Angewandte
Chemie
Combined, these results clearly demonstrate the effi-
ciency and operational simplicity of the biotransformations
that use 1 as the amine donor. A further significant advantage
of this system lies in the formation of intensely colored by-
products, which presumably arise from spontaneous polymer-
ization of isoindole 5 and offer a simple high-throughput
screening strategy to identify desired transaminase activity.
We sought to employ this assay to rapidly evaluate a panel of
commercially available w-TAs for their ability to utilize 1 as
an amine donor. This panel included a series of six (S)-
selective (Figure 2a; L2, L4, and L6 A–F) and six (R)-
selective (L1, L3, and L5 A–F) w-TAs from the Codex ATA
screening kit v2, which have been specifically engineered to
operate using high concentrations of isopropyl amine donor
and are currently the most widely utilized w-TA biocatalysts
for chiral amine synthesis. In the absence of an amine
acceptor, low levels of color change were observed with
a number of these biocatalysts following incubation with
diamine 1 (5 mm) for three hours (L1 and L2), which is
presumably due to the conversion of enzyme-bound PLP into
PMP, although the presence of small quantities of ketone
impurities in the commercial enzyme preparation cannot be
excluded. The addition of benzylacetone (10; 5 mm, 1.0 equiv)
as an amine acceptor resulted in significant color changes in
a number of the biotransformations after only 15 minutes (L3
and L4). Following overnight incubation, reactions with all
Codexis biocatalysts resulted in the formation of intensely
colored solutions and significant quantities of dark precipitate
(L5 and L6), demonstrating the broad utility of this amine
donor for the synthesis of both the (S) and (R) enantiomers of
chiral amines. In all cases, significant conversion of 10 into the
corresponding amine was confirmed by GC-FID analysis (see
the Supporting Information, Tables S1 and S2), demonstrat-
ing the reliability of this simple screening strategy, which
provides an ideal high-throughput platform for evaluating
panels of w-TA biocatalysts for their activity towards large
libraries of ketone or aldehyde substrates. In general, higher
conversions were achieved with the (S)-selective biocatalysts
than with their (R)-selective counterparts. Comparable levels
of conversion of 10 were achieved using the corresponding
immobilized w-TA biocatalysts that were recently commer-
cialized by Purolite/Codexis.
Further evidence for the reliability of this colorimetric
screening method was obtained by using commercial w-TAs
supplied by Almac (L7 A–F). Whereas the biotransforma-
tions with TAm106/TAm107 (L7A and L7B, respectively)
proceeded with moderate conversion and gave intensely
colored reaction mixtures, the reactions with TAm121 (L7D)
and TAm140 (L7F) gave no conversion (< 1%), and the
wells remained pale yellow. Significantly, low but detectable
levels of color change were observed in the reactions with
TAm115 (L7C) and TAm125 (L7E), which were shown to
proceed with < 5% and 5% conversion, respectively (see
Table S3).
Aside from providing a high-throughput method to
evaluate the activity of commercial w-TAs, the colorimetric
screening strategy described above offers an ideal platform
for the development of a new generation of w-TA biocatalysts
that are engineered to efficiently utilize 1 as an amine donor.
Unfortunately, the protein sequence of commercial biocata-
lysts is rarely disclosed, necessitating the identification and
application of suitable wild-type enzymes as starting points
for directed evolution. We have recently reported the use of
a transaminase from Pseudogulbenkiania ferrooxidans (pf-
ATA) for the regio- and stereoselective amination of dike-
tones using l-alanine as the amine donor.[17] We now
demonstrate that this wild-type biocatalyst displays modest
activity towards diamine 1 (see the Supporting Information).
Significantly, this has allowed the development of a single-
enzyme, colony-based assay to identify desired w-TA activity.
Figure 2. a) Conversion of 10 (5 mm) into the corresponding amine
using commercially available w-TAs and diamine 1 (5 mm). L1, L3, and
L5A–F contain the (R)-selective Codexis enzymes ATA025, 303, 013,
301, 415, and 117, respectively. L2, L4, and L6 contain the (S)-selective
Codexis enzymes ATA254, G05, 260, 256, 234, and 113, respectively.
L1/L2: diamine 1 only, 3 h; L3/L4: 15 min after addition of substrate
10; L5/L6: 24 h after addition of substrate 10; L7: A–F=Almac
TAm106, 107, 115, 121, 125, and 140, respectively, substrate 10,
diamine 1, 24 h. b) Colony-based screen with ortho-xylylenediamine (1).
Cells expressing the pf-ATA gene turn dark in color after 30 min (right).
Cells lacking the pf-ATA gene remain colorless (left).
Angew. Chem. Int. Ed. 2014, 53, 1 – 5
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
3
These are not the final page numbers!