β-aminopeptidase-catalyzed biotransformations of β2- dipeptides: Kinetic resolution and enzymatic coupling

Article abstract of DOI:10.1002/cbic.200900757

We have previously shown that the β-aminopeptidases BapA from Sphingosinicella xenopeptidilytica and DmpA from Ochrobactrum anthropi can catalyze reactions with non-natural β³-peptides and β³-amino acid amides. Here we report that these exceptional enzymes are also able to utilize synthetic dipeptides with N-terminal β²-amino acid residues as substrates under aqueous conditions. The suitability of a β²-peptide as a substrate for BapA or DmpA was strongly dependent on the size of the Cα substituent of the N-terminal β²-amino acid. BapA was shown to convert a diastereomeric mixture of the β²-peptide H-β²hPhe- β²hAla-OH, but did not act on diastereomerically pure β²,β³-dipeptides containing an N-terminal β²-homoalanine. In contrast, DmpA was only active with the latter dipeptides as substrates. BapA-catalyzed transformation of the diastereomeric mixture of H-β²hPhe-β²hAla-OH proceeded along two highly S-enantioselective reaction routes, one leading to substrate hydrolysis and the other to the synthesis of coupling products. The synthetic route predominated even at neutral pH. A rise in pH of three log units shifted the synthesis-to-hydrolysis ratio (v_S/v_H) further towards peptide formation. Because the equilibrium of the reaction lies on the side of hydrolysis, prolonged incubation resulted in the cleavage of all peptides that carried an N-terminal β-amino acid of S configuration. After completion of the enzymatic reaction, only the S enantiomer of β²-homophenylalanine was detected (ee>99% for H-(S)-β²-hPhe-OH, E>500); this confirmed the high enantioselectivity of the reaction. Our findings suggest interesting new applications of the enzymes BapA and DmpA for the production of enantiopure β²-amino acids and the enantioselective coupling of N-terminal β²-amino acids to peptides.

Full text of DOI:10.1002/cbic.200900757

DOI: 10.1002/cbic.200900757
b-Aminopeptidase-Catalyzed Biotransformations of b2-
Dipeptides: Kinetic Resolution and Enzymatic Coupling
Tobias Heck,[a, b] Artur Reimer,[a, c] Dieter Seebach,[b] James Gardiner,[b] Gildas Deniau,[b]
Aneta Lukaszuk,[b] Hans-Peter E. Kohler,*[a] and Birgit Geueke[a]
We have previously shown that the b-aminopeptidases BapA
from Sphingosinicella xenopeptidilytica and DmpA from Ochro-
bactrum anthropi can catalyze reactions with non-natural b3-
peptides and b3-amino acid amides. Here we report that these
exceptional enzymes are also able to utilize synthetic dipep-
tides with N-terminal b2-amino acid residues as substrates
under aqueous conditions. The suitability of a b2-peptide as a
substrate for BapA or DmpA was strongly dependent on the
size of the Ca substituent of the N-terminal b2-amino acid.
BapA was shown to convert a diastereomeric mixture of the
b2-peptide H-b2hPhe-b2hAla-OH, but did not act on diastereo-
merically pure b2,b3-dipeptides containing an N-terminal b2-ho-
moalanine. In contrast, DmpA was only active with the latter
dipeptides as substrates. BapA-catalyzed transformation of the
diastereomeric mixture of H-b2hPhe-b2hAla-OH proceeded
along two highly S-enantioselective reaction routes, one lead-
ing to substrate hydrolysis and the other to the synthesis of
coupling products. The synthetic route predominated even at
neutral pH. A rise in pH of three log units shifted the synthe-
sis-to-hydrolysis ratio (vS/vH) further towards peptide formation.
Because the equilibrium of the reaction lies on the side of hy-
drolysis, prolonged incubation resulted in the cleavage of all
peptides that carried an N-terminal b-amino acid of S configu-
ration. After completion of the enzymatic reaction, only the S
enantiomer of b2-homophenylalanine was detected (ee>99%
for H-(S)-b2-hPhe-OH, E>500); this confirmed the high enantio-
selectivity of the reaction. Our findings suggest interesting
new applications of the enzymes BapA and DmpA for the pro-
duction of enantiopure b2-amino acids and the enantioselec-
tive coupling of N-terminal b2-amino acids to peptides.
Introduction
b-Amino acids with proteinogenic side chains are backbone-
elongated homologues of the naturally occurring proteinogen-
ic a-amino acids (Scheme 1).1 In contrast to their a-peptidic
b,a-peptides revealed novel hairpin-turn-like structures that re-
semble the b-turns of conventional a-peptides.[4] These proper-
ties give rise to interesting new biomedical applications for b-
peptides as proteolytically stable mimics of bioactive natural
peptides.[5] Some recent examples of b-peptide-based designs
of peptidomimetics include inhibitors of protein–protein inter-
actions and viral cell entry,[6] ligands of the somatostatin recep-
tor and the major histocompatibility complex (MHC),[7] as well
as b-peptides with antifungal and antimicrobial activities.[8] Fur-
thermore, a very recent study has suggested mixed a,b-pep-
tides to be a new class of water-soluble nanoporous materials
that adsorb nitrogen gas.[9]
Scheme 1. Structural comparison of a- and b-amino acid residues in pep-
tides.
counterparts, peptides comprising b-amino acids are character-
ized by high resistance to enzymatic degradation and metabol-
ic breakdown.[2] Properly designed b-peptides composed of b3-
amino acid residues with proteinogenic side chains fold into
stable secondary structures, such as the 314-helix, which was
found to be the predominant secondary structure of b3-pep-
tides.[1,3] b2-Peptides, with side chains attached to the a- in-
stead of the b-carbon (Scheme 1), are conformationally more
flexible than their b3-peptide isomers. Synthetic peptides with
alternating b2- and b3-amino acid residues as well as mixed
[a] Dipl. Biol. T. Heck, A. Reimer, Dr. H.-P. E. Kohler, Dr. B. Geueke
Eawag, Swiss Federal Institute of Aquatic Science and Technology
ꢀberlandstrasse 133, 8600 Dꢁbendorf (Switzerland)
Fax: (+41)44-823-5547
E-mail: kohler@eawag.ch
[b] Dipl. Biol. T. Heck, Prof. Dr. D. Seebach, Dr. J. Gardiner, Dr. G. Deniau,
Dr. A. Lukaszuk
Department of Chemistry and Applied Biosciences
Laboratory of Organic Chemistry, ETH Zꢁrich
Hçnggerberg HCI, Wolfgang-Pauli-Strasse 10, 8093 Zꢁrich (Switzerland)
[c] A. Reimer
Department of Biochemical and Chemical Engineering, TU Dortmund
Emil-Figge-Strasse 66, 44221 Dortmund (Germany)
1
The b-amino acids used in this study are named according to the nomencla-
ture for b-amino acids with proteinogenic side chains (bhXaa).[1]
Supporting information for this article is available on the WWW under
http://dx.doi.org/10.1002/cbic.200900757.
ChemBioChem 2010, 11, 1129 – 1136
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
1129

H.-P. E. Kohler et al.
b2-Amino acids are valuable compounds for extending the
repertoire of building blocks for peptide design because their
incorporation into peptides induces unique secondary-struc-
ture elements and leads to increased stability against proteo-
lytic breakdown.[10] Unlike b3-amino acids, most of which are
by now commercially available, enantiopure b2-amino acids
with proteinogenic side chains are less easily accessible by
chemical synthesis. They cannot be prepared by stereoselec-
tive homologation of the naturally occurring a-amino acid ana-
logues, the strategy most commonly applied to the synthesis
of enantiopure b3-amino acids (Arndt–Eistert reaction).[1,3,10,11]
Instead, the synthesis of enantiomerically pure b2-amino acids
frequently involves multistep transformations with the use of
chiral auxiliaries. Classical chemical-resolution strategies to
obtain enantiopure b2-amino acids from racemic starting mate-
rials have so far been unsuccessful.[12]
Results and Discussion
Enzymatic conversion of a diastereomeric mixture of the b2-
dipeptide H-b2hPhe-b2hAla-OH (1)
Dissolved BapA and DmpA were tested for their ability to act
on peptides composed of non-natural b2-amino acids under
aqueous conditions. We synthesized a diastereomeric mixture
of the b2-dipeptide H-b2hPhe-b2hAla-OH ((S,S)-1, (S,R)-1, (R,S)-1,
and (R,R)-1), which allowed us to investigate the b-aminopepti-
dase-catalyzed breakdown of the compound and the stereose-
lectivity of the reaction. We separated the diastereomers of 1
on a reversed stationary phase, and quantified substrates and
products of the enzymatic conversions by HPLC-UV. As the
mixture of 1 contains two diastereomeric pairs of enantiomers
((S,S)-1, (R,R)-1 and (S,R)-1, (R,S)-1), two reversed-phase HPLC
peaks were obtained. According to the elution of a pure stan-
dard of (S,S)-1 (tR =17.1 min) and a standard of (S,S)-1 that con-
tained traces of (S,R)-1 (tR =17.7 min), peak 1 (tR =17.1 min)
could be attributed to (S,S)-1/(R,R)-1 and peak 2 (tR =17.7 min)
to (S,R)-1/(R,S)-1.
In our investigation, we detected substrate transformation
only with BapA, whereas DmpA left the mixture of 1 un-
touched. During the assay period at pH 7.2 and 378C, the de-
crease of 1 leveled off at approximately 57% of the initial sub-
strate concentration of 5 mm (Figure 1). It is possible that the
enzymatic reaction slowed down at the end of the 11-day incu-
bation due to partial inactivation of the enzyme. However, the
fact that both peaks of 1 decreased to approximately half of
their initial areas indicated a high stereoselectivity of the BapA-
catalyzed reaction. The initial decrease of the area under
peak 2 ((S,R)-1/(R,S)-1) was about 1.3 times faster than the de-
crease of the area under peak 1 ((S,S)-1/(R,R)-1). Appropriate
control experiments showed that the substrate 1 was chemi-
cally stable under the reaction conditions over the assay
period.
The b-aminopeptidases BapA from Sphingosinicella xenopep-
tidilytica and DmpA from Ochrobactrum anthropi recently
aroused interest because they can catalyze the degradation of
b3-peptides with proteinogenic side chains that are otherwise
stable to proteolytic breakdown.[2,13] Despite sharing an amino
acid sequence identity of 42%, BapA and DmpA differ with
respect to their substrate specificities. DmpA cleaves both a-
and b3-amino acids from the N termini of peptides, but shows
a distinct preference for N-terminal amino acids that are un-
substituted or carry only small side chains (e.g., -CH3). In con-
trast, BapA only catalyzes the removal of N-terminal b3-amino
acids and does not cleave peptides with N-terminal a-amino
acids. However, BapA has a broader substrate specificity than
DmpA as it accepts a wide variety of peptides that carry N-ter-
minal b3-amino acids with different proteinogenic side
chains.[14] These findings demonstrated for the first time that
the supposed proteolytic stability of b-peptides is not abso-
lute. Further studies with BapA and DmpA revealed their use-
fulness as biocatalysts for the efficient production of enantio-
pure aliphatic b3-amino acids by kinetic resolution of racemic
b3-amino acid amides,[15] and for the enzyme-catalyzed synthe-
sis of various b3-amino acid-containing peptides from C-termi-
nally activated b-amino acid residues under kinetic control.[16]
The promising results obtained from the b-aminopeptidase-
catalyzed conversions of b3-peptides encouraged us to con-
duct enzymatic studies with b2-amino acid-containing pep-
tides, which have been less explored due to their limited avail-
ability. In fact, only very few enzyme-catalyzed conversions of
b2-amino acid-containing compounds have been reported,
most of which describe the enantioselective production of b2-
amino acids from racemic starting materials.[17] To the best of
our knowledge, the proteolytic breakdown of b2-peptides has
not yet been investigated. In this investigation, we address the
potential of the b-aminopeptidases BapA and DmpA to de-
grade b2-peptidic substrates as well as the stereoselectivity of
these reactions. Furthermore, we describe the b-aminopepti-
dase-catalyzed formation of b2-peptides under kinetic control.
Figure 1. Conversion of H-b2hPhe-b2hAla-OH (1, +) by BapA (0.3 mg of pro-
tein per mL) at pH 7.2 and 378C. Two peaks of the diastereomeric mixture
~
that correspond to (S,S)-1, (R,R)-1 (tR =17.1 min, ) and to (S,R)-1, (R,S)-1
~
(tR =17.7 min, ) were detected by reversed-phase HPLC. The original HPLC
traces are shown in the Supporting Information (Figure S1).
1130
www.chembiochem.org
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemBioChem 2010, 11, 1129 – 1136

Biotransformations of b2-Dipeptides
Previous studies with b3-peptides and b3-amino acid amides
showed that BapA and DmpA strongly discriminated between
N-terminal b3-amino acids that carry side chains of different
sizes.[14,16a] While BapA accepts a broad range of N-terminal b3-
amino acids with different side-chain lengths and functionali-
ties, DmpA only catalyzes reactions with small, sterically unde-
manding b3-amino acids, such as N-terminal b3-homoalanine
and b-homoglycine.2 According to our present results, this ob-
servation applies to the degradation of the b2-dipeptide 1 as
well. Structural information on the active-site compositions of
BapA (unpublished results) and DmpA (PDB ID: 1B65)[18] reveal
that their different substrate specificities are possibly caused
by different topologies of the enzyme active sites. Whereas the
active-site pocket of BapA is relatively wide, the DmpA sub-
strate binding site is constricted by a loop region, ranging
from Gln131 to Trp137, that is likely to hinder the accessibility
of the active site for substrates with bulky N-terminal amino
acids.
2
2
*
Figure 2. Formation of H-(S)-b hPhe-OH ((S)-2, ) and H-(R)-b hPhe-OH ((R)-
*
2
2
2, ) from the diastereomeric mixture of 5 mm H-b hPhe-b hAla-OH (1, +)
catalyzed by BapA (0.3 mg of protein per mL) at pH 7.2 and 378C. The origi-
nal HPLC traces are shown in Figure S2.
Kinetic resolution of the diastereomeric pairs of
enantiomers of the b2-dipeptide 1 by BapA
The BapA-catalyzed release of H-b2hPhe-OH (2) from
a 5 mm solution of the b2-dipeptide 1 at pH 7.2 was
analyzed by HPLC on a chiral stationary phase. For
the separation of the enantiomers of 2 we used the
teicoplanin stationary phase Chirobiotic T2, which
provides good separation of many chiral b3- and b2-
amino acids.[19] We achieved baseline separation of
the enantiomers of 2, which was a prerequisite for ex-
amining the stereoselectivity of the enzyme-catalyzed
reaction. Under the applied conditions, (R)-2 (tR =
11.3 min) eluted prior to (S)-2 (tR =15.4 min). The
enantiomers of the second hydrolysis product H-
b2hAla-OH (3) could not be separated. The four ste-
reoisomers of 1 were partially separated under the
applied conditions. According to standards of (S,S)-1 and (S,R)-
1, peak 2 (tR =31.6 min) could be attributed to (S,R)-1 and
peak 4 (tR =36.0 min) to (S,S)-1.
Over the reaction period of 11 days, (S)-2 accumulated to a
concentration of 2.3 mm, which corresponds to a conversion
of 46% of the initially employed b2-dipeptide 1 (Figure 2). The
reaction was highly S-enantioselective (ee>99% and E>
500),[20] and the release of (R)-2 over the assay period remained
below the detection limit of 0.01 mm.
These results indicate almost quantitative conversion of
both diastereomers with an N-terminal (S)-b2-homophenylala-
nine residue ((S,S)-1 and (S,R)-1) (Scheme 2). Analysis of the
same samples by reversed-phase HPLC showed that (S,R)-1
was converted 1.3-times faster than (S,S)-1 (Figure 1). Hence,
we conclude that the configuration of the N-terminal b2-homo-
phenylalanine residue is crucial for the successful conversion
of 1 by BapA, whereas the configuration of the second amino
acid in the b2-dipeptide is only of minor importance.
Scheme 2. Kinetic resolution of the diastereomeric pairs of enantiomers of the b2-dipep-
tide 1 by BapA. a) BapA, 11 days, pH 7.2, 378C.
BapA-catalyzed peptide coupling reactions with the
b2dipeptide 1
The BapA-catalyzed conversion of the b2-dipeptide 1 was char-
acterized by simultaneously occurring hydrolysis and coupling
reactions among substrates and products (Scheme 3). This
complex interplay of competing reactions is based on the
postulated general catalytic mechanism of b-aminopeptidas-
es,[21] which was previously described for other enzyme mem-
bers of the N-terminal nucleophile (Ntn) hydrolase family, in-
cluding DmpA.[18,22] Accordingly, the first step in substrate con-
version is the formation of a characteristic acyl enzyme com-
plex through nucleophilic attack of the enzyme’s catalytically
active serine residue on the carbonyl carbon atom of the sub-
strate’s amide bond. The acyl moiety is subsequently released
from the acyl enzyme by the action of a nucleophile. The nu-
cleophile can be either water, which leads to hydrolysis of the
acyl enzyme, or the deprotonated N terminus of an amino acid
or a peptide, which results in peptide bond formation. The
creation of a potent nucleophile requires alkaline conditions to
keep the attacking amino group in its deprotonated state.
2
b-Homoglycine is commonly referred to as b-alanine.
ChemBioChem 2010, 11, 1129 – 1136
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chembiochem.org
1131

H.-P. E. Kohler et al.
Scheme 3. BapA-catalyzed routes for the conversion of the b2-dipeptide H-b2hPhe-b2hAla-OH (1) at the starting point of the reaction. All the products formed
by the initial hydrolysis and coupling reactions can further react as nucleophiles. This leads to the formation of additional peptidic coupling products (5–7,
see Table 1).
For the BapA-catalyzed coupling reactions with dipeptide 1,
we used an alkaline aqueous reaction system at pH 10, which
had previously been shown to support peptide couplings by
b-aminopeptidases.[16] At the initial stage of the reaction we
observed substrate hydrolysis and the formation of the b2-tri-
peptide H-[b2hPhe]2-b2hAla-OH (4; Scheme 3). This is due to
the fact that water and the substrate itself were the only nu-
cleophiles to release the acyl moiety from the active site of
BapA. As the reaction proceeded, all products created by hy-
drolysis and coupling reactions could further react as nucleo-
philes themselves, which gave rise to the additional coupling
products 5–7 (Table 1 and Figure 3).
Table 1. HPLC-MS analysis of the hydrolysis and coupling products that
accumulated after 48 hours of the BapA-catalyzed conversion of the dia-
stereomeric mixture of 1 at pH 10 and 378C (see also Figure 3).
Figure 3. HPLC-UV trace of a sample taken after 48 hours from the BapA-cat-
alyzed conversion of the diastereomeric mixture of H-b2hPhe-b2hAla-OH (1)
at pH 10 and 378C. The chromatogram shows the hydrolysis product H-(S)-
b2hPhe-OH ((S)-2) as well as the coupling products 4–7 (see also Table 1).
Compound
tR
[min]
Calcd
mass
m/z[a]
[M+Na]+
[M+H]+
H-b2hPhe-OH (2)
H-b2hPhe-b2hAla-OH (1)
16.6
17.1
17.7
22.4
22.8
23.1
23.2
25.7
25.9
26.1
26.4
26.8
179.1
264.2
180.2
265.1
265.1
426.2
426.1
426.2
341.4
587.5
587.5
587.3
587.4
502.6
n.d.[b]
287.2
287.3
448.2
448.3
448.3
363.2
609.8
609.5
609.7
609.8
524.2
H-[b2hPhe]2-b2hAla-OH (4)
425.2
could not be analyzed by reversed-phase HPLC, as the com-
pound coeluted from the column with the flow through. Simi-
larly to the reaction at pH 7.2, only (S)-2 was detected by HPLC
on chiral column material after completion of the enzymatic
reaction at pH 10, and the concentration of (R)-2 remained
below the detection limit. After the reaction period of 10 days,
the final concentration of (S)-2 (2.3 mm) corresponded to sub-
strate conversion of almost 50%.
H-[b2hPhe]2-OH (5)
H-[b2hPhe]3-b2hAla-OH (6)
340.2
586.3
H-[b2hPhe]3-OH (7)
501.3
[a] The detected mass/charge ratios (m/z) of the hydrogen and sodium
adducts of the molecular ions are given. Further fragmentations of the
peptides are not shown. [b] Not detected.
From these data we conclude 1) that the unconverted dia-
stereoisomers correspond to (R,S)-1 and (R,R)-1 and 2) that the
acyl enzyme could only be formed when the N-terminal b2-
homophenylalanine residue of the peptide had the S-configu-
ration (Scheme 3). Consequently, all coupling products 4–7
formed through enzymatic conversion must share the pres-
ence of one or more N-terminal b2-homophenylalanine resi-
dues of S configuration. For the first coupling product H-(S)-
b2hPhe-b2hPhe-b2hAla-OH (4), only three peaks were detected
for the four possible diastereoisomers ((S,S,S)-4, (S,S,R)-4,
(S,R,S)-4 and (S,R,R)-4) by reversed-phase HPLC (Figure 3 and
Since the overall equilibrium of the BapA-catalyzed reaction
lies on the side of the hydrolysis products, all the observed
coupling products that temporarily accumulated in the re-
action mixture were finally hydrolyzed to their amino acid
constituents b2-homophenylalanine (2) and b2-homoalanine (3;
Figure 4). In contrast to 2, the second hydrolysis product 3
1132
www.chembiochem.org
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemBioChem 2010, 11, 1129 – 1136

Biotransformations of b2-Dipeptides
precursor tripeptide 4 must have been present in the reaction
mixture. This is supported by the fact that the ratio of the
three peak areas of 4 was approximately 1:2:1; this suggests
that two of the four diastereoisomers of 4 coeluted from the
column. Only one peak was detected by reversed-phase HPLC
for each of the b2-homophenylalanine oligomers 5 and 7,
which temporarily accumulated in the reaction mixture
(Figure 3). As (R)-2 is not present in the reaction mixture to
attack the acyl enzyme, we conclude that 5 and 7 are solely
composed of b2-amino acid residues of S configuration.
pH-Dependence of BapA-catalyzed hydrolysis and coupling
reactions with the b2-dipeptide 1
Figure 4. Conversion of the b2-dipeptide H-b2hPhe-b2hAla-OH (1, +) at a
concentration of 5 mm by BapA (0.3 mg of protein per mL) at pH 10 and
The influence of pH on the BapA-catalyzed conversion of the
diastereomeric mixture of 1 (5 mm) was investigated at pH 7.2,
8 and 10. From the initial data points obtained by reversed-
phase HPLC analysis, we calculated the enzymatic rates for the
overall conversion of the substrate 1 (vC) as well as the rates
for the formation of the hydrolysis product 2 (vH) and of the
coupling product 4 (vS; Table 2). As expected from a previous
378C. Competing hydrolysis and coupling reactions gave rise to the forma-
2
*
tion of H-(S)-b hPhe-OH ((S)-2, ) and various peptidic coupling products
&
(4–7, ), respectively. Due to a lack of standards for the quantification of
compounds 4–7, the overall concentration of coupling products was esti-
mated on the basis of the experimentally determined concentrations of 1
and 2, assuming that one equivalent of coupling product is formed from
two equivalents of 1.
Table 2. Initial rates [mmolminÀ1 per mg protein] of the BapA-catalyzed
hydrolysis and coupling reactions with substrate 1 (5 mm) at 378C and at
different pH values.[a]
pH
7.2
8
10
vC
vH
vS
vS/vH
cmax [mm]
0.013
0.021
0.017
0.0022
0.0026
0.00075
0.0053
0.0094
0.0081
2.4
3.6
10.8
0.28
0.46
0.60
[a] The conversion rate (vC) relates to the initial decrease in substrate,
whereas the hydrolysis (vH) and synthesis (vS) rates relate to the initial in-
crease of free 2 and coupling product 4, respectively. cmax, which repre-
sents the maximum accumulation of all coupling products 4–7 in the re-
action mixture, was estimated by assuming that 1 equiv of the coupling
product is formed by the condensation of 2 equiv of 1.
investigation on the pH profile of BapA,[13c] the highest conver-
sion rate was observed at pH 8 (vC =0.021 mmolminÀ1 per mg
of protein). The vS/vH ratio was strongly dependent on the pH
of the reaction system. Interestingly, peptide synthesis predo-
minated over substrate hydrolysis in the initial stage of the
reaction under all pH conditions. A rise in pH from 7.2 to 10
caused a further shift towards coupling product formation, as
indicated by an increase in vS/vH from 2.4 to 10.8. Generally,
the conversion of 1 catalyzed by BapA was one to two orders
of magnitude slower than the conversion of b3-amino acid-
containing substrates under similar reaction conditions.[14]
While the accumulation of the coupling products 4–7 was
very low at pH 7.2 (0.28 mm), we observed maximum coupling
product concentrations of 0.46 mm at pH 8 and 0.60 mm at
pH 10. These concentrations correspond to conversions of 18.4
and 24%, respectively, of the initial dipeptide 1 (Table 2). Cou-
pling product accumulation peaked in a clear-cut maximum,
after which peptide hydrolysis became the predominant reac-
tion (see Figure 4). After completion of the reactions, the only
remaining detectable compounds were (S)-2 and the uncon-
Table 1). However, the appearance of four peaks for the diaste-
reoisomers of the tetrapeptide H-[(S)-b2hPhe]2-b2hPhe-b2hAla-
OH (6) implies that the four possible diastereoisomers of the
ChemBioChem 2010, 11, 1129 – 1136
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chembiochem.org
1133

H.-P. E. Kohler et al.
verted stereoisomers of the substrate, which correspond to
(R,S)-1 (tR =28.8 min) and (R,R)-1 (tR =33.5 min) as determined
by HPLC on the chiral column material.
Conclusions
In this investigation we have studied transformations of non-
natural peptides with N-terminal b2-amino acid residues under
aqueous reaction conditions by using the b-aminopeptidases
BapA and DmpA as catalysts. Our study describes for the first
time how peptides composed of b2-amino acids with proteino-
genic side chains can be converted by hydrolytic enzymes,
though with low catalytic rates when compared to b3-peptidic
substrates.[14] As previously observed for the conversion of b3-
peptides, the enzymes had distinct substrate preferences de-
pending on the size of the side chain on the N-terminal amino
acid. Only BapA acted on substrate 1 with its bulky N-terminal
b2-homophenylalanine residue, whereas the conversion of pep-
tides 8 and 9, which carry N-terminal b2-homoalanine and C-
terminal b3-homoleucine or b3-homophenylalanine residues,
was only catalyzed by DmpA. While DmpA converted sub-
strates with N-terminal amino acids of S and R configuration,
the BapA-catalyzed reaction was highly S enantioselective.
BapA simultaneously catalyzed hydrolysis and synthesis reac-
tions with the diastereomeric mixture of 1, leading to the hy-
drolysis of the substrate and to the formation of various cou-
pling products. The ratio of the rate of peptide formation to
the rate of peptide hydrolysis was positively correlated with an
increase in the pH.
Enzymatic degradation of b2,b3-dipeptides with N-terminal
b2-homoalanine (8 and 9)
In addition to the diastereomeric mixture of the b2-dipeptide
1, we performed experiments with the b-aminopeptidases
BapA and DmpA and four diastereomerically pure b2,b3-dipep-
tides carrying an N-terminal b2-homoalanine and a C-terminal
(S)-b3-homoamino acid residue ((S,S)-8, (R,S)-8 and (S,S)-9, (R,S)-
9).
The availability of enantiopure b2-amino acids is limited be-
cause chemical preparations of these compounds involve
labor- and cost-intensive multistep reactions.[12] Hence, alterna-
tive enzymatic approaches to the enantioselective production
of b2-amino acids are highly desirable to complement chemical
asymmetric-synthesis strategies. The highly S enantioselective
reactions of BapA with a model b2-dipeptide indicate that b-
aminopeptidases could become useful for the biocatalytic pro-
duction of enantiopure b2-amino acids by the kinetic resolu-
tion of racemic b2-amino acid derivatives, for example, b2-
amino acid amides or esters. Furthermore, the enzymes could
be used to catalyze the introduction of a b2-amino acid residue
as an N-terminal protecting group into a peptide, thereby sta-
bilizing the peptide against degradation by other exopeptidas-
es. To fully assess the biocatalytic potential of BapA and DmpA
for enantioselective conversion of b2-peptidic substrates, it will
be necessary to 1) optimize the present reaction system with
the aim of suppressing uncontrolled cross-couplings among
substrates and products and 2) carry out a detailed analysis of
the enzymes’ substrate specificities for peptides and amides
carrying N-terminal b2-amino acids with different side chains.
Interestingly, we did not detect any conversion of substrates
8 and 9 with BapA, whereas DmpA cleaved all four diastereo-
isomers. This is in contrast to our observations with substrate
1 and to previous results with H-(S)-b3hAla-(S)-b3hLeu-OH ((S,S)-
10), which is the b3-dipeptide isomer of (S,S)-8.[13c,14] The results
shown in Table 3 indicate that the catalytic rates for the
DmpA-catalyzed conversion of (S,S)-8 and (R,S)-8 were 70 and
90 times lower than the conversion rate for the corresponding
b3-dipeptide (S,S)-10 determined in this study under the same
conditions. An exchange of the C-terminal b3-homoleucine res-
idue of 8 by the more hydrophobic b3-homophenylalanine of
9 had an accelerating effect on substrate conversion by DmpA.
DmpA converted dipeptides 8 and 9 with N-terminal b2-homo-
alanine residues of both S and R configuration with low stereo-
selectivity, which is in contrast to the previously reported
highly S-enantioselective kinetic resolution of racemic b3-
homoalanine amide and b3-homoalanine p-nitroanilide by the
same enzyme.[15]
Experimental Section
Table 3. Conversion rates [mmolminÀ1 per mg of protein] of the diaste-
reomerically pure substrates 8, 9 and 10 (5 mm) by DmpA at pH 7.2 and
378C.
General remarks: The amino acids and peptides 1–10 were ana-
lyzed on a Dionex HPLC system equipped with a P680 pump, an
ASI-100 automated sample injector, an UltiMate 3000 thermostat-
ted column compartment, and a UVD 340U photodiode array de-
tector (Dionex, Sunnyvale, CA, USA). Enantiomers of the b2-amino
acid H-b2hPhe-OH (2) were separated without further derivatization
on the chiral teicoplanin stationary phase Chirobiotic T2 (250ꢁ
4.6 mm; Astec, Whippany, NJ, USA) at a constant temperature of
Substrate
Conversion rate
Substrate
Conversion rate
(S,S)-8
(S,S)-9
(S,S)-10
0.029
0.099
2.5
(R,S)-8
(R,S)-9
0.037
0.48
1134
www.chembiochem.org
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemBioChem 2010, 11, 1129 – 1136

Biotransformations of b2-Dipeptides
108C and quantified by relating the absorbance at 205 nm to a
standard curve of enantiopure (R)-2. The mobile phase was metha-
nol/H2O (90:10), and the applied flow rate was 1 mLminÀ1. Under
the described separation and detection conditions, the detection
limit for 2 was 0.01 mm.
Acknowledgements
This work was supported by the Swiss National Science Founda-
tion (Project No. 3152A0-100770, B.G.), the Deutsche Bundesstif-
tung Umwelt (Project No. 13176-32, T.H.) and the New Zealand
Foundation for Research Science and Technology (No. SWSS0401,
J.G.).
Additionally, all samples were analyzed by reversed-phase HPLC on
a Nucleodur C18-Pyramid stationary phase (250ꢁ4 mm, 5 mm parti-
cle size; Macherey–Nagel, Dꢂren, Germany), which was equilibrated
with 0.1% aqueous trifluoroacetic acid (TFA) at a column tempera-
ture of 208C. All substrates and products were separated with a
linear gradient of 0 to 40% acetonitrile at a constant flow rate of
1 mLminÀ1 and detected at a wavelength of 205 nm. Coupling
products were quantified on basis of the experimentally deter-
mined concentrations of the compounds 1 and 2 on the assump-
tion that one molecule of coupling product is formed from two
molecules of substrate. Mass spectra of all compounds were
recorded with an API 4000 liquid chromatography/tandem MS
system connected to an Agilent 1100 LC system. For protein deter-
mination, we used fivefold concentrated Bradford reagent (Bio-
Rad, Rheinach, Switzerland) and bovine serum albumin (BSA) as a
standard; absorbance measurements were performed at 595 nm
with a Specord S100 spectrophotometer (Analytik Jena, Jena, Ger-
many).
Keywords: aminopeptidases · hydrolases · kinetic resolution ·
nucleophiles · peptides
[1] D. Seebach, A. K. Beck, D. J. Bierbaum, Chem. Biodiversity 2004, 1, 1111–
1239.
[2] a) D. Seebach, S. Abele, J. V. Schreiber, B. Martinoni, A. K. Nussbaum, H.
Schild, H. Schulz, H. Hennecke, R. Woessner, F. Bitsch, Chimia 1998, 52,
734–739; b) J. Frackenpohl, P. I. Arvidsson, J. V. Schreiber, D. Seebach,
ChemBioChem 2001, 2, 445–455; c) H. Wiegand, B. Wirz, A. Schweitzer,
G. P. Camenisch, M. I. Rodriguez Perez, G. Gross, R. Woessner, R. Voges,
P. I. Arvidsson, J. Frackenpohl, D. Seebach, Biopharm. Drug Dispos. 2002,
23, 251–262; d) H. Wiegand, B. Wirz, A. Schweitzer, G. Gross, M. I. Rodri-
guez Perez, H. Andres, T. Kimmerlin, M. Rueping, D. Seebach, Chem. Bio-
diversity 2004, 1, 1812–1828.
[3] D. Seebach, M. Overhand, F. N. M. Kꢂhnle, B. Martinoni, L. Oberer, U.
Hommel, H. Widmer, Helv. Chim. Acta 1996, 79, 913–941.
Enzyme expression and purification: The genes bapA and dmpA,
which code for the b-aminopeptidases BapA from S. xenopeptidily-
tica 3-2W4 and DmpA from O. anthropi LMG7991, respectively,
were cloned into the expression plasmid pET9c (Novagen, Madi-
son, USA). Plasmid p3BapA[13c] was digested with BamHI and NdeI.
The bapA-containing fragment was cloned into pET9c, which was
cut at the same sites to yield pAR116. The dmpA gene was cloned
from pODmpA[14] into pET9c according to the same procedure
with the restriction enzymes BamHI and SalI to yield pAR114. The
two obtained expression plasmids were transformed into E. coli
BL21(DE3) pLysS (Novagen, Madison, USA) and recombinantly ex-
pressed in the presence of kanamycin and chloramphenicol. BapA
and DmpA were purified according to established procedures.[13b,14]
The lyophilized enzyme powders were dissolved in a potassium
phosphate buffer (10 mm, pH 7.2), and the protein contents of the
enzyme stocks were determined spectrophotometrically.
[4] a) D. Seebach, S. Abele, K. Gademann, G. Guichard, T. Hintermann, B.
Jaun, J. L. Matthews, J. V. Schreiber, L. Oberer, U. Hommel, H. Widmer,
Helv. Chim. Acta 1998, 81, 932–982; b) D. Seebach, B. Jaun, R. Sebesta,
R. I. Mathad, O. Flçgel, M. Limbach, H. Sellner, S. Cottens, Helv. Chim.
Acta 2006, 89, 1801–1825.
[5] a) J. A. Patch, A. E. Barron, Curr. Opin. Chem. Biol. 2002, 6, 872–877;
b) D. L. Steer, R. A. Lew, P. Perlmutter, A. I. Smith, M.-I. Aguilar, Curr. Med.
Chem. 2002, 9, 811–822; c) M.-I. Aguilar, A. W. Purcell, R. Devi, R. Lew, J.
Rossjohn, A. I. Smith, P. Perlmutter, Org. Biomol. Chem. 2007, 5, 2884–
2890; d) D. Seebach, J. Gardiner, Acc. Chem. Res. 2008, 41, 1366–1375.
[6] a) J. A. Kritzer, J. D. Lear, M. E. Hodsdon, A. Schepartz, J. Am. Chem. Soc.
2004, 126, 9468–9469; b) J. A. Kritzer, O. M. Stephens, D. A. Guarracino,
S. K. Reznik, A. Schepartz, Bioorg. Med. Chem. 2005, 13, 11–16; c) O. M.
Stephens, S. Kim, B. D. Welch, M. E. Hodsdon, M. S. Kay, A. Schepartz, J.
Am. Chem. Soc. 2005, 127, 13126–13127; d) E. P. English, R. S. Chuma-
nov, S. H. Gellman, T. Compton, J. Biol. Chem. 2006, 281, 2661–2667;
e) R. Akkarawongsa, T. B. Potocky, E. P. English, S. H. Gellman, C. R.
Brandt, Antimicrob. Agents Chemother. 2008, 52, 2120–2129.
[7] a) K. Gademann, M. Ernst, D. Hoyer, S. Seebach, Angew. Chem. 1999,
111, 1302–1304; Angew. Chem. Int. Ed. 1999, 38, 1223–1226; b) K. Gade-
mann, T. Kimmerlin, D. Hoyer, D. Seebach, J. Med. Chem. 2001, 44,
2460–2468; c) D. Seebach, E. Dubost, R. I. Mathad, B. Jaun, M. Limbach,
M. Lçweneck, O. Flçgel, J. Gardiner, S. Capone, A. K. Beck, H. Widmer, D.
Langenegger, D. Monna, D. Hoyer, Helv. Chim. Acta 2008, 91, 1736–
1786; d) S. Krebs, J. R. Lamas, S. Poenaru, G. Folkers, J. A. Lꢃpez de Cas-
tro, D. Seebach, D. Rognan, J. Biol. Chem. 1998, 273, 19072–19079; e) S.
Poenaru, J. R. Lamas, G. Folkers, J. A. Lꢃpez de Castro, D. Seebach, D.
Rognan, J. Med. Chem. 1999, 42, 2318–2331; f) S. Dedier, S. Krebs, J. R.
Lamas, S. Poenaru, G. Folkers, J. A. Lꢃpez de Castro, D. Seebach, D.
Rognan, J. Recept. Signal Transduction 1999, 19, 645–657; g) A. I. Webb,
M. A. Dunstone, N. A. Williamson, J. D. Price, A. de Kauwe, W. Chen, A.
Oakley, P. Perlmutter, J. McCluskey, M.-I. Aguilar, J. Rossjohn, A. W. Pur-
cell, J. Immunol. 2005, 175, 3810–3818.
General procedure for the enzyme-catalyzed conversion of the
diastereomeric mixture of H-b2hPhe-b2hAla-OH (1): The reaction
mixtures contained the diastereomeric mixture of 1 (5 mm) in solu-
tions of different buffering systems (100 mm, potassium phosphate
pH 7.2, Tris/HCl pH 8, sodium carbonate pH 10). Enzymatic conver-
sions of the substrates were initiated by the addition of an appro-
priate amount of BapA or DmpA. Samples were withdrawn at inter-
vals from the reaction mixtures, and the enzymatic reaction was
quenched by heating the samples at 908C for 3 min. The reaction
products were analyzed by chiral and reversed-phase HPLC. A
sample taken from the enzymatic assay at pH 10 after 48 h was
subjected to HPLC-MS analysis.
General procedure for the enzyme-catalyzed conversion of the
enantiopure substrates 8, 9 and 10: The reaction mixtures con-
tained 5 mm of the diastereomerically pure b2,b3-dipeptides H-(S)-
b2hAla-(S)-b3hLeu-OH ((S,S)-8), H-(R)-b2hAla-(S)-b3hLeu-OH ((R,S)-8),
H-(S)-b2hAla-(S)-b3hPhe-OH ((S,S)-9), H-(R)-b2hAla-(S)-b3hPhe-OH
((R,S)-9), or the b3-dipeptide H-(S)-b3hAla-(S)-b3hLeu-OH ((S,S)-10)[13c]
in potassium phosphate buffer (100 mm, pH 7.2) containing 10%
DMSO. Enzymatic conversions of the substrates were carried out as
described above, and samples were analyzed by reversed-phase
HPLC.
[8] a) E. A. Porter, X. Wang, H.-S. Lee, B. Weisblum, S. H. Gellman, Nature
2000, 404, 565; b) P. I. Arvidsson, J. Frackenpohl, N. S. Ryder, B. Liechty,
F. Petersen, H. Zimmermann, G. P. Camenisch, R. Woessner, D. Seebach,
ChemBioChem 2001, 2, 771–773; c) E. A. Porter, B. Weisblum, S. H. Gell-
man, J. Am. Chem. Soc. 2002, 124, 7324–7330; d) P. I. Arvidsson, N. S.
Ryder, H. M. Weiss, G. Gross, O. Kretz, R. Woessner, D. Seebach, Chem-
BioChem 2003, 4, 1345–1347; e) E. A. Porter, B. Weisblum, S. H. Gellman,
J. Am. Chem. Soc. 2005, 127, 11516–11529; f) A. J. Karlsson, W. C. Pom-
erantz, B. Weisblum, S. H. Gellman, S. P. Palecek, J. Am. Chem. Soc. 2006,
ChemBioChem 2010, 11, 1129 – 1136
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chembiochem.org
1135

H.-P. E. Kohler et al.
128, 12630–12631; g) M. A. Schmitt, B. Weisblum, S. H. Gellman, J. Am.
Chem. Soc. 2007, 129, 417–428.
[9] S. Guha, T. Chakraborty, A. Banerjee, Green Chem. 2009, 11, 1139–1145.
[10] G. Lelais, D. Seebach, Biopolymers 2004, 76, 206–243.
[11] E. Juaristi, V. Soloshonok, Enantioselective Synthesis of b-Amino Acids,
Wiley, Hoboken, 2005.
terman, J. G. T. Kierkels, T. Milcent (DSM IP Assets B.V.), WO 2005/085462
A1, 2005; d) A. Liljeblad, L. Kanerva, Tetrahedron 2006, 62, 5831–5854;
e) D.-Y. Ma, D.-X. Wang, Q.-Y. Zheng, M.-X. Wang, Tetrahedron: Asymme-
try 2006, 17, 2366–2376; f) M. Fitz, E. Forrꢃ, E. Vigꢃczki, L. Lꢅzꢅr, F.
Fꢂlçp, Tetrahedron: Asymmetry 2008, 19, 1114–1119; g) A. I. Martꢆnez-
Gꢃmez, S. Martꢆnez-Rodrꢆguez, J. Pozo-Dengra, D. Tessaro, S. Servi, J. M.
Clemente-Jimꢇnez, F. Rodrꢆguez-Vico, F. J. L. Heras Vꢅzquez, Appl. Envi-
ron. Microbiol. 2009, 75, 514–520.
ˇ
[12] D. Seebach, A. K. Beck, S. Capone, G. Deniau, U. Groselj, E. Zass, Synthe-
sis 2009, 1, 1–32, and references therein.
[13] a) L. Fanuel, C. Goffin, A. Cheggour, B. Devreese, G. van Driessche, B.
Joris, J. van Beeumen, J.-M. Frꢄre, Biochem. J. 1999, 341, 147–155; b) B.
Geueke, K. Namoto, D. Seebach, H.-P. E. Kohler, J. Bacteriol. 2005, 187,
5910–5917; c) B. Geueke, T. Heck, M. Limbach, V. Nesatyy, D. Seebach,
H.-P. E. Kohler, FEBS J. 2006, 273, 5261–5272.
[18] C. Bompard-Gilles, V. Villeret, G. J. Davies, L. Fanuel, B. Joris, J.-M. Frꢄre,
J. van Beeumen, Structure 2000, 8, 153–162.
[19] a) A. Pꢇter, L. Lꢅzꢅr, F. Fꢂlçp, D. W. Armstrong, J. Chromatogr. A 2001,
926, 229–238; b) A. Pꢇter, A. ꢈrki, E. Vꢇkes, D. Tourwꢇ, L. Lꢅzꢅr, F. Fꢂlçp,
D. W. Armstrong, J. Chromatogr. A 2004, 1031, 171–178; c) Z. Pataj, I.
Ilisz, R. Berkecz, A. Misicka, D. Tymecka, F. Fꢂlçp, D. W. Armstrong, A.
Pꢇtꢇr, J. Sep. Sci. 2008, 31, 3688–3697.
[14] T. Heck, M. Limbach, B. Geueke, M. Zacharias, J. Gardiner, H.-P. E. Kohler,
D. Seebach, Chem. Biodiversity 2006, 3, 1325–1348.
[15] T. Heck, D. Seebach, S. Osswald, M. K. J. ter Wiel, H.-P. E. Kohler, B.
Geuek, ChemBioChem 2009, 10, 1558–1561.
[20] A. J. J. Straathof, J. A. Jongejan, Enzyme Microb. Tech. 1997, 21, 559–
571.
[16] a) T. Heck, H.-P. E. Kohler, M. Limbach, O. Flçgel, D. Seebach, B. Geueke,
Chem. Biodiversity 2007, 4, 2016–2030; b) T. Heck, V. S. Makam, J. Lutz,
L. M. Blank, A. Schmid, D. Seebach, H.-P. E. Kohler, B. Geueke, Adv. Synth.
Catal. 2010, 352, 407–415; c) J. Heyland, N. Antweiler, J. Lutz, T. Heck,
B. Geueke, H.-P. E. Kohler, L. M. Blank, A. Schmid, Microb. Biotechnol.
2010, 3, 74–83.
[21] B. Geueke, H.-P. E. Kohler, Appl. Microbiol. Biotechnol. 2007, 74, 1197–
1204.
[22] H. J. Duggleby, S. P. Tolley, C. P. Hill, E. J. Dodson, G. Dodson, P. C. E.
Moody, Nature 1995, 373, 264–268.
[17] a) G. M. Salamonczyk, K. Han, Z.-w. Guo, C. J. Sih, J. Org. Chem. 1996, 61,
6893–6900; b) M. Solymꢅr, A. Liljeblad, L. Lꢅzꢅr, F. Fꢂlçp, L. T. Kanerva,
Tetrahedron: Asymmetry 2002, 13, 1923–1928; c) B. Kaptein, Q. B. Brox-
Received: December 11, 2009
Published online on March 25, 2010
1136
www.chembiochem.org
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemBioChem 2010, 11, 1129 – 1136