Microbial enantioselective removal of the N-benzyloxycarbonyl amino protecting group

Article abstract of DOI:10.1016/j.molcatb.2012.03.005

In order to deprotect N-carbobenzoxy-l-aminoacids (Cbz-AA) and related compounds, a series of microorganisms was selected from soil by enrichment cultures with Cbz-l-Glu as sole nitrogen source. A lyophilized whole-cell preparation of two Arthrobacter sp. strains grown on Cbz-Glu or Cbz-Gly exhibited a high cleavage activity. The conditions of hydrolysis have been optimized and a quantitative enantioselective deprotection of several Cbz-dl-amino acids was obtained, as well as the deprotection of N-carbamoylester derivatives of several synthetic amino compounds. The preparation of Cbz-d-allylglycine and l-allylglycine in high yield and high optical purity is described as an application of this method.

Full text of DOI:10.1016/j.molcatb.2012.03.005

Journal of Molecular Catalysis B: Enzymatic 84 (2012) 22–26
Contents lists available at SciVerse ScienceDirect
Journal of Molecular Catalysis B: Enzymatic
journal homepage: www.elsevier.com/locate/molcatb
Microbial enantioselective removal of the N-benzyloxycarbonyl amino
protecting group
Michèle Maurs, Francine Acher, Robert Azerad∗
Laboratoire de Chimie et Biochimie Pharmacologiques et Toxicologiques, UMR8601 CNRS, Université Paris Descartes, 45 rue des Saints-Pères, 75006 Paris, France
a r t i c l e i n f o
a b s t r a c t
Article history:
Available online 17 March 2012
In order to deprotect N-carbobenzoxy-l-aminoacids (Cbz-AA) and related compounds, a series of microor-
ganisms was selected from soil by enrichment cultures with Cbz-l-Glu as sole nitrogen source. A
lyophilized whole-cell preparation of two Arthrobacter sp. strains grown on Cbz-Glu or Cbz-Gly exhib-
ited a high cleavage activity. The conditions of hydrolysis have been optimized and a quantitative
enantioselective deprotection of several Cbz-dl-amino acids was obtained, as well as the deprotection
of N-carbamoylester derivatives of several synthetic amino compounds. The preparation of Cbz-d-
allylglycine and l-allylglycine in high yield and high optical purity is described as an application of this
method.
Keywords:
Protected aminoacids
Enzymatic hydrolysis
Arthrobacter
Resolution
© 2012 Elsevier B.V. All rights reserved.
1. Introduction
comparative cleavage of N-Cbz-GluOH and various N-Cbz-amino
acids with crude lyophilized preparations, their ability to react
The amino group protection during synthetic transformations
is generally ensured by acylation or carbamoylation. A num-
ber of such protecting groups may be subjected to deprotection
by mild enzymatic methods using acylases [1] or carbamate
(urethane) hydrolases [2,3], respectively. In most cases such an
enzymatic deprotection is enantioselective. The benzyloxycar-
bonyl group (Cbz) has been widely used to protect amino groups
in organic synthetic reactions. Deprotection is usually achieved
by catalytic hydrogenation, with the drawbacks introduced by
the presence of reactive groups, such as carbon–carbon double
bonds, nitro or sulphide groups, in the hydrogenolysis condi-
tions. Moreover, some products are resistant or need extreme
conditions for deprotection by hydrogenation. In addition, such a
deprotection is not enantioselective. Only a few examples of enzy-
matic hydrolytic deprotection of N-benzyloxycarbonyl derivatives,
mainly N-␣-benzyloxycarbonyl aminoacids and oligopeptides have
been reported, through the use of microbial whole cells[4], cell
extracts [5] or purified urethane hydrolases [6–14], directed to
various classes of ␣-aminoacids and reported to be highly enan-
tioselective.
with other N-oxycarbonyl esters of chemical interest and to carry
out deprotection and optical resolution of these derivatives.
2. Experimental
2.1. Chemicals
Phenylmethylsulfonyl fluoride (PMSF), DTT, EDTA, benzylchlo-
roformate and Cbz-l-amino acids were purchased from Sigma
or Fluka. All other reagents were of the highest purity avail-
able. Cbz-dl-allylglycine was synthetized from dl-allylglycine
(2-amino-4-pentanoic acid, Sigma) by the Shotten–Baumann reac-
tion: 1 eq. of amino acid solubilized in water containing sodium
hydrogen carbonate (1.1 eq.) was shaken with a benzyloxycar-
bonylchloride solution (1.1 eq., 50% in toluene). After several hours,
the mixture was washed three times with ether then the aqueous
solution was acidified with HCl and extracted with ethyl acetate.
Evaporation of solvent afforded a pure colorless oil (95% yield).
1H NMR (500 MHz, CD3CN), ı ppm, J Hz: 7.36 (5H, m, ArH), 5.88
(1H, br.d; JNH-2 = 7.0, NH), 5.78 (1H, ddt, J4-5a = 17.3, J4-5b = 10.4,
J4-3 = 7.3, H-4), 5.15 (1H, dd, J5a-5b = 1.6, J5a-4 = 17.0, H-5a), 5.10 (1H,
dd, J5b-5a = 1.3, J5b-4 = 10.4, H-5b), 5.08 (2H, s, Ar-CH2), 4.23 (1H, dt,
J2-NH = 5.0, J2-3 = 8.2, H-2), 2.57 (1H, dt, J3a-3b = 14.2, J3a-4 = 6.0, H-3a),
2.43 ((1H, dt, J3b-3a = 14.2, J3b-4 = 7.9, H-3b).
In the present study, we have screened isolated microorgan-
isms for their ability to hydrolyze N-Cbz-l-glutamic acid and
identified some promising strains. We describe in this paper the
13C NMR (500 MHz, CD3CN), ı ppm: 174,44 (CO2H), 157,11
(OCONH), 138.09 (ArC-1), 134.29 (C-4), 129.46 (ArC-3, ArC-5),
128.81 (ArC-4), 128.71 (ArC-2, ArC-6), 118.97 (C-5), 67.19 (ArCH2),
54.45 (C-2), 36.57 (C-3). The corresponding dicyclohexylammo-
nium (DCHA) salt was obtained by solubilization in methanol,
∗
Corresponding author. Tel.: +33 0 1 42862171; fax: +33 0 1 42868387.
E-mail addresses: michele.maurs@parisdescartes.fr (M. Maurs),
francine.acher@parisdescartes.fr (F. Acher), robert.azerad@parisdescartes.fr
(R. Azerad).
1381-1177/$ – see front matter © 2012 Elsevier B.V. All rights reserved.
doi:10.1016/j.molcatb.2012.03.005

M. Maurs et al. / Journal of Molecular Catalysis B: Enzymatic 84 (2012) 22–26
23
addition of a slight excess of dicyclohexylamine, and crystallization
by addition of ethyl ether at −20 ◦C. m.p. 150–151 ◦C.
Bradford protein assay (Bradford reagent, Sigma) was used to
determine the protein concentration [16]. The assays were per-
formed on 5 mg of lyophilized powders pretreated at 25 ◦C with 2 M
NaOH (1 mL) during 72 h. The assay mixture contained 10–100 ␮L
of solubilized powder, and 3 mL of Bradford reagent. After mixing,
the absorbance of the solution was measured at 595 nm. Bovine
serum albumin, treated in the same conditions, was used for cali-
bration.
Cbz-l-4-nitrophenylalanine (free acid, m.p. 103–104 ◦C) was
prepared in the same way. Cbz-dl-␣-methylglutamic acid,
bis(DCHA) salt (m.p. 152–153 ◦C) and Cbz-homocycloleucine (Cbz-
1-aminocyclohexanecarboxylic acid), DCHA salt (m.p. 142–143 ◦C)
were prepared by the same method, using a longer reaction time
(48–72 h), with successive additions in small amounts of NaHCO3
and benzyloxycarbonylchloride solution.
2.5. Kinetic measurement of hydrolytic activity
2.2. Selection of microorganisms
Enzyme activity in the lyophilized powder was measured at
30 ◦C and orbital shaking (200 rpm) in 0.1 M potassium phosphate
buffer pH 7.8, containing 1 mM PMSF and 10 mM EDTA, in a total
volume of 2.5 mL. The final carbamate concentration was 4 mM.
The reaction was started by addition of 10 mg of lyophilized prepa-
ration. Suitable aliquots of the reaction mixture were periodically
diluted with one volume of acetonitrile containing 0.1% formic acid,
centrifuged and the microfiltered supernatants were analyzed by
HPLC, as described in Section 2.4.
Hydrolytic activity, measured by the substrate concentration
linear decrease (up to 40–50%) by HPLC/UV at 250 nm, is expressed
as nmol min−1/mg of lyophilized powder or per mg of protein in the
lyophilized preparation under the assay conditions. The lyophilized
powder contained 0.06 mg protein/mg.
Microorganisms were obtained from local soil samples by
enrichment liquid culture in minimal medium with Cbz-l-Glu as
unique nitrogen source. A soil sample (1 g) was vigourously shaken
in 10 mL of sterile saline and after filtration, a few drops of filtrate
were introduced in 250 mL vials containing 100 mL of sterilized
medium A (glucose 20 g, Cbz-l-glutamate 10 g, KH2PO4 2 g, K2HPO4
2 g, MgSO4·7H2O 0.2 g, FeSO4·7H2O 0.18 g, ZnSO4 0.01 g in 1 L of
deionized water, adjusted to pH 7.0). After orbital shaking at 27 ◦C
and 200 rpm (Infors multitron) during 120 h, diluted samples of
the cultures were transferred to agar plates containing medium
A added with 1.5% agar. After 6 days at 25 ◦C, six different types
of colonies were recovered which were maintained on the same
solid medium at 4 ◦C and as suspensions in 25% glycerol at −80 ◦C.
Strains 7C1 and 3C/1J were selected for further studies. Cultures
were sent to the German Collection of Microorganisms and Cell
Cultures (DSMZ GmbH, Braunschweig, Germany) for identification
by 16S rDNA sequencing. Strain 7C1 was identified as an unknown
Arthrobacter sp., and strain 3C/1J, from partial sequence, as closely
related to Arthrobacter nicotinovorans. Both strains are freely avail-
able from the authors under a usual Material Transfer Agreement
procedure.
2.6. Hydrolysis and resolution of Cbz-dl-allylglycine
Cbz-dl-allylglycine (257 mg, 1.03 mmoles) in 0.1 M potassium
phosphate buffer pH 7.8 containing 1 mM PMSF and 10 mM EDTA
(total volume: 100 mL), was incubated at 40 ◦C with 670 mg of
lyophilized 7C1 for 6 h. After acidification to pH 2 with 5 N
HCl and centrifugation, the supernatant was diluted with water
(100 mL) and adsorbed on a 50WX4 column (Biorad, 50–100 mesh,
2.4 cm × 19 cm, H+ form). Cbz-d-allylglycine was eluted with water
(250 mL) and then l-allylglycine was eluted with 1 M NH4OH.
Fractions to be pooled were assayed by HPLC or ninhydrin stain-
ing. Evaporation of solvents under vacuum afforded pure oily
Cbz-d-allylglycine (97 mg, 82% calculated on the theoretical 50%
yield) which was crystallized in hexane-ether as its dicyclohexy-
2.3. Cultivation of microorganisms
The strains were grown with orbital shaking (200 rpm) at 27 ◦C
in medium B (glucose 20 g, yeast extract Difco 0.15 g, KH2PO4
2 g, K2HPO4 2 g, MgSO4·7H2O 0.1 g, NaCl 2 g in 1 L of permuted
water, adjusted to pH 7.0) containing 3.6 mM of Cbz-l-GluOH or
Cbz-GlyOH. Growth was followed by measuring by HPLC the dis-
appearance of the Cbz-amino acid along time. After about 50%
consumption (96 h), the cultures were centrifuged for 30 min at
13,000 rpm and 4 ◦C. The bacterial mass was washed twice in a
0.1 M potassium phosphate buffer pH 7.8 containing 10 mM EDTA,
then centrifuged again for lyophilization at −50 ◦C. The powder
obtained is stored on dry silicagel at −20 ◦C.
lammonium salt, m.p. 155–156 ◦C, [˛]D – 14.8 (c 0.68, MeOH),
20
and a l-allylglycine fraction (131 mg) contaminated with vari-
ous ammonium salts; the amino acid fraction was adsorbed on
an AG 1X4 column (200–400 mesh, 1.5 cm × 14 cm, OH− form)
and successively eluted with water (150 mL) and 0.1 N HCl. Frac-
tions were assayed by ninhydrin staining, pooled and evaporated
to give 54.7 mg (0.475 mmoles, 92% calculated on the theoreti-
cal 50% yield) of l-allylglycine, m.p. 240–245 ◦C (dec), [˛]D – 38
20
23
(c 0.4, H2O); lit. [˛]D
– 37.2 (c 4, H2O)[17]. Optical purity
of l-allylglycine (>98%) was determined by HPLC separation
of diastereoisomers on reverse phase C18 column after
derivatisation with Marfey’s reagent (see Supplementary Material,
Fig. S7).
2.4. Analytical methods
a
HPLC was performed at 40 ◦C using Gilson 306 and 307 pumps,
a Gilson 155 UV detector and a Gilson 235P autoinjector equipped
with a 20 ␮L sample loop. Samples were applied to a Symmetry C18
column (Waters, 4 × 250 mm) then eluted with a water-acetonitrile
(containing 0.1% formic acid) gradient, from 10:90 to 35:65 (v/v)
in 15 min. The flow rate was 1 mL/min. The detection wave-
lengths were 230 and 250 nm for the Cbz-amino compounds and
264 and 340 nm for the N-dinitrophenyl-l-alanylamide derivatives
produced by reaction with 1-fluoro-2,4-dinitrophenyl-5-l-alanine
amide (Marfey’s reagent) [15].
3. Results
3.1. Selection of strains
At least six different types of colonies were identified from a gar-
den soil by enrichment process and agar plate selection on medium
A containing Cbz-l-Glu. Whole cells grown in such conditions were
all able to hydrolyse Cbz-l-Glu, either as freshly grown cells or
after freezing. Excellent results were obtained with lyophilized
cells while an acetonic powder did not give reproducible results.
HPLC/MS analysis of hydrolysis mixtures showed that benzyl alco-
hol was formed in a quasi-stoichiometric amount compared to
NMR spectra were recorded on a Bruker Avance500 instrument
(500.13 MHz and 125.77 MHz for 1H and 13C, respectively) at 27 ◦C.
˛
was measured at 589 nm and 20 ◦C on a Perkin Elmer 341
[D]
spectropolarimeter. Melting points were determined on a Buchi
capillary instrument.

24
M. Maurs et al. / Journal of Molecular Catalysis B: Enzymatic 84 (2012) 22–26
10
9
8
7
6
5
4
3
2
1
0
10
10000
9
8
7
6
5
4
3
2
1
0
8000
6000
4000
2000
0
3
5
7
pH
9
11
0
20
40
60
80
100
Fig. 2. Effect of pH on hydrolysis of Cbz-l-Glu by strain 7C1. Different buffers used:
50 mM Na acetate, pH 4–5; 50 mM NaK phosphate buffer, pH 5.5–8; 50 mM glycine-
NaOH, pH 8.5–10, all of them containing 10 mM EDTA.
Hours
Fig. 1. Course of Cbz-cleavage activity production during growth of strain 7C1.
100 mL of medium B were inoculated with 0.5 mL of seed culture and incubated
at 27 ◦C. A sample of cultured broth (10 mL) was periodically withdrawn for mea-
surement of O.D.; the cells were harvested by centrifugation, washed with 25 mM K
phosphate buffer containing 10 mM EDTA, and lyophilized. Protein determinations
and kinetic measurements of activity were carried out as described in sections 2.4
and 2.5, respectively. (ꢀ) bacterial growth; (ꢀ) residual Cbz-l-GLu; and (ꢁ) enzyme
activity on CBz-l-Glu.
hydrolysis of other alkyloxycarbonyl-l-amino acids such as Boc-
l-Phe or methoxycarbonyl-l-Phe, and a significant activity was
detected with Boc-Phe-Leu. Unexpectedly, Cbz-l-Glutamine was
not a subtrate for the enzyme, while Cbz-l-Asparagine was actively
deprotected. Interestingly, no significant conversion of d-amino
acids derivatives was observed, as shown for Cbz-d-Glu (see
Supplementary Material, Fig. S4), Cbz-d-Ala, or Cbz-d-Phe, demon-
strating that the enzyme preparation is fully enantioselective and
can be used for the resolution of racemic Cbz-amino acids. Cbz-
l-Glu-␣-O-methyl ester was not a substrate indicating the need
for a free ␣-carboxylic acid function. N␣,Nε-diCbz-l-Lysine was
regioselectively deprotected in the ␣-position, as demonstrated by
HPLC analysis (data not shown), indicating again a specificity for
the presence of a free ␣-carboxylic group.
Cbz-l-Glu decrease, but a small amount of benzoic acid, formed by
oxidation during the incubations [10], was also present. The reac-
tion progress was thus analyzed for the substrate concentration
decrease by HPLC/UV (see Supplementary Material, Figs. S1 and
S2).
3.2. Selectivity of the lyophilized whole-cell preparation
The enzymatic activity observed for both strains looks very
similar, but 7C1 appeared to be slightly more active with most sub-
strates. Induction of activity by Cbz-Gly was lower with 7C1, but
very similar to that of Cbz-l-Glu on strain 3C/1J (Table 1).
Our results on the cleavage activity of the isolated strains look
noticeably different from previously published results obtained
with other microorganisms. No special specificity was observed
for a single or a few Cbz-amino acid substrates as described
by the Japanese or the German groups [2,3,6,14]. Moreover,
Cbz-l-Proline is not a substrate for our enzyme, contrarily to
the Sphingomonas paucimobilis SC16113 enzyme [5,11]. However
our Cbz-deprotecting enzyme shares with all others previously
described a strict enantioselectivity toward Cbz-l-amino acids.
With the lyophilized preparations, which were used through-
out all the following study, it was shown that two strains (7C1
and 3C/1J) showed the highest hydrolytic activity on Cbz-l-Glu
and were thus selected for further studies. It was immediately
apparent that the cleavage activity was induced by this substrate
(see Supplementary Material, Table S1) but also, to a lesser extent,
by other Cbz-amino acids, such as Cbz-Gly (see Supplementary
Material, Fig. S3). In these experiments, several additives such as
DTT, EDTA and PMSF were added, in order to protect the enzyme in
the crude preparation after suspension in buffer. However, a careful
examination of the individual effect of such additives showed that
only EDTA and PMSF were beneficial for the enzyme activity. Inter-
estingly, 0.5 mM DTT exerted a significant inhibition effect and was
thus omitted in the incubation media (see Supplementary Material,
Table S2).
3.3. Properties of the 7C1 lyophilized preparation
The hydrolytic specific activity on Cbz-l-Glu or Cbz-Gly of Strain
7C1 (grown on Cbz-l-Glu) was higher at the beginning of the cul-
ture and slightly decreased during growth. However a maximal
total activity yield was obtained in 70–80 h when Cbz-l-Glu was
approximately 30% consumed (Fig. 1).
Various N-Cbz-amino acids and related derivatives were evalu-
ated as substrates for deprotection using lyophilized preparations
of both strains (Table 1). Initial hydrolysis rates at 30 ◦C were main-
tained up to high conversion rates (more than 75% for the best
substrates) indicating high affinities.
The best substrates were Cbz-l-Asp, Cbz-l-Met, Cbz-Gly,
Cbz-l-Ala, Cbz-l-Phe, and Cbz-l-Tyr, followed by Cbz-l-Glu, Cbz-
l-His, N-␣-Cbz-l-Lys, Cbz-l-Trp, N-Cbz-l-Ser independently of
the inducer used (Cbz-l-Glu or Cbz-Gly). N-Cbz-l-Thr, Cbz-l-
Val, Cbz-l-Leu, Cbz-l-Arg, and particularly Cbz-l-Ileu were poor
substrates. Cbz-l-Pro was not a substrate. Cbz-peptides such as
Cbz-GlyGly, or Cbz-l-PheGly were good substrates for deprotec-
tion without any detectable proteolytic activity. Some activity
(at least one half compared to Cbz-Glu) was observed in the
This preparation was remarkably stable. When kept at −20 ◦C,
after more than one year, no significant decrease of activity toward
Cbz-l-Glu could be observed. The optimum hydrolytic activity was
observed at pH 7.8 (Fig. 2) then decreases rapidly to zero at pH
9–9.5. It is still 50% at pH 8.5, which should be profitable to solubi-
lize some of the CbzAA substrates.
The enzyme was fairly stable when incubated for 20 min up to
50 ◦C in the presence of 1 mM PMSF (Fig. 3). Stability promptly
decreased at temperatures above 35 ◦C in the absence of PMSF,
indicating the presence of some protease activities in the prepa-
ration (Fig. 3). Cbz-cleavage activity increased with temperature
with a maximum at 50 ◦C, then a decrease was observed due to
inactivation (see Supplementary Material, Table S3).
Divalent ions were poorly effective (Table 2), excepted for
a slight activation in the presence of 0.5–2 mM Mg++ (data not
shown) and a significant inhibition by Cu++, nearly complete above
1 mM (Fig. 4). That could be due to the well known complexation
of ␣-amino-␣-carboxylic acids with Cu++ ions. These properties

M. Maurs et al. / Journal of Molecular Catalysis B: Enzymatic 84 (2012) 22–26
25
Table 1
Compared hydrolysis rates of various Cbz-amino acids and related derivatives by strains 7C1 and 3C/1J grown on Cbz-Glu or Cbz-Gly as nitrogen source.
Hydrolysisa of
Strain 7C1 grown
on Cbz-l-Glu
Strain 3C/1J grown
on Cbz-l-Glu
Strain 7C1 grown
on Cbz-Gly
Strain 3C/1J grown
on Cbz-Gly
N-Cbz-l-GluOH
N-Cbz-d-GluOH
N-Cbz-l-AspOH
N-Cbz-l-MetOH
N-Cbz-l-LeuOH
N-Cbz-l-IleuOH
N-Cbz-l-ValOH
N␣-Cbz-l-LysOH
N␣,Nε-Cbz-l-LysOH
N-Cbz-l-PheOH
N-Cbz-d-PheOH
N-Cbz-l-TyrOH
N-Cbz-l-TrpOH
N-Cbz-GlyOH
N-Cbz-l-AlaOH
N-Cbz-d-AlaOH
N-Cbz-l-SerOH
N-Cbz-l-ThrOH
N-Cbz-l-ArgOH
N-Cbz-l-HisOH
N-Cbz-l-ProOH
N-Cbz-Gly-GlyOH
N-Cbz-Phe-GlyOH
N-Cbz-Glu-␣-OMe
N-Cbz-Glutamine
N-Cbz-Asparagine
N-Boc-l-PheOH
N-Methoxycarbonyl-l-Phe
N-Boc-Phe-LeuOH
100b
1.4
410
248
10
100c
0
206
171
46
100d
0
258
111
10
6
59
120
163f
600
0
133
80
500
106
0
159
39
14
89
0
320
206
6
100e
0
169
80
23
17
60
116
198f
180
0
97
57
353
104
0
57
82
7
134
0
343
241
6
1
47
85
0
53
38
209f
260
1
188f
270
2
155
96
223
154
200
149
0
100
27
16
198
0
275
275
0
0
309
75
250
154
0.6
121
27
32
96
0
215
120
0
0
396
48
6
0
454
25
–
182
84
–
62
9
–
–
–
–
a
Initial rates were measured for substrate consumption by HPLC during incubation (2–5 h, 30 ◦C, 200 rpm) of 10 ␮mol Cbz- or Boc-substrates with 10 mg lyophilized
microorganism in a 2.5 mL volume of 0.1 M Na, K phosphate buffer pH 7.8 containing 1 mM PMSF and 10 mM EDTA.
b
Initial rate:3.0 0.2 nmol min−1 mg−1 lyophilized powder (50 3 nmol min−1 mg−1 protein).
c
Initial rate:1.4 0.2 nmol min−1 mg−1 lyophilized powder (23 3 nmol min−1 mg−1 protein).
d
Initial rate:1.4 0.2 nmol min−1 mg−1 lyophilized powder (23 3 nmol min−1 mg−1 protein).
e
Initial rate:1.3 0.2 nmol min−1 mg−1 lyophilized powder (22 3 nmol min−1 mg−1 protein).
f
Only ␣-Cbz group hydrolyzed.
completely differ from most previously isolated carbamoylases,
which show a marked dependence on Co++ ion for maximal activity
[2,7].
3.4. Deprotection and resolution of non-natural Cbz-amino acids
and Cbz-amino compounds
In the perspective to use the enzyme to deprotect non natural
Cbz-aminoacids or diverse Cbz-amino compounds less soluble in
aqueous conditions, the effect of added solvent in the incubation
mixture was experimented (see Supplementary Material, Fig. S5):
acetonitrile was inhibitory even at low concentrations (50% inhi-
bition at 5% concentration), while DMSO exerted less inhibition, as
50% activity was still measured at 10% concentration.
A series of non-proteinic Cbz-amino acids and Cbz-amino
compounds were assayed as substrates to evaluate the scope
of the deprotection reaction and its enantioselectivity. Typical
results are shown in Table 3 and illustrated in the resolution
of Cbz-dl-allylglycine (see Supplementary Material, Fig. S6), as
a model for unsaturated amino acids. Cbz-␤-Alanine was not a
substrate, but N-methyl-N-Cbz-l-alanine was significantly cleaved.
Cbz-l-4-nitrophenylalanine was efficiently cleaved, but Cbz-d-or
l-phenylglycine were very poorly hydrolyzed. The resolution by
deprotection of ␣-disubstituted dl-amino acids was attempted:
30
25
20
15
10
5
Table 2
Effect of metal ions on 7C1 activity.
Additivesa
Relative rates (%)
Noneb
100
105
70
CoSO4·7H2O
CuSO4·7H2O
FeSO4·7H2O
MgSO4·7H2O
MnSO4·H2O
CaCl2·2H2O
ZnSO4·H2O
BaCO3
97
121
106
97
105
100
0
0
20
30
40
50
60
70
T°C
Hg(OAc)2
Fig. 3. Heat stability of 7C1 lyophilized powder. Initial rates of hydrolysis of Cbz-
l-Glu were measured at 30 ◦C in the standard kinetic conditions (see Section 2.5)
and expressed per 10 mg of lyophilized powder, after20-min incubation at different
temperatures: (᭹) with 1 mM PMSF and (ꢀ) PMSF omitted.
a
Final concentration of salts: 0.5 mM; EDTA was omitted in the incubation media.
Incubation at 30 ◦C in 2.5 mL of 0.1 M K phosphate buffer pH 7.8 containing 1 mM
b
PMSF, 10 mM EDTA, 10 ␮mol Cbz-l-Glu and 10 mg lyophilized 7C1.

26
M. Maurs et al. / Journal of Molecular Catalysis B: Enzymatic 84 (2012) 22–26
30,00
25,00
20,00
15,00
10,00
5,00
the amide linkage in the benzyloxycarbamate is another possibility
and may lead to the amino acid and the monoester of carbonic acid,
which likewise would decompose into benzyl alcohol and carbon
dioxide. In both cases, a volatile reaction product is formed pulling
the reaction to completion. Apparently, the size of the carbamoyl
ester is not critical for hydrolysis, as shown by similar activities
on Cbz, Boc or methoxycarbonyl derivatives. On the other hand, a
very high activity was observed in the hydrolysis of some N-benzoyl
amino acids (Bn-Gly, Bn-l-Phe, or Bn-dl-allylglycine) when assayed
as substrates of the 7C1 lyophilized powder (data not shown), thus
demonstrating a l-amino acid acylase activity.
0,00
0
0,2
0,4
mM Cu++
0,6
0,8
1
5. Conclusion
A strong alcoxycarbamoyl hydrolysing activity was obtained
from two isolated soil microorganisms. After optimization of
hydrolytic conditions, it was shown that a crude cell lyophilisate
could be used for the deprotection and resolution of natural and
unnatural Cbz-dl-amino acids. Work is in progress to purify the
enzymic activity responsible for carbamate hydrolysis in Arthrobac-
ter 7C1 and eventually obtain recombinant strains expressing
higher activities which might be useful for the resolution and
protection/deprotection of unnatural amino acids in organic
synthesis.
Fig. 4. Effect of Cu++ concentration on hydrolysis of Cbz-l-Glu by 7C1. Activity is
measured in the standard kinetic conditions (see Section 2.5) and expressed per
10 mg of lyophilized powder (0.6 mg of proteins).
Table 3
Relative rates of Cbz-cleavage of various N-Cbz-derivatives by strain 7C1 or 3C/1J
grown on Cbz-l-Glu.
Hydrolysis of
Strain 7C1 grown
Strain 3C/1J grown
on Cbz-l-Glu
on Cbz-l-Glu
N-Cbz-l-GluOH
100
35
2
1.5
200
37
12
0
100
10
0
0
352
7
N-Cbz-␣-methyl-dl-GluOH
N-Cbz-l-phenylglycine
N-Cbz-d-phenylglycine
N-Cbz-dl-allylglycine
N-Cbz-N-methyl-l-AlaOH
N-Cbz-homocycloleucine
N-Cbz-␤-alanine
Appendix A. Supplementary data
Supplementary data associated with this article can be found, in
the online version, at doi:10.1016/j.molcatb.2012.03.005.
–
1
N-Cbz-l-4-nitrophenylalanine
149
–
References
[1] H. Waldmann, D. Sebastian, Chem. Rev. 94 (1994) 911–937.
[2] E. Lewke, M.-R. Kula, Biotechnol. J. 9 (1988) 29–38.
[3] C. Sambale, M.-R. Kula, Biotechnol. J. 7 (1988) 49–60.
[4] T. Ohshiro, M. Shinji, Y. Morita, Y. Takayama, Y. Izumi, Appl. Microbiol. Biotech-
nol. 48 (1997) 546–548.
[5] R. Patel, V. Nanduri, D. Brzozovsk, C. McNamee, A. Banerjee, Adv. Synth. Catal.
345 (2003) 830–834.
[6] S. Murao, E. Matsumura, T. Shin, T. Kawano, Agric. Biol. Chem. 48 (1984)
1673–1675.
[7] S. Murao, E. Matsumura, T. Kawano, Agric. Biol. Chem. 49 (1985)
967–972.
[8] E. Matsumura, T. Shin, S. Murao, E. Yamamoto, T. Kawano, Chem. Pharm. Bull.
33 (1985) 408–411.
[9] E. Matsumura, T. Shin, S. Murao, S. Yomoda, T. Kawano, Chem. Pharm. Bull. 33
(1985) 1739–1744.
[10] E. Matsumura, T. Shin, S. Murao, M. Sakaguchi, T. Kawano, Agric. Biol. Chem. 49
(1985) 3643–3645.
[11] V.B. Nanduri, S. Goldberg, R. Johnston, R.N. Patel, Enzyme Microb. Technol. 34
(2004) 304–312.
[12] G. Alvaro, J.A. Feliu, G. Caminal, J. Lopez-Santin, P. Clapes, Biocatal. Biotransform.
18 (2000) 253–258.
[13] Y. Kimura, N. Yasuda, Agric. Biol. Chem. 53 (1989) 497–504.
[14] E. Matsumura, T. Shin, S. Murao, T. Kawano, Agric. Biol. Chem. 49 (1985)
973–979.
[15] R. Bhushan, H. Brückner, Amino Acids 27 (2004) 231–247.
[16] M.M. Bradford, Anal. Biochem. 72 (1976) 248.
[17] S. Black, N.G. Wright, Biol. Chem. J. 213 (1955) 39–50.
[18] S.J. Ryan, Y. Zhang, A.J. Kennan, Org. Lett. 7 (2005) 4765–4767.
[19] T.P. Boyle, J.B. Brenner, J.A. Coates, P.A. Keller, S.G. Pyne, Tetrahedron 61 (2005)
7271–7276.
Cbz-dl-␣-methyl-Glu
and
Cbz-dl-homocycloleucine
(1-
aminocyclohexanecarboxylic acid) were significantly hydrolyzed.
Allylglycine enantiomers and their N-protected derivatives
have been used as general synthetic intermediates for the access
to glutamic acid side chain homologues [18], lysine and argi-
nine homologues [19] or isoxazoline derived amino acids [20].
When Cbz-dl-allylglycine was incubated in preparative conditions
with 7C1 lyophilized powder, in optimized conditions, hydroly-
sis stopped at 50% conversion as expected (see Supplementary
Material, Fig. S6) and the acidified and the centrifuged incuba-
tion mixture was treated by adsorption on a cation exchanger to
obtain residual Cbz-d-allylglycine (82% yield). Repurification of the
aminoacid fraction on an anion exchanger afforded l-allylglycine
(92% yield) with >98% optical purity, as determined by HPLC
after derivatization with the Marfey’s reagent (see Supplementary
Material, Fig. S7).
4. Discussion
The mechanism of the cleavage enzymatic reaction is not clear.
One possibility is the hydrolysis of the ester linkage in the Cbz-
amino acid leading to benzyl alcohol and to the corresponding
unstable carbonic acid amide, which would spontaneously decom-
pose into amino acid and carbon dioxide. Enzymatic hydrolysis of
[20] S.D. Abbott, P. Lane-Bell, K.P.S. Sidhu, J.C. Vederas, J. Am. Chem. Soc. 116 (1994)
6514–6520.