Structures of down syndrome kinases, DYRKs, reveal mechanisms of kinase activation and substrate recognition

Article abstract of DOI:10.1016/j.str.2013.03.012

Dual-specificity tyrosine-(Y)-phosphorylation-regulated kinases (DYRKs) play key roles in brain development, regulation of splicing, and apoptosis, and are potential drug targets for neurodegenerative diseases and cancer. We present crystal structures of one representative member of each DYRK subfamily: DYRK1A with an ATP-mimetic inhibitor and consensus peptide, and DYRK2 including NAPA and DH (DYRK homology) box regions. The current activation model suggests that DYRKs are Ser/Thr kinases that only autophosphorylate the second tyrosine of the activation loop YxY motif during protein translation. The structures explain the roles of this tyrosine and of the DH box in DYRK activation and provide a structural model for DYRK substrate recognition. Phosphorylation of a library of naturally occurring peptides identified substrate motifs that lack proline in the P+1 position, suggesting that DYRK1A is not a strictly proline-directed kinase. Our data also show that DYRK1A wild-type and Y321F mutant retain tyrosine autophosphorylation activity.

Full text of DOI:10.1016/j.str.2013.03.012

Structure
Article
Structures of Down Syndrome Kinases, DYRKs,
Reveal Mechanisms of Kinase Activation
and Substrate Recognition
Meera Soundararajan,1,6 Annette K. Roos,1,7 Pavel Savitsky,1 Panagis Filippakopoulos,1 Arminja N. Kettenbach,3
Jesper V. Olsen,5 Scott A. Gerber,3,4 Jeyanthy Eswaran,1,8 Stefan Knapp,1,2 and Jonathan M. Elkins1,
*
1Structural Genomics Consortium
2Target Discovery Institute
Nuffield Department of Clinical Medicine, University of Oxford, Old Road Campus, Roosevelt Drive, Oxford OX3 7DQ, UK
3Department of Genetics
4Department of Biochemistry
Geisel School of Medicine at Dartmouth, Lebanon, NH 03756, USA
5Department of Proteomics, Novo Nordisk Foundation Center for Protein Research, Copenhagen DK-2200, Denmark
6Present address: School of Life Sciences, Northumbria University, Newcastle upon Tyne, NE1 8ST, UK
7Present address: Department of Cell and Molecular Biology, Uppsala University, Biomedical Center, SE-75124 Uppsala, Sweden
8Present address: Department of Biochemistry and Molecular Biology, George Washington University, 2300 Eye Street, Washington, DC
20037, USA
*Correspondence: jon.elkins@sgc.ox.ac.uk
http://dx.doi.org/10.1016/j.str.2013.03.012
SUMMARY
The best-studied member of the DYRK family is DYRK1A, owing
to its role in the pathology of Down syndrome and the early onset
Dual-specificity tyrosine-(Y)-phosphorylation-regu- of neurodegeneration. DYRK members have been clearly shown
to participate in important signaling pathways that control post-
lated kinases (DYRKs) play key roles in brain devel-
embryonic neurogenesis, developmental processes, cell sur-
vival, differentiation, and death (Arron et al., 2006; Mercer
et al., 2005; Tejedor et al., 1995). In addition, recent studies
opment, regulation of splicing, and apoptosis, and
are potential drug targets for neurodegenerative dis-
eases and cancer. We present crystal structures of
one representative member of each DYRK subfamily:
DYRK1A with an ATP-mimetic inhibitor and
consensus peptide, and DYRK2 including NAPA
and DH (DYRK homology) box regions. The current
show DYRK1A and DYRK2 phosphorylate NFATc, countering
the effect of calcium signaling and maintaining inactive NFATc
(Arron et al., 2006; Gwack et al., 2006; Lee et al., 2009).
The first evidence for the key role of DYRK1A in neural prolif-
eration and neurogenesis of the developing brain was provided
activation model suggests that DYRKs are Ser/Thr
kinases that only autophosphorylate the second
by mutational analysis of the DYRK Drosophila ortholog mini-
brain (mnb), where loss-of-function mutations resulted in
tyrosine of the activation loop YxY motif during pro- reduced brain size (Tejedor et al., 1995). DYRK1A is localized
in the Down syndrome (DS) critical region of chromosome 21
that has been linked to the development of DS phenotypes
tein translation. The structures explain the roles of
this tyrosine and of the DH box in DYRK activation
when triplicated (Delabar et al., 1993; Sinet et al., 1994). Indeed,
triplication of the DYRK1A locus in DS results in overexpression
of DYRK1A in the fetal as well as adult brain and strongly impli-
and provide a structural model for DYRK substrate
recognition. Phosphorylation of a library of naturally
occurring peptides identified substrate motifs that
lack proline in the P+1 position, suggesting that
cates DYRK1A in neurodevelopmental alterations linked to some
DS pathologies and disease predispositions (Dowjat et al.,
2007). These links prompted studies on the role of DYRK1A in
DYRK1A is not a strictly proline-directed kinase.
Our data also show that DYRK1A wild-type and
age-associated neurodegeneration and suggested DYRK1A as
Y321F mutant retain tyrosine autophosphorylation
activity.
a target for the development of inhibitors (Mazur-Kolecka
et al., 2012; Park et al., 2009). The binding modes of the inhibitors
INDY and Harmine in DYRK1A have recently been published
(Ogawa et al., 2010).
INTRODUCTION
Apart from the well-studied DYRK1A isozyme, studies have
provided evidence for the roles of DYRK1B in the development
The dual-specificity tyrosine-phosphorylation-regulated kinases of various sarcomas (Deng et al., 2006) and in skeletal muscle
(DYRKs) are an evolutionarily conserved family of kinases with differentiation (Deng et al., 2003, 2005). DYRK2 is reported to
five human members (DYRK1A, DYRK1B, DYRK2, DYRK3, regulate key developmental and cellular processes such as neu-
and DYRK4). They belong to the CMGC family of serine/threo- rogenesis, cell proliferation, cytokinesis, and cellular differentia-
nine (S/T) kinases and are categorized as class I (DYRK1A and tion (Taira et al., 2007; Woods et al., 2001; Yoshida, 2008).
DYRK1B) and class II (DYRK2, DYRK3, and DYRK4) DYRKs. Notably, DYRK2 may function in DNA damage signaling
986 Structure 21, 986–996, June 4, 2013 ª2013 Elsevier Ltd All rights reserved

Structure
Activation and Substrate Recognition of DYRKs
substrates and provides a model for the structure-based design
of selective DYRK inhibitors.
RESULTS
Structures of the DYRK1A and DYRK2 Catalytic Domain
and DH Box
The crystal structures of DYRK1A and DYRK2 comprising the
catalytic kinase domain and DH box were determined. The
DYRK1A structure was determined from a construct expressing
residues 127–485 of human DYRK1A (National Center for
Biotechnology Information [NCBI] genInfo identifier [gi] number
18765758). Residues Val135–Lys480 comprising the DH box
and kinase domain were resolved in the electron density. The
structure was determined in complex with the ATP-competitive
Figure 1. Domain Arrangement of Human DYRK Family Kinases
The construct boundaries for the crystallized DYRK1A and DYRK2 proteins are
indicated. NLS, nuclear localization signal; PEST, PEST domain. See also
Figure S1.
pathways, because it phosphorylates p53 at Ser46 in response inhibitor (S)-N-(5-(4-amino-2-(3-chlorophenyl)butanamido)-1H-
˚
to DNA damage, which induces cellular apoptosis after geno- indazol-3-yl)benzamide (DJM2005) at 2.40 A resolution (Fig-
toxic stress (Taira et al., 2007). In addition, ataxia telangiectasia ure 2A; Table 1). The inhibitor DJM2005 was kindly provided
mutated was shown to phosphorylate nuclear DYRK2 upon DNA by the laboratory of Kevan Shokat; the chemical structure is
damage, which appeared to enable DYRK2 to protect itself from shown in Figure S2.
degradation that occurs due to its association with MDM2 under
The DYRK2 structure was determined from a construct ex-
normal conditions (Taira et al., 2010). Emerging studies show pressing residues 74–479 of human DYRK2 (NCBI gi number
DYRK2 has important roles in protein proteolysis, proteosomal 4503427). Residues Gly74–Pro470 comprising NAPA1 (N-termi-
degradation, and tumor progression (Varjosalo et al., 2008; nal autophosphorylation accessory 1), NAPA2, DH box, and
Maddika and Chen, 2009; Taira et al., 2012). As for kinase domain were resolved in the electron density, as well as
DYRK3 and DYRK4, their physiological functions remain poorly part of the N-terminal purification tag. The structure was deter-
˚
understood. mined in the absence of inhibitor (apo form) at 2.36 A resolution
All DYRKs contain a conserved catalytic kinase domain pre- (Figure 2B; Table 1).
ceded by the DYRK-characteristic DYRK homology (DH) box
For both DYRK1A and DYRK2 the entire catalytic domain was
(Figure 1A; for a sequence alignment, see Figure S1 available on- well ordered, including a long hairpin-like structure for the N-ter-
line). DYRKs rapidly autoactivate during folding by phosphoryla- minal DH box and an active kinase conformation with a fully or-
tion on the second tyrosine residue of the conserved activation dered activation segment (Figure 2). Mass spectrometry showed
loop YxY motif (Tyr321 of DYRK1A). This tyrosine corresponds that the purified DYRKs were heterogeneously phosphorylated
to the secondary activation loop phosphorylation site in the in solution (data not shown). However, the electron density
TxY motif in MAPKs. It was reported based on studies with maps only showed clear evidence of phosphorylation of
Drosophila melanogaster DYRKs that this phosphorylation event DYRK1A at the second tyrosine of the dual-phosphorylation
occurs in cis while DYRK is still bound to the ribosome, and sub- motif YxY (Tyr321) and double phosphorylation of DYRK2 at
sequently DYRKs lose tyrosine phosphorylation ability and retain Ser159 of the glycine-rich loop and Tyr309 of the activation
only S/T phosphorylation ability (Lochhead et al., 2005). For the loop. The other phosphorylation sites might either have had
human DYRK1A, mutation of Tyr321 or dephosphorylation did low occupancy or were located in unstructured regions of the
not abolish kinase activity (Adayev et al., 2007).
protein.
The DYRK1A and DYRK2 structures superimpose with a
DYRKs were initially assumed to be proline-directed S/T
˚
kinases with specificity for proline and arginine at P+1 and root-mean-square deviation (rmsd) of 1.03 A over 297 Ca atoms
PÀ3 positions, respectively (Himpel et al., 2000). However, (using chain A of the DYRK1A structure). In DYRK1A, the ATP-
further investigations revealed DYRK cellular substrates (e.g., mimetic inhibitor DJM2005 binds to the ATP binding site, form-
synuclein; Kim et al., 2006) with a wide variation in phosphoryla- ing three hydrogen bonds with the hinge backbone and an
tion motifs (Aranda et al., 2011).
additional two hydrogen bonds from the inhibitor’s primary
To understand the molecular mechanism of DYRK1A activa- amine with the side chains of Asn292 and the DFG motif aspar-
tion, the roles of Tyr321 phosphorylation and regulatory ele- tate Asp307 (Figure 2C). There is also an electrostatic interac-
ments located N-terminal to the catalytic domain, as well as tion via an ion (modeled as chloride) linking an inhibitor amide
substrate recognition, we determined the structure of the phos- nitrogen to the backbone nitrogen of Asp307 from the DFG
phorylated DYRK1A and DYRK2 catalytic domain and N-termi- motif, and hydrogen bonding via a water molecule to the back-
nal regulatory DH box sections. The autophosphorylation bone carbonyl of Glu291. There are various favorable hydropho-
behavior of DYRK1A was analyzed, and the substrate specificity bic interactions with DYRK1A active site residues, including at
of DYRK1A, DYRK1B, and DYRK2 was investigated using a the entrance to the ATP site, where the side chain of Tyr243
novel mass spectrometry methodology (Kettenbach et al., packs against the inhibitor’s phenyl ring. All of the DYRK1A res-
2012). The structure of a ternary substrate complex of DYRK1A, idues involved in hydrogen bonding to the inhibitor are
the ATP-mimetic inhibitor DJM2005, and a consensus substrate conserved in DYRK2 (Figure S3); there are, however, some po-
peptide (RARPGT*PALRE) reveals how DYRK1A recognizes tential differences in the hydrophobic interactions, such as the
Structure 21, 986–996, June 4, 2013 ª2013 Elsevier Ltd All rights reserved 987

Structure
Activation and Substrate Recognition of DYRKs
Figure 2. Structures of DYRK1A and DYRK2
(A) Structure of DYRK1A kinase domain and DYRK
homology box with the inhibitor DJM2005 bound in
the ATP binding site. The DH box and CMGC-spe-
cific inset are shown in magenta and the activation
segment in orange.
(B) Similar view of DYRK2 showing the NAPA region
and DH box in magenta.
(C) Active site of DYRK1A with inhibitor DJM2005
bound. Part of strand b1 has been removed for
clarity. The inhibitor is colored yellow, the protein is
colored as in (A), and hydrogen bonds are shown as
dashed red lines.
See also Figure S3.
(D) Correlation of the binding of various kinase in-
hibitors (measured by DTm) to DYRK1A and DYRK2
showing that whereas some inhibitors bind both
proteins, the active site differences allow for
DYRK1A- or DYRK2-specific inhibitors.
in comparison with other CMGC family
members such as CLK1, CLK3 (Bullock
et al., 2009), GSK3b (Dajani et al., 2001),
or MAPKs (Canagarajah et al., 1997). In
DYRK1A, this insert forms an elaborate
subdomain of 40 residues comprising
two short helices followed by an antipar-
allel b sheet that is conserved in vertebrate
DYRK1 family members but not in
Drosophila mnb. In DYRK2, this insert
also forms a distinctive subdomain with
two short helices and three short antipar-
allel b sheets. Along with the structural
divergence between DYRK1A and DYRK2
replacement of Tyr243 with Met233 in DYRK2 as well as differ- in this insertion, this region is also the place of greatest diver-
ences at the back of the pocket and the hydrophobic residue gence among the other DYRK family members.
preceding the DFG motif. Analysis of changes in DYRK1A and
DYRK2 temperature shift values (DTm) in the presence of a set Regulatory Role of the N-Terminal Region
of potential kinase inhibitors showed only weak correlation, Deletion of the entire region of DYRK1A N-terminal to the kinase
and therefore that it is possible to have DYRK1A- or DYRK2- domain (1–148) has been shown to decrease catalytic activity
specific inhibitors, as shown by some of the inhibitors screened (Himpel et al., 2001). In Drosophila, the DH box was required
that give changes in Tm with only DYRK1A or only DYRK2 for phosphorylation of SNR1 by DYRK2 but not by DYRK1 (Kin-
(Figure 2D).
strie et al., 2006). With Drosophila DYRKs, the NAPA regions are
Interestingly, the inhibitor’s primary amine also interacts with a required for the transient intramolecular tyrosine kinase activity
sulfate molecule from the DYRK1A crystallization buffer that is of DYRKs (Kinstrie et al., 2010) and are conserved across a
found in a similar location as an autophosphorylated serine res- wide range of eukaryotes, including for Trypanosoma brucei
idue in DYRK2 (pS159; Figure S3). This sulfate is also bound by DYRK2, where the NAPA1 and NAPA2 regions are required for
the side chains of Asp307 of the DFG motif, Ser169 of the tyrosine autophosphorylation (Han et al., 2012). In the following
glycine-rich loop, and Lys289 of the catalytic loop, and is in a analysis, the DH boxes and NAPA regions are those defined by
similar position as that of a hydrolyzed g-phosphate from ATP Kinstrie et al. (2010).
bound to PKA (Protein Data Bank [PDB] ID code 1RDQ; Yang
In both DYRK1A and DYRK2, the N-terminal region containing
et al., 2004) or a bound phosphate in the structure of Haspin the DH box is positioned on top of the N-terminal lobe of the
with a 5-iodotubercidin ligand (PDB ID code 3IQ7; Eswaran kinase domain and forms a large network of interactions with
et al., 2009). Addition of negatively charged groups to inhibitors all five strands of the N lobe b sheet, providing considerable sta-
to exploit this conserved binding pocket may help inhibitor bilization (Figures 3A and 3B). The most highly conserved resi-
design for some of these kinases.
dues in the DH box are those essential for stabilization of its
The C-terminal lobe reveals several unique features that folded state, in particular the two central tyrosines, Tyr140 and
define the DYRK family. The MAP kinase characteristic inser- Tyr147, in DYRK1A (Figure 3A). This compact folded DH box
tion observed in the C lobe of DYRKs (Figure 2) is extended appears essential for the formation of tertiary structure in the
988 Structure 21, 986–996, June 4, 2013 ª2013 Elsevier Ltd All rights reserved

Structure
Activation and Substrate Recognition of DYRKs
Table 1. Data Collection and Refinement Statistics
DYRK1A-Inhibitora
DYRK1A-Inhibitora-Peptide
2WO6
DYRK2
3K2L
PDB ID code
2VX3
Crystallization conditions
4% (v/v) PEG 300, 0.1 M Li2SO4, 0.2 M sodium formate, 20% (w/v) 1.26 M (NH4)2SO4, 0.2 M Li2SO4,
0.1 M Tris, pH 8.5
PEG 3350, 10% ethylene glycol
0.1 M Tris, pH 8.5
Space group
C2
P65
P42212
No. of molecules in the asymmetric
unit
4
2
1
Unit cell dimensions
˚
a, b, c (A)
264.2, 65.1, 140.3
90.0, 115.44, 90.0
168.4, 168.4, 62.4
90.0, 90.0, 120.0
84.3, 84.3, 148.5
90.0, 90.0, 90.0
a, b, g (ꢀ)
Data Collection
Beamline
SLS X10SA
27.24–2.40 (2.53–2.40)
85,770 (12,458)
3.4 (3.2)
Diamond I02
55.13–2.50 (2.64–2.50)
35,283 (5,093)
7.5 (6.9)
Diamond I03
42.19–2.36 (2.49–2.36)
22,801 (3,275)
6.2 (6.4)
b
˚
Resolution range (A)
Unique observationsb
Average multiplicityb
Completeness (%)b
99.9 (99.9)
100.0 (100.0)
0.18 (0.57)
99.8 (100.0)
0.08 (0.90)
b
Rmerge
Mean (I)/s(I)b
0.10 (0.82)
9.5 (1.9)
10.8 (3.7)
12.2 (2.1)
Refinement
˚
Resolution range (A)
26.00–2.40
0.19, 0.23
53
40.00–2.50
0.19, 0.23
23.1c
42.19–2.36
0.23, 0.29
33.8c
R value, Rfree
2
Mean protein B values (A )
˚
2
Mean ligand B values (A )
˚
46 (inhibitor)
34 (inhibitor)
71 (peptide)
0.014
˚
Rmsd from ideal bond length (A)
0.014
1.52
0.15
96.1
0.014
1.60
0.0
Rmsd from ideal bond angle (ꢀ)
1.53
Ramachandran outliers (%)
Most favored (%)
0.0
96.3
95.0
aThe inhibitor structure is shown in Figure S2.
bValues within parentheses refer to the highest resolution shell.
cResidual after TLS parameterization.
remainder of the N terminus, especially for class II DYRKs, which inactive kinase structures (Figure 3A). Interestingly, the interac-
have NAPA1 and NAPA2 regions (Figure 3C). Although many of tion appears not to be charge dependent, unlike for the interac-
the DH box interactions are conserved between DYRK1A and tion of the N-terminal hairpin of CLK3 (Bullock et al., 2009) with
DYRK2, we observed more hydrogen-bonding interactions in the kinase N-terminal lobe (Figure S4).
the DYRK1A structure. In particular, in DYRK1A, the central tyro-
The DYRK2 NAPA1 and NAPA2 regions, which are separated
sine, Tyr147, interacts with the DYRK1A equivalent of the NAPA2 in sequence, fold together into a small domain that stabilizes the
region, Glu153 and Trp155. These residues are not present in the N-terminal lobe of the kinase domain (Figures 3C and 3D). The
standard NAPA2 region of DYRK2, and DYRK2 does not have an larger NAPA1 region folds around the five residues of the
equivalent interaction between the DH box and NAPA2 regions NAPA2 region. As with the DH box, the most highly conserved
(Figure 3B). Recent evidence suggests phosphorylation of residues (marked in Figure 3D) are those forming the core of
DYRK1A at Tyr145 and Tyr147 may have important regulatory this folded subdomain, in particular His144 and Tyr147 from
roles (Kida et al., 2011). Tyr145 is solvent exposed, but phos- NAPA2. It is notable that the residues on the end of strand b4
phorylation of Tyr147 would change its interactions significantly, (DYRK2: Phe218, Phe220, Arg221), which interact with the
although it is not possible to predict whether this would be favor- NAPA2 region (Figure 3D), are conserved across all human
able or unfavorable, because pTyr147 could maintain interac- DYRKs (Figure S1). For DYRK1A, an early folded intermediate
tions with Arg231 and replace the interactions of Glu153.
is implicated in enabling transient Tyr autophosphorylation in
As well as the stabilization provided by the N-terminal region, cis (Lochhead et al., 2005). The presence of small N-terminal do-
the 11 residues of the DH box itself interact with the loop linking mains (DH/NAPA1/NAPA2) capable of folding independently
aC with b3, providing stabilization to an ‘‘aC-in’’ active kinase and that stabilize the kinase domain may explain how during
conformation by fixing the N-terminal end of aC in position and translation a stabilized and catalytically active conformation
so preventing aC from moving outward, as sometimes seen in can be achieved before translation is complete.
Structure 21, 986–996, June 4, 2013 ª2013 Elsevier Ltd All rights reserved 989

Structure
Activation and Substrate Recognition of DYRKs
Figure 3. The N-Terminal NAPA and DH Box
Regions
The DH box region is in dark blue, the NAPA1 re-
gion for DYRK2 is in green, and the NAPA2 regions
for DYRK1A and DYRK2 are in cyan. Residues in
these motifs that are highly conserved across
DYRK kinases from different species (Kinstrie
et al., 2010) are labeled.
(A) DH box region of DYRK1A.
(B) DH box region of DYRK2.
(C) Overview of the N terminus of DYRK2 showing
NAPA and DH box motifs.
(D) NAPA1 and NAPA2 regions of DYRK2 showing
their folded, assembled state.
See also Figure S4.
there is an interaction with aC that is inde-
pendent of phosphorylation; neverthe-
less, the direct contribution of DYRK1A
Tyr319 toward catalytic activity and tyro-
sine autophosphorylation has been found
to be negligible (Himpel et al., 2001). The
phosphate moiety of the second YxY
motif tyrosine is therefore the major acti-
The DYRK Activation Segment Is Stabilized by Tyr
Phosphorylation
vation loop phosphorylation site for DYRKs; it forms similar inter-
actions as those of the secondary phosphorylation site of ERK2
The structures of DYRK1A and DYRK2 both show a completely at Tyr185.
ordered activation segment in a similar conformation (Figure 2).
The second tyrosine of the YxY dual-phosphorylation motif DYRK1A Autophosphorylates Serine and Threonine as
(DYRK1A: Tyr321; DYRK2A: Tyr309) is the main mediator of a Well as Tyrosine In Vitro
network of interactions that stabilize the active conformation, The intact mass spectra of purified DYRK1A and DYRK2 clearly
including with Arg325 and Arg328 (numbering for DYRK1A) indicated multiple phosphorylation states. We coexpressed
that precede the APE motif, and with the backbone carbonyl of DYRK1A with l-phosphatase in bacteria, yielding DYRK1A
the catalytically important Gln323 (Figure 4A).
singly phosphorylated at Tyr321. This protein was subjected to
Related CMGC kinases such as ERK2 have a TxY motif, and autophosphorylation in vitro, generating up to three additional
phosphorylation of the first residue of this motif, the threonine, al- phosphorylation sites after the reaction ran to completion, as
lows formation of salt bridges with the arginine of the usually measured by mass spectrometry. The resulting sites were map-
conserved activation segment HRD motif, neutralizing its charge ped by liquid chromatography-tandem mass spectrometry
(Canagarajah et al., 1997; Dajani et al., 2001). However, in all (LC-MS/MS) (Table 2; Figure S5); sites were identified at
DYRK kinases, the HRD arginine is replaced by a cysteine, sug- Tyr140 and Tyr159 in the DH box region, Tyr177 in b2 of the
gesting that phosphorylation of the primary phosphorylation site N-terminal lobe, Ser310 immediately C-terminal of the DFG
is not required for activity. In both DYRKs, the HCD motif motif, Tyr319 of the YxY motif, and Tyr449 located near the
cysteine (DYRK1A: Cys286; DYRK2: Cys274) is within range of C terminus of the molecule (Figure 5C). To rule out that tyrosine
potential disulfide-bond formation with a cysteine at the begin- autophosphorylation was due to the short construct used for
˚
ning of the activation loop (DYRK1A: Cys312, 4.4 A distant; crystallization, we generated additional constructs of DYRK1A
˚
DYRK2: Cys300, 4.2 A distant), raising the possibility that 1–485 and 37–485 and again looked for autophosphorylation
DYRK kinase activity might be regulated by the cellular redox sites. LC-MS/MS confirmed Tyr autophosphorylation on
state (Figure 4A). No disulfide bonds were observed in the Tyr111, which was not present in the shorter constructs. The
crystallized proteins, which were prepared under reducing mapped sites demonstrate the capability of DYRK1A to auto-
conditions.
phosphorylate on tyrosine after the formation of mature, folded
Comparison with the diphosphorylated ERK2 structure (PDB protein in vitro, which is contradictory to previous data based
ID code 2ERK) reveals only small differences in the overall acti- on experiments with D. melanogaster DYRKs, which reported
vation loop conformation, mainly due to sequence and loop tyrosine phosphorylation as a one-time-only event during trans-
length variations (Figures 4A and 4B). However, the positions lation (Lochhead et al., 2005). However, these data agree with re-
of the TxY/YxY motifs are well conserved. The primary phos- ports of weak Tyr phosphorylation observed by Adayev et al.
phorylation site pT183 in ERK2 links aC (Arg68) with the activa- (2007) using a specific pTyr antibody. The distant location of
tion segment and with the catalytic loop, the HRD motif. In many sites from the active site also suggests that most of these
DYRKs, the first tyrosine residue (DYRK1A: Tyr319; DYRK2A: phosphorylations were carried out in trans rather than in cis.
Tyr307) forms a hydrogen bond with Gln199 from aC. Therefore, Although all the identified phosphorylation sites are conserved
990 Structure 21, 986–996, June 4, 2013 ª2013 Elsevier Ltd All rights reserved

Structure
Activation and Substrate Recognition of DYRKs
Figure 4. Activation Loop and Active State
Stabilization
Comparison of the activation segment arrange-
ments of (A) DYRK1A, activation segment in
orange, and (B) dual-phosphorylated ERK2, acti-
vation segment in green. The ERK2 structure is
from PDB ID code 2ERK. Both DYRK1A and ERK2
have a completely ordered activation loop and
glycine-rich loop, and active aC conformations.
The activation loop in dual-phosphorylated active
ERK2 forms an extensive hydrogen-bonding
network around pT183. Phosphorylated Y185 is
also stabilized through an extensive interaction
network that is similar to the pY321 network
formed by DYRK1A.
within the DYRK family and also across various species, the Tyr preference for PÀ3 arginine was shown to be a feature of
phosphorylation is weak, and any physiological significance of DYRK1A but not DYRK2 or DYRK4 (Papadopoulos et al., 2011).
phosphorylation at these phosphorylation sites is yet to be es-
The substrate specificities identified from HeLa cell extracts
tablished. Interestingly, phosphorylation on Tyr145 and Tyr147 correlate well with previous studies that have identified DYRK1A
has recently been identified as a modification that determines substrates such as Tau (Ryoo et al., 2007; Liu et al., 2008), am-
nuclear localization of DYRK1A in neurons (Kida et al., 2011).
phiphysin (Murakami et al., 2006), and caspase 9 (Seifert et al.,
To analyze the importance of Tyr321 phosphorylation for 2008) that conform to the above definition (proline at P+1 posi-
activity, we also measured autophosphorylation kinetics for a tion and arginine at PÀ positions), whereas other studies report
DYRK1A Y321F mutant. This mutant was not phosphorylated substrates such as spliceosomal protein SF3b1 (de Graaf et al.,
after coexpression with l-phosphatase, but under the same 2006) that contain proline in the P+1 position but not basophilic
in vitro reaction conditions we observed autophosphorylation residues at PÀ positions. Furthermore, DYRK1A substrates
activity (Figures 5A and 5B). To verify the activity of our DYRK1A a-synuclein (Kim et al., 2006) and p53 (Park et al., 2010) do not
126–485 constructs and the suitability of the consensus peptide contain either of these substrate specificity determinants, indi-
used for cocrystallization (see below), we measured activity of cating the flexibility of DYRK1A in substrate recognition.
the 126–485 wild-type and Y321F mutant in an in vitro assay (Fig-
ure S6). Y321F had 72% of wild-type activity against the Ternary Complex Structure of DYRK1A Comprising
consensus peptide, in general agreement with previous activity Consensus Substrate Peptide
measurements on other DYRK1A constructs (Himpel et al., To further explore the substrate recognition of DYRKs, we deter-
2001). To assess the importance of pTyr321 for DYRK1A stabil- mined the crystal structure of the ternary complex of DYRK1A
ity, we measured DTm data for DYRK1A 126-485 wild-type and with the consensus substrate peptide RARPT*PALRE and
Y321F; the Y321F mutant melted at an approximately 12ꢀC lower DJM2005 (Figure 6B). Although several consensus peptides
temperature compared to wild-type (Figure 5D).
were used in cocrystallization experiments, crystals were only
obtained with this substrate peptide. Electron density for 8 of
the 11 peptide amino acids was visible, from PÀ4 Ala to P+3
Leu. The peptide binds in an extended conformation in the cleft
The Phosphorylation Substrate Recognition Motifs of
DYRKs
Initially, DYRKs were considered to be proline-directed kinases formed by the activation segment and catalytic loop. There were
with a similar recognition motif as ERKs. However, subsequent no major protein conformational changes upon peptide binding.
biochemical studies identified substrates with a variety of recog- The hydroxyl group of the P0 threonine forms hydrogen bonds
nition sequences. We employed an in vitro kinase substrate with the catalytic loop aspartate (Asp287) and the highly
screening method using naturally occurring substrates from conserved Lys289 of the catalytic loop flanking region and
HeLa cells (Kettenbach et al., 2012) to identify the substrate Ser324 of the activation loop (Figure 6D). The P+2 leucine and
recognition motifs for class I and II DYRKs, DYRK1A and P+3 alanine bind against a hydrophobic pocket formed by
DYRK2. The results showed that DYRK1A phosphorylates sub- Phe196 from aC and Phe170 from the glycine-rich loop, explain-
strates exclusively on serine or threonine residues, on peptides ing the preference for hydrophobic residues at these substrate
that have smaller hydrophobic residues at the P+1 position (Fig- positions. The P+1 proline was bound adjacent to the catalyti-
ure 6A). Another notable preference observed was arginine at cally important Gln323 (Wiechmann et al., 2003) and also against
PÀ2 to PÀ4, positions poised for occupying the C lobe electro- the aromatic ring of pTyr321 (Figure 6E). The small size and lack
negative pocket. In the case of DYRK2, proline is strongly of charge in this pocket explains why only substrates with small
preferred at P+1, recognizing S/TP motifs and S/TPxP motifs, aliphatic residues at P+1 were acceptable. The PÀ3 arginine was
but the arginine preference at PÀ2 to PÀ4 is not as strong as bound in a negatively charged pocket formed by Glu291, Tyr327,
with DYRK1A (Figure S7). Arginine at PÀ3 has been previously Tyr246, and Glu353 (Figures 6E and 6F). Interestingly, it faces to-
shown to be more favorable than at PÀ2 on a small selection ward the activation segment in the C lobe, not the glycine-rich
of artificial peptides with DYRK2 and DYRK3 (Campbell loop, but stabilizes the aD helix through the interactions with
and Proud, 2002). On a larger set of nonendogenous peptides, Tyr246 on aD. However, the binding pocket is extended and
Structure 21, 986–996, June 4, 2013 ª2013 Elsevier Ltd All rights reserved 991

Structure
Activation and Substrate Recognition of DYRKs
protein. In vitro autophosphorylation at Tyr111 has been re-
ported previously (Himpel et al., 2001), but mutation of this res-
idue was shown to have no influence on catalytic activity.
Tyr111 is located N-terminal to the DH box and was not included
in the construct used for the crystal structure. The importance of
phosphorylation of Tyr321 and Tyr319 for activity has been re-
ported by several groups (Himpel et al., 2001; Lochhead et al.,
2005). These studies showed that Tyr321 is the main phosphor-
ylation site observed in recombinant protein expressed in both
bacteria and eukaryotic cells. Mutation Y321F dramatically re-
duces catalytic activity, whereas Y319F does not alter activity
(Adayev et al., 2007; Wiechmann et al., 2003). Some other muta-
tions (e.g., double mutants Y319Q/Y321Q and Y319H/Y321H)
also retain considerable activity (Adayev et al., 2007). The
DYRK1A structure provides explanations for these observations,
as the mutations to polar side chains would still allow the activa-
tion segment to retain favorable polar interactions with residues
that coordinate pTyr321 (Arg325, Arg328, Glu366). By analogy,
Table 2. Phosphopeptides Identified following
Autophosphorylation of DYRK1A
Sequence
Residue Range
176–188
305–317
318–325
318–325
398–406
438–451
135–150
155–167
155–167
106–114
ApYDRVEQEWVAIK
IVDFGpSSCQLGQR
IpYQpYIQSR
IYQpYIQSR
LPDGpTWNLK
RAGESGHTVADpYLK
VYNDGpYDDDNYDYIVK
WM(ox)DRpYEIDSLIGK
WMDRpYEIDSLIGK
HINEVpYYAKa
Related to Figure S5.
aOnly observed for DYRK1A 1–485 or 37–485, not for DYRK1A 127–485.
so it may also accommodate substrate arginines located in the the activity of nonphosphorylated Tyr321 (Adayev et al., 2007)
À2 and À4 positions, again fitting with the observed DYRK1A is also explained by the continued ability of Tyr321 to retain
substrate profile.
some favorable polar interactions that would stabilize an active
conformation of the activation loop.
DISCUSSION
Substrate peptide profiling did not reveal any tyrosine phos-
phorylation, suggesting that tyrosine kinase activity is limited to
It has been established that DYRKs autoactivate shortly after autophosphorylation events. Comparison of the DYRK1A sub-
translation through autophosphorylation of the YxpY motif of strate complex with substrate-bound tyrosine kinase structures
the activation segment (Becker et al., 1998; Becker and Joost, shows little similarity, and the substrate binding site in DYRK1A
1999; Lochhead et al., 2005). The DYRK structures revealed has the typical appearance of those in S/T kinases. Kinase auto-
that phosphorylation of Tyr321 stabilizes the activation segment phosphorylation has been shown to frequently target noncon-
in a similar way to that observed for phosphorylated ERK2 (Can- sensus sequences (Pike et al., 2008; Oliver et al., 2007). More-
agarajah et al., 1997). The rapid autoactivation of DYRK kinases over, our data revealed that DYRK1A S/T phosphorylation
immediately after translation raises the question of how DYRK activity is not stringently proline directed and that substrates
activity is regulated. Mouse models and truncations of DYRK1A with small hydrophobic residues such as valine or alanine in
in patients with microcephaly demonstrated that deregulation of the P+1 position can be recognized.
DYRK1A activity has severe phenotypic consequences (Fotaki
Finally, as well as explaining the observed substrate specific-
et al., 2002; Møller et al., 2008). Several reports point to transcrip- ities, the presented structures will serve as a model for the devel-
tional regulation, and the presence of PEST sequences (Figure 1) opment of more potent and selective ligands that might find
suggests a quick turnover (Maenz et al., 2008; Arron et al., 2006). application in the treatment of neurodegenerative diseases.
The current model suggests that Tyr321 phosphorylation is a
‘‘one-time-only’’ event that happens during maturation of
DYRK1A while bound to the ribosome. For class II DYRKs
EXPERIMENTAL PROCEDURES
Cloning
(DYRK2, 3, and 4), this process requires the N-terminal NAPA
regions, which are absent/modified in class I DYRKs (DYRK1A
and 1B). The DYRK2 structure shows how the DH box and
NAPA regions bind across the N-terminal lobe of the kinase
domain, via significant conserved interactions at the kinase
domain loops between b3 and aC and between b4 and b5.
Thus, the DH box and NAPA regions come together to form a
domain that can fold before translation of the kinase domain is
complete and assist in the folding of the kinase domain. Presum-
ably, this enables a partially folded intermediate with the previ-
ously observed transient tyrosine kinase capability.
DNA for DYRK1A residues 127–485 (NCBI gi number 18765758) or DYRK2A
residues 74–479 (NCBI gi number 4503427) was PCR amplified and subcloned
into a pET-based vector carrying kanamycin resistance, pNIC28-Bsa4 (Gen-
Bank accession number EF198106), using ligation-independent cloning. The
resulting plasmids expressed the kinase domains of DYRK1A or DYRK2,
with an N-terminal hexahistidine tag and tobacco etch virus protease tag
cleavage site (extension MHHHHHHSSGVDLGTENLYFQ*SM-). DYRK1A
1–485 and 37–485 constructs were generated similarly. DYRK1A 127–485
Y321F and K188R mutants were generated by the overlapping PCR product
method.
Expression and Purification
However, our data, with human enzymes as opposed to the
D. melanogaster enzymes used previously (Lochhead et al.,
2005), also showed that tyrosine autophosphorylation can occur
weakly in vitro and is not restricted to the Tyr321 site (DYRK1A)
and that this activity is independent of Tyr321 phosphorylation,
although a possible caveat is that the constructs used were
the same as those used for crystallization, and not the full-length
Constructs were used to express protein in Escherichia coli BL21 (DE3) cells,
and protein was purified using standard methods (see Supplemental Experi-
mental Procedures).
Crystallization and Data Collection
All crystals were obtained using the sitting-drop vapor-diffusion method
at 4ꢀC. Data collection statistics and crystallization conditions can be
992 Structure 21, 986–996, June 4, 2013 ª2013 Elsevier Ltd All rights reserved

Structure
Activation and Substrate Recognition of DYRKs
Figure 5. Phosphomapping and Autophos-
phorylation
(A) Autophosphorylation kinetics of DYRK1A
Y321F showing electrospray ionization-MS
spectra recorded after 0 hr, 4 hr, and overnight.
(B) Phosphorylation capability of DYRK1A Y321F
mutant. The top panel shows a western blot of
DYRK1A Y321F autophosphorylation probed by
anti-phosphotyrosine antibody after the reaction
times indicated. The bottom panel shows a quan-
titative control with equal amounts of sample run
on the gel. Related to Figure S6.
(C) The autophosphorylation sites mapped for
wild-type DYRK1A are shown on the structure as
green sticks.
(D) Thermal unfolding of DYRK1A wild-type and
Y321F mutant.
See also Figure S6.
(Wako) was added 1:100 (w/w) and the lysates
were digested for 4 hr at room temperature. The
resulting peptide mixtures were diluted 4-fold
with deionized water to achieve
a final urea
concentration below 2 M. Trypsin (modified
sequencing grade; Promega) was added 1:100
(w/w) and the sample was digested overnight.
Trypsin and Lys-C activity was quenched by acid-
found in Table 1. More detail is available in Supplemental Experimental
Procedures.
ification of the reaction mixtures with a 10% TFA solution to pH ꢁ2. The
peptide mixture was desalted and concentrated on a reverse-phase C18
StageTip (Rappsilber et al., 2007) and eluted with 2 3 20 ml of 60% acetonitrile
in 0.3% TFA.
Structure Determination
All diffraction data were indexed and integrated using MOSFLM (Leslie and
Powell, 2007) and scaled using SCALA (Evans, 2006). All models were refined
with REFMAC5 (Murshudov et al., 2011).
In-Gel Digestion
The excised gel plugs with the DYRK1A protein were digested in situ with
trypsin as previously described (Shevchenko et al., 2006).
The DYRK1A structure was solved by molecular replacement using Phaser
(McCoy et al., 2007) and a search ensemble of the coordinates from two CLK
kinases (PDB ID codes 2EXE and 1Z57). Four molecules were present in the
asymmetric unit and, after NCS averaging and density modification in dm
(Cowtan, 1994), the resulting phases could be utilized in ARP/wARP (Langer
et al., 2008) to autobuild the main parts of one of the molecules in the asym-
metric unit. After further model building in Coot (Emsley et al., 2010), this mole-
cule was used to generate the other three molecules for restrained refinement
with tight NCS restraints. Rebuilding and refinement (including TLS parame-
ters) resulted in the final model.
Phosphopeptide Enrichment and Analysis
Phosphopeptides were enriched using titansphere chromatography as
described previously (Olsen et al., 2006), and analyzed by online nanoflow
LC-MS/MS as described previously (Olsen et al., 2006) with
a few
modifications. More detail is available in Supplemental Experimental
Procedures.
Kinase Assays
The DYRK1A peptide complex and the DYRK2 structure were both solved
by molecular replacement using Phaser, with the structure of the inhibitor-
bound DYRK1A as a search molecule.
Peptide substrates were chemically synthesized by Thermo Peptides or
Generon. The phosphorylation reactions were measured using
a spec-
trophotometric assay (Adams et al., 1995) in which ADP production is
coupled to NADH oxidation by pyruvate kinase and lactate dehydroge-
nase (LDH). The reaction was followed by the decrease in NADH
fluorescence (excitation 350 nm, emission 460 nm). The assay mixture
contained 30 U/ml LDH, 12 U/ml pyruvate kinase, 1 mM phosphoenol-
pyruvate, 0.2 mM NADH, 25 mM HEPES (pH 7.5), 150 mM NaCl, 5 mM
MgCl2, and DYRK1A (wild-type or mutants). The concentration of the
consensus substrate peptide (RARPGTPALRE) was 0.25 mM. After
incubation for 1–5 min, the reaction was initiated by the simultaneous
addition of 1 mM ATP. The initial reaction rate was used to compare
activities of wild-type and mutants. Control reactions in the absence of
Analysis of DTm upon Inhibitor Binding
Changes in Tm caused by small-molecule binding were correlated for 433
compounds that caused an increase in Tm of >2ꢀC for either DYRK1A or
DYRK2 and for which a measurement was available for both proteins. The
measurements were made according to established protocols (Niesen et al.,
2007).
Autophosphorylation
DYRK1A proteins were mixed with ATP (1 mM) and Mg2+ (2 mM) and incubated
at room temperature. Western blot analysis of phosphotyrosine was per-
formed using rabbit anti-pTyr antibody (Cell Signaling Technology).
substrate were used to detect ATPase activity for basal correction.
A
K188R mutant was also measured to control for any peptide-stimulated
ATPase activity.
In-Solution Digestion
The DYRK1A protein from in vitro autophosphorylation was diluted in 100 ml of
an 8 M urea buffer (6 M urea, 2 M thiourea in 10 mM HEPES [pH 8]). The pro-
tein was reduced for 30 min at room temperature with 1 mM dithiothreitol and
then alkylated for 15 min by 5.5 mM iodoacetamide. Endoproteinase Lys-C
Determination of Peptide Phosphorylation Specificity
Peptides derived from HeLa cell-lysate digests that were phosphorylated
by DYRKs were identified as in published procedures (Kettenbach et al.,
2012).
Structure 21, 986–996, June 4, 2013 ª2013 Elsevier Ltd All rights reserved 993

Structure
Activation and Substrate Recognition of DYRKs
Figure 6. Substrate Binding of DYRK1A
(A) Representative panels of DYRK1A substrate specificity defining peptide residues identified using in vivo isolation. Each panel represents a separate clustering
of peptide sequences, with the most commonly observed residues at each position at the top of each letter stack. Within each clustering, the fraction of the height
occupied by each residue represents its predominance at that position.
(B) Ternary complex of DYRK1A substrate peptide and inhibitor DJM2005 bound to DYRK1A. The substrate peptide is shown bound between the two lobes of the
kinase in the binding cleft extending from the ATP site toward helix aC. The DH box and kinase domain of DYRK1A are shown as an electrostatic surface
representation, with the substrate peptide in white balls and sticks and residues of substrates labeled with reference to the phosphoacceptor residue (threonine
T0).
(C) Electron density map of the DYRK1A substrate peptide for the 8 out of 11 residues for which the density was visible in the structure, numbered with respect to
phosphoacceptor threonine T0.
(D) Close-up view of the atomic arrangement around the phosphoacceptor residue threonine T0.
(E) Close-up view of the arginine binding pocket of DYRK1A, where the arginine at position À3 of the substrate binds to negatively charged residues of the C lobe
(red-colored surface) and forms an extensive bonding network with residues from the activation loop, aD and aF.
(F) Stick representation similar to (E).
See also Figure S7.
SUPPLEMENTAL INFORMATION
REFERENCES
Supplemental Information includes seven figures and Supplemental Experi-
mental Procedures and can be found with this article online at http://dx.doi.
org/10.1016/j.str.2013.03.012.
Adams, J.A., McGlone, M.L., Gibson, R., and Taylor, S.S. (1995).
Phosphorylation modulates catalytic function and regulation in the cAMP-
dependent protein kinase. Biochemistry 34, 2447–2454.
Adayev, T., Chen-Hwang, M.-C., Murakami, N., Lee, E., Bolton, D.C., and
Hwang, Y.-W. (2007). Dual-specificity tyrosine phosphorylation-regulated
kinase 1A does not require tyrosine phosphorylation for activity in vitro.
Biochemistry 46, 7614–7624.
ACKNOWLEDGMENTS
We thank Prof. Walter Becker, Institute of Pharmacology and Toxicology, Aa-
chen University, for discussions. The Structural Genomics Consortium is a
registered charity (number 1097737) that receives funds from the Canadian In-
stitutes for Health Research, the Canada Foundation for Innovation, Genome
Canada, GlaxoSmithKline, Pfizer, Eli Lilly, Takeda, AbbVie, the Novartis
Research Foundation, the Ontario Ministry of Research and Innovation, and
the Wellcome Trust (092809/Z/10/Z).
Aranda, S., Laguna, A., and de la Luna, S. (2011). DYRK family of protein
kinases: evolutionary relationships, biochemical properties, and functional
roles. FASEB J. 25, 449–462.
Arron, J.R., Winslow, M.M., Polleri, A., Chang, C.-P., Wu, H., Gao, X., Neilson,
J.R., Chen, L., Heit, J.J., Kim, S.K., et al. (2006). NFAT dysregulation by
increased dosage of DSCR1 and DYRK1A on chromosome 21. Nature 441,
595–600.
Received: August 30, 2012
Revised: February 28, 2013
Accepted: March 19, 2013
Published: May 9, 2013
Becker, W., and Joost, H.G. (1999). Structural and functional characteristics of
Dyrk, a novel subfamily of protein kinases with dual specificity. Prog. Nucleic
Acid Res. Mol. Biol. 62, 1–17.
994 Structure 21, 986–996, June 4, 2013 ª2013 Elsevier Ltd All rights reserved

Structure
Activation and Substrate Recognition of DYRKs
Becker, W., Weber, Y., Wetzel, K., Eirmbter, K., Tejedor, F.J., and Joost, H.G.
(1998). Sequence characteristics, subcellular localization, and substrate spec-
ificity of DYRK-related kinases, a novel family of dual specificity protein
kinases. J. Biol. Chem. 273, 25893–25902.
(2001). Identification of the autophosphorylation sites and characterization of
their effects in the protein kinase DYRK1A. Biochem. J. 359, 497–505.
Kettenbach, A.N., Wang, T., Faherty, B.K., Madden, D.R., Knapp, S., Bailey-
Kellogg, C., and Gerber, S.A. (2012). Rapid determination of multiple linear
kinase substrate motifs by mass spectrometry. Chem. Biol. 19, 608–618.
´
Bullock, A.N., Das, S., Debreczeni, J.E., Rellos, P., Fedorov, O., Niesen, F.H.,
Guo, K., Papagrigoriou, E., Amos, A.L., Cho, S., et al. (2009). Kinase domain
insertions define distinct roles of CLK kinases in SR protein phosphorylation.
Structure 17, 352–362.
Kida, E., Walus, M., Jarza˛ bek, K., Palminiello, S., Albertini, G., Rabe, A.,
Hwang, Y.W., and Golabek, A.A. (2011). Form of dual-specificity tyrosine-
(Y)-phosphorylation-regulated kinase 1A nonphosphorylated at tyrosine 145
and 147 is enriched in the nuclei of astroglial cells, adult hippocampal progen-
itors, and some cholinergic axon terminals. Neuroscience 195, 112–127.
Campbell, L.E., and Proud, C.G. (2002). Differing substrate specificities of
members of the DYRK family of arginine-directed protein kinases. FEBS
Lett. 510, 31–36.
Kim, E.J., Sung, J.Y., Lee, H.J., Rhim, H., Hasegawa, M., Iwatsubo, T., Min,
D.S., Kim, J., Paik, S.R., and Chung, K.C. (2006). Dyrk1A phosphorylates
a-synuclein and enhances intracellular inclusion formation. J. Biol. Chem.
281, 33250–33257.
Canagarajah, B.J., Khokhlatchev, A., Cobb, M.H., and Goldsmith, E.J. (1997).
Activation mechanism of the MAP kinase ERK2 by dual phosphorylation. Cell
90, 859–869.
Cowtan, K. (1994). ‘dm’: an automated procedure for phase improvement by
density modification. Joint CCP4 and ESF-EACBM Newsletter on Protein
Crystallography 31, 34–38.
Kinstrie, R., Lochhead, P.A., Sibbet, G., Morrice, N., and Cleghon, V. (2006).
dDYRK2 and Minibrain interact with the chromatin remodelling factors SNR1
and TRX. Biochem. J. 398, 45–54.
Dajani, R., Fraser, E., Roe, S.M., Young, N., Good, V., Dale, T.C., and Pearl,
L.H. (2001). Crystal structure of glycogen synthase kinase 3b: structural basis
for phosphate-primed substrate specificity and autoinhibition. Cell 105,
721–732.
Kinstrie, R., Luebbering, N., Miranda-Saavedra, D., Sibbet, G., Han, J.,
Lochhead, P.A., and Cleghon, V. (2010). Characterization of a domain that
transiently converts class 2 DYRKs into intramolecular tyrosine kinases. Sci.
Signal. 3, ra16.
de Graaf, K., Czajkowska, H., Rottmann, S., Packman, L.C., Lilischkis, R.,
Lu¨ scher, B., and Becker, W. (2006). The protein kinase DYRK1A phosphory-
lates the splicing factor SF3b1/SAP155 at Thr434, a novel in vivo phosphory-
lation site. BMC Biochem. 7, 7.
Langer, G., Cohen, S.X., Lamzin, V.S., and Perrakis, A. (2008). Automated
macromolecular model building for X-ray crystallography using ARP/wARP
version 7. Nat. Protoc. 3, 1171–1179.
Lee, Y., Ha, J., Kim, H.J., Kim, Y.-S., Chang, E.-J., Song, W.-J., and Kim, H.-H.
(2009). Negative feedback inhibition of NFATc1 by DYRK1A regulates bone
homeostasis. J. Biol. Chem. 284, 33343–33351.
Delabar, J.M., Theophile, D., Rahmani, Z., Chettouh, Z., Blouin, J.L., Prieur, M.,
Noel, B., and Sinet, P.M. (1993). Molecular mapping of twenty-four features of
Down syndrome on chromosome 21. Eur. J. Hum. Genet. 1, 114–124.
Leslie, A.G.W., and Powell, H.R. (2007). Processing diffraction data with
mosflm. In Evolving Methods for Macromolecular Crystallography, Vol. 245,
R.J. Read and J.L. Sussman, eds. (Berlin: Springer), pp. 41–51.
Deng, X., Ewton, D.Z., Pawlikowski, B., Maimone, M., and Friedman, E. (2003).
Mirk/dyrk1B is a Rho-induced kinase active in skeletal muscle differentiation.
J. Biol. Chem. 278, 41347–41354.
Liu, F., Liang, Z., Wegiel, J., Hwang, Y.W., Iqbal, K., Grundke-Iqbal, I.,
Ramakrishna, N., and Gong, C.X. (2008). Overexpression of Dyrk1A contrib-
utes to neurofibrillary degeneration in Down syndrome. FASEB J. 22, 3224–
3233.
Deng, X., Ewton, D.Z., Mercer, S.E., and Friedman, E. (2005). Mirk/dyrk1B
decreases the nuclear accumulation of class II histone deacetylases during
skeletal muscle differentiation. J. Biol. Chem. 280, 4894–4905.
Deng, X., Ewton, D.Z., Li, S., Naqvi, A., Mercer, S.E., Landas, S., and
Friedman, E. (2006). The kinase Mirk/Dyrk1B mediates cell survival in pancre-
atic ductal adenocarcinoma. Cancer Res. 66, 4149–4158.
Lochhead, P.A., Sibbet, G., Morrice, N., and Cleghon, V. (2005). Activation-
loop autophosphorylation is mediated by a novel transitional intermediate
form of DYRKs. Cell 121, 925–936.
Dowjat, W.K., Adayev, T., Kuchna, I., Nowicki, K., Palminiello, S., Hwang,
Y.W., and Wegiel, J. (2007). Trisomy-driven overexpression of DYRK1A kinase
in the brain of subjects with Down syndrome. Neurosci. Lett. 413, 77–81.
Maddika, S., and Chen, J. (2009). Protein kinase DYRK2 is a scaffold that
facilitates assembly of an E3 ligase. Nat. Cell Biol. 11, 409–419.
Maenz, B., Hekerman, P., Vela, E.M., Galceran, J., and Becker, W. (2008).
Characterization of the human DYRK1A promoter and its regulation by the
transcription factor E2F1. BMC Mol. Biol. 9, 30.
Emsley, P., Lohkamp, B., Scott, W.G., and Cowtan, K. (2010). Features and
development of Coot. Acta Crystallogr. D Biol. Crystallogr. 66, 486–501.
Eswaran, J., Patnaik, D., Filippakopoulos, P., Wang, F., Stein, R.L., Murray,
J.W., Higgins, J.M.G., and Knapp, S. (2009). Structure and functional charac-
terization of the atypical human kinase haspin. Proc. Natl. Acad. Sci. USA 106,
20198–20203.
Mazur-Kolecka, B., Golabek, A., Kida, E., Rabe, A., Hwang, Y.-W., Adayev, T.,
Wegiel, J., Flory, M., Kaczmarski, W., Marchi, E., and Frackowiak, J. (2012).
Effect of DYRK1A activity inhibition on development of neuronal progenitors
isolated from Ts65Dn mice. J. Neurosci. Res. 90, 999–1010.
Evans, P. (2006). Scaling and assessment of data quality. Acta Crystallogr. D
McCoy, A.J., Grosse-Kunstleve, R.W., Adams, P.D., Winn, M.D., Storoni, L.C.,
and Read, R.J. (2007). Phaser crystallographic software. J. Appl. Crystallogr.
40, 658–674.
Biol. Crystallogr. 62, 72–82.
Fotaki, V., Dierssen, M., Alca´ ntara, S., Martı´nez, S., Martı´, E., Casas, C., Visa,
J., Soriano, E., Estivill, X., and Arbone´ s, M.L. (2002). Dyrk1A haploinsufficiency
affects viability and causes developmental delay and abnormal brain
morphology in mice. Mol. Cell. Biol. 22, 6636–6647.
Mercer, S.E., Ewton, D.Z., Deng, X., Lim, S., Mazur, T.R., and Friedman, E.
(2005). Mirk/Dyrk1B mediates survival during the differentiation of C2C12
myoblasts. J. Biol. Chem. 280, 25788–25801.
Gwack, Y., Sharma, S., Nardone, J., Tanasa, B., Iuga, A., Srikanth, S.,
Okamura, H., Bolton, D., Feske, S., Hogan, P.G., and Rao, A. (2006). A
genome-wide Drosophila RNAi screen identifies DYRK-family kinases as reg-
ulators of NFAT. Nature 441, 646–650.
Møller, R.S., Ku¨ bart, S., Hoeltzenbein, M., Heye, B., Vogel, I., Hansen, C.P.,
Menzel, C., Ullmann, R., Tommerup, N., Ropers, H.-H., et al. (2008).
Truncation of the Down syndrome candidate gene DYRK1A in two unrelated
patients with microcephaly. Am. J. Hum. Genet. 82, 1165–1170.
Han, J., Miranda-Saavedra, D., Luebbering, N., Singh, A., Sibbet, G.,
Ferguson, M.A.J., and Cleghon, V. (2012). Deep evolutionary conservation of
an intramolecular protein kinase activation mechanism. PLoS One 7, e29702.
Murakami, N., Xie, W., Lu, R.C., Chen-Hwang, M.-C., Wieraszko, A., and
Hwang, Y.W. (2006). Phosphorylation of amphiphysin I by minibrain kinase/
dual-specificity tyrosine phosphorylation-regulated kinase, a kinase impli-
cated in Down syndrome. J. Biol. Chem. 281, 23712–23724.
Himpel, S., Tegge, W., Frank, R., Leder, S., Joost, H.G., and Becker, W. (2000).
Specificity determinants of substrate recognition by the protein kinase
DYRK1A. J. Biol. Chem. 275, 2431–2438.
Murshudov, G.N., Skuba´ k, P., Lebedev, A.A., Pannu, N.S., Steiner, R.A.,
Nicholls, R.A., Winn, M.D., Long, F., and Vagin, A.A. (2011). REFMAC5 for
the refinement of macromolecular crystal structures. Acta Crystallogr. D
Biol. Crystallogr. 67, 355–367.
Himpel, S., Panzer, P., Eirmbter, K., Czajkowska, H., Sayed, M., Packman,
¨
L.C., Blundell, T., Kentrup, H., Grotzing
er, J., Joost, H.G., and Becker, W.
Structure 21, 986–996, June 4, 2013 ª2013 Elsevier Ltd All rights reserved 995

Structure
Activation and Substrate Recognition of DYRKs
Niesen, F.H., Berglund, H., and Vedadi, M. (2007). The use of differential scan-
ning fluorimetry to detect ligand interactions that promote protein stability.
Nat. Protoc. 2, 2212–2221.
Shevchenko, A., Tomas, H., Havlis, J., Olsen, J.V., and Mann, M. (2006). In-gel
digestion for mass spectrometric characterization of proteins and proteomes.
Nat. Protoc. 1, 2856–2860.
´
Sinet, P.M., Theophile, D., Rahmani, Z., Chettouh, Z., Blouin, J.L., Prieur, M.,
Ogawa, Y., Nonaka, Y., Goto, T., Ohnishi, E., Hiramatsu, T., Kii, I., Yoshida, M.,
Ikura, T., Onogi, H., Shibuya, H., et al. (2010). Development of a novel selective
inhibitor of the Down syndrome-related kinase Dyrk1A. Nat. Commun. 1, 86.
Noel, B., and Delabar, J.M. (1994). Mapping of the Down syndrome phenotype
on chromosome 21 at the molecular level. Biomed. Pharmacother. 48,
247–252.
Oliver, A.W., Knapp, S., and Pearl, L.H. (2007). Activation segment exchange:
a common mechanism of kinase autophosphorylation? Trends Biochem. Sci.
32, 351–356.
Taira, N., Nihira, K., Yamaguchi, T., Miki, Y., and Yoshida, K. (2007). DYRK2 is
targeted to the nucleus and controls p53 via Ser46 phosphorylation in the
apoptotic response to DNA damage. Mol. Cell 25, 725–738.
Olsen, J.V., Blagoev, B., Gnad, F., Macek, B., Kumar, C., Mortensen, P., and
Mann, M. (2006). Global, in vivo, and site-specific phosphorylation dynamics in
signaling networks. Cell 127, 635–648.
Taira, N., Yamamoto, H., Yamaguchi, T., Miki, Y., and Yoshida, K. (2010). ATM
augments nuclear stabilization of DYRK2 by inhibiting MDM2 in the apoptotic
response to DNA damage. J. Biol. Chem. 285, 4909–4919.
Papadopoulos, C., Arato, K., Lilienthal, E., Zerweck, J., Schutkowski, M.,
Chatain, N., Mu¨ ller-Newen, G., Becker, W., and de la Luna, S. (2011). Splice
variants of the dual specificity tyrosine phosphorylation-regulated kinase 4
(DYRK4) differ in their subcellular localization and catalytic activity. J. Biol.
Chem. 286, 5494–5505.
Taira, N., Mimoto, R., Kurata, M., Yamaguchi, T., Kitagawa, M., Miki, Y., and
Yoshida, K. (2012). DYRK2 priming phosphorylation of c-Jun and c-Myc mod-
ulates cell cycle progression in human cancer cells. J. Clin. Invest. 122,
859–872.
Tejedor, F., Zhu, X.R., Kaltenbach, E., Ackermann, A., Baumann, A., Canal, I.,
Heisenberg, M., Fischbach, K.F., and Pongs, O. (1995). minibrain: a new pro-
tein kinase family involved in postembryonic neurogenesis in Drosophila.
Neuron 14, 287–301.
Park, J., Song, W.-J., and Chung, K.C. (2009). Function and regulation of
Dyrk1A: towards understanding Down syndrome. Cell. Mol. Life Sci. 66,
3235–3240.
Park, J., Oh, Y., Yoo, L., Jung, M.-S., Song, W.-J., Lee, S.-H., Seo, H., and
Chung, K.C. (2010). Dyrk1A phosphorylates p53 and inhibits proliferation of
embryonic neuronal cells. J. Biol. Chem. 285, 31895–31906.
¨
¨
Varjosalo, M., Bjorklund, M., Cheng, F., Syvanen, H., Kivioja, T., Kilpinen, S.,
Sun, Z., Kallioniemi, O., Stunnenberg, H.G., He, W.W., et al. (2008).
Application of active and kinase-deficient kinome collection for identification
of kinases regulating hedgehog signaling. Cell 133, 537–548.
Pike, A.C.W., Rellos, P., Niesen, F.H., Turnbull, A., Oliver, A.W., Parker, S.A.,
Turk, B.E., Pearl, L.H., and Knapp, S. (2008). Activation segment dimerization:
a mechanism for kinase autophosphorylation of non-consensus sites. EMBO
J. 27, 704–714.
Wiechmann, S., Czajkowska, H., de Graaf, K., Gro¨ tzinger, J., Joost, H.-G., and
Becker, W. (2003). Unusual function of the activation loop in the protein kinase
DYRK1A. Biochem. Biophys. Res. Commun. 302, 403–408.
Rappsilber, J., Mann, M., and Ishihama, Y. (2007). Protocol for micro-purifica-
tion, enrichment, pre-fractionation and storage of peptides for proteomics
using StageTips. Nat. Protoc. 2, 1896–1906.
Woods, Y.L., Cohen, P., Becker, W., Jakes, R., Goedert, M., Wang, X., and
Proud, C.G. (2001). The kinase DYRK phosphorylates protein-synthesis initia-
tion factor eIF2Bε at Ser539 and the microtubule-associated protein tau at
Thr212: potential role for DYRK as a glycogen synthase kinase 3-priming
kinase. Biochem. J. 355, 609–615.
Ryoo, S.-R., Jeong, H.K., Radnaabazar, C., Yoo, J.-J., Cho, H.-J., Lee, H.-W.,
Kim, I.-S., Cheon, Y.-H., Ahn, Y.S., Chung, S.-H., and Song, W.J. (2007).
DYRK1A-mediated hyperphosphorylation of Tau. A functional link between
Down syndrome and Alzheimer disease. J. Biol. Chem. 282, 34850–34857.
Yang, J., Ten Eyck, L.F., Xuong, N.-H., and Taylor, S.S. (2004). Crystal struc-
˚
ture of a cAMP-dependent protein kinase mutant at 1.26 A: new insights into
the catalytic mechanism. J. Mol. Biol. 336, 473–487.
Seifert, A., Allan, L.A., and Clarke, P.R. (2008). DYRK1A phosphorylates cas-
pase 9 at an inhibitory site and is potently inhibited in human cells by harmine.
FEBS J. 275, 6268–6280.
Yoshida, K. (2008). Role for DYRK family kinases on regulation of apoptosis.
Biochem. Pharmacol. 76, 1389–1394.
996 Structure 21, 986–996, June 4, 2013 ª2013 Elsevier Ltd All rights reserved