
[
[
[
48] Stambolieva N, Mincheva Z, Galunsky B (1998) Biocat Biotrans 16: 225
49] Collins AN, Sheldrake GN, Crosby J (eds) (1992) Chyrality in Industry. Wiley, Chichester
50] Rakels JLL, Paffen MT, Straathof AJJ, Heijnen JJ (1994) Enzyme Microb Technol 16: 791
51] Zmiewski MJ, Briggs, BS Tompson AR, Wright IG (1991) Tetrahedron Lett 32: 1621
52] Bornscheuer UT, Altenbuchner J, Meyer HH (1998) Biotechnol Bioeng 58: 554
53] Galunsky B, Ignatova S, Kasche V (1997) Biochim Biophys Acta 1343: 130
54] Suyama TJ (1965) Chem Abstr 63: 7095
55] Daumy GO, McColl AS, Apostolakos D (1982) J Bacteriol 152: 104
56] Kasche V, Gottschlich N, Lindberg A, Niebuhr-Redder C, Schmieding J (1994) J Chromatog 660:
137
[
57] Kutzbach C, Rauenbusch E (1974) Hoppe-Seyler's Z Physiol Chem 354: 45
58] Kasche V (1985) Biotechnol Lett 7: 877
[
Received December 20, 1999. Accepted (revised) February 4, 2000