J.-W. Youn, et al.
MolecularCatalysisxxx(xxxx)xxxx
2.3. Biotransformation
(94:6:0.2) as a mobile phase at 30 °C and a flow rate of 1 ml min-1. As
enantiomers could be thus resolved, we identified the corresponding R-
enantiomers, e.g. R-p-chloromandelate, R-fluoromandelate, and R-p-
hydroxymandelate by a commercially available R-specific mandelate
dehydrogenase (Applichem, Darmstadt Germany). Our attempts to re-
solve DOMA, o-fluoromandelic acid and 3,5-difluoromandelic acid as S-
and R-mandelic acids derivatives were hampered by either the lack of
standards or by difficulties in chiral resolution as described also by
The washed cells were resuspended in 200 mM Tris−HCl buffer (pH
7.5) and the OD600 of the culture was set to ∼30. The biotransforma-
tion was started by adding the test substrate to a final concentration of
approximately 5 mM to the reaction volume (10 ml). Cells were in-
cubated in 100 ml Erlenmeyer flasks at 37 °C with 150-rpm agitation.
Samples (180 μl) were withdrawn every hour (until 11 h, final sample
after 20 h), and immediately mixed with 1 N HCl (20 μl) to stop the
reaction. Samples were then centrifuged at 22,000 g for 10 min and the
supernatants were removed and stored at – 20 °C until further use.
2.6. qRT-PCR
2.4. Analytical methods
To determine the transcriptional expression strength of aspC, hms
and tyrB a qRT-PCR was performed. Total RNA were isolated from
growing cells in LB before induction and 4 h after induction with IPTG
by using RNeasy Kit (Qiagen, Germany). The cDNA synthesis and the
following quantitative PCR was performed as described previously
[48]. The used primers are listed in Table 1. The quantification was
relative to the expression level of the housekeeping gene gapA per-
formed via cycle threshold (ΔΔ CT).
The progress of biotransformation was routinely analyzed and
quantified by high-pressure liquid chromatography (HPLC, 1100 series
Agilent Technologies); analysis was carried out by a diode array de-
tector (G1315B, 1100 series Agilent Technologies). The conversions of
o-, m-fluorophenylalanine, L-phenylalanine and m-F-tyrosine were
analyzed by using 5 μl sample and using a Prontosil ace-eps column
(250 × 4 mm, CS-Chromatographie, Langerwehe, Germany) with
40 mM Na2SO4 (pH 2.65 was adjusted with methanesulfonic acid) as
mobile phase at a flow rate of 0.8 ml min−1. All other conversions were
analyzed by using a Lichrospher RP18 column (250 × 4 mm, Trentec
Analysentechnik, Rutesheim Germany). The mobile phase consisted of
solvent 1 (aqueous solution with 0.1 % trifluoric acid (v/v)) and solvent
2 (acetonitrile with 0.1 % trifluoric acid (v/v)). A gradient was applied
with a flow rate of 1 ml min-1 starting with 98 % solvent 1 / 2 % solvent
2–90% solvent 1 / 10 % solvent 2 after 5 min followed with a gradient
of 50 % solvent 1 / 50 % solvent 2 after 25 min followed by isocratic
period of 10 min with the starting conditions. All separated components
were detected at a wavelength of 210 nm.
The following standard commercial compounds were used: L-phe-
nylalanine, L-tyrosine, DL-o-fluoro-phenylalanine, DL-m-fluoro-phenyla-
lanine, DL-p-fluoro-phenylalanine, DL-m-fluoro-tyrosine, L-o-chloro-phe-
nylalanine, DL-p-chloro-phenylalanine, L-3,4-dihydroxyphenylalanine,
p-methyl-L-phenylalanine, o-propargyl-L-tyrosine, phenylpyruvate, p-
hydroxyphenylpyruvate, phenylacetate, phenylglyoxylate, pheny-
lethanol, indole-3-pyruvate, imidazole pyruvic acid, o-nitrophenylpyr-
uvate, R- and S-mandelic acid, R,S-hydroxymandelic acid, R,S-o-fluoro-
mandelic, R,S-m-fluoro-mandelic acid, R- and R,S-p-fluoro-mandelic
acid, R,S-o-chloro-mandelic acid, S- and R,S-p-chloro-mandelic acid, R,S
-3,4-dihydroxy-mandelic acid (DOMA) and R-p-methylmandelate.
These compounds were of the highest available purity and were pur-
chased either from Alfa Aesar (Karlsruhe, Germany), Iris-Biotech
(Marktredwitz, Germany) or Sigma-Aldrich (Taufkirchen, Germany).
No standards were available for enantiomeric pure p-hydro-
3. Results
3.1. Biotransformation of L-tyrosine to p-hydroxymandelic acid
Our initial efforts to produce stably active HMS enzyme from E. coli
BL21 (DE3) pET28a-hms for in-vitro transformations showed that the
recombinant His-tagged enzyme could be easily overexpressed and
purified to near homogeneity by IMAC (data not shown). However,
although HMS was active in conversion of HPP or PP into the respective
mandelic acids, the enzyme was very unstable and lost activity rapidly,
even in the presence of the antioxidant ascorbic acid (at 50 mM) as
recommended by an earlier publication [37]). We therefore decided to
work with hms-recombinant resting cells of E. coli as whole cell bioca-
talysts. This should offer a protective compartment to the enzyme due
to the reducing properties of E. coli cells. Moreover, E. coli cells are
known to harbor aromatic aminotransferases which are able to convert
L-tyr or L-phe into the HMS substrates, HPP and PP, respectively
[41–43]. Thus a two-step biocatalysis from aromatic amino acids via
keto acids to mandelic acids could occur.
Recombinant cells of E. coli BL21 (DE3)/ pET28a-hms were grown in
LB medium and IPTG was added to induce the formation of HMS for 4 h
at 37 °C. Thereafter, the cells were harvested by centrifugation, washed
in buffer and resuspended to an OD600 of ∼30 into a buffer containing
L-tyr at a final concentration of 5 mM. To ensure oxygen input for the
HMS dioxygenase, the cells were shaken at 37 °C and samples were
taken hourly up to 20 h. As a control, cells of E. coli BL21 (DE3)/
pET28a were treated likewise.
xymandelate,
o-fluoromandelate,
m-fluoromandelate,
3,5-di-
fluoromandelate, p-chloromandelate, o-nitromandelate and DOMA.
As it can be seen from Fig. 2, both strains started to consume L-tyr
from the medium, the hms-recombinant strain having already con-
sumed L-tyr completely after 6 h, whereas the control showed residual L-
tyr (3 mM) in the supernatant at the end of incubation. The control
strain did not form S-HMA (detected by chiral chromatography),
whereas the hms-recombinant started immediately to produce S-HMA
and yielded ca. 4.5 mM of S-HMA after 6 h. This shows that the resting
cells of hms-recombinant E. coli are able to convert exogenously added
L-tyr to p-hydroxymandelate at high yields (∼ 85 % of theoretical
value). We assume that L-tyr is taken up and converted (transaminated
by aminotransferases) into HPP which then serves as a substrate for
HMS; HPP was not detected in the supernatant. The aspC and tyrB gene
expression was not alternate by the plasmid-borne hms overexpression
compared with the empty plasmid control (Fig. S1). The maximum
HMA space time yield (STY) was 177 μmol h-1 gCDW-1 (Fig. 4). The
remainder of L-tyrosine may have been used for protein biosynthesis or
other metabolism.
2.5. Chiral determination of mandelic acid derivatives
The enantiomeric excess (ee) of the products was determined
quantitatively by using HPLC with comparison to standard curves of
available racemic and enantiomeric pure MAs. The pH of the samples
was adjusted to ∼ 1 by using HCl. Subsequently, the MAs were ex-
tracted by using a double volume of ethyl acetate.
The separation of the enantiomers of mandelic-, p-hydro-
xymandelic-, o-chloromandelic, p-methylmandelic and m- and p-fluor-
omandelic acid was achieved by using a Chirex 3126 (D)-penicillamine
column (150 × 4.6 mm, Phenomenex, Aschaffenburg, Germany) with
2 mM copper (II) sulfate and 15 % acetonitrile as mobile phase at 30 °C.
The flow rate was set to 1 ml min-1 and 4 μl was injected. The chiral
separation of 3,5-difluoro- and p-chloromandelic acid was achieved by
using Lux 5μ Amylose-2 column (250 × 4.6 mm, Phenomenex,
Aschaffenburg, Germany) with n-heptane
/ isopropanol / TFA
3