Class III Polyphosphate Kinase 2 Enzymes Catalyze the Pyrophosphorylation of Adenosine-5′-Monophosphate

Article abstract of DOI:10.1002/cbic.201900303

Polyphosphate kinase 2 (PPK2) transfer phosphate from inorganic polyphosphate to nucleotides. According to their activity, PPK2 enzymes are classified into three groups. Among them, class III enzymes catalyze both the phosphorylation of nucleotide mono- to diphosphates and di- to triphosphates by using polyphosphate, which is a very inexpensive substrate. Therefore, class III enzymes are very attractive for use in biotechnological applications. Despite several studies on class III enzymes, a detailed mechanism of how phosphate is transferred from the polyphosphate to the nucleotide remains to be elucidated. Herein, it is reported that PPK2 class III enzymes from two different bacterial species catalyze the phosphorylation of adenosine mono- (AMP) into triphosphate (ATP) not only through step-by-step phosphorylation, but also by pyrophosphorylation. These are the first PPK2 enzymes that have been shown to possess polyphosphate-dependent pyrophosphorylation activity.

Full text of DOI:10.1002/cbic.201900303

Accepted Article
Title: Class III polyphosphate kinase 2 enzymes catalyze the
pyrophosphorylation of adenosine-5'-monophosphate
Authors: Marin Ogawa, Atsuko Uyeda, Kazuo Harada, Yu Sato,
Yasuhiko Kato, Hajime Watanabe, Kohsuke Honda, and
Tomoaki Matsuura
This manuscript has been accepted after peer review and appears as an
Accepted Article online prior to editing, proofing, and formal publication
of the final Version of Record (VoR). This work is currently citable by
using the Digital Object Identifier (DOI) given below. The VoR will be
published online in Early View as soon as possible and may be different
to this Accepted Article as a result of editing. Readers should obtain
the VoR from the journal website shown below when it is published
to ensure accuracy of information. The authors are responsible for the
content of this Accepted Article.
To be cited as: ChemBioChem 10.1002/cbic.201900303
Link to VoR: http://dx.doi.org/10.1002/cbic.201900303
A Journal of
www.chembiochem.org

10.1002/cbic.201900303
ChemBioChem
Class
III
polyphosphate
kinase
2
enzymes
catalyze
the
pyrophosphorylation of adenosine-5'-monophosphate
Marin Ogawa1, Atsuko Uyeda1, Kazuo Harada2, Yu Sato1, Yasuhiko Kato1,
Hajime Watanabe1, Kohsuke Honda1, Tomoaki Matsuura1*
1Department of Biotechnology, Division of Advance Science and Biotechnology,
Graduate School of Engineering, Osaka University, 2-1 Yamadaoka, Suita,
Osaka 565-0871, Japan.
2Department of Applied Environmental Biology, Graduate School of
Pharmaceutical Sciences, Osaka University, 1-6 Yamadaoka, Suita, Osaka
565-0871
Corresponding Author
*Tomoaki Matsuura, Department of Biotechnology, Graduate School of
Engineering, Osaka University, 2-1 Yamadaoka, Suita, Osaka 565-0871, Japan.
Tel:
+81-6-6879-4172;
Fax:
+81-8-6879-7428;
E-mail:
matsuura_tomoaki@bio.eng.osaka-u.ac.jp
Short title: Pyrophosphorylation by PPK2 class III enzymes
1
This article is protected by copyright. All rights reserved.

10.1002/cbic.201900303
ChemBioChem
Abstract
Polyphosphate kinase 2 (PPK2) transfer phosphate from inorganic
polyphosphate to nucleotides. According to their activity, PPK2 enzymes are
classified into three groups. Among them, class III enzymes catalyze both the
phosphorylation of nucleotide mono- to di-phosphates and di- to tri-phosphates
using polyphosphate which is a very inexpensive substrate. Therefore, class III
enzymes are very attractive for use in biotechnological applications. Despite
several studies on class III enzymes, the detailed mechanism of how phosphate
is transferred from the polyphosphate to the nucleotide remains to be elucidated.
Here, we report that PPK2 class III enzymes from two different bacterial species
catalyze the phosphorylation of adenosine mono- (AMP) to triphosphate (ATP)
not only through step-by-step phosphorylation but also by pyrophosphorylation.
To the best of our knowledge, these are the first PPK2 enzymes that have been
shown to possess polyphosphate-dependent pyrophosphorylation activity.
2
This article is protected by copyright. All rights reserved.

10.1002/cbic.201900303
ChemBioChem
Introduction
Adenosine- and guanosine-5’-triphosphate (ATP and GTP, respectively) are
used in various biochemical reactions inside cells, including protein synthesis,
metabolism, signaling, and cell motility. In most cases, ATP is hydrolyzed into
adenosine-5’-diphosphate (ADP) or monophosphate (AMP), and the energy
obtained through the release of the phosphate is used for subsequent
biochemical reactions. Similarly, GTP is used as an energy source through
hydrolysis to guanosine-5’-diphosphate (GDP). The hydrolyzed products (i.e.,
AMP, ADP, and GDP) need to be regenerated into ATP and GTP for the living
system to sustain its activity. This regeneration is also important for
biotechnological applications using enzymes that require ATP and GTP as
substrates [1].
Polyphosphate kinase 2 (PPK2) is one of the enzymes involved in AMP, ADP,
and GDP regeneration [2]. PPK2 transfers phosphate from inorganic
polyphosphate to nucleotides. PPK2 enzymes are present in many bacteria
[3]
species and have been classified into three groups . Class I PPK2 enzymes
catalyze the phosphorylation of nucleotide diphosphates into triphosphates,
class II enzymes catalyze the phosphorylation of nucleotide monophosphates
into diphosphates, and class III enzymes catalyze both the phosphorylation of
nucleotide monophosphates into diphosphates and diphosphates into
triphosphates. Among these three classes, class III enzymes have attracted
much attention, since a single enzyme can catalyze the regeneration of AMP,
ADP, and GDP using polyphosphate, which is a very inexpensive compound,
and thus, PPK2 class III enzymes are very attractive for biotechnological
applications [3-4]
.
Since the first report about the class III enzymes by Motomura et al. in 2014 [3],
several crystal structures of class III enzymes in complex with substrates have
been identified [5]. The structural studies revealed the polyphosphate- and
nucleotide-binding sites and essential residues for interaction with these
substrates. In addition, the role of magnesium ion has been revealed. Kinetic
analysis has also been performed. Interestingly, phylogenic analysis showed
3
This article is protected by copyright. All rights reserved.

10.1002/cbic.201900303
ChemBioChem
that PPK2 class III enzymes were further divided into three clusters, and the
enzyme belonging to each cluster showed the phosphorylation of AMP, ADP,
GMP, and GDP but different substrate preferences among the clusters [5A]. While
these studies contributed significantly to understanding the features of PPK2
class III enzymes, our understanding is still limited. Only a single enzyme from
each of the three clusters has been characterized, and different substrate
preferences were observed among the clusters. Whether the observed substrate
preference is true for other enzymes remains unclear. Additionally, the detailed
mechanism underlying the transfer of phosphate from the polyphosphate to the
nucleotide remains unclear.
Here, we studied the properties of two PPK2 class III enzymes from Delftia
tsuruhatensis (DtCs) and Deinococcus radiodurans R1 (DrR1), both of which
belong to cluster 2 based on phylogenic analysis. The cluster number is based
on the definition described by Nocek et al. [5A]. We report that DtCs catalyzes the
phosphorylation of AMP to ATP via the transfer of inorganic phosphate from
polyphosphate as with other enzymes reported previously, but the mechanism of
phosphorylation was found to differ from what has been proposed. When
phosphorylating AMP to ATP, not only the step-by-step monophosphorylation
but also a transfer of pyrophosphate to AMP (pyrophosphorylation) resulting in
the generation of ATP was observed. This reaction was also observed with
another PPK2 class III enzyme, DrR1. In addition, DtCs showed similar
substrate preference as another previously characterized enzyme belonging to
the same cluster. Our results provide a further understanding of the activity of
the PPK2 class III enzymes and provide clues to the underlying mechanism of
the synthesis of adenosine polyphosphate observed in a recent study [6].
Results
PPK2 class III enzyme from Delftia tsuruhatensis (DtCs)
We first selected the following two different PPK2 class III enzymes: one from D.
radiodurans R1 (gi|499189240) and one from D. tsuruhatensis (gi|503566024),
both of which are mesophilic bacteria. The mesophilic enzymes were chosen for
their suitability for future applications in nucleotide mono- and diphosphate
4
This article is protected by copyright. All rights reserved.

10.1002/cbic.201900303
ChemBioChem
regeneration for reactions such as cell-free protein translation systems [7], which
[8]
requires the regeneration of GDP to GTP, and of AMP and ADP to ATP
.
Amino acid sequences of the enzymes examined in this study are shown in Fig.
1. Both sequences show the Walker A and B motifs and the lid module present
[9]
on enzymes from all PPK2 classes , and also show the signature residue for
this class of enzymes (green box) [5A]. PPK2 class III enzymes were previously
shown to be divided into three clusters [5A], and DtCs and DrR1 were found to
belong to cluster 2 based on phylogenic analysis (Fig. S1). The cluster number
is in accordance with the definition described by Nocek et al. [5A]. Both enzyme
sequences were PCR-amplified from the original host and cloned into an
Escherichia coli expression plasmid. A poly-histidine tag was attached to the
C-terminus of the enzyme and immobilized metal ion chromatography was used
to purify the enzymes from E. coli (Fig. S2).
D. radi odurans
D. tsuruhatensi s
MD--------------------IDNYRVKPGKRVKLSDWAT---NDDAGLSKEEGQAQTA
MDAFNPFSSRDAAVQKLWEQWQPRSAKSDKDKPIALADFDPGAKPFSNGNGKTEDKAAVE
37
60
**
. : . .* : *:*:
. * .* *.:* .
WalkerA
D. radi odurans
D. tsuruhatensi s
KLAGELAEWQERLYAEGKQSLLLILQARDAAGKDGAVKKVIGAFNPAGVQITSFKQPSAE
ALAEEIDALQNLLYADRRYKLLVVLQGTDTAGKDGTIRGVFGRTSALGMNAVGWKAPTDA
97
120
** *:
*: ***: : .**::**. *:*****::: *:* . *:: ..:* *:
WalkerB
D. radi odurans
D. tsuruhatensi s
ELSHDFLWRIHQKAPAKGYVGVFNRSQYEDVLVTRVYDMIDDKTAKRRLEHIRHFEELLT
ERAHDYLWRIHQRVPGAGEITVFNRSHYEDVLVPVVNGWITPEQQRQRFAHINDFERMLS
* :**:******:.*. * : *****:****** * . * : ::*: **..**.:*:
Lid module
157
180
D. radi odurans
D. tsuruhatensi s
DNATRIVKVYLHISPEEQKERLQARLDNPGKHWKFNPGDLKDRSNWDKFNDVYEDALTT-
ETGTVVVKFLLHISADEQRARLQERLDDPAKHWKFDENDMKVRTQWKDYQQAYGQLLGAT
:..* :**. **** :**: *** ***:*.*****: .*:* *::*..:::.* : * :
216
240
D. radi odurans
D. tsuruhatensi s
STDDAPWYVVPADRKWYRDLVLSHILLGALKDMNPQFPAIDYDPSKVVIH
HTPWAPWTIVPADSKTHRNLMIATLLREVMRNLDLRFPEGDPALKDYKIT
266
290
* *** :**** * :*:*::: :* .::::: :** *
.. *
Figure 1: Sequence alignment of PPK2 class III enzymes used in this study. Both
sequences contain Walker A (red) and B motifs (blue), and a lid module (black). The
signature residue for class III PPK2 is indicated by the green box.
Enzymatic reaction with DtCs using AMP and polyphosphate as substrates
We first investigated whether AMP and ADP could act as substrates of DtCs. Fig.
2A shows the chromatogram for reverse-phase chromatography before and
after a 5-min reaction at 37C with 4 mM AMP and 1 mM polyphosphate
(polyP60: an average length of 60-mer, thereby equivalent to 60 mM
orthophosphate) using 13 µM of DtCs enzyme. The results showed that AMP
5
This article is protected by copyright. All rights reserved.

10.1002/cbic.201900303
ChemBioChem
was almost completely converted to ADP or ATP in 5 min. Fig. 2B shows the
representative time course of AMP, ADP, and ATP concentrations. The majority
of AMP was converted to ADP or ATP after 1 min and the amount then remained
fairly constant up to 5 min. After 5 min of reaction, 4-fold more ATP was
produced than the amount of ADP (Fig. 2B).
Based on the proposed mechanism of the PPK2 class III enzymes [3, 5-6], we
assumed that AMP was first phosphorylated to ADP, followed by an additional
phosphorylation to generate ATP. As more ATP was generated than ADP, we
hypothesized that not only AMP, but also ADP, is a favorable substrate of DtCs.
We thus performed the reaction using ADP as a starting material (substrate).
When the substrates were changed to 5 mM ADP and 1 mM polyP60, the amount
of ADP decreased during the first 1 min, and both ATP and AMP were produced,
suggesting that both phosphorylation and dephosphorylation took place (Fig.
2C). After 1 min incubation, the amount of ATP reached a maximum of 1.3 mM,
which was less than the amount obtained when using AMP as a substrate (3.4
mM). Furthermore, after 5 min incubation, the amount of ATP reduced to almost
[5A]
zero because of dephosphorylation. A recent study
identified adenylate
kinase (ADK) activity of PPK2 class III enzymes, although the activity was three
orders of magnitude lower than polyphosphate-dependent ADP phosphorylation.
Therefore, the result shown in Fig. 2C could partially be caused by ADK activity
of DtCs.
These results showed that DtCs can catalyze both the phosphorylation of AMP
to ADP, and of ADP to ATP, consistent with what has been reported for other
enzymes in this class (Fig. 2B,C). However, as the phosphorylation of ADP to
ATP takes place much less efficiently than that of AMP to ATP, the assumption
that AMP was first phosphorylated to ADP, followed by an additional
phosphorylation to generate ATP might be incorrect.
6
This article is protected by copyright. All rights reserved.

10.1002/cbic.201900303
ChemBioChem
■ 0 min
■ 5 min
16
14
12
10
8
A
AMP
ATP
6
ADP
4
2
0
10
15
20
25
30
35
Retention time /min
C
B
Figure 2: Phosphorylation of adenosine-5’-mono- or diphosphate by DtCs. (A)
Chromatogram of the phosphorylation product using 13 µM DtCs with 4 mM AMP and 1 mM
polyP60. (B) Time course of the reaction product analyzed by high-performance liquid
chromatography (HPLC) using 4 mM AMP and 1 mM polyP60 as substrates. (C) Time
course of the reaction product analyzed by HPLC using 5 mM ADP and 1 mM polyP60 as
substrates.
Kinetic properties of DtCs
We then aimed to determine the kinetic parameters of DtCs with AMP and ADP
in the presence of 1 mM polyP60. We measured the rate of decrease in the
substrate (AMP or ADP) in the presence of various concentrations of AMP or
ADP. We obtained a kcat/Km value of 467 (mM-1 s-1) for AMP (Table 1). As for
ADP, a linear increase in the rate was observed even when the ADP
concentration was increased to 50 mM, indicating that the Km for ADP was larger
7
This article is protected by copyright. All rights reserved.

10.1002/cbic.201900303
ChemBioChem
than 50 mM. Nevertheless, assuming that the substrate concentration is far
smaller than the Km value, we obtained a kcat/Km value of 1.9 (mM-1 s-1) for ADP
(Table 1). Similar experiments were performed with GMP and GDP, and we
obtained kcat/Km values of 0.29 (mM-1 s-1) and 0.099 (mM-1 s-1) for GMP and GDP,
respectively (Table 1).
We found that the kcat/Km value for AMP was 180-fold larger than that of ADP. In
addition, the kcat/Km for AMP was 1200-fold larger than that of GMP. In addition,
the kcat/Km values for GMP and GDP were similar. The relationship between
these values and the phylogenic analysis will be described in the Discussion.
Table 1: Kinetic parameters of DtCs
AMP
0.12±0.02
56±2.0
467
ADP
> 50 mM
N.D.
GMP
8.5±2.8
2.5±0.2
0.29
GDP
> 50 mM
N.D.
Km (mM)
kcat (s-1)
kcat/Km (mM-1・s-1)
1.9
0.099
N.D.: not determined.
DtCs directly convert AMP to ATP by pyrophosphorylation
From the above observations, we assumed that DtCs not only phosphorylate
AMP to ATP via ADP but convert AMP directly to ATP by adding pyrophosphate
from the polyP60, i.e., via pyrophosphorylation (Fig. 3, red arrow).
ADP+polyPn-1
AMP+polyPn
ATP+polyPn-2
Figure 3: Schematic of the phosphorylation of AMP with polyphosphate catalyzed by PPK2
class III enzymes.
To test this idea, we performed the reaction with AMP and triphosphate (P3) as
substrates. Both ADP and ATP were produced by incubating DtCs with AMP and
8
This article is protected by copyright. All rights reserved.

10.1002/cbic.201900303
ChemBioChem
P3 (Fig. S3A), although monophosphate or pyrophosphate could not serve as a
substrate for AMP phosphorylation (Fig. S3B,C). Under the condition where
[P3]>>[AMP], if AMP is directly pyrophosphorylated to generate ATP,
monophosphate should be produced, whereas if ATP is generated via ADP,
monophosphate should not appear, but pyrophosphate should be produced.
We directly measured the amount of monophosphate after the enzymatic
reaction. A malachite green phosphate assay detected only monophosphate, but
not pyrophosphate or triphosphate (Fig. S4A). We performed the reaction using
different concentrations of triphosphate (1, 2, and 5 mM) with a fixed
concentration of AMP (1 mM). Fig. 4A shows the relationship between the
produced monophosphate and ATP concentration. A clear linear correlation was
observed, and the slope was approximately 0.8. These results support the
hypothesis that most of the ATP (80%) was converted from AMP to ATP by
taking the pyrophosphate from the triphosphate, i.e., pyrophosphorylation (Fig. 3,
red arrow), while the remaining 20% was produced by a different pathway,
presumably by two monophosphorylations of AMP using a different P3 molecule
as a substrate.
We performed the same experiment as that shown in Fig. 4A with a PPK2 class
III enzyme from D. radiodurans R1 (DrR1), another mesophilic bacterium (Fig.
S2). Again, a clear linear correlation was observed between the produced
monophosphate and ATP concentration (Fig. 4B). The slope was approximately
0.3. This result suggests that ATP production from AMP can occur by
pyrophosphorylation. Quantitatively, 30% of ATP was converted from AMP by
pyrophosphorylation, while the remaining 70% was produced by a different
pathway. The results indicate that the pathway preference differs among
enzymes from different species.
9
This article is protected by copyright. All rights reserved.

10.1002/cbic.201900303
ChemBioChem
A
B
Figure 4: Phosphorylation using triphosphate (P3) as a substrate. Correlation between the
concentration of monophosphate and ATP using (A) DtCs and (B) DrR1. The reaction was
performed using 1 mM AMP and 1, 2, or 5 mM P3. Enzyme concentration was 13 µM and
incubation was performed at 37C for 30 min. The gray line shows the result of linear
regression. The slope of the line was 0.78 and 0.33 in A and B, respectively. The
monophosphate concentrations shown were calculated from the absorbance obtained after
subtracting the data without the enzyme from the one with the enzyme (Fig. S5).
DtCs can produce adenosine tetraphosphate
In a recent study, four PPK2 enzymes derived from class I, II, and III were shown
to produce adenosine 5’-polyphosphate molecules such as adenosine tetra-
(AP4) and pentaphosphate (AP5) using AMP or ADP and polyphosphate as
[6]
substrates . We also observed a similar result with DtCs. Fig. 5A shows the
chromatogram of the products after incubation with AMP and polyP60 as
substrates for 30 min at 37C. We observed a peak with a retention time later
than that for ATP (orange arrow). This product was collected and subjected to
mass spectrometry (MS) analysis. We found a peak of m/z with 586, whose size
matched with authentic AP4 (Fig. S6). This product was further subjected to
tandem mass spectrometry (MS/MS), and the obtained spectrum matched well
with that of authentic AP4 (Fig. 5B). These results suggest that DtCs can also
produce AP4 using AMP and polyphosphate as substrates.
10
This article is protected by copyright. All rights reserved.

10.1002/cbic.201900303
ChemBioChem
A
160809_AP4_STD_NO_HPLC_MSMS 7 (0.118) Sm (SG, 2x0.75); Cm (4:29)
B
239.20
100
487.83
488.35
AP4
158.78
ATP
408.21
177.05
426.08
505.83
256.95
290.82
235.43
432.84
353.16
79.25
133.75
0
ADP
160809_FRACTION_MSMS 12 (0.202) Sm (SG, 2x0.75); Cm (2:28)
487.96
100
Fraction
AMP
238.94
408.08
488.48
505.96
159.17
426.28
506.4
353.22
334.96
176.85
158.00
150
213.72
291.60
300
398.14
470.15
256.63
250
72.55
0
50
100
200
350
400
450
500
m/z
Figure 5: Enzymatic reaction with extended time with DtCs using polyP60 and AMP as
substrates. (A) Chromatogram of the reaction product. Peak fraction indicated by orange
arrow was collected for MS analysis. Enzyme at a concentration of 13 µM was used and
incubation was performed at 37C for 30 min. (B) MS/MS analysis of the compound with m/z
of 586 (Fig. S6). Red arrows indicate the common peaks between the authentic AP4 and
that fractionated by HPLC.
Discussion
We studied the properties of a PPK2 class III enzyme from D. tsuruhatensis
(DtCs), and found that this enzyme not only catalyzes the monophosphorylation
of both AMP to ADP and ADP to ATP using polyphosphate as a substrate, but
also directly converts AMP to ATP by pyrophosphorylation. The latter reaction
was also observed with another PPK2 class III enzyme from D. radiodurans R1
(DrR1), while the efficiency of the two enzymes differed. In addition, we also
found that DtCs can produce adenosine tetraphosphate (AP4) using AMP and
polyP60 as substrates.
11
This article is protected by copyright. All rights reserved.

10.1002/cbic.201900303
ChemBioChem
In a recent study, PPK2 class III enzymes were shown to be divided into three
[5A]
clusters, and only one enzyme from each cluster was characterized in detail
(Fig. S1). The trend in the kinetic parameters differed among the three clusters.
The kcat/KM values for ADP and AMP were nearly identical for both cluster 1 and
3 enzymes, which was in the order of 10 mM-1・s-1. The difference between
cluster 1 and 3 was seen in the kcat/KM value for GMP and GDP. While the
kcat/KM values were similar for GMP and GDP for both enzymes, the cluster 1
enzyme showed a 10-fold higher kcat/KM value compared to that from cluster 3.
For an enzyme from cluster 2, kcat/KM values for AMP were higher (350 mM-1・s-1)
compared to that from clusters 1 and 3, and the value for ADP was 100-fold
lower than that of AMP. Hence, a larger kcat/KM value for AMP than ADP is the
signature for the cluster 2 enzyme. As this observed substrate preference was
based on a single enzyme from each cluster, it remains unclear whether this is
true for other enzymes. In this study, DtsC showed a kcat/KM value of 467 mM-1・
s-1 for AMP and a 246-fold lower value for ADP. These features were very similar
to an enzyme from cluster 2. Indeed, phylogenic analysis showed that DtsC
belonged to cluster 2 (Fig. S1). These results suggest that substrate preference
is similar among the enzymes belonging to the same cluster. Class III enzymes
are defined as enzymes that catalyze the phosphorylation of nucleotide mono-
into diphosphate and di- into triphosphate; however, the cluster 2 enzyme
showed much lower activity for the latter reaction. More enzymes should be
characterized to clarify the difference among the enzymes belonging to each
cluster.
The linear correlation between the concentration of ATP and monophosphate
(Fig. 4) suggests that some, but not all of the ATP produced from AMP and
polyphosphate is generated by pyrophosphorylation. In the case of DtsC and
DrR1, the slope of the correlation between the monophosphate and ATP
concentration was approximately 0.8 and 0.3, respectively. These results
indicate that 80 and 30% of the ATP produced by DtsC and DrR1 occur via
pyrophosphorylation, while the rest occur via step-by-step monophosphorylation
using different P3 molecules. Although several enzymes, such as
12
This article is protected by copyright. All rights reserved.

10.1002/cbic.201900303
ChemBioChem
ribose-phosphate diphosphokinase, thiamine diphosphokinase, and nucleotide
diphosphokinase, are known to catalyze pyrophosphorylation of their substrates,
all of them employ nucleotide triphosphate as a phosphate donor [10]. The PPK1
from E. coli has been shown to convert GDP to GTP, but also produce guanidine
tetraphosphate (GP4) as a minor product using polyphosphate as a substrate [11]
While this study did not provide direct evidence, as GP4 did not appear when
GTP and polyphosphate were used as substrates, it is reasonable to assume
that GP4 was a product of pyrophosphorylation of GDP. Therefore, to the best of
our knowledge, DtsC and DrR1 are the first PPK2 enzymes shown to have
polyphosphate-dependent pyrophosphorylation activity. However, it should be
noted that our observation does not necessarily indicate that reactions catalyzed
by these enzymes occur via the direct pyrophosphate transfer from P3. It is
possible that the enzyme binds to the P3 and AMP, and that the two phosphates
are transferred to the AMP one-by-one from the single bound P3. In this case,
monophosphate would also be released. This mechanism and pyrophosphate
transfer cannot be distinguished. Nevertheless, in this study, we defined both
mechanisms as pyrophosphorylation since these two mechanisms are not
distinguishable based on our data. Studying a reaction intermediate, for example
by obtaining a structure in complex with nonhydrolyzable P3 or by other means,
will be necessary to clarify the detailed mechanism.
.
In another recent study [6], adenosine polyphosphate was shown to be produced
by several PPK2 enzymes. Interestingly, this was the case for all classes of
PPK2 enzymes (e.g., classes I, II, and III). Not only the synthesis of AP4, but
also of AP5, was observed. When co-crystallizing the PPK2 enzyme from
Francieslla tularensis (class I) with the nonhydrolyzable ATP analog
β,γ-methylene adenosine 5ʹ-triphosphate (AMPPCP) and polyphosphate,
β,γ-methylene adenosine 5ʹ-pentaphosphate (AMPPCPPP) was found in the
crystal [5B], indicating that two phosphates were added to the AMPPCP. We also
observed the production of AP4 with DtsC (Fig. 5). Synthesis of adenosine
polyphosphate seems to be a general feature of the PPK2 enzymes; however,
the detailed mechanism remains to be elucidated. As we observed the transfer
of pyrophosphate to AMP, we postulate that not all but part of the adenine
13
This article is protected by copyright. All rights reserved.

10.1002/cbic.201900303
ChemBioChem
polyphosphate is generated by the transfer of di-, tri- or longer polyphosphate,
rather than a step-by-step monophosphorylation.
Conclusion
In summary, we found that PPK2 class III enzymes belonging to cluster 2 (DtCs
and DrR1) catalyze the pyrophosphorylation of AMP to ATP. To the best of our
knowledge, these are the first PPK2 enzymes shown to have
polyphosphate-dependent pyrophosphorylation activity. In addition, DtCs
showed similar substrate preference as a previously characterized enzyme
belonging to cluster 2, suggesting that substrate preference is similar among the
enzymes belonging to the same cluster.
Experimental Section
Cloning and enzyme purification
DNA encoding PPK2 class III enzymes were PCR-amplified using primers for
each enzyme from the genome extracted from D. tsuruhatensis (gi|503566024)
and D. radiodurans R1 (gi|499189240) obtained from NITE Biological Resource
Center (Chiba, Japan). PCR was performed using KOD plus DNA polymerase
(Toyobo, Japan) according to the manufacturer’s instructions and the primers
were designed to carry EcoRI and HindIII sites at N and C termini of the PPK2
enzyme, respectively. PCR products were purified using QIAquick (Qiagen),
digested with EcoRI and HindIII, and ligated in between EcoRI and HindIII sites
of the pET-gusA plasmid [12]. Thus, a histidine tag was introduced into the
C-terminus of the enzyme.
After confirming the DNA sequence, the plasmids were transformed into
BL21(DE3). The enzymes were overexpressed, purified by immobilized metal
affinity chromatography, dialyzed against stock buffer (50 mM HEPES-KOH
(pH7.6), 100 mM KCl, 10 mM MgCl2, 30% glycerol, 7mM 2-mercaptethanol), and
stored at -80C.
Enzyme reaction and assay
14
This article is protected by copyright. All rights reserved.

10.1002/cbic.201900303
ChemBioChem
An enzyme assay with PPK2 enzymes from D. tsuruhatensis (DtCs) and D.
radiodurans R1 (DrR1) was performed in the reaction buffer (50 mM MOPS (pH
7.0) and 10 mM MnCl2). UniProt accession IDs of enzymes used in this study
are UP000002524 (DrR1) and UP000234143 (DtCs). Polyphosphate, nucleotide,
and enzyme concentrations are as indicated in the Results section.
Polyphosphate with an average length of 60-mer (polyP60) was prepared
according to methods described by Honda et al. [13]. PolyP60 chain length was
estimated by comparing its electrophoretic mobility with those of commercial
standards (Polyphosphate glass 45, and 65 from Sigma) as described previously
[14]. When measuring the reaction rate, the reaction mixture was incubated at
37C, and aliquots were collected at 0, 1, 2, and 5 min, respectively. The aliquot
was mixed with equal volume of 5% trichloroacetate, centrifuged for 10 min at
4C and 20,000 × g, and the supernatant was diluted 2-fold with distilled water. A
diluted sample was centrifuged for 10 min at 4C and 20,000 × g, and the
supernatant was used for HPLC analysis.
HPLC analysis was performed using a reverse-phase column (TSKgel
ODS-100V (Tosoh)) on an HPLC instrument (Prominence, Shimadzu). The
mobile phase buffer A was 20 mM tert-butylamine-phosphate (pH 6.8) and buffer
B was 10% methanol with 90% buffer A. Flow rate and column temperature were
fixed to 1 mL/min and 40ºC, respectively. The concentration of mobile phase
buffer B was increased from 0% to 100% during 35 min, and the absorbance at
260 nm was recorded (Fig. 2A). Regarding the data shown in Figs. 5A and S3,
mobile phase buffer B was increased from 33% to 100% during 25 min.
Authentic nucleotides were used to determine the nucleotide retention time.
Nucleotide concentrations were determined from the peak area obtained from
the chromatogram.
Phosphate quantification
The amount of monophosphate was measured using a Malachite Green
Phosphate Assay Kit (BioAsssay Systems, California, U.S.A.) according to the
manufacturer’s instructions. After mixing the reagents from the kit with the
15
This article is protected by copyright. All rights reserved.

10.1002/cbic.201900303
ChemBioChem
phosphate solution, samples were incubated for 30 min, and the absorbance at
630 nm was measured using a plate reader (Synergy H1, Biotek).
Mass spectrometry analysis
A sample containing adenosine tetraphosphate (AP4) was prepared using AMP
and polyP60 as substrates under the reaction conditions described above, except
that the incubation time was extended to 30 min. When attempting to detect the
nucleotide products of the enzymatic reaction by mass spectrometry (MS), only
the polyP60 present in excess to the nucleotides was detected. Therefore, we
removed the polyP60 from the nucleotide solution before subjecting the sample
to MS analysis. To remove the polyP60 from the sample, the reaction mixture
was loaded to a Vivaspin500 (3000 NMCO) concentrator and the flow-through
was collected. The flow-through was then subjected to a metal-free column
(Inertsustain AQ-C18) on an HPLC instrument (Prominence, Shimadzu). The
mobile phase buffer A was 20 mM tert-butylamine-acetate (pH 6.8) and buffer B
was 15% methanol with 85% buffer A. Flow rate and the column temperature
were fixed to 1 mL/min and 40C, respectively. The concentration of mobile
phase buffer B was fixed at 33% during the first 15 min, followed by an increase
to 100% during the next 35 min. The peak fraction corresponding to AP4 was
collected and freeze-dried under a vacuum. Authentic AP4 was purchased from
Jena Bioscience (Germany).
Mass spectrometry (MS) analysis was carried out using Quattro Premier XE
(Waters). The sample solution was delivered with a syringe pump at a flow rate
of 10 L/min and, prior to injecting the MS probe, mixed with a flow of the 0.1%
formic acid in water/acetonitrile (50:50) at a flow rate of 0.2 mL/min from Acquity
UPLC binary solvent manager (Waters). The mass spectrometer was operated
in the negative electrospray ionization (ESI-) mode. Capillary and cone voltages
were set at 3.0 kV and 30 V, respectively. MS scan was conducted for monitoring
AP4 deprotonated ions with a range from m/z 50 to 700. Product ion scan of AP4
was conducted at collision energy of 25 eV.
Acknowledgments
16
This article is protected by copyright. All rights reserved.

10.1002/cbic.201900303
ChemBioChem
This work was funded by KAKENHI grants 17H00888 (T.M.) from the Japan
Society for the Promotion of Science.
Keywords: polyphosphate kinase 2, PPK2 class III, pyrophosphorylation,
enzyme kinetics
Supporting Information
Supporting information for this article is given via a link at the end of the
document.
References
[1]
[2]
J. N. Andexer, M. Richter, Chembiochem 2015, 16, 380-386.
a) B. Nocek, S. Kochinyan, M. Proudfoot, G. Brown, E. Evdokimova, J. Osipiuk,
A. M. Edwards, A. Savchenko, A. Joachimiak, A. F. Yakunin, Proc Natl Acad
Sci U S A 2008, 105, 17730-17735; b) H. Zhang, K. Ishige, A. Kornberg, Proc
Natl Acad Sci U S A 2002, 99, 16678-16683.
[3]
K. Motomura, R. Hirota, M. Okada, T. Ikeda, T. Ishida, A. Kuroda, Appl
Environ Microbiol 2014, 80, 2602-2608.
[4]
[5]
S. Suzuki, R. Hara, K. Kino, J Biosci Bioeng 2018, 125, 644-648.
a) B. P. Nocek, A. N. Khusnutdinova, M. Ruszkowski, R. Flick, M. Burda, K.
Batyrova, G. Brown, A. Mucha, A. Joachimiak, L. Berlicki, A. F. Yakunin, Acs
Catalysis 2018, 8, 10746-10760; b) A. E. Parnell, S. Mordhorst, F. Kemper, M.
Giurrandino, J. P. Prince, N. J. Schwarzer, A. Hofer, D. Wohlwend, H. J. Jessen,
S. Gerhardt, O. Einsle, P. C. F. Oyston, J. N. Andexer, P. L. Roach, Proc Natl
Acad Sci U S A 2018, 115, 3350-3355.
[6]
[7]
S. Mordhorst, J. Singh, M. K. F. Mohr, R. Hinkelmann, M. Keppler, H. J. Jessen,
J. N. Andexer, Chembiochem 2019, 20, 1019-1022.
a) E. D. Carlson, R. Gan, C. E. Hodgman, M. C. Jewett, Biotechnol Adv 2012,
30, 1185-1194; b) Y. Endo, T. Sawasaki, Curr Opin Biotechnol 2006, 17,
373-380; c) H. Ohashi, T. Kanamori, Y. Shimizu, T. Ueda, Curr Pharm
Biotechnol 2010, 11, 267-271.
[8]
H. C. Kim, D. M. Kim, J. Biosci. Bioeng. 2009, 108, 1-4.
17
This article is protected by copyright. All rights reserved.

10.1002/cbic.201900303
ChemBioChem
[9]
a) D. D. Leipe, E. V. Koonin, L. Aravind, J Mol Biol 2003, 333, 781-815; b) J. E.
Walker, M. Saraste, M. J. Runswick, N. J. Gay, EMBO J 1982, 1, 945-951.
[10] a) C. Azevedo, A. Saiardi, Biochem Soc Trans 2016, 44, 13-17; b) N. B. Ray, C.
K. Mathews, Curr Top Cell Regul 1992, 33, 343-357.
[11] A. Kuroda, A. Kornberg, Proc Natl Acad Sci U S A 1997, 94, 439-442.
[12] T. Matsuura, K. Hosoda, N. Ichihashi, Y. Kazuta, T. Yomo, J Biol Chem 2011,
286, 22028-22034.
[13] K. Honda, N. Hara, M. Cheng, A. Nakamura, K. Mandai, K. Okano, H. Ohtake,
Metab Eng 2016, 35, 114-120.
[14] E. Restiawaty, Y. Iwasa, S. Maya, K. Honda, T. Omasa, R. Hirota, A. Kuroda, H.
Ohtake, Process Biochem 2011, 46, 1747-1752.
18
This article is protected by copyright. All rights reserved.

10.1002/cbic.201900303
ChemBioChem
Graphical Abstract
Class III polyphosphate kinase 2 can
ADP+polyPn-1
ATP+polyPn-2
transfer pyrophosphate from
polyphosphate to convert AMP to
ATP.
AMP+polyPn
19
This article is protected by copyright. All rights reserved.