J. Yu et al.
CatalysisCommunications120(2019)28–32
reaction mixture was stirred under N2 at 220 rpm for 12 h after the
complete addition.
conversions of all substrates by 7% at least, in particularly, the con-
versions of DMPPA and PPA were 91% and 94%, respectively. The
variants 10b, 261b could effectively catalyze all methoxy substituted
phenylpyruvates and phenylpyruvate, except 2-methoxy substituted
phenylpyruvate. The results illustrated the mutations may influence the
enzymatic activities, thus providing the effective mutation sites for
further improving AspAT activity.
L-DM-Phe precipitated with progressing bioconversion reaction, and
the post-treatment of L-DM-Phe was similar to that of L-tyrosine
[24,25]. The reaction mixture was centrifuged, and the collected L-DM-
Phe was purified according to the difference of solubility at the different
pH (see supporting information). The spectral characterization of L-DM-
Phe was shown as follows: [α]D = −4.3o (lit11 = −4.4o, c = 1, 4 M
25
hydrochloric acid), 1H NMR (D2O) (Bruker DRX600, Germany) δ: 6.88
(d, 1H, J = 8.2 Hz), 6.81 (s, 1H), 6.75 (dd, 1H, J = 1.2 Hz, 8.2 Hz), 3.76
(s, 3H), 3.74 (s, 3H), 3.38 (dd, 1H, J = 5.9 Hz, 6.8 Hz), 2.84 (dd, 1H,
J = 5.5 Hz, 13.6 Hz), 2.70 (dd, 1H, J = 7.2 Hz, 13.6 Hz). 13C NMR
(D2O, Bruker DRX600, Germany) δ: 182.4, 147.8, 146.7, 131.3, 122.0,
112.9, 111.8, 57.3, 55.7, 55.6, 40.3 (Fig. S1–2). Based on the similar
operation, the other methoxy substituted phenylalanines were ob-
tained, and their structures were confirmed by 1H NMR and 13C NMR as
shown in Supporting information (Fig. S3–10).
3.2. The properties of the engineered 170a AspAT
To elucidate the properties of the engineered 170a AspAT, the
overall characterization including the apparent kcat and Km of the wild-
type and AspAT variant for five phenylpyruvate substrates has been
performed (Table 1). The SDS-PAGE analysis of the purified AspAT was
shown in Fig. S21. The apparent kinetic parameters were determined as
described in Supporting information (Fig. S22–23). Based on the ap-
parent kcat and Km of the wild type and 170a variant, kcat/Km values of
the examined variant increased by 3–24 fold. The 170a variant AspAT
showed the highest increase in kcat/Km values for DMPPA and sig-
nificantly improved its conversion from 72.5% to 91.6%. This result
demonstrated that these four substituted amino acid positions were
functionally important.
3. Results and discussion
3.1. AspAT mutants obtained by error-prone PCR
According to HPLC chromatogram of D, L-Phe and L-Phe (Fig.
S28–30), L-Phe isomer was produced from the conversion of phe-
nylpyruvate under the catalysis of the variant 170a AspAT. For L-DM-
Phe, its L-(−) isomer characteristics was confirmed by HPLC chroma-
togram (Fig. S31) and the optical rotation determination [12]. Com-
pared with HPLC results of the corresponding racemates, the unique
product peaks in chromatograms for the other Phe derivatives indicated
that they were all L-isomers (Fig. S32–S37). The HPLC results indicated
that the ee % values of all the phenylalanine derivatives were > 99%,
the products were all L-configuration and consequently the exchanged
sites in 170a did not influence the chiral centers of products.
Three clones 10b, 170a and 261b were selected from about 4000
independent variants library, which exhibited higher catalytic effi-
ciency than the wild-type enzyme (see Supporting information 1.3).
Sequencing analysis clearly revealed that 10b was mutated at two
amino acid positions (F170S, R316H), 170a was mutated at four amino
acid positions (N165Y, R316H, A381Tand P399L), 261b was mutated at
two amino acid positions (R316H, T366A). The conversions of methoxy
substituted phenylpyruvates were analyzed under the catalysis of these
variants (Fig. 1). The synthesis of phenylpyruvic acid and its derivatives
were described in Supporting information and the structures were
confirmed by 1H NMR and 13C NMR (Fig. S11–20).
As shown in Fig. 1, all the conversions of DMPPA catalyzed by the
AspAT variants increased in comparison to the wild type AspAT. The
wild type AspAT catalyzed the transamination reactions with all the
methoxy substituted phenylpyruvates in approximately 70% conversion
but phenylpyruvate in 83% conversion. The variant 170a increased the
3.3. Synthesis of L-DM-Phe
So far, AspAT has never been engineered to improve the catalytic
efficiency from DMPPA to L-DM-Phe. Therefore, the synthesis of L-DM-
Phe was further studied among these products. After library screening,
the variant 170a was selected with an approximately 91.6% conversion,
which suggested DMPPA was not completely converted into the target
product. DMPPA substrate in the mixture could be detected readily by
HPLC, and the remaining substrate was determined to be < 0.01%.
Therefore, an approximately 8% loss of the product could be attributed
to the instability of DMPPA, which decomposed during the conversion
process as it was observed with PPA [29]. Considering the stability of
the substrate, we improved the conversion process by feeding DMPPA
in batches and performing the reaction under inert gas. It turned out
that the best addition method of DMPPA (9 mol) was an initial addition
of DMPPA (3 mol), followed by the subsequent addition in batch, 1 h
(3 mol), 2 h (1.5 mol), and 3 h (1.5 mol) on a preparative scale. As
shown in Fig. 2, the conversion could reach 95.4%, which was higher
than that of the one-batch addition of DMPPA (91%). The concentration
of DMPPA gradually decreased during the scale-up biotransformation,
and finally L-DM-Phe increased to 572 mM.
Additionally, as shown in Fig. S38 and scheme 1, the by-product,
oxaloacetate produced from L-aspartate, subsequently undergoes spon-
taneous decarboxylation to pyruvate, and then pyruvate was converted
into L-alanine by AspAT [22,23]. Nevertheless, the variant 170a pro-
Fig. 1. The conversions of phenylpyruvate and its derivatives with the purified
AspATs as catalyst. Phenylpyruvate substrates (200 mM, 125 μL, pH = 8), L-
aspartate (220 mM, 125 μL, pH = 8) and 2 μL PLP (1%, m/v) were mixed in
100 mM Tris-HCl buffer (200 μL, pH = 8) and separately catalyzed by the
purified AspAT solution from the variants (50 μL, 0.003 μg/μL) at 40 °C with
170 rpm shaking for 8 h.
PPA = Phenylpyruvic acid; 2-MPPA = 2-Methoxy phenylpyruvic acid; 3-
MPPA = 3-Methoxy phenylpyruvic acid; 4-MPPA = 4-Methoxy phenylpyruvic
acid; DMPPA = 3, 4-dimethoxy phenylpyruvic acid.
duced much less L-alanine (9.4%
conversion of L-aspartate into L-alanine to be 100%.) than the wild type
AspAT (32.3% 1.2%) (Fig. S38) after 8 h, which indicated the in-
0.7%, we supposed the complete
crease in DMPPA conversion for 170a AspAT was related to the de-
crease in alanine production.
In general, the scale-up biocatalysis might have anticipated diffi-
culty in the separation of the target compound from the reaction
30