Identification of two active functional domains of human adenylate kinase 5
Nicola Solaroli*, Christakis Panayiotou, Magnus Johansson, Anna Karlsson
Department of Laboratory Medicine, F68, Karolinska Institute, S-14186 Huddinge, Sweden
a r t i c l ei n f o
Received 6 July 2009
Revised 23 July 2009
Accepted 27 July 2009
Available online 3 August 2009
Edited by Miguel De la Rosa
a b s t r a c t
A full length cDNA that partially corresponded to human adenylate kinase 5 (AK5) was identified
and shown to encode for two separate domains. The full length protein could be divided in two dis-
tinct functional domains, a previously unidentified domain of 338 amino acids and a second domain
of 198 amino acids that corresponded to the protein characterized as AK5, now called AK5p2. The
first domain, AK5p1, phosphorylated AMP, CMP, dAMP and dCMP with ATP or GTP as phosphate
donors similarly to AK5p2. Our data demonstrate that human AK5 has two separate functional
domains and that both have enzymatic activity.
? ? 2009 Federation of European Biochemical Societies. Published by Elsevier B.V. All rights reserved.
Phosphoryl transfer reactions are critical in cellular metabolism
and to generate and modulate metabolic signals . Enzymes that
catalyze the reversible phosphoryl transfer between nucleoside tri-
phosphates and nucleoside monophosphates, resulting in the con-
version of the monophosphate to its corresponding diphosphate
form are named nucleoside monophosphate kinases (NMPKs).
Based on the differences in substrate specificity, NMPKs are
divided into subgroups consisting of adenylate kinases (AKs),
uridylate–cytidylate (UMP–CMP), guanylate kinases and thymidyl-
ate (dTMP). While there are several AKs and guanylate kinases
present in mammalian tissues there was until recently only one
identified UMP-CMP kinase and one dTMP kinase. A second
UMP–CMP kinase has been identified and shown to have a mito-
chondrial location and there still may be a mitochondrial dTMP
kinase to be found [2,3].
Adenylate kinase catalyzes the nucleotide phosphoryl exchange
reaction AMP + ATP 2 ADP, and thus regulates adenine nucleotide
ratios in different intracellular compartments. There are seven dif-
ferent adenylate kinases identified and six of these enzymes have
been thoroughly characterized. AK1, AK2 and AK3 are expressed
at higher levels than the other AKs and have a more general
expression in most tissues although AK1 is dominant in skeletal
muscle . AK2 is the main mitochondrial adenylate kinase and
is located in the intermembrane space , while the other mito-
chondrial adenylate kinases, AK3 and AK4 are located in the mito-
chondrial matrix . AK5 is a cytosolic isoform initially detected
exclusively in brain, when only a limited number of tissues were
investigated, but low levels of AK5 expression is also detected in
other tissues . The recently characterized isoform AK6 is a nucle-
ar enzyme but without any experimentally confirmed expression
profile . AK7 is reported to be associated with ciliar function
since a mutation in the AK7 gene in a rat model recently was re-
ported to cause ciliar dysfunction and a phenotype similar to the
human syndrome called primary ciliary dyskinesia .
Thus, the AK isoforms have been found in different intracellular
compartments and with tissue specific levels of expression to serve
the specific needs of different cellular functions. The AKs also differ
in molecular weight, kinetic properties and nucleotide specificity
. The common substrate recognized by all AKs is AMP but
AK5 and AK6 also phosphorylate CMP and dCMP [11,8]. In addition,
all isoforms are coded by separate genes, localized to different
chromosomes. In the present study we have identified yet another
specific characteristic of one of the AK isoforms. We demonstrate
that the human AK5 has two separate functional domains and that
both domains show similar adenylate kinase activity.
2. Materials and methods
2.1. Cloning and sequencing
The IMAGE clone 4816351 was from Geneservice. The encoded
proteins (AK5, AK5p1 and AK5p2) were expressed in Escherichia
coli. Nine primers flanking the three possible open reading frames
were designed. Three forward primers for pET16b vector were: 50-
AAATGGGAGGTTTCATGGAAGA-30. Two reverse primers for pET16b
0014-5793/$36.00 ? 2009 Federation of European Biochemical Societies. Published by Elsevier B.V. All rights reserved.
* Corresponding author. Address: Karolinska Institute, Department of Laboratory
Medicine, F68, Karolinska University Hospital, S-14186 Huddinge, Sweden. Fax: +46
E-mail address: email@example.com (N. Solaroli).
FEBS Letters 583 (2009) 2872–2876
journal homepage: www.FEBSLetters.org
vector were: 50-CAGCCGGATCCTCGATTCAGAAAATAGAGTCAATAG-
Two forward primers for pEGFP-N1 vector were: 50-CTCAA-
AATTTATGTGTTCTAAGCCCGAAGA-30. Three reverse primers for
pEGFP-N1 vector were: 50-GTCGACTGCAGAATTCGGAAAATAGAGT-
ATA-30, 50- GTCGACTGCAGAATTCGCAGTCCATTCAAAAGGCTCT-30.
The PCR products were cloned into the XhoI site of pET-16b vec-
tor or the EcoRI site of pEGFP-N1 vector using the In-Fusion Dry-
Down Cloning. The plasmids were sequenced by MWG-Biotech to
verify the DNA sequences.
2.2. Expression and purification
The plasmids were transformed into the E. coli strain Rosetta
(DE3) and single colonies inoculated into LB-medium with
100 lg/ml ampicillin. The bacteria were grown at 37 ?C and pro-
tein expression was induced at OD600 0.8 with 1 mM isopropyl-
1-thio-b-D-galacto-pyranoside for 12 h at 27 ?C. The expressed pro-
teins were purified using TALON Metal Affinity Resin as described
by the manufacturer. Purity of the protein was verified by SDS–
PAGE gel and concentration was determined with Bradford Protein
Assay using BSA as standard. The protein was aliquoted and stored
at ?80 ?C.
2.3. Enzyme assay
The substrate specificity of the purified enzymes was assayed
by thin layer chromatography (TLC) as described . The assay
was performed with 1 lCi [c-32P]ATP (Perkin–Elmer), recombinant
enzyme and concentrations of xMP or dxMP indicated. The TLC
sheets were autoradiographed (Personal Molecular Imager FX)
and analyzed/edited with Quantity One v4.6 software. For
Michaelis–Menten kinetics, non-radiolabeled products were sepa-
rated and quantified by reversed-phase HPLC using a ChromolithTM
column as described .
2.4. In vitro translation
Fresh precursor proteins (AK5, AK5p1 and AK5p2) were synthe-
sized (PURExpress In Vitro Protein Synthesis Kit) in the presence of
[35S]methionine following the manufacturer’s instructions. For
protein quantification 2 ll of each reaction was loaded in a 4–
12% SDS–PAGE gel. The gel was washed in Gel-Dry Drying solution,
dried with Dry Easy Membrane and used for autoradiographic
2.5. Cell culture and transfection
HeLa cells were seeded in Ibidi l-Slides and cultured in Dul-
becco’s Modified Eagle’s Medium with 10% (v/v) fetal bovine serum,
100 U/ml penicillin and 0.1 mg/ml streptomycin. Cells were grown
at 37 ?C in a humidified incubator with 5% CO2. Plasmids were
transfected into cells using FuGENE HD transfection reagent as de-
scribed by the manufacturer. GFP fluorescence was observed in a
Leica DMI-6000B with Leica Application Suite AF software v1.8.
3.1. Identification of two functional domains of AK5
A cDNA with the Walker A motif that partially corresponded to
the previously characterized human adenylate kinase 5 (AK5) was
Fig. 1. (A) Map of the full length AK5 protein (two arrows show where the two variants start). (B) Alignment of AK5p1 and AK5p2 with human adenylate kinases (black boxes
indicate completely conserved amino acid residues and different shades of grey boxes indicate different levels of conserved residues. * Indicate amino acids involved in
substrate binding or in the active site pocket definition). (C) Phylogenetic tree of the human adenylate kinases.
N. Solaroli et al./FEBS Letters 583 (2009) 2872–2876
identified in a human genome library. The gene was localized on
chromosome 1 and consisted of 14 exons with a total length of
277.92 kb. Two main variants were reported (BC036666 and
BC033896) which differed in the N-terminus where one variant
had an additional short sequence of 26 amino acids (TAG) (Fig. 1A).
The full length protein of 536 amino acids contained two dis-
tinct functional domains; a previously unidentified domain of
338 amino acids and a second domain of 198 amino acids that cor-
responded to the protein already characterized as AK5 . The
open reading frame contained three main ATG start codons, which
with the common single stop codon, could result in transcripts of
different sizes (Fig. 1A). The sequence of the two distinct proteins
suggested to be closely related to known AKs. The first domain
was named AK5p1 and the second domain, previously considered
to be the full length AK5, was renamed AK5p2. AK5p1, differently
from AK5p2, contained 107 amino acids after the second ATG but
without any recognizable targeting signal or other known func-
tional domain. After the AK5p1 domain there was a second stretch
of 60 amino acids before the ATG of AK5p2. The sequence compar-
ison between AK5p1 and AK5p2 showed an identity of 36% and a
general homology of 64% (Fig. 1B). AK5p1 also showed an identity
of 35%, 24%, 26% and 27%, respectively, with AK1, AK2, AK3 and
AK4, as represented by the phylogenetic tree (Fig. 1C). The highest
level of identity of the novel AK5p1 domain with known adenylate
kinases was within the glycin rich p-loop structure, the NMP bind-
ing domain and the LID domain. The two domain gene structure of
AK5 was also found to be conserved in other mammals (in mice the
predicted amino acid sequence has 94% identity, data not shown).
3.2. Expression and characterization of the different forms of AK5
The full length protein and the AK5p1 protein domain alone
were expressed in E. coli, with and without the initial fragment
of 26 amino acids, as described in Methods. For an initial screening
of substrate specificity AMP, CMP, GMP, UMP, dAMP, dCMP, dGMP,
TMP and dUMP were tested in a TLC assay using ATP or GTP as
phosphate donors (Table 1). AK5p1 phosphorylated AMP, CMP,
dAMP and dCMP with ATP or GTP as phosphate donors. At the as-
say conditions used, and at lower concentrations of substrate,
AK5p1 showed generally a higher affinity for AMP compared to
dAMP (Fig. 2). The relative efficiency of CMP and dCMP was
?15% compared to AMP. No significant differences were found be-
tween the enzymes with or without the initial 26 amino acid
To investigate if there were synergic effects of the two subunits
the full length AK5, AK5p1 and AK5p2 were synthesized by in vitro
translation. We used [35S]methionine to visualize the synthesized
proteins and to compare the relative amount. As expected the
[35S]methionine present in each sequence (Fig. 3A). The three en-
zymes, at similar concentrations, were tested for their capacity to
phosphorylate AMP, dAMP, CMP and dCMP. Full length AK5
showed an increased capacity to phosphorylate AMP as compared
to the single subunits (Fig. 3B), while with dAMP and CMP as phos-
showedanintensity thatwasproportional to the
Substrate specificity of the recombinant enzymes using ATP or GTP as phosphate
Substrate (1 mM) Phosphate donor (5 mM)
ATPGTPATP GTPATP GTP
aVan Rompay et al. .
Fig. 2. Affinity assay of AK5p1 with variable concentrations of AMP or dAMP in
[c-32P]ATP phosphotransferase assay. (The control reaction, lane C, is without
addition of substrate.)
Fig. 3. (A) SDS–PAGE gel of the in vitro synthesized proteins. Lane 1, 2, 3 contain,
respectively, AK5 (63.2 kDa), AK5p1 (41.3 kDa) and AK5p2 (21.9 kDa). (B) TLC assay
of AK5 (lane 1, 4, 7), AK5p1 (lane 2, 5, 8) and AK5p2 (lane 3, 6, 9) without substrate
or with AMP/dAMP. (Relative levels of phosphorylation expressed in relation to% of
AK5 with AMP/dAMP phosphorylation.) (C) TLC assay of AK5 (lane 1, 4, 7), AK5p1
(lane 2, 5, 8) and AK5p2 (lane 3, 6, 9) without substrate or with CMP/dCMP.
(Relative levels of phosphorylation expressed in relation to percentage of AK5 with
N. Solaroli et al./FEBS Letters 583 (2009) 2872–2876
phate acceptor the difference of full length AK5 and the AK5p1 and
AK5p2 enzymes was minimal (Fig. 3B and C).
The Kmand Vmaxkinetic parameters for the substrates AMP and
dAMP for AK5p1 showed the highest affinity for AMP with a Kmof
172 lM while the Kmfor dAMP was >2000 lM. The Vmaxfor AMP as
substrate was 226 pmol/lg/min while the Vmaxfor dAMP was only
39 pmol/lg/min, resulting in AMP to be more than 60-fold more
efficient as substrate for AK5p1 as compared to dAMP (Table 2).
We were not able to measure the AK5p1 kinetic parameter for
the substrates CMP or dCMP using ATP as phosphate donor, and
not for AMP, dAMP, CMP or dCMP using GTP as phosphate donor.
3.3. AK5. subcellular localization
The subcellular prediction software used did not indicate any
putative targeting signal in the cDNA sequences (data not shown).
This was confirmed by in vivo expression of the enzyme as fusion
protein with the Green fluorescent protein (GFP). The previously
studied AK5p2 was shown to have a cytosolic/nuclear localization
and therefore we decided to investigate the full length protein and
the AK5p1 with and without the 26 amino acid initial TAG
(Fig. 4A). The cells transfected with wild type GFP, full length
AK5 and AK5p1 with the initial TAG showed a uniform distribution
in the cytosol/nucleus; on the other side the lacking of the 26 aa
TAG produced proteins exclusively localized into the cytosol
The many members of the AK family are likely to reflect not
only the necessity for cells in general to maintain cellular adenine
nucleotide pools but also that a fine tuning of this demand is
important for specialized cell functions. The differences in enzy-
matic activity, subcellular location, tissue distribution and expres-
sion levels suggest both general and very specialized needs of AK
activity in specific cellular processes. Recent studies have shown
the involvement of AKs in severe medical conditions . Heart
function in AK1 knockout mice displayed accelerated loss of car-
diac contractions on ischemic challenge . Mutations of the gene
encoding for AK2 have been identified in individuals with reticular
dysgenesis, an autosomal recessive form of severe combined
immunodeficiency . Another recent study has shown that sen-
sorineural deafness is associated with reticular dysgenesis caused
by AK2 deficiency .
In contrast to the multi-tissue expression profiles of most other
AKs, AK5 has the unique property of being highly expressed in
brain . Autoantibodies targeted against AK5 have been impli-
cated in two cases of limbic encephalitis . In the present study
we analyzed two transcript variants of the AK5 gene; the first one
represents a shorter transcript (3526 bp) that encodes the longest
isoform (562 aa) and the second transcript has instead a different,
and longer 50sequence, and use a downstream start codon, as com-
pared to variant 1, which leads to a shorter isoform (536 aa). Iso-
form 1 has 26 amino acids extra in the N-terminal compared to
isoform 2, while the remaining part of the sequence is identical
in both isoforms. These 26 amino acids and the different isoforms
Kinetic properties of recombinant AK5p1 with ATP as phosphate donor (2.5 mM) and
AMP or dAMP as acceptor. (Data from 4 to 5 measurements ± S.E.)
172 ± 29
226 ± 6
39 ± 3
Fig. 4. (A) Plasmid vector constructs (pEGFP-N1) used to express AK5, AK5 w/o TAG, AK5p1 and AK5p1 w/o TAG. (B) HeLa cells expressing pEGFP-N1 (control), AK5, AK5 w/o
TAG, AK5p1 and AK5p1 w/o TAG (cells were contra-stained with Mitotracker).
N. Solaroli et al./FEBS Letters 583 (2009) 2872–2876
have been analyzed with different software for the prediction of Download full-text
protein subcellular localization, but none of them could identify
possible targeting signals, such as the nuclear localization
As far as enzymatic activity is concerned, both the full length
AK5 and its first functional domain AK5p1 phosphorylated AMP,
dAMP, CMP and dCMP with ATP and GTP as phosphate donors.
The previously characterized AK5p2 phosphorylated AMP, CMP
and dAMP when ATP was used as phosphate donor and AMP,
CMP and dCMP with GTP as phosphate donor. Interestingly,
AK5p2 cannot phosphorylate dAMP in the presence of GTP which
is different from AK5p1. Our results show that the substrate spec-
ificity of AK5, in contrast to all the human AKs characterized so far,
is dependent on the phosphate donor. All the other AKs have a sub-
strate specificity that is independent of the phosphate donor as
they phosphorylate predominantly AMP irrespective of phosphate
The AKs are characterized by their high level of conservation in
three domains; the glycine rich domain (p-loop), which is involved
in the binding of the beta-gamma phosphate moiety of the bound
nucleotide (ATP or GTP) and the Mg2+ion, the NMP binding-LID do-
mains with the amino acids involved in the interaction with the
phosphate substrate or with the phosphate donors. Interestingly,
AK5p1 and AK5p2 contain a shorter conserved LID domain, simi-
larly to AK1. The high degree of identity and close evolutionary
relationship of AK5p1 and AK5p2 with AK1 is clearly shown by a
phylogenetic tree analysis of all the known human adenylate
kinases (Fig. 1C).
Although a high level of similarity between the different iso-
forms of adenylate kinases is a dominant feature of this enzyme
family, important differences have also been shown in previous
studies. In the present study we have identified yet another spe-
cific characteristic of one of the AK isoforms. Our data demonstrate
that human adenylate kinase 5 has two separate functional do-
mains and that both domains have similar adenylate kinase activ-
ity. The advantage of different activities in one single enzyme is not
obvious, but the property of having two transcripts from the AK5
gene, resulting in one or two functional and independently active
proteins, may be physiologically significant. To our knowledge this
is the only AK with this particular feature. In the nucleotide metab-
olism of Trypanosoma brucei there is gene duplication with tandem
repeats of the adenosine kinase gene. However, a comparison with
this parasite may be difficult since it lacks the de novo synthesis
and has a different pathway of nucleotide metabolism .
If this novel mechanism will prove to regulate the level of AK
activity in specific cells will be a subject for future investigations.
This work was supported by grants from the Swedish Cancer
Society and the Swedish Research Council.
 Okajima, T., Fukamizo, T., Goto, S., Fukui, T. and Tanizawa, K. (1998) Exchange
of nucleoside monophosphate-binding domains in adenylate kinase and UMP/
CMP kinase. J. Biochem. 124, 359–367.
 Van Rompay, A.R., Johansson, M. and Karlsson, A. (1999) Phosphorylation of
deoxycytidine analog monophosphates by UMP–CMP kinase: molecular
characterization of the human enzyme. Mol. Pharmacol. 56, 562–569.
 Xu, Y., Johansson, M. and Karlsson, A. (2008) Human UMP–CMP kinase 2, a
novel nucleoside monophosphate kinase localized in mitochondria. J. Biol.
Chem. 283, 1563–1571.
 Janssen, E., de Groof, A., Wijers, M., Fransen, J., Dzeja, P.P., Terzic, A. and
Wieringa, B. (2003) Adenylate kinase 1 deficiency induces molecular and
structural adaptations to support muscle energy metabolism. J. Biol. Chem.
 Fukami-Kobayashi, K., Nosaka, M., Nakazawa, A. and Go, M. (1996) Ancient
divergence of long and short isoforms of adenylate kinase: molecular
evolution of the nucleoside monophosphate kinase family. FEBS Lett. 385,
 Noma, T., Fujisawa, K., Yamashiro, Y., Shinohara, M., Nakazawa, A., Gondo, T.,
Ishihara, T. and Yoshinobu, K. (2001) Structure and expression of human
mitochondrial adenylate kinase targeted to the mitochondrial matrix.
Biochem. J. 358, 225–232.
 Van Rompay, A.R., Johansson, M. and Karlsson, A. (1999) Identification of a
chromosome localization and characterization of the recombinant protein.
Eur. J. Biochem. 261, 509–517.
 Ren, H. et al. (2005) The crystal structure of human adenylate kinase 6: An
adenylate kinase localized to the cell nucleus. Proc. Natl. Acad. Sci. USA 102,
 Fernandez-Gonzalez, A., Kourembanas, S., Wyatt, T.A. and Mitsialis, S.A. (2009)
Mutation of murine adenylate kinase 7 underlies a primary ciliary dyskinesia
phenotype. Am. J. Respir. Cell Mol. Biol. 40, 305–313.
 Dzeja, P. and Terzic, A. (2009) Adenylate kinase and AMP signaling networks:
metabolic monitoring, signal communication and body energy sensing. Int. J.
Mol. Sci. 10, 1729–1772.
 Van Rompay, A.R., Johansson, M. and Karlsson, A. (2000) Phosphorylation of
monophosphate kinases. Pharmacol. Ther. 87, 189–198.
 Lust, J. and Sahud, M.A. (1972) Contributions of anion-exchange thin-layer
chromatography to the study of free nucleotides in tissues. J. Chromatogr. 71,
 Curbo, S., Amiri, M., Foroogh, F., Johansson, M. and Karlsson, A. (2003) The
Drosophila melanogaster UMP–CMP kinase cDNA encodes an N-terminal
mitochondrial import signal. Biochem. Biophys. Res. Commun. 311, 440–
 Pucar, D. et al. (2002) Adenylate kinase AK1 knockout heart: energetics and
functional performance under ischemia-reperfusion. Am. J. Physiol. Heart Circ.
Physiol. 283, H776–H782.
 Pannicke, U. et al. (2009) Reticular dysgenesis (aleukocytosis) is caused by
mutations in the gene encoding mitochondrial adenylate kinase 2. Nat. Genet.
 Lagresle-Peyrou, C. et al. (2009) Human adenylate kinase 2 deficiency causes a
profound hematopoietic defect associated with sensorineural deafness. Nat.
Genet. 41, 106–111.
 Tuzun, E., Rossi, J.E., Karner, S.F., Centurion, A.F. and Dalmau, J. (2007)
Adenylate kinase 5 autoimmunity in treatment refractory limbic encephalitis.
J. Neuroimmunol. 186, 177–180.
 Luscher, A., Onal, P., Schweingruber, A.M. and Maser, P. (2007) Adenosine
kinase of Trypanosoma brucei and its role in susceptibility to adenosine
antimetabolites. Antimicrob. Agents Chemother. 51, 3895–3901.
CDNA cloning, expressionanalysis,
by mammalian nucleoside
N. Solaroli et al./FEBS Letters 583 (2009) 2872–2876