Putrescine Transaminases for the Synthesis of Saturated Nitrogen Heterocycles from Polyamines

Article abstract of DOI:10.1002/cctc.201600075

Putrescine transaminase (pATA; EC 2.6.1.82) catalyzes the transfer of an amino group from terminal diamine donor molecules to keto acid acceptors by using pyridoxal-5′-phosphate as a cofactor. The ygjG genes from Escherichia coli K12, Bacillus megaterium, and Bacillus mycoides were successfully cloned and expressed in E. coli BL21(DE3) cells. The three putrescine transaminases were all shown to prefer diaminoalkanes as substrates and thereby generated cyclic imines from the ω-amino aldehyde intermediates. The addition of a mild chemical reducing agent rapidly reduced the imine intermediate in situ to furnish a range of N-heterocycle products. We applied pATA in a biomimetic synthesis of 2,3-dihydro-1H-indolizinium-containing targets, notably the bioactive alkaloid ficuseptine.

Full text of DOI:10.1002/cctc.201600075

DOI: 10.1002/cctc.201600075
Communications
Putrescine Transaminases for the Synthesis of Saturated
Nitrogen Heterocycles from Polyamines
Iustina Slabu,[a] James L. Galman,[a] Nicholas J. Weise,[a] Richard C. Lloyd,[b] and
Nicholas J. Turner*[a]
Putrescine transaminase (pATA; EC 2.6.1.82) catalyzes the trans-
fer of an amino group from terminal diamine donor molecules
to keto acid acceptors by using pyridoxal-5’-phosphate as a co-
factor. The ygjG genes from Escherichia coli K12, Bacillus mega-
terium, and Bacillus mycoides were successfully cloned and ex-
pressed in E. coli BL21(DE3) cells. The three putrescine transa-
minases were all shown to prefer diaminoalkanes as substrates
and thereby generated cyclic imines from the w-amino alde-
hyde intermediates. The addition of a mild chemical reducing
agent rapidly reduced the imine intermediate in situ to furnish
a range of N-heterocycle products. We applied pATA in a bio-
mimetic synthesis of 2,3-dihydro-1H-indolizinium-containing
targets, notably the bioactive alkaloid ficuseptine.
mercial products such as plastics, fabric softeners, and
petroleum additives.[10] Recently, researchers have elegantly
shown the potential to engineer overproducing strains of bac-
teria such as Corynebacterium glutamicum and E. coli to pro-
vide an alternative to polyamine production,[11] which at pres-
ent is largely achieved by chemical synthesis. Three catabolic
routes for the degradation of putrescine have been proposed,
namely, oxidative deamination by amine oxidases, transamina-
tion by transaminases, and glutamylation of putrescine with
subsequent oxidation steps to g-aminobutyric acid (GABA).[12]
Amongst these, the transaminase pathway for catabolism of
putrescine has not previously been investigated for potential
application in biocatalysis. Transaminases are pyridoxal-5’-
phosphate(PLP)-dependent enzymes that have shown remark-
able versatility in the asymmetric synthesis of chiral amino
building blocks.[8] Terminal diamines present themselves as at-
tractive alternative substrates for transaminases. In addition to
their relatively low cost and high availability, upon transamina-
tion diamines are converted into reactive amino aldehydes[13]
(Scheme 1), which spontaneously cyclize to cyclic imines thus
driving the reaction towards product formation. We envisaged
using putrescine transaminases (pATAs) to generate imines by
this process for further application in the synthesis of N-hetero-
cycles as described herein.
Saturated nitrogen-containing heterocycles are prevalent in
organic chemistry; they are frequently found in natural
products[1] and are used as intermediates for target molecule
synthesis. In particular, substituted pyrrolidines, piperidines,
and azabicyclic motifs are found in bioactive natural prod-
ucts[2,3] as well as in currently marketed agrochemicals, fine
chemicals, and pharmaceuticals. A number of synthetic strat-
egies have been developed for the preparation of N-heterocy-
clic compounds including hydroamination, CÀH amination,
and cyclization reactions.[4] During the last decade, microbial
and enzymatic catalysis have gained considerable interest for
the production of enantiopure amine-containing building
blocks.[5] Biocatalysts are typically highly regio- and stereoselec-
tive and can be used under mild conditions. Successful biocat-
alytic strategies have been reported for the synthesis of chiral
cyclic amines from different enzyme classes, including mono-
amine oxidases,[6] imine reductases,[7] and w-transaminases.[8]
Polyamines such as putrescine and cadaverine are biogenic
diamines present in almost all living cells, in which they modu-
late cellular proliferation and protein synthesis.[9] These poly-
amines are also used industrially for the production of com-
Scheme 1. Selected biocatalytic strategies for the synthesis of cyclic imines.
MAO-N: monoamine oxidase, w-TA: w-transaminase.
[a] I. Slabu, Dr. J. L. Galman, N. J. Weise, Prof. N. J. Turner
School of Chemistry
To identify suitable candidate transaminases with activity to-
wards diamines, we initially screened w-transaminases from
Chromobacterium violaceum,[14] Alcaligenes denitrificans,[15]
Arthrobacter citreus,[16] commercially available ATA-113, and
Arthrobacter sp. (ATA-117)[17] against a panel of polyamines by
using our well-developed l/d-amino acid oxidase colorimetric
assay.[18] However, none of these w-transaminases showed any
activity towards either putrescine (1c) or cadaverine (1d) as
the amino donor with pyruvate as the keto acceptor. We there-
University of Manchester, Manchester Institute of Biotechnology
131 Princess Street, Manchester, M1 7DN (UK)
Fax: (+44)161-445-0093
E-mail: Nicholas.Turner@manchester.ac.uk
[b] Dr. R. C. Lloyd
Dr. Reddy’s Laboratories, Chirotech Technology Centre
410, Cambridge Science Park, Milton Road, Cambridge, CB4 0PE (UK)
Supporting Information and the ORCID identification number(s) for the
author(s) of this article can be found under http://dx.doi.org/10.1002/
cctc.201600075.
ChemCatChem 2016, 8, 1038 – 1042
1038
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Communications
fore targeted the ygjG gene coding for putrescine transami-
nase from E. coli K12 (Ec-ygjG),[19,20] for which little data has
been reported regarding substrate scope. A basic local align-
ment search tool (BLAST) analysis of the ygjG gene revealed
two other putative transaminase genes from Bacillus mega-
terium (BM-ygjG) and Bacillus mycoides (BMy-ygjG) with protein
sequence identities of 57 and 58%, respectively. The ygjG
genes were subcloned into a pET-28b vector followed by over-
expression in E. coli BL21(DE3) cells. Purified recombinant His6-
tag protein was obtained and employed in specific activity ex-
periments. In parallel reactions, either l-amino acid oxidase
(AAO) from Crotalus adamanteus or d-AAO from porcine
kidney was added to reactions containing the diamine and
Ec-ygjG with pyruvate as the amine acceptor (Scheme 2). In the
Table 1. Specific activities of Ec-ygjG pATA on polyamines 1–6.[a]
Substrate
Specific activity
[mUmgÀ1
[b]
]
1a, n=2
0.62
1a–i
1b, n=3
1c, n=4
1d, n=5
1e, n=6
1 f, n=7
1g, n=8
1h, n=9
1i, n=10
2a, X=N
2b, X=O
2c, X=S
0.83
270.04
251.13
164.42
57.70
27.04
18.90
8.10
62.10
124.21
13.05
2a–c
3a–b
3a, X=H
3b, X=(CH2)2NH2
121.51
16.20
4
40.50
5
6
0.04
0.04
[a] The rates were determined by using a horse radish peroxidase/l-AAO
colorimetric assay with purified enzyme. [b] Assay conditions: 10 mm
amine, pATA (purified enzyme) 0.2 mgmlÀ1, pH 8.0, 308C.
Scheme 2. Putrescine transaminase (pATA)/amino acid oxidase (AAO) liquid-
phase colorimetric assay with the use of amino acid oxidase (l-AAO), horse-
radish peroxidase (HRP), 4-aminoantipyrine (4-AAP), and 2,4,6-tribromo-3-hy-
droxybenzoic acid (TBHBA). Assay conditions: diamine 10 mm, pyruvate
10 mm, TBHBA 4.5 mm, 4-AAP 1.5 mm, PLP 0.1 mgmLÀ1, HRP 0.1 mgmLÀ1, l-
AAO 4.5 UmLÀ1, pATA (purified enzyme) 5 mgmLÀ1
.
a highly hydrophobic channel to the active-site cavity relative
to that of other class III transaminases.[22]
l-AAO experiment, the generation of l-alanine was confirmed
as the product of transamination, whereas the use of d-AAO
did not lead to the production of d-alanine. We subsequently
used this method to screen all three putrescine transaminases,
Ec-ygjG, BMy-ygjG, and BM-ygjG, against a broad range of dia-
mines and polyamines to assess substrate scope (see Table 1
and Table S2 in the Supporting Information).
Interestingly, putrescine transaminase was found to exhibit
higher selectivity for diamines than for monoamines of compa-
rable size; for example, putrescine (1c) showed high activity
with Ec-ygjG pATA, whereas corresponding monoamine butyl
amine 6 showed no detectible activity. Ornithine 5, mono N-
tert-butoxycarbonyl (Boc)-protected derivatives of 1a and 1b
(amines 10 and 11, respectively), and diamines 7a, 7b, 8, and
9 were also found to be inactive substrates, which suggests
that the ygjG enzyme is strictly involved in the degradation of
linear diamines (see Table S2).
In the diaminoalkane series 1a–I, the length of the alkyl
chain was found to be an important parameter in determining
the relative activities of the substrates (Table 1). No activity was
observed with diamines possessing alkyl chains with two or
three C atoms, that is, 1a and 1b, possibly as a result of the
in situ generation of a bridged bicyclic complex with PLP,
which would thereby effectively inhibit the catalytic role of the
cofactor.[21] The highest activity was observed with putrescine
(1c, set to 100%), and thereafter, the activity subsequently de-
creased in a stepwise manner with increasing chain length
from 1d to 1 f. Diamines containing greater than eight C
atoms, that is, 1g–i, had correspondingly lower measureable
activity. We also examined diamines 2a–c, 3a, and 3b contain-
ing heteroatoms in addition to branched diamine 4. All of
these diamines were shown to be active, although the lower
activities of the bulky substrates can possibly be rationalized
by the presence of a small substrate binding-site entrance with
A recent report suggested that the N2 atom of 1c is stabi-
lized by a hydrogen bond to a conserved Q119 residue (Q=
glutamine),[22] which is also found in a range of GABA/a-keto-
glutarate aminotransferases with low putrescine amino donor
specificity.[23] The catalytic role of Q119 was therefore examined
by mutating this residue to alanine, which resulted in a 60%
loss of activity with 1c as the substrate (Figure 1). This reduced
activity is consistent with the role of Q119 in stabilizing the ori-
entation of putrescine bound as the PLP–substrate imine, al-
though it is clearly not crucial for catalysis. Furthermore, for
substrates with increasing chain lengths from C5 to C8, the
effect of the mutation on activity was less pronounced.
Further insight into the factors affecting substrate recogni-
tion was gained by mutation of other active-site residues of
ChemCatChem 2016, 8, 1038 – 1042
www.chemcatchem.org
1039
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Communications
pletely the factors responsible for substrate binding and cataly-
sis (Figure 1b).
To develop preparative-scale reactions by using putrescine
transaminase, we initially focused on using 1d as a model sub-
strate. Transamination of 1d yielded 1-amino-5-pentanal,
which underwent cyclization to 1-piperideine followed by
rapid pH-dependent polymerization resulting in a mixture of
oligomers that was unable to revert back to the imine mono-
mer.[25,26] The imine is only stable at low pH values, which are
not compatible with the transamination reaction. An alterna-
tive approach was therefore pursued, involving in situ reduc-
tion of the imine intermediate to prevent polymerization. Sev-
eral chemical reducing agents were tested, including NH3BH3,
NaCNBH3, and Na(AcO)3BH with pyruvate and a-ketoglutarate
as the keto acceptors (see Table 2 and Table S3). In all cases,
GC analysis revealed clean monomeric product peaks. Both
Na(AcO)3BH and NaCNBH3 reduced the imines with high con-
versions with no loss of transaminase activity. Using Na(A-
cO)3BH (10 equiv.), the highest conversions were observed for
1c–e, whereas the presence of a heteroatom in similar sub-
strates 2a and 2b resulted in lower conversion values
(Table 2). Interestingly, substrate 3a gave bicyclic product 18.
Preparative-scale studies resulted in isolation of open-chain N-
substituted pyrroline 17 from acidic workup, which confirmed
that transamination occurred at the amine of the C4 terminus
as expected. Conjugate addition of the C3 terminal amine of
17 was pH dependent and generated fused bicyclic structure
18 in 60% yield under basic conditions.
As an alternative to chemical reduction of the imine, we ex-
amined the use of imine reductases. One-pot enzyme cascades
with 1c and 1d with the use of Ec-ygjG with (R)- or (S)-imine
reductase (IRED) from Streptomyces sp. GF3587[27] and Strepto-
myces sp. GF3546[28] were separately examined. Interestingly,
both (S)-IRED and (R)-IRED showed good to excellent conver-
sions (up to 99%), highlighting the potential for combining
these two enzymes in vivo for the conversion of polyamines
into N-heterocycles (see Table S3). Finally, to demonstrate the
utility of putrescine transaminase in synthesis, we developed
a biomimetic one-pot route to 2,3-dihydro-1H-indolizinium
motifs, notably the antibacterial and antifungal alkaloid ficu-
septine (Scheme 3).[29] Starting from 1c, we prepared 20a
(40% yield) and analogue 20b (50%), both of which were pre-
viously reported to require a five-step chemical synthesis[29] or
a similar biomimetic route under harsh acidic conditions.[30]
Figure 1. a) PyMol representation of the pATA enzyme (PDB: 4UOX), high-
lighting the important residues in the enzyme substrate cleft. b) The effect
of mutation to alanine of substrate-recognition residues on the specific ac-
tivities of the Ec-ygjG transaminase in the 1c–g series. Substrates were nor-
malized with respect to activity of the wild-type (WT) enzyme.
the Ec-ygjG transaminase. According to the published crystal
structure of this enzyme (PDB: 4UOX), the substrate-binding
residues are not conserved between Ec-ygjG and its structural
homologues,[22] unlike the PLP-binding residues. A combination
of bulky residues that protrude into the active-site cleft (i.e.,
F327, F91, and L419; F=phenylalanine, L=leucine) ensures
that the entrance to the active site is hydrophobic and narrow,
and this allows the enzyme to discriminate between diamines
1c and 5, which only differ by the presence of a carboxylate
group. These residues are unique to this transaminase, and it
has been suggested that they are responsible for substrate
binding[24] (Figure 1a).
A series of point mutations to alanine were per-
formed to examine the effect of each of these resi-
dues (i.e., F91A, F327A, L419A; A=alanine) on the
transamination of diaminoalkanes 1c–g. Interestingly,
mutation of F to A at position 91 resulted in a com-
plete loss of activity, and similarly, this effect was also
observed for L419A. Substitution of F with A at posi-
tion 327 resulted in an approximately 50% loss of ac-
tivity for shorter-chain diamines 1c–e; the decrease
was more pronounced for longer-chain diamines 1 f
and 1g. These results indicate that the F91 and L419
residues are important in catalysis; however, further
Scheme 3. Biomimetic formation of 2,3-dihydro-1H-indolizinium motifs by using two
mutagenesis studies are required to elucidate com- molecules of aldehyde 19 and in situ product 1-pyrroline (21) to generate 20a and 20b.
ChemCatChem 2016, 8, 1038 – 1042
www.chemcatchem.org
1040
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Communications
Table 2. Transamination of polyamines by using pyruvate/a-ketoglutarate as keto acceptors.
Substrate
Product
Conversion [%]
Ec-ygjG
BM-ygjG
BMy-ygjG
Acceptor A[a]
98
Acceptor B[b]
98
Acceptor A[a]
Acceptor B[b]
Acceptor A[a]
Acceptor B[b]
1c
97
97
95
95
97
95
95
12
13
1d
99
97
99
93
97
97
1e
97
95
95
14
15
16
2a
2b
15
60
15
53
10
72
14
72
15
65
12
64
3a[c]
51
72
56
58
39
25
18
[a] Keto acceptor: pyruvate. [b] Keto acceptor: a-ketoglutarate. Reaction conditions: 400 mL reaction, 5 mm diamine; 10 mm keto acceptor, buffer: 50 mm
TRIS, 300 mm NaCl, 1 mm PLP; 10 equiv. Na(AcO)3BH; cell-free lysate. [c] Reducing reagent was not added. Conversions were measured by GC-FID by com-
parison with commercially available standards by using an empty vector control as baseline.
In summary, three putrescine transaminases, with broad spe-
Keywords: alkaloids · amino aldehydes · biocatalysis · diamine
cificity for terminal aliphatic diamines, were characterized, and
transaminase · enzymes
their potential for biocatalysis was subsequently investigated.
These transaminases used pyruvate and a-ketoglutarate as
keto acceptors and were combined with a chemical reducing
reagent to generate a range of different N-heterocycles in
good yields. Further engineering of these enzymes should lead
to new biocatalysts for the synthesis of precursors for pyrroli-
dine- and piperidine-derived alkaloids.
[1] T. J. Ritchie, S. J. F. Macdonald, R. J. Young, S. D. Pickett, Drug Discovery
Today 2011, 16, 164.
[2] J. P. Michael, Nat. Prod. Rep. 2008, 25, 139–165.
[3] K. Eller, E. Henkes, R. Rossbacher, U. Hçke, Ullmann’s Encyclopedia of In-
dustrial Chemistry, Wiley-VCH, Weinheim, 2012, pp. 647–698.
[4] C.-V. T. Vo, J. W. Bode, J. Org. Chem. 2014, 79, 2809–2815.
[5] D. Ghislieri, N. J. Turner, Top. Catal. 2014, 57, 284–300; H. Kohls, F. Stef-
fen-Munsberg, M. Hçhne, Curr. Opin. Chem. Biol. 2014, 19, 180–192.
[6] D. Ghislieri, A. P. Green, M. Pontini, S. C. Willies, I. Rowles, A. Frank, G.
Grogan, N. J. Turner, J. Am. Chem. Soc. 2013, 135, 10863–10869; N. J.
Turner, R. S. Heath, M. Pontini, B. Bechi, ChemCatChem 2014, 6, 996–
1002.
[7] F. Leipold, S. Hussain, S. P. France, N. J. Turner in Science of Synthesis:
Biocatalysis in Organic Synthesis 2 (Eds.: K. Faber, W.-D. Fessner, N. J.
Turner), Thieme Verlag, Stuttgart, 2015, pp. 359–382; D. Gamenara, P.
Domínguez de Maria, Org. Biomol. Chem. 2014, 12, 2989–2992; J.
Schrittwieser, S. Velikogne, W. Kroutil, Adv. Synth. Catal. 2015, 357,
1655–1685.
Acknowledgements
I.S. acknowledges a CASE award from Biotechnology and Biologi-
cal Sciences Research Council (BBSRC) and Dr. Reddy’s (Grant
agreement BB/K013076/1). J.L.G. acknowledges support of the BI-
OINTENSE project, financed through the European Union 7th
Framework Programme FP7/2007–2013 (Grant agreement no.
312148). N.J.W. was funded by the European Union’s 7th Frame-
work Programme FP7/2007–2013 (Grant agreement no. 289646-
KYROBIO). N.J.T thanks the Royal Society for a Wolfson Research
Merit Award.
[8] M. Fuchs, J. E. Farnberger, W. Kroutil, Eur. J. Org. Chem. 2015, 6965–
6982; R. C. Simon, N. Richter, E. Busto, W. Kroutil, ACS Catal. 2014, 4,
129–143; N. J. Turner, E. O’Reilly, C. Iglesias, D. Ghislieri, J. Hopwood,
ChemCatChem 2016, 8, 1038 – 1042
www.chemcatchem.org
1041
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Communications
J. L. Galman, R. C. Lloyd, Angew. Chem. Int. Ed. 2014, 53, 2447–2450;
Angew. Chem. 2014, 126, 2479–2482.
[21] M. H. O’Leary, Biochim. Biophys. Acta Enzymol. 1971, 242, 484–492.
[22] H. J. Cha, J.-H. Jeong, C. Rojviriya, Y.-G. Kim, PLoS One 2014, 9, e113212.
[23] Y. Kim, J. Biol. Chem. 1964, 239, 783–786.
[24] F. Steffen-Munsberg, C. Vickers, H. Kohls, H. Land, H. Mallin, A. Nobili, L.
Skalden, T. van den Bergh, H.-J. Joosten, P. Berglund, M. Hçhne, U. T.
Bornscheuer, Biotechnol. Adv. 2015, 33, 566–604.
[25] C. Schçpf, F. Braun, A. Komzak, Chem. Ber. 1956, 89, 1821–2833.
[26] E. Poupon, E. Gravel, Chem. Eur. J. 2015, 21, 10604–10615.
[27] K. Mitsukura, M. Suzuki, S. Shinoda, T. Kuramoto, T. Yoshida, T. Nagasa-
wa, Biosci. Biotechnol. Biochem. 2011, 75, 1778–1782; S. Hussain, F. Lei-
pold, H. Man, E. Wells, S. P. France, K. R. Mulholland, G. Grogan, N. J.
Turner, ChemCatChem 2015, 7, 579–583.
[9] C. W. Tabor, H. Tabor, Microbiol. Rev. 1985, 49, 81–99.
[10] J. Schneider, V. F. Wendisch, Appl. Microbiol. Biotechnol. 2011, 91, 17–30.
[11] Z. G. Qian, X. X. Xia, S. Y. Lee, Biotechnol. Bioeng. 2009, 104, 651–662.
[12] R. Voellym, T. Leisinger, J. Bacteriol. 1976, 128, 722–729.
[13] D. Schiroli, A. Peracchi, Biochim. Biophys. Acta Proteins Proteomics 2015,
1854, 1200–1211.
[14] U. Kaulmann, K. Smithies, M. E. B. Smith, H. C. Hailes, J. M. Ward, Enzyme
Microb. Technol. 2007, 41, 628–637.
[15] H. Yun, S. Lim, B.-K. Cho, B.-G. Kim, Appl. Environ. Microbiol. 2004, 70,
2529–2534.
[16] R. Hanson, B. L. Davis, Y. Chen, S. L. Goldberg, W. L. Parker, T. P. Tully, M.
Montana, R. N. Patel, Adv. Synth. Catal. 2008, 350, 1367–1375.
[17] D. Koszelewski, I. Lavandera, D. Clay, D. Rozzell, W. Kroutil, Adv. Synth.
Catal. 2008, 350, 2761–2766.
[28] K. Mitsukura, T. Kuramoto, T. Yoshida, N. Kimoto, H. Yamamoto, T. Naga-
sawa, Appl. Microbiol. Biotechnol. 2013, 97, 8079–8086; F. Leipold, S.
Hussain, D. Ghislieri, N. J. Turner, ChemCatChem 2013, 5, 3505–3508.
[29] F. Bracher, J. Daab, Eur. J. Org. Chem. 2002, 2288–2291.
[18] J. Hopwood, M. D. Truppo, N. J. Turner, R. C. Lloyd, Chem. Commun.
2011, 47, 773–775.
[30] B. B. Snider, B. J. Neubert, Org. Lett. 2005, 7, 2715–2718.
[19] N. N. Samsonova, S. V. Smirnov, A. E. Novikova, L. R. Ptitsyn, FEBS Lett.
2005, 579, 4107–4112.
Received: January 22, 2016
[20] N. N. Samsonova, S. V. Smirnov, I. B. Altman, L. R. Ptitsyn, BMC Microbiol.
2003, 3, 2.
Published online on February 18, 2016
ChemCatChem 2016, 8, 1038 – 1042
www.chemcatchem.org
1042
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim