Efficient Enzymatic Preparation of13N-Labelled Amino Acids: Towards Multipurpose Synthetic Systems

Article abstract of DOI:10.1002/chem.201602471

Nitrogen-13 can be efficiently produced in biomedical cyclotrons in different chemical forms, and its stable isotopes are present in the majority of biologically active molecules. Hence, it may constitute a convenient alternative to Fluorine-18 and Carbon-11 for the preparation of positron-emitter-labelled radiotracers; however, its short half-life demands for the development of simple, fast, and efficient synthetic processes. Herein, we report the one-pot, enzymatic and non-carrier-added synthesis of the¹³N-labelled amino acids l-[¹³N]alanine, [¹³N]glycine, and l-[¹³N]serine by using l-alanine dehydrogenase from Bacillus subtilis, an enzyme that catalyses the reductive amination of α-keto acids by using nicotinamide adenine dinucleotide (NADH) as the redox cofactor and ammonia as the amine source. The integration of both l-alanine dehydrogenase and formate dehydrogenase from Candida boidinii in the same reaction vessel to facilitate the in situ regeneration of NADH during the radiochemical synthesis of the amino acids allowed a 50-fold decrease in the concentration of the cofactor without compromising reaction yields. After optimization of the experimental conditions, radiochemical yields were sufficient to carry out in vivo imaging studies in small rodents.

Full text of DOI:10.1002/chem.201602471

DOI: 10.1002/chem.201602471
Full Paper
&
Radiochemistry
13
Efficient Enzymatic Preparation of N-Labelled Amino Acids:
Towards Multipurpose Synthetic Systems
Eunice S. da Silva,[a] Vanessa Gꢀmez-Vallejo,[b] ZuriÇe Baz,[a] Jordi Llop,*[a] and
Fernando Lꢀpez-Gallego*[c, d]
Abstract: Nitrogen-13 can be efficiently produced in bio-
medical cyclotrons in different chemical forms, and its stable
isotopes are present in the majority of biologically active
molecules. Hence, it may constitute a convenient alternative
to Fluorine-18 and Carbon-11 for the preparation of posi-
tron-emitter-labelled radiotracers; however, its short half-life
demands for the development of simple, fast, and efficient
synthetic processes. Herein, we report the one-pot, enzymat-
ic and non-carrier-added synthesis of the 13N-labelled amino
acids l-[13N]alanine, [13N]glycine, and l-[13N]serine by using l-
alanine dehydrogenase from Bacillus subtilis, an enzyme that
catalyses the reductive amination of a-keto acids by using
nicotinamide adenine dinucleotide (NADH) as the redox co-
factor and ammonia as the amine source. The integration of
both l-alanine dehydrogenase and formate dehydrogenase
from Candida boidinii in the same reaction vessel to facilitate
the in situ regeneration of NADH during the radiochemical
synthesis of the amino acids allowed a 50-fold decrease in
the concentration of the cofactor without compromising re-
action yields. After optimization of the experimental condi-
tions, radiochemical yields were sufficient to carry out in
vivo imaging studies in small rodents.
In contrast, the use of Nitrogen-13 (13N, T1/2 =9.97 min) has
been historically more restricted, and only a few studies that
describe new synthetic strategies for the incorporation of this
radioisotope into bioactive molecules have been reported to
date.[4] Interestingly, 13N can be efficiently produced in biomed-
ical cyclotrons in different chemical forms, and its stable iso-
topes (14N and 15N) are present in the majority of biologically
active molecules. Hence, the inclusion of 13N in the toolbox of
PET chemists might become a valuable alternative to 11C and
18F, either for the preparation of new labelled compounds or
the incorporation of the label in different positions. The prepa-
ration of 13N-labelled amino acids for the investigation of
tumour-related metabolic alterations is particularly relevant,[5]
because the synthesis of natural amino acids that use other
positron emitters is usually challenging.
Introduction
Positron Emission Tomography (PET) is a molecular imaging
technique that relies on the administration of trace amounts of
compounds labelled with positron emitters that enable exter-
nal, non-invasive detection with unparalleled sensitivity. In
recent years, PET has proved valuable in the diagnosis and
prognosis of different diseases,[1] the determination of pharma-
cokinetic properties of new chemical entities,[2] and the investi-
gation of the underlying molecular mechanisms of pathologi-
cal processes.[3]
The application of PET requires radiolabelling of the bio-
marker or molecule under investigation with a positron emit-
ter. Fluorine-18 (18F) and Carbon-11 (11C) have been extensively
utilized because of their relatively long half-lives (T1/2
=
109.8 min and 20.4 min, respectively) and versatile chemistries.
The short half-life of 13N calls for the development of simple,
fast, and efficient synthetic processes. In this context, biocata-
lysis can offer attractive solutions because enzymes present an
exquisite selectivity and high turnover numbers, which enables
fast chemical conversions and yields highly pure products
under extremely mild conditions. The use of enzymes for the
preparation of 13N-labelled radiotracers was first described in
the 70s and applied to the preparation of amino acids. For ex-
ample, l-[13N]glutamate was synthesized by incubation of a-ke-
toglutarate (a-KG) with [13N]NH3 in the presence of the enzyme
l-glutamate dehydrogenase and NADH as the redox cofactor.[6]
Likewise, l-[13N]glutamine was synthesized by glutamine syn-
thetase by using glutamic acid as the substrate and adenosine
triphosphate (ATP) as the cofactor.[6b,7] These two amino acids
were synthesized with incorporation values close to 90% after
15 min of incubation.
[a] E. S. da Silva, Z. Baz, Dr. J. Llop
Radiochemistry and Nuclear Imaging Group, CIC biomaGUNE
Paseo Miramꢀn 182, 20009 San Sebastiꢁn, Guipfflzcoa, (Spain)
E-mail: jllop@cicbiomagune.es
[b] Dr. V. Gꢀmez-Vallejo
Radiochemistry Platform CIC biomaGUNE
Paseo Miramꢀn 182, 20009 San Sebastiꢁn, Guipfflzcoa (Spain)
[c] Dr. F. Lꢀpez-Gallego
Heterogeneous Biocatalysis GroupCIC biomaGUNE
Paseo Miramꢀn 182, 20009, San Sebastiꢁn, Guipfflzcoa, (Spain)
[d] Dr. F. Lꢀpez-Gallego
IKERBASQUE, Basque Foundation for Science
Maria Diaz de Haro 3, 48013 Bilbao, Bizkaia (Spain)
E-mail: flopez.ikerbasque@cicbiomagune.es
Supporting information for this article can be found under
http://dx.doi.org/10.1002/chem.201602471.
Chem. Eur. J. 2016, 22, 1 – 9
1
ꢁ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
&
These are not the final page numbers! ÞÞ

Full Paper
Immobilization of the enzymes on solid supports resulted in
a significant qualitative improvement by simplifying the purifi-
cation process. For example, immobilization of glutamate de-
hydrogenase enabled the preparation of l-[13N]valine and l-
[13N]leucine.[8] Through a more sophisticated process, the pro-
duction of l-[13N]alanine was achieved by immobilization of
the glutamine dehydrogenase and alanine transaminase on
silica beads.[9] In this tandem reaction, the reductive amination
of a-KG by using [13N]NH3 and NADH as the amine and hydride
sources, respectively, yielded l-[13N]glutamate; this was se-
quentially used as a co-substrate by the transaminase to trans-
fer the radiolabelled amine to the pyruvic acid, which resulted
in the formation of l-[13N]alanine. Yields close to 70% were ob-
tained in 4 min. Unfortunately, the investigation of enzyme-
mediated synthesis of 13N-labelled amino acids was unexpect-
edly discontinued, and no further works have been reported
so far.
Scheme 1. Schematic representation of: a) enzyme-mediated preparation of
l-[13N]alanine, [13N]glycine, and l-[13N]serine by using l-alanine dehydrogen-
ase (l-AlaDH) from Bacillus subtilis as the enzyme. b) One-pot, multi-enzyme
reaction for the preparation of l-[13N]alanine with regeneration of NADH.
With the widespread installation of biomedical cyclotrons
and small animal-dedicated cameras around the world, scien-
tists are increasingly identifying PET as an accessible and valu-
able tool for the investigation of physiological or biological
problems in the pre-clinical setting. This fact, together with the
significant advances in the investigation of enzymatic reactions
and the understanding of the precise mechanisms underlying
the catalytic reactions, may open new avenues for the prepara-
tion and utilization of radiolabelled amino acids in multiple
biomedical applications.
l-alanine, l-serine, and glycine are biosynthetically linked, and
together, they provide the essential precursors for the synthe-
sis of proteins, nucleic acids, and lipids that are crucial to
cancer cell growth.[5b,11] Hence, these three amino acids were
identified as the priority target compounds. Additionally, with
the aim of investigating the versatility of our approach, the
synthesis of two additional amino acids, l-phenylalanine and l-
norvaline, was also investigated.
Here, in continuation of our previous works related to the
application of enzyme-mediated reactions to 13N-radiochemis-
try,[10] we report the one-pot/one-step radiosynthesis of l-
[13N]alanine, [13N]glycine, and l-[13N]serine, which is catalysed
by a versatile l-alanine dehydrogenase (l-AlaDH) from Bacillus
subtilis (Scheme 1a). As demonstrated in this work, this
enzyme catalyses the reductive amination of different a-keto
acids by using NADH as the redox cofactor and cyclotron-pro-
duced, no-carrier-added [13N]NH3 as the amine source. Chroma-
tographic yields (conversion of [13N]NH3 into the corresponding
amino acid, measured from the chromatographic profile and
expressed in percentage) above 90% were obtained after ap-
propriate selection of the experimental conditions. Additional-
ly, simultaneous utilization of the aforementioned enzyme with
formate dehydrogenase (FDH) from Candida boidinii led to the
in situ regeneration of the cofactor (Scheme 1b), which en-
abled a 50-fold decrease in the concentration of the cofactor
without compromising the reaction rates. This synthetic strat-
egy resulted in ready-to-inject 13N-labelled amino acids in suffi-
cient amount to approach in vivo studies in small rodents, and
paves the way towards future solid-supported, multi-purpose,
flow synthetic processes.
Selection, expression, and purification of an active l-AlaDH
in E. coli
l-Alanine dehydrogenase (l-AlaDH) is an oxidoreductase (EC
1.4.1.1) that catalyses the reversible oxidative deamination of
l-alanine to pyruvic acid and the reductive amination of pyru-
vic acid. The kinetic properties of l-AlaDHs from different mi-
crobial sources have been extensively investigated, and some
of these enzymes have been cloned in heterologous hosts.[12]
Although the physiological role of l-AlaDH remains controver-
sial, it has been reported that this enzyme is involved in the
utilization of alanine as an energy source during the germina-
tion process of bacillus species.[13] However, inspection of dif-
ferent data bases (BRENDA, KEEGG, MetaCyc) evidences that l-
AlaDH is more versatile than anticipated from its physiological
role, and this suggests that l-AlaDH from Bacillus subtilis and
Bacillus sphaericus may enable the preparation of different 13N-
labelled amino acids because they present broad substrate
scope for the reductive amination of a-keto acids. Among l-
AlaDHs, the l-AlaDH from Bacillus subtilis was selected here
owing to its better kinetic properties for a wider variety of
amino acids. The enzyme was expressed in E. coli BL21 and pu-
rified by affinity chromatography based on the histidine tag in-
serted at its N-terminus. According to the literature,[12,14] the
pure l-AlaDH is a hexamer with a subunit mass of 44 KDa,
which results in a total mass of 264 KDa (see the Supporting
Information, Figure S1). The pure l-AlaDH showed a specific
enzyme activity of 106Æ5 UmgÀ1 towards pyruvate.
Results and Discussion
The ultimate goal of the work reported here was the develop-
ment of an efficient synthetic procedure for the preparation of
13N-labelled amino acids, which will be applied in future works
to the assessment of tumour-associated metabolic alterations
in a mouse model of prostate cancer. The natural amino acids
&
&
Chem. Eur. J. 2016, 22, 1 – 9
www.chemeurj.org
2
ꢁ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÝÝ These are not the final page numbers!

Full Paper
Synthesis of amino acids under non-radioactive conditions
The amination activity towards a-keto acids with different side
chains was first tested in non-radioactive conditions. With that
aim, the synthesis of l-alanine, glycine, l-serine, l-phenylala-
nine, and l-norvaline was achieved by incubation of the appro-
priate substrates with ammonium chloride in the presence of
NADH as the cofactor and l-AlaDH from Bacillus subtilis
(Table 1).
Table 1. Specific enzyme activity (SEA) values for amination and deami-
nation reactions under non-radioactive conditions.[a]
Amino acid
SEAdeamination
SEAamination
Figure 1. Activity–pH profile of l-AlaDH for the production of l-alanine, gly-
cine, l-serine, l-phenylalanine, and l-norvaline. The reactions were carried
out in the presence of NADH (0.5 mm), ammonium chloride (450 mm), and
the corresponding a-keto acid (10 mm) in buffer solution (300 mm) at 258C.
Acetate buffer was used for pH 5–6, sodium phosphate buffer for pH 7–8,
and sodium bicarbonate buffer for pH 9–10. Enzyme activity measurements
were carried out for 2 min, and blank experiments (without AlaDH) were
performed to discard any stability issues of NADH that may lead to underes-
timation of the enzyme activity. Data are expressed as mean Æstandard de-
viation (n=3); (*) denotes the optimal pH value for the synthesis of the cor-
responding amino acid.
[mmolminÀ1 mgÀ1
]
[mmolminÀ1 mgÀ1
]
alanine
glycine
serine
5.4Æ0.1 (l)
N.D.
0.039Æ0.001 (l)
0.003Æ0.001 (d)
0.002Æ0.001 (l)
N.D.
106Æ5
34.3Æ1.6
22.5Æ3.1
norvaline
phenylalanine
18.6Æ1.2
21.4Æ2.1
[a] Carried out at pH 8. N.D.=not detected.
Synthesis of l-[13N]amino acids
As expected, the highest activity was observed for pyruvic
acid, but the enzyme also catalysed the reductive amination of
other a-keto acids, such as glycolic acid, 3-hydroxypyruvic
acid, 2-oxopentanoic acid, and 2-phenylpyruvic acid, to yield
glycine, serine, norvaline, and phenylalanine, respectively. The
specific enzyme activity towards these a-keto acids ranged
from 17 to 32% of the maximum specific enzyme activity that
was measured with pyruvic acid (Table 1). In general terms,
bulky a-keto acids were worse substrates than small ones, as
previously observed in the vast majority of the studied l-
AlaDHs.[15]
Translation of the reaction parameters to radioactive condi-
tions is often not straightforward. If a carrier-free radioactive
precursor is used (in our case [13N]NH3), the concentration of
this species is extremely low (around 15 mm in typical experi-
ments conducted in our set up). Hence, the kinetics of the re-
action might be significantly altered, and careful refinement of
the experimental conditions is usually required. In this work,
we first attempted the synthesis of l-[13N]alanine: Pyruvic acid
(75 mm) was incubated with cyclotron-produced [13N]NH3
(aqueous solution, 20–25 MBq, 100 mL) and l-AlaDH in the
presence of NADH as the cofactor (0.5 mm) in buffered solu-
tion at pH 8.0Æ0.1. The reactions were carried out in tubes
with 50 KDa cut-off membranes to enable easy separation of
the enzymes from the final products by fast centrifugation.
Chromatographic yield was monitored over time (Figure 2,
solid line; see the Supporting Information, Figure S2A for ex-
ample of chromatographic profiles), and values were then cor-
rected by radioactive decay to the reaction starting time
(Figure 2, dotted line).
Despite the low concentration of [13N]NH3, chromatographic
yields of 68Æ5% were obtained after 5 min under the reaction
conditions; this value increased to 95Æ1% when the reaction
was carried out for 20 min. Taking into account the short half-
life of 13N, incubation times of 10, 15, and 20 min proved to be
less desirable than the 5 min period, because the gain in yield
was negligible compared to the decay of Nitrogen-13
(Figure 2, dotted line). Notably, when the reaction time was set
to 20 min, the amount of radiochemical impurities (mainly un-
reacted [13N]NH3) was <5%, which enabled putative in vivo
studies without the need for a purification step to remove un-
reacted [13N]NH3.
As aforementioned, l-alanine dehydrogenase from Bacillus
subtilis presented a broad substrate scope for the reductive
amination towards different a-keto acids. In contrast, such sub-
strate versatility was not observed in the oxidative deami-
nation of l-amino acids (Table 1). The enzyme was inactive to-
wards glycine and phenylalanine and poorly active towards
serine and norvaline. These results are in good agreement with
those previously reported both for the recombinant and native
enzymes purified from E. coli and B. subtilis, respectively.[12,14]
The effect of the pH on the reductive amination activity of
l-AlaDH towards different a-keto acids was subsequently in-
vestigated (Figure 1). Optimal specific enzyme activity values
for the production of all tested amino acids were observed at
pH 8, except for l-serine, which were achieved at pH 9. Nota-
bly, the specific enzyme activity of l-AlaDH towards 3-hydroxy-
pyruvic acid at pH 9 was only 1.24 times lower than the activity
towards pyruvic acid at pH 8.
The activity–pH profile of l-AlaDH for the reductive ami-
nation of pyruvic acid was wider than for other amino acid de-
hydrogenases.[16]
Chem. Eur. J. 2016, 22, 1 – 9
www.chemeurj.org
3
ꢁ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
&
These are not the final page numbers! ÞÞ

Full Paper
values increased to >95% when the enzyme concentration
was raised to 180 and 300 mgmLÀ1 for the synthesis of
[13N]glycine and l-[13N]serine, respectively (Table 2). The radio-
chemical synthesis of glycine and serine required higher
enzyme concentrations because the specific enzyme activity of
l-AlaDH was significantly lower towards glycolic acid and 3-hy-
droxypyruvic acid than towards pyruvic acid (Table 1). These re-
sults are in good agreement with our experiments under non-
radioactive conditions, and they confirm that l-AlaDH from Ba-
cillus subtilis is able to catalyse the reductive amination of gly-
colic acid and 3-hydroxypyruvic acid, although reaction rates
were lower than those observed for pyruvic acid. Interestingly,
the synthesis of l-[13N]serine at pH 8 resulted in chromato-
graphic yields of 7% (Table 2). The results corroborate that the
synthesis of l-serine is both kinetically and thermodynamically
favoured under alkaline conditions, as observed under non-ra-
dioactive conditions.
Figure 2. Chromatographic yields (CY, solid line) and decay-corrected CY
(dotted line) at different reaction times for the production of l-[13N]alanine.
The reactions were carried out in the presence of [13N]NH4OH (20–25 MBq),
l-AlaDH (15 mgmLÀ1), NADH (0.5 mm), and pyruvic acid (75 mm) in sodium
phosphate buffer solution (300 mm) at T=258C.
The preparation of l-[13N]phenylalanine and l-[13N]norvaline
led to unexpected results. Although the enzyme was active to-
wards the corresponding a-keto acids under non-radioactive
conditions, no conversion was observed under radioactive con-
ditions, regardless of the reaction time. The absence of forma-
tion of l-[13N]phenylalanine and l-[13N]norvaline and the rela-
tively low reaction rates obtained for [13N]glycine and l-
[13N]serine could be initially attributed to an equilibrium be-
tween amination and deamination reactions. However, the low
specific enzyme activity that was observed for the deamination
reactions towards the different amino acids (Table 1) discarded
this hypothesis and confirmed that the reductive amination is
thermodynamically favoured at pH 8, as previously re-
ported.[15c]
To gain further knowledge on the effect of the amount of
biocatalyst in the reaction kinetics, we performed the radio-
chemical synthesis of l-[13N]alanine at different enzyme con-
centrations (see the Supporting Information, Figure S3). As ex-
pected, for a reaction time of 5 min, higher enzyme concentra-
tions led to higher yields. Values of 68Æ5, 78Æ4, 85Æ9, and
98Æ1% were achieved when concentrations of 15, 30, 45, and
60 mgmLÀ1 of the enzyme were used, respectively.
One of the major goals of the current work was to develop
a versatile synthetic process that was suitable for the efficient
production of different amino acids. In addition to l-
[13N]alanine, the preparation of [13N]glycine and l-[13N]serine
was a priority for future use in animal studies. Hence, and to
investigate the versatility of our approach, the synthesis of
these two amino acids as well as l-[13N]phenylalanine and l-
[13N]norvaline were also attempted by using the same experi-
mental setup that was successfully employed for l-[13N]alanine.
The reactions were carried out for 5 min (optimal reaction
time, Figure 2) by using 60 mgmLÀ1 of l-AlaDH and a saturating
concentration of the corresponding a-keto acid at the optimal
pH values (Figure 1). Chromatographic yields were calculated
from chromatographic profiles (Table 2; see the Supporting In-
formation, Figure S2).
To further investigate the kinetics of the enzymatic reactions,
steady-state kinetic parameters for ammonia in the reductive
amination of the different a-keto acids were determined under
non-radioactive conditions (Table 3).
Table 3. Kinetic parameters (Michaelis–Menten model) for ammonia, for
the different amino acids.
Amino acid[a]
Vmax [UmgÀ1
]
KM [mm]
alanine
glycine
serine
norvaline
phenylalanine
100.46Æ4.30
51.96Æ6.05
74.63Æ1.26
114.29Æ9.86
N.D.
47.0Æ11.4[a]
527Æ149[a]
150Æ21[b] 2810Æ613[a]
2231Æ544
Chromatographic yields of 49Æ3 and 42Æ5% were ob-
tained for [13N]glycine and l-[13N]serine, respectively. These
N.D.
[a] pH 8. [b] pH 9. N.D.=not detected.
Table 2. Chromatographic yields (CY) for the synthesis of different 13N-la-
belled amino acids.
All Vmax values fell in the range 50–120 UmgÀ1; however, KM
values for ammonia varied significantly from one amino acid
to the next. KM values for the synthesis of l-alanine and glycine
at pH 8 were 47.0Æ11.4 and 527Æ149, respectively, while the
KM value for l-serine at pH 9 was 150Æ21. However, the
KM(ammonia) value for the synthesis of l-norvaline was 2231Æ
544 mm, and the same parameter could not be determined for
the synthesis of l-phenylalanine because the substrate satura-
tion point was never reached, regardless of the concentration
Amino acid
pH
CY [%]
alanine
glycine
serine
8
8
8
9
8
8
98Æ1[a]
49Æ3[a]/98Æ1[b]
7[a]
42Æ5[a]/99Æ1[c]
norvaline
phenylalanine
N.D.[a]
N.D.[a]
[a] 60 mgmLÀ1 of protein. [b] 180 mgmLÀ1 of protein. [c] 300 mgmLÀ1 of
protein. N.D.=not detected.
&
&
Chem. Eur. J. 2016, 22, 1 – 9
www.chemeurj.org
4
ꢁ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÝÝ These are not the final page numbers!

Full Paper
of ammonia. These values suggest that the activity of the l-
AlaDH is very sensitive to the concentration of ammonia, and
such sensitivity strongly depends on the nature of the a-keto
acid.[17] The KM(ammonia) values explain that high radiochemical
yields are difficult to achieve with the short reaction times that
are employed for [13N]glycine and l-[13N]serine. In the case of
l-norvaline and l-phenylalanine, a high concentration of am-
monia is required for the reactions to occur at reasonable
rates; consequently, no formation of the amino acid can be ob-
served under radioactive (no-carrier-added) conditions, which
is due to the low concentration of [13N]NH3 in the reaction
media. A similar effect was observed for the synthesis of l-
serine at pH 8, which resulted in a KM(ammonia) as high as the one
found for l-norvaline, which supports the fact that, unlike l-
[13N]alanine, l-[13N]serine was poorly synthesized at pH 8. The
different KM(ammonia) values of l-AlaDH by using different a-keto
acids as amine acceptors suggest that the affinity of l-AlaDH
for ammonia depends on the positioning of the a-keto acid
into the active site. This fact is crucial for the success of the ra-
diochemical syntheses, for which extremely low concentrations
of ammonia are used. In the light of these results, we can con-
clude that the enzymes or the substrates that facilitate the
binding of the labelling agent to the active site (low KM values)
will lead to the best results in radiochemistry.
Figure 3. Chromatographic yields for the production of l-[13N]alanine at dif-
ferent concentrations of NADH (black bars). Grey bars correspond to values
obtained under identical experimental conditions in the absence of FDH.
Values are expressed as mean Æstandard deviation (n=3). The reactions
were carried out in the presence of [13N]NH3 (20–25 MBq), l-AlaDH
(60 mgmLÀ1), NADH (0.01–0.5 mm), and pyruvic acid (75 mm) in sodium
phosphate buffer (300 mm) at T=258C and pH 8.
In vivo studies
To prove the suitability of the labelling strategy for in vivo in-
vestigations, we carried out PET imaging with l-[13N]alanine in
healthy mice. Computerized Tomography (CT) images were se-
quentially acquired for proper localization of the radioactive
signal. Starting from 220Æ20 MBq of cyclotron-produced
[13N]NH4OH (volume=100 mL), we obtained 85Æ6 MBq of
pure l-[13N]alanine by using the bi-enzymatic approach in an
overall reaction time of 9 min (decay-corrected RCY=73%;
non-corrected RCY=39%). Specific activity values were in the
range 15–35 GBqmmolÀ1. Acquisitions were started concomi-
tantly with the intravenous administration of the labelled
amino acid (100 mL, 16.5Æ3.1), and images were acquired in
dynamic mode for 50 min.
Overall, the above one-pot methodology is clearly superior
to previously published enzymatic methods,[18] and it enables
the preparation of l-[13N]alanine, [13N]glycine, and l-[13N]serine
with chromatographic yields of >95% in short reaction times.
Synthesis of l-[13N]alanine with in situ NADH regeneration
With the aim of simplifying the purification process, we envi-
sioned the possibility of using lower concentrations of the co-
factor by introducing a second enzyme into the reaction
media that is capable of regenerating NADH in situ. The select-
ed enzyme was formate dehydrogenase (FDH) from Candida
boidinii, which is known to catalyse the oxidation of formate to
carbon dioxide by donating the electrons to a second sub-
strate, such as NAD+, and regenerating NADH (Scheme 1b).
The one-pot, bi-enzymatic reaction delivered exceptional re-
sults. The concentration of NADH in the starting solution could
be decreased by 50-fold without a significant impact in chro-
matographic yields (Figure 3). The simultaneous addition of l-
AlaDH and FDH led to the formation of l-[13N]alanine with
chromatographic yields of 95Æ2% by using NADH concentra-
tions as low as 10 mm. These values are three times higher
than those observed without FDH under identical experimental
conditions. The FDH is essential for replenishing the pool of
NADH by consumption of the formate as a sacrificial substrate;
in this manner the reduced form remains available for the l-
AlaDH to catalyse the reductive amination of pyruvate. These
results substantially improve the synthetic scheme described
by Baumgartner et al.;[18] with our new configuration, similar
yields can be obtained but the concentration of cofactor is
100-fold lower.
As it can be seen in Figure 4, the presence of radioactivity is
detected in the heart at short times after administration, which
confirms the presence of the labelled amino acid in the blood.
At longer time points, a progressive accumulation of radioac-
tivity was observed in the liver, which is consistent with the
anabolic function of this organ. The highest accumulation of
radioactivity was found in the abdominal region, especially in
the intestine; this result is consistent with biodistribution data
that was previously reported for 11C-labelled natural amino
acids.[19] Low accumulation of radioactivity was found in the
brain, which may facilitate the potential localization of tumours
in this organ. Importantly, elimination by urine was negligible
in the time window of the study. This is an essential finding
considering the intended future application of this labelled
amino acid as a potential marker for the investigation of meta-
bolic processes in animal models of prostate cancer.
Conclusions
We have presented the fast, efficient, one-pot enzymatic syn-
thesis of l-[13N]alanine, [13N]glycine, and l-[13N]serine by using
Chem. Eur. J. 2016, 22, 1 – 9
www.chemeurj.org
5
ꢁ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
&
These are not the final page numbers! ÞÞ

Full Paper
Bio-Rad Laboratories (Madrid, Spain). All others reagents were of
analytical grade.
Production of l-AlaDH and FDH in E. coli
Expression
An overnight culture of E. coli BL21 (1 mL), transformed with the
respective plasmid (pET28b-BsL-AlaDH or pET28b-CbFDH), was ino-
culated in a Luria–Bertani (LB) broth (50 mL) that contained kana-
mycin (final concentration of 30 mgmLÀ1). For pET28b-BsL-AlaDH,
the resulting culture was incubated at 378C with energetic shaking
until the OD600nm reached 0.6. 1-Thio-b-d-galactorpyranoside (IPTG)
was added to the culture (final concentration of 1 mm) to induce
protein synthesis. Cells were grown at 378C for 3 h and harvested
by centrifugation at 1290 g over 30 min. For pET28b-CbFDH, the
resulting culture was incubated at 218C with energetic shaking
until the OD600nm reached 0.6. 1-Thio-b-d-galactorpyranoside (IPTG)
was added to the culture (final concentration of 1 mm) to induce
protein synthesis. Cells were grown at 218C for 18 h and harvested
by centrifugation at 1290 g over 30 min.
Figure 4. Representative PET-CT images obtained after intravenous adminis-
tration of l-[13N]alanine in mice. Coronal projections of PET images have
been co-registered with representative CT slices for co-localization of the ra-
dioactive signal. Images correspond to data averaged in the following time
frames: a) 0–30 s, b) 31–70 s, c) 71–230 s, d) 231–740 s, and e) 741–3020 s.
l-AlaDH and NADH as the cofactor. The synthesis of l-
[13N]alanine was significantly more efficient than the synthesis
of [13N]glycine and l-[13N]serine; however, yields obtained for
[13N]glycine and l-[13N]serine should be sufficient to approach
subsequent preclinical investigations in rodents. The prepara-
tion of l-norvaline and l-phenylalanine did not work under no-
carrier-added conditions owing to the high KM values for am-
monia. In the case of l-[13N]alanine, the addition of FDH to the
reaction media to regenerate NADH enabled a 50-fold de-
crease in the required concentration of NADH without compro-
mising the radiochemical yields. The resulting labelled amino
acid could be obtained in sufficient amount to carry out bio-
distribution studies in healthy mice by using PET-CT. This re-
ported one-pot, multi-enzyme reaction opens new avenues for
the enzymatic preparation of 13N-labelled amino acids. Future
efforts will pursue the co-immobilization of both enzymes on
the same carrier to enable continuous flow reactions, the engi-
neering of l-AlaDH to broaden its substrate scope, and the ap-
plication of the here-reported labelled amino acids to the in-
vestigation of metabolic alterations in a mouse model of pros-
tate cancer.
Purification
Both FDH and l-AlaDH were purified as follows: The resulting
pellet was resuspended in one-tenth of its original volume of
sodium phosphate buffer solution (25 mm, pH 7, binding buffer).
Cells were broken by sonication (LABSONIC P, Sartorius Stedim bio-
tech) at amplitude=1 and cycle=1 over 5 min. The suspension
was centrifuged at 10528 g over 30 min. The supernatant that con-
tained the enzyme was collected and passed through a TALON
resin that was equilibrated with a binding buffer. Following the
protein binding to the column, the column was washed three
times with binding buffer prior to the protein elution with elution
buffer (binding buffer supplemented with 300 mm imidazole). The
eluted protein was gel-filtered by using PD-10 columns (GE health-
care) to remove the imidazole and exchange the enzyme buffer.
SDS-page and Bradford assay were carried out after each produc-
tion to determine the purity, concentration, and specific activity of
the enzymes. Gel filtration was performed by FLPC by using a Su-
perdex TM 200, 10/300 GL.
Protein and enzyme colorimetric assays
Experimental Section
Determination of the protein concentration
Reagents
The protein concentration of soluble l-AlaDH and FDH was carried
out by Bradford’s method by using bovine serum albumin (BSA) as
protein standard. The soluble enzyme (5 mL) and Bradford reagent
(200 mL) were mixed. The absorbance was measured at 595 nm
after 5 min of reaction at T=258C. For the standard curve, a series
of BSA solutions (concentration=0.06, 0.12, 0.25, 0.5, and
1 mgmLÀ1) was used.
The genes of l-alanine dehydrogenase (l-AlaDH) from Bacillus sub-
tilis (l-alanine: NAD oxidoreductase, EC1.4.1.1) and formate dehydro-
genase (FDH) from Candida boidinii (Formate: NAD+ oxidoreduc-
tase, EC1.2.1.2) were synthesized and cloned into pET28b by Gen-
Script company (Piscataway, USA). b-Nicotinamide adenine dinucle-
otide, reduced disodium salt hydrate (b-NADH; purity>97%), py-
ruvic acid (CH3COCOOH, purity ꢀ98%), ammonium chloride
(NH4Cl, for molecular biology, purity ꢀ99.5%), sodium phosphate
dibasic (Na2HPO4, for molecular biology, purity ꢀ98.5%), sodium
phosphate monobasic (NaH2PO4, purity ꢀ99.0%), l-alanine (purity
ꢀ98%), d-alanine (purity ꢀ98%), glycine (purity ꢀ99%), l-serine
(purity ꢀ99.5%), d-serine (purity ꢀ98%), l-phenylalanine (purity
ꢀ98%), d-phenylalanine (purity ꢀ98%), l-norvaline, and d-norva-
line (purity ꢀ99%) were purchased from Sigma Chemical Co. (St.
Louis, IL, USA) and used without further purification. TALON resin
was purchased from Clontech (Saint-Germain-en-Laye, France). Bio-
Rad Protein Assay Dye Reagent Concentrate was purchased from
Determination of enzyme activity
The enzymatic activities were spectrophotometrically measured in
96-well plates by monitoring the absorbance at 340 nm, which
varied depending on the concomitant production or consumption
of NADH.
L-AlaDH: Reductive amination: l-AlaDH (10 mL), NADH (0.5 mm,
190 mL), and pyruvic acid (75 mm, or other a-keto acids) were incu-
bated in ammonium chloride (450 mm) at pH 5–10 and T=258C.
Oxidative deamination: L-AlaDH (10 mL), NAD+ (1 mm, 200 mL), and
&
&
Chem. Eur. J. 2016, 22, 1 – 9
www.chemeurj.org
6
ꢁ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÝÝ These are not the final page numbers!

Full Paper
alanine (10 mml/d, or other amino acids) were incubated in
sodium phosphate buffer (300 mm) at pH 8 and T=258C.
column, 50ꢃ4.6 mm; Micron Phenomenex, Madrid, Spain) was
used as the stationary phase. As the mobile phase, 1 mm CuSO4 so-
lution (1 mLminÀ1) was used for alanine, glycine, and norvaline;
2 mm CuSO4 solution (0.2 mLminÀ1) was used for serine; and 2 mm
CuSO4/methanol solution (70:30, 1 mLminÀ1) was used for phenyl-
alanine. Pure commercial enantiomers were used as standards.
For the determination of the specific activity (l-[13N]alanine), the
concentration of amino acid in the purified fraction was deter-
mined by HPLC-MS by using an AQUITY UPLC separation module
coupled to an LCT TOF Premier XE mass spectrometer (Waters,
Manchester, UK). An Acquity UPLC Glycan BEH amide column
(1.7 mmꢃ50 mmꢃ2.1 mm) was used as the stationary phase. The
elution buffers were acetonitrile (A) and 0.1m ammonium formate
(B). The column was eluted with a linear gradient: t=0 min, 95%
B; t=3 min, 50% B; t=3.5 min, 95% B; t=5 min, 95% B. Total run
was 5 min, injection volume was 10 mL, and the flow rate was
500 mLminÀ1. The detection was carried out in positive ion mode,
monitoring the most abundant isotope peaks from the mass spec-
tra. Alanine was detected as a protonated molecule (m/z 89.9) with
retention time=1.81 min.
FDH: NAD+ (1 mm, 200 mL) and formic acid (100 mm) in sodium
phosphate buffer (25 mm, pH 7) were incubated with enzymatic
solution (5 mL) at T=258C. One unit of activity was defined as the
amount of enzyme that was needed to either reduce or oxidize
1 mmol of the corresponding nicotinamide cofactor at T=258C
and the corresponding pH.
Enzyme kinetic parameters
The kinetic parameters, Michaelis constant value (KM), and maxi-
mum rate (Vmax) were determined by activity colorimetric assay at
pH 8 (pH 9 for serine) and T=258C by following the method de-
scribed above for l-AlaDH, with 0–1.8m ammonium chloride as the
amine source.
Production of the radiolabelling agent [13N]NH3
Nitrogen-13 (13N) was produced in an IBA Cyclone 18/9 cyclotron
by using the 16O(p,a)13N nuclear reaction. The target system con-
sisted of an aluminium insert (2 mL) covered with havar foil (thick-
ness=25 mm, 1 29 mm) and with an aluminium vacuum foil
(thickness 25 mm, 1 23 mm). The target (containing 1.7 mL of
5 mm EtOH in H2O) was irradiated with 18 MeV protons. The beam
current was maintained at 22 mA (pressure in the range 5–10 bar
into the target during bombardment) to reach the desired inte-
grated currents (0.1–1 mAh). The resulting solution was transferred
to a 10 mL vial, and the activity was measured in a dose calibrator
(Capintec CRCꢂ-25 PET, New Jersey, USA).
Animal experimentation
General considerations
Animals were maintained in a temperature and humidity-con-
trolled room with a 12 h light–dark cycle. Food and water were
available ad libitum. Animals were cared for and handled in ac-
cordance with the Guidelines for Accommodation and Care of Ani-
mals (European Convention for the Protection of Vertebrate Ani-
mals Used for Experimental and Other Scientific Purposes), and
ethical protocols were approved by the internal Ethical Committee
and by regional authorities.
Synthesis of l-[13N]amino acids
General procedure
Radioactive enzymatic reactions of the amino acids were carried
out by adding 100 mL of a mixture that contained NADH (final con-
centration 0.5 mm), a-keto acid (final concentration 75 mm), and
sodium phosphate buffer solution (300 mm, pH 8) to a solution
that contained l-AlaDH (20 mL, concentration in the range 15–
300 mgmLÀ1) and [13N]NH3 (100 mL, 20–450 MBq). Reactions were
conducted in tubes with 50 KDa cut-off membranes. For the prepa-
ration of l-[13N]serine, sodium bicarbonate buffer (300 mm, pH 9)
was used. The mixture was incubated at 258C under mild stirring
for 2.5–20 min and filtered under centrifugation to remove the
enzyme.
In vivo studies
PET studies were performed by using an eXploreVista-CT small
animal PET-CT system (GE Healthcare). Mice (20–22 g, C57BL/6JRJ,
Janvier, France) were used (n=2). Mice were anesthetized with
a mixture of 3–4% isoflurane in O2 for induction, and this was re-
duced to 1–1.5% for maintenance. A nose cone was used to facili-
tate regular breathing, which was monitored by an SA Instrument
Inc. (NY, USA) pressure sensor. Respiration and body temperature
were monitored throughout the scan. The temperature, measured
rectally, was maintained at 37Æ18C by using a water heating blan-
ket (Homeothermic Blanket Control Unit, Bruker, Germany). For ad-
ministration of 13N-labelled amino acid, the tail vein was catheter-
ized with a C10SS-MTV1301 29-gauge catheter (Instech, Laborato-
ries Inc., Plymouth Meeting, PA, USA), and 16.5Æ3.1 MBq (approxi-
mately 100 mL) was injected in tandem with the start of a PET dy-
namic acquisition. The cannula was then rinsed with physiological
saline solution (50 mL), and dynamic images (33 frames: 8ꢃ5 s, 6ꢃ
15 s, 6ꢃ30 s, 6ꢃ60 s, 7ꢃ120 s) were acquired in 2 bed positions in
the 400–700 keV energy window, with a total acquisition time of
50 min 20 s. After each PET scan, CT acquisitions were also per-
formed, which provided anatomical information as well as the at-
tenuation map for later image reconstruction. Dynamic acquisitions
were reconstructed (decay, random and CT-based attenuation cor-
rected) with OSEM-2D (2 iterations, 16 subsets). PET images were
visually analysed by using PMOD image analysis software (PMOD
Technologies Ltd., Zꢄrich, Switzerland).
Enzymatic cascade reactions for regeneration of the
cofactor
The reactions were conducted as above, but FDH at 5:1 (FDH/l-
AlaDH) molar ratio and formic acid (100 mm) were added. 100 mL
of this soluble preparation was incubated with cyclotron-produced
[13N]NH3 solution (100 mL). All the reactions were carried out in
tubes with 50 KDa cut-off membranes to easily separate the en-
zymes from the final products by fast centrifugation.
HPLC and radio-HPLC analysis
The determination of all radioactive/non-radioactive compounds
was carried out by HPLC by using an Agilent 1100 Series system
equipped with a quaternary pump and diode array detector con-
nected in series with a radiometric detector (Gabi, Raytest, Strau-
benhardt, Germany). A chiral column (Phenomenex Chirexꢂ3126
Chem. Eur. J. 2016, 22, 1 – 9
www.chemeurj.org
7
ꢁ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
&
These are not the final page numbers! ÞÞ

Full Paper
[7] M. B. Cohen, L. Spolter, N. S. McDonald, B. Cassen, J. Nucl. Med. 422.
[8] J. R. Barrio, F. J. Baumgartner, E. Henze, M. S. Stauber, J. E. Egbert, N. S.
MacDonald, H. R. Schelbert, M. E. Phelps, F. T. Liu, J. Nucl. Med. 1983, 24,
937–944.
[9] M. B. Cohen, L. Spolter, C. C. Chang, N. S. MacDonald, J. Takahashi, D. D.
Bobinet, J. Nucl. Med. 1974, 15, 1192–1195.
[10] E. S. da Silva, V. Gomez-Vallejo, J. Llop, F. Lopez-Gallego, Catal. Sci. Tech-
nol. 2015, 5, 2705–2713.
[11] I. Amelio, F. Cutruzzola, A. Antonov, M. Agostini, G. Melino, Trends Bio-
chem. Sci. 2014, 39, 191–198.
Acknowledgements
All the work has been supported by the RADIOMI project (EU
FP7-PEOPLE-2012-ITN-RADIOMI). We would like to thank IKER-
BASQUE foundation for funding F.L.-G. and the support of
COST Action CM1303 Systems Biocatalysis. The authors would
like to thank Susana Velasco for her support in HPLC-MS.
[12] W. Ye, G. Huo, J. Chen, F. Liu, J. Yin, L. Yang, X. Ma, Microbiol. Res. 2010,
165, 268–275.
[13] a) K. J. Siranosian, K. Ireton, A. D. Grossman, J. Bacteriol. 1993, 175,
6789–6796; b) E. Freese, S. W. Park, M. Cashel, Proc. Natl. Acad. Sci. USA
1964, 51, 1164–1172.
[14] A. Yoshida, E. Freese, Biochim. Biophys. Acta 1965, 96, 248–262.
[15] a) P. J. Baker, Y. Sawa, H. Shibata, S. E. Sedelnikova, D. W. Rice, Nat.
Struct. Mol. Biol. 1998, 5, 561–567; b) P. Fernandes, H. Aldeborgh, L. Car-
lucci, L. Walsh, J. Wasserman, E. Zhou, S. T. Lefurgy, E. C. Mundorff, Pro-
tein Eng. Des. Sel. 2015, 28, 29–35; c) T. Ohashima, K. Soda, Eur. J. Bio-
chem. 1979, 100, 29–30.
Keywords: amino acids · biocatalysis · nitrogen-13 · positron
emission tomography · radiochemistry
[1] Positron Emission Tomography: Clinical Practice, Springers–Vera,
London, 2006.
[2] a) C. M. Gomes, A. J. Abrunhosa, P. Ramos, E. K. J. Pauwels, Adv. Drug De-
livery Rev. 2011, 63, 547–554; b) N. Gupta, P. M. Price, E. O. Aboagye,
Eur. J. Cancer 2000, 38, 2094–2107; c) E. Fernandes, Z. Barbosa, G.
Clemente, F. Alves, A. J. Abrunhosa, Curr. Radiopharm. 2012, 5, 90–98.
[3] K. A. Frey, in Basic Neurochemistry: Molecular, Cellular and Medical As-
pects (Ed.: B. W. A. G. J. Siegel, R. W. Albers, S. K. Fisher, M. D. Uhler), Lip-
pincott–Raven, Philadelphia, 1999.
[16] N. M. Brunhuber, J. S. Blanchard, Crit. Rev. Biochem. Mol. Biol. 1994, 29,
415–467.
[17] T. Ohshima, K. Soda, in Bioprocesses and Applied Enzymology, Springer,
Heidelberg, 1990, pp. 187–209.
[4] V. Gꢀmez-Vallejo, V. Gaja, K. B. Gona, J. Llop, J. Labelled Compd. Radio-
pharm. 2014, 57, 244–254.
[18] F. J. Baumgartner, J. R. Barrio, E. Henze, H. R. Schelbert, N. S. MacDonald,
M. E. Phelps, D. E. Kuhl, J. Med. Chem. 1981, 24, 764–766.
[19] P. L. Jager, W. Vaalburg, J. Pruim, E. G. de Vries, K. J. Langen, D. A. Piers,
J. Nucl. Med. 2001, 42, 432–445.
[5] a) M.-B. Tessem, M. G. Swanson, K. R. Keshari, M. J. Albers, D. Joun, Z. L.
Tabatabai, J. P. Simko, K. Shinohara, S. J. Nelson, D. B. Vigneron, I. S.
Gribbestad, J. Kurhanewicz, Magn. Reson. Med. 2008, 60, 510–516;
b) M. J. Albers, R. Bok, A. P. Chen, C. H. Cunningham, M. L. Zierhut, V. Y.
Zhang, S. J. Kohler, J. Tropp, R. E. Hurd, Y.-F. Yen, S. J. Nelson, D. B. Vi-
gneron, J. Kurhanewicz, Cancer Res. 2008, 68, 8607–8615.
[6] a) N. Lembares, R. Dinwoodie, I. Gloria, P. Harper, K. Lathrop, J. Nucl.
Med. 1972, 13, 786–786; b) M. G. Straatmann, M. J. Welch, Radiat. Res.
1973, 56, 48–56.
Received: May 24, 2016
Published online on && &&, 0000
&
&
Chem. Eur. J. 2016, 22, 1 – 9
www.chemeurj.org
8
ꢁ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÝÝ These are not the final page numbers!

Full Paper
FULL PAPER
13N-labelled radiotracers: l-Alanine de-
hydrogenase enables the fast and effi-
cient synthesis of l-[13N]alanine,
[13N]glycine, and l-[13N]serine under
mild conditions.
&
Radiochemistry
E. S. da Silva, V. Gꢀmez-Vallejo, Z. Baz,
J. Llop,* F. Lꢀpez-Gallego*
&& – &&
Efficient Enzymatic Preparation of 13N-
Labelled Amino Acids: Towards
Multipurpose Synthetic Systems
Chem. Eur. J. 2016, 22, 1 – 9
www.chemeurj.org
9
ꢁ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
&
These are not the final page numbers! ÞÞ