Content uploaded by Antje Gohla
Author content
All content in this area was uploaded by Antje Gohla on Aug 21, 2018
Content may be subject to copyright.
Contents lists available at ScienceDirect
BBA - Molecular Basis of Disease
journal homepage: www.elsevier.com/locate/bbadis
Improved cognition, mild anxiety-like behavior and decreased motor
performance in pyridoxal phosphatase-deficient mice
Elisabeth Jeanclos
a,⁎
, Monique Albersen
b,1
, Rúben J.J. Ramos
b
, Annette Raab
a,c
,
Christian Wilhelm
a
, Leif Hommers
a,c,d
, Klaus-Peter Lesch
c,d,e,f,g
, Nanda M. Verhoeven-Duif
b
,
Antje Gohla
a,⁎
a
Institute of Pharmacology and Toxicology, University of Würzburg, Germany
b
Department of Genetics, University Medical Center Utrecht, the Netherlands
c
Interdisciplinary Center for Clinical Research, University Hospital Würzburg, Germany
d
Comprehensive Heart Failure Center, University Hospital Würzburg, Germany
e
Division of Molecular Psychiatry, Center of Mental Health, University of Würzburg, Germany
f
Laboratory of Psychiatric Neurobiology, Institute of Molecular Medicine, I.M. Sechenov First Moscow State Medical University, Moscow, Russia
g
Department of Neuroscience, School for Mental Health and Neuroscience, Maastricht University, Maastricht, the Netherlands
ARTICLE INFO
Keywords:
Pyridoxal phosphatase
Vitamin B6
γ-Aminobutyric acid (GABA)
Motor performance
Neuropsychiatric diseases
Neurotransmitter biosynthesis
ABSTRACT
Pyridoxal 5′-phosphate (PLP) is an essential cofactor in the catalysis of ~140 different enzymatic reactions. A
pharmacological elevation of cellular PLP concentrations is of interest in neuropsychiatric diseases, but whole-
body consequences of higher intracellular PLP levels are unknown. To address this question, we have generated
mice allowing a conditional ablation of the PLP phosphatase PDXP. Ubiquitous PDXP deletion increased PLP
levels in brain, skeletal muscle and red blood cells up to 3-fold compared to control mice, demonstrating that
PDXP acts as a major regulator of cellular PLP concentrations in vivo. Neurotransmitter analysis revealed that the
concentrations of dopamine, serotonin, epinephrine and glutamate were unchanged in the brains of PDXP
knockout mice. However, the levels of γ-aminobutyric acid (GABA) increased by ~20%, demonstrating that
elevated PLP levels can drive additional GABA production. Behavioral phenotyping of PDXP knockout mice
revealed improved spatial learning and memory, and a mild anxiety-like behavior. Consistent with elevated
GABA levels in the brain, PDXP loss in neural cells decreased performance in motor tests, whereas PDXP-defi-
ciency in skeletal muscle increased grip strength. Our findings suggest that PDXP is involved in the fine-tuning of
GABA biosynthesis. Pharmacological inhibition of PDXP might correct the excitatory/inhibitory imbalance in
some neuropsychiatric diseases.
1. Introduction
Pyridoxal 5′-phosphate (PLP), the co-enzymatically active form of
vitamin B6, is one of the most versatile cofactors found in nature. In
mammals, PLP-dependent enzymes catalyze ~140 different types of
biochemical transformations, including reactions that are required for
neurotransmitter-, amino acid- and glycogen metabolism [1]. PLP is a
highly reactive aldehyde that transiently forms a Schiffbase with the ε-
amino group of the active site lysine in its cognate apo-enzymes [2]. It
is thought that intracellular PLP concentrations and PLP trafficking are
tightly controlled to ensure sufficient PLP supply to apo-enzymes, while
minimizing free PLP levels. This is important to prevent non-specific
reactions of PLP with cellular nucleophiles and proteins that are not B6
enzymes [3–5].
Intracellular PLP levels depend on the availability and -transport of
PLP precursors [6], the biosynthetic activities of pyridoxal 5′-kinase
(PDXK) and pyridox(am)ine 5′-phosphate oxidase (PNPO), the extent of
PLP scavenging by proteins and small molecules, PLP binding to carriers
such as PROSC, PDXK and PNPO, and PLP hydrolysis by pyridoxal 5′-
phosphate phosphatase (PDXP) [3,4,7]. Specifically, PDXK catalyzes the
phosphorylation of pyridoxine (PN), pyridoxamine (PM) or pyridoxal (PL)
to their 5′-phosphorylated variants PNP, PMP or PLP. PNPO additionally
(re)generates PLP from PNP or PMP by catalyzing the oxidation of their
respective hydroxyl or amino groups to the co-enzymatically essential
https://doi.org/10.1016/j.bbadis.2018.08.018
Received 29 June 2018; Received in revised form 13 August 2018; Accepted 14 August 2018
⁎
Corresponding authors.
1
Current address: Department of Clinical Chemistry, VU Medical Center, Amsterdam, the Netherlands.
E-mail addresses: elisabeth.jeanclos@uni-wuerzburg.de (E. Jeanclos), antje.gohla@uni-wuerzburg.de (A. Gohla).
BBA - Molecular Basis of Disease xxx (xxxx) xxx–xxx
0925-4439/ © 2018 The Authors. Published by Elsevier B.V. This is an open access article under the CC BY-NC-ND license
(http://creativecommons.org/licenses/BY-NC-ND/4.0/).
Please cite this article as: Jeanclos, E., BBA - Molecular Basis of Disease (2018), https://doi.org/10.1016/j.bbadis.2018.08.018
aldehyde moiety of PLP [3](Fig. 1). The pathways of PLP biosynthesis
have been established for decades, yet a dedicated PLP phosphatase
(PDXP) was cloned more recently [8]. PDXP activity may play a significant
role in the regulation of cellular PLP concentrations [7,8]. Nonetheless, the
relative importance of PDXP for the regulation of PLP homeostasis and for
PLP-dependent functions in vivo is unknown.
A pharmacological elevation of cellular PLP concentrations is of
interest in neuropsychiatric disorders and inflammation [9–11]. Inborn
errors of vitamin B6 metabolism can lead to inadequate levels of PLP in
the brain and cause childhood epilepsy [4,12–17]. In these cases, the
administration of PLP or PN leads to prompt seizure cessation. How-
ever, high doses of PN can cause respiratory depression, and impose
treatment withdrawal [4]. Individuals with B6-dependent epilepsy
often require continuous oral PN supplementation, yet long-term
treatment with high doses of PN or PLP can be toxic [18–20]. A phar-
macological inhibition of PDXP [21] might circumvent toxic effects
caused by elevated extracellular B6 vitamer levels, and provide a novel
approach to increase cellular PLP concentrations.
The aim of the present study was to analyze the relative importance
of PDXP for the regulation of tissue PLP concentrations, and to explore
the biochemical and behavioral consequences of PDXP deletion in gene-
targeted mice. Our results demonstrate that PDXP is a major factor in
the control of cellular PLP concentrations in vivo, and we uncover a
previously unexploited approach to increase GABA levels in the brain.
PDXP loss improved spatial learning and memory, but also caused a
mild anxiety-like phenotype and central nervous system-dependent
muscle weakness. Further research is necessary to evaluate whether
pharmacological PDXP inhibition might be an advantageous strategy to
increase cellular PLP levels.
2. Materials and methods
2.1. Materials
Unless otherwise specified, all reagents were of the highest available
purity and purchased from Sigma-Aldrich.
2.2. Generation and breeding of Pdxp knockout mice
Floxed Pdxp mice (Pdxp
tm1Goh
) were generated on a C57Bl/6J
background by Ozgene Pty Ltd., Australia. The neomycin resistance
cassette was removed by breeding with the global FLPe deleter strain
B6.129S4-Gt(ROSA)26Sor < tm1(FLP1)Dym > /RainJ. Whole-body,
skeletal muscle or neural cell-directed Pdxp knockouts were achieved
by breeding with B6.FVB-Tg(EIIa-cre)C5379Lmgd/J (EIIa-Cre), B6.Cg-
Tg(ACTA1-cre)79Jme/J (Acta1-Cre), or B6.Cg-Tg(Nes-cre)1Kln/J
(Nestin-Cre) transgenic mice obtained from The Jackson Laboratory.
Mouse experiments were approved by the local authorities (Regierung
Unterfranken), and all analyses were carried out in strict accordance
with all German and European Union applicable laws and regulations
concerning care and use of laboratory animals.
2.3. Genotyping
Genomic DNA was isolated with DNeasy (Qiagen) and digested with
EcoRV. Southern blotting was performed according to standard proce-
dures. Briefly, DNA was subjected to agarose gel electrophoresis, de-
natured, blotted onto nitrocellulose and crosslinked with UV light. The
probe was generated by PCR using Pfx polymerase (Invitrogen) and
labeled with [α-
32
P] dCTP (Hartmann Analytic), employing the
DecaPrime II DNA labeling kit (Ambion). Unincorporated labeled nu-
cleotides were removed with Illustra ProbeQuant G-50 micro columns
(GE Healthcare). Genotyping of Acta1-Cre and Nestin-Cre transgenic
mice was performed by PCR as described by The Jackson Laboratory.
2.4. Tissue sample preparation
Mice were sacrificed by cervical dislocation and perfused with PBS.
Skeletal muscle samples were taken from the upper back hind limb, and
contained musculus biceps femoris,musculus semimembranosus and mus-
culus semitendinosus. Small intestine samples (comprising duodenum,
jejunum and ileum) were cut into small segments and washed ex-
tensively with phosphate-buffered saline (PBS). For liver analysis, the
entire organ was used. Brains were either used in toto,ordifferent brain
regions were dissected. To this end, the brain was placed on an ice-cold
adult mouse brain slicer (Harvard Apparatus). A 3 mm coronal section
was cut starting from the basis of the circle of Willis towards the frontal
part of the brain (Bregma −3.16 to Bregma 0.02), and the brain cortex
and the midbrain region (containing hippocampus, hypothalamus and
thalamus) were dissected from this slice. All brain parts including cer-
ebellum and hindbrain were weighed (wet weight). Pups were sacri-
ficed by decapitation on postnatal day 6 (P6), and brains were
PL PMPN
PLP PMPPNP
HH H
PO32-
PNPO PNPO
PNPO PNPO
PDXK
PDXP PDXK
PDXP
PO32- PO32-
PA
AOX DH
PDXK
PDXP
Fig. 1. Enzymes involved in mammalian
PLP metabolism. PA, 4′-pyridoxic acid;
PN(P), pyridoxine (-5′-phosphate); PL(P),
pyridoxal (-5′-phosphate); PM(P), pyr-
idoxamine (-5′-phosphate). PDXP, pyr-
idoxal phosphatase; PDXK, pyridoxal
kinase; PNPO, pyridox(am)ine oxidase;
AOX/DH, aldehyde oxidase/β-NAD de-
hydrogenase. PL is oxidized to PA in the
liver and excreted into the urine.
Catalytic efficiencies (k
cat
/K
M
) of PDXP:
PLP, 6.1 × 10
5
M
−1
×s
−1
; PNP, 2.7 ×
10
4
M
−1
×s
−1
; PMP, 5.6 × 10
3
M
−1
×s
−1
[8].
E. Jeanclos et al. BBA - Molecular Basis of Disease xxx (xxxx) xxx–xxx
2
dissected, weighed and flash-frozen in liquid nitrogen. For analysis,
organs and tissue parts were homogenized in ice-cold PBS with a Potter
S homogenizer (Sartorius). Proteins were precipitated with 10% (w/v)
6.1-N trichloroacetic acid (TCA) for 15 min on ice with occasional
vortexing. After centrifugation (20,000 ×g, 15 min, 4 °C), supernatants
were flash-frozen in liquid nitrogen and kept at −80 °C.
2.5. Plasma and red blood cell preparation
Heparinized blood was centrifuged at room temperature (800 ×g,
10 min). The supernatant (plasma) was centrifuged again, and proteins
in the supernatant were removed by precipitation, using 10% (w/v)
6.1-N TCA. Packed red blood cells (RBCs) were washed by centrifuga-
tion and resuspension in 0.9% (w/v) NaCl. Cells were lysed and pro-
teins were precipitated by adding 10% 6.1-N TCA. Plasma and RBC
extracts were flash-frozen in liquid nitrogen and kept at −80 °C until
analysis. To calculate the intracellular concentrations of B
6
vitamers,
the intracellular RBC volume was estimated to correspond to 52% of the
packed RBC volume.
2.6. Quantification of B
6
vitamers, GABA and amino acids
B
6
vitamers, GABA and other amino acids were quantified using
UPLC-MS/MS as described [6,22–24].
2.7. Neurotransmitter analysis by HPLC
Mice were sacrificed by cervical dislocation, brains were dissected,
weighed and flash-frozen in liquid nitrogen. The entire procedure was
performed in < 3 min. Frozen brains were sonicated (2 × 30 s) in a
0.3 M perchloric acid solution at 4 °C, and samples were centrifuged
(16,000 ×g, 10 min, 4 °C). Supernatants were analyzed on a Dionex
Ultimate 3000 HPLC (ThermoScientific), using 75 mM NaH
2
PO
4
,
1.5 mM octane sulfonic acid, 10% (v/v) acetonitrile; pH 3 as a mobile
phase. Neurotransmitters were separated on a 3 μm reverse phase
column (BDS-HYPERSIL-C18, ThermoScientific) and detected by elec-
trochemical reaction at 450 mV.
2.8. Western blot analysis
Tissue or cell homogenates were prepared as detailed above and
extracted with RIPA buffer [50 mM Tris, pH 7.5; 150mM NaCl, 1% (v/v)
Triton X-100, 0.5% (v/v) sodium deoxycholate, 0.1% (w/v) SDS, 1 mM
4-(2-aminoethyl)benzenesulfonyl fluoride (Pefabloc), 5 μg/mL aprotinin,
1μg/mL leupeptin, 1 μg/mL pepstatin] for 15 min at 4 °C under rotation,
and lysates were clarified by centrifugation (20,000 ×g, 15 min, 4 °C).
Protein concentrations in the supernatants were determined using the
Micro BCA Protein Assay Kit (Thermo Scientific). Samples were solubi-
lized in Laemmli buffer and subjected to standard immunoblotting. The
following antibodies were used: Cell Signaling Technology, α-PDXP
(clone C85E3) and α-GAPDH (clone 14C10); Merck/Millipore, α-actin
(mAb1501) and α-FAK (clone 4.47); ThermoFisher Scientific, α-PNPO
(#PA5-26400); and Sigma Aldrich, α-PDXK (AB1, #AV53615). To ana-
lyze Ser3-phosphocofilin and cofilin levels, tissues were dissected from 4-
months old male mice, and immediately flash-frozen in liquid nitrogen as
described above. Frozen tissues were then pulverized using a pre-cooled
porcelain mortar and pestle. About 50–100 mg of this powder was so-
lubilized in 250 μLice-coldlysisbuffer [20 mM Tris, pH 7.5; 150 mM
NaCl, 1% (v/v) Triton X-100, 1 mM β-glycerophosphate, 2.5 mM sodium
pyrophosphate, 1 mM sodium orthovanadate, 1 mM Pefabloc, 1 μg/mL
leupeptin/aprotinin/pepstatin; 1/1000 phosphatase inhibitor cocktail I
and II (both from Sigma Aldrich, #P2850 and P5726)] for 2 h at 4 °C
under rotation. Lysates were cleared by centrifugation (20,000 ×g,
15 min, 4 °C), and protein concentrations in the supernatants were de-
termined and adjusted to ~20 μg/μL using the Micro BCA Protein Assay
Kit. Samples were diluted to a final concentration of ~2 μg/μLin2×
Laemmli buffer, and ~40–50 μgofproteinswereseparatedona12%
SDS-PAGE gel. For hippocampal extracts, ~200 μg of proteins were
analyzed. Proteins were semi-dry transferred to nitrocellulose and im-
munoblotted using α-Ser3-phosphocofilin antibodies (clone 77G2, Cell
Signaling Technology). Blots were stripped and reprobed using α-cofilin
antibodies (#3312, Cell Signaling Technology or #ACFL02, Cytoskeleton
Inc.).
2.9. Behavioral studies
2.9.1. Animals
All animals were housed at the Center for Experimental Molecular
Medicine at the University of Würzburg, and were kept on a regular 14/
10 h light-dark cycle, in a temperature (21 ± 0.5 °C)- and humidity
(50 ± 5%)-controlled environment with ad libitum access to food and
water. All experiments were performed by the same experienced female
investigator between 9 a.m. and 3 p.m. All mice were males housed in
groups of 2–5 animals per cage. The age of the mice is specified in the
respective figures. All animal protocols were in line with the provisions
of the Animal Protection Law according to the Directive of the
European Communities Council of 1986 (86/609/EEC). Unless other-
wise indicated, behavioral tests were performed according to standard
procedures using devices from TSE Systems. Mice were imaged with an
infrared light-sensitive CCD camera, and data were collected and ana-
lyzed using the automated tracking software VideoMot2 (TSE Systems).
2.9.2. Tests for anxiety-like behavior, exploratory behavior, and spatial
learning and memory
Mouse behavior in the open field arena, the elevated plus maze, the
dark-light box and in the Barnes maze was evaluated as described [25].
An independent cohort of mice was analyzed in the Barnes maze; bright
light (~100 lx) was used as a mildly aversive stimulus. Marble burying
was performed as published [26]. Briefly, a standard mouse cage
(42.5 × 26.5 × 15 cm) was filled with ~5 cm bedding material. Eigh-
teen glass marbles were spaced evenly in a 3 × 6 grid-fashion on the
surface of the bedding. Mice were placed in the cage for 30 min, and the
number of marbles buried up to 2/3 of their depth was counted.
2.9.3. Sucrose preference test
In a two-bottle free choice test measuring the preferred intake of
either sucrose solution or regular water, mice typically prefer swee-
tened water. A reduction in the sucrose preference ratio in mice is
considered to be indicative of anhedonia, a core symptom of depression
in humans [27]. Mice were presented with two drinking bottles in their
home cage, one containing plain tap water and the other filled with a
2% (w/v) sucrose solution as described elsewhere [27]. The beginning
of the test started with the onset of the dark (active) phase of the ani-
mals' cycle. No food or water deprivation was applied prior to the test.
The position of the two bottles was switched daily to reduce any pos-
sible confounding effect due to a side-bias, and mice were tested for a
total of nine days. The consumption of water and sucrose solution was
estimated simultaneously in control and experimental groups by
weighing the bottles. The sucrose preference ratio was calculated as the
percentage of sucrose intake relative to the total volume of fluid intake
using the following equation: Volume (sucrose solution) / [volume
(sucrose solution) + volume (water)] × 100 [28].
2.9.4. Inverted screen test
Each mouse was placed on a standard cage lid and allowed to obtain
afirm grip. The lid was then swiftly inverted over an empty mouse cage
filled with bedding. The time until the mouse fell into the empty cage
was measured using a stopwatch; maximum test duration was 5 min.
Mice were given three trials each day with a 20 s recovery phase in
between trials. The test was repeated on three consecutive days. The
cage lid was cleaned before testing another mouse. Latency to fall into
the cage was used to evaluate limb strength.
E. Jeanclos et al. BBA - Molecular Basis of Disease xxx (xxxx) xxx–xxx
3
2.9.5. Weights test
The weights test was performed as described [29], with the excep-
tion that mice were allowed to hold the wire mesh with all four limbs.
Briefly, mice were held by the middle of their tails and allowed to grasp
afine wire mesh with an increasing number of steel chain links attached
to it. The mice grasping the wire mesh with the attached weights were
lifted, and the time they held on to the weight was measured with a
stopwatch. The criterion was a 3 s hold; if the mouse dropped the
weight earlier, the test was repeated twice after a recovery phase of
10 s. Failure to hold a given weight three times terminated the trial. If
the mouse succeeded in holding the weight, the next heaviest weight
was tested. The test score indicates the number of links in the heaviest
chain that the mouse held for 3 s multiplied by the time (in s) it held on
to it. For example, a score of 15 indicates a mouse holding a 5-link
weight for 3 s; a score of 16 indicates that it held a 5-link weight for 3 s
and a 6-link weight for 1 s. The weight of each chain was 13 g, the
weight of the mesh attached to the chains links was 5 g.
2.9.6. Grip strength
Hind paw strength was determined using an automated Grip
Strength Meter (Columbus Instruments) as reported previously [30]. In
short, mice were trained to grasp a bar with either their front or hind
paws. A pull was exerted, and the maximal strength that the mice ex-
hibited before releasing the grip bar was measured. For each mouse, ten
measurements per day were acquired on three consecutive days. The
average grip strength value (in N) was determined for each mouse, and
values were expressed relative to the mean values of the respective
control cohort.
2.10. Image quantification and statistical analysis
Western blots were quantified with NIH ImageJ, version 1.45i. For
statistical analysis, two-tailed, unpaired Student's t-tests were per-
formed using GraphPad Prism version 6.0. All samples that were sub-
jected to statistical analysis were biological replicates. The number of
independent samples or individual mice per group and genotype is
specified in each figure or figure legend. Reprobes of Western blot ex-
periments were performed using the same membrane that was probed
for the primary protein under study.
3. Results
3.1. PDXP is a dominant regulator of pyridoxal 5′-phosphate levels in vivo
To study the functions of PDXP in vivo, we generated conditionally
Pdxp-deficient mice (Supplementary Fig. S1A). Southern blot analysis
demonstrated homologous recombination after breeding Pdxp
flx/flx
mice
with the whole-body Cre deleter strain EIIa-Cre (Supplementary Fig.
S1B). Homozygous Pdxp-deficient (KO) mice on a C57BL/6J back-
ground were born at the expected Mendelian frequencies, and were
indistinguishable from wildtype (WT) controls in terms of viability,
growth and fertility. Fig. 2A demonstrates efficient PDXP deletion in all
investigated organs and cells at the protein level. In accordance with
the important role of the PDXP substrate pyridoxal 5′-phosphate (PLP)
for neurotransmitter biosynthesis and metabolism, PDXP was highly
expressed in the brain. The identity of the ~25 kDa band that specifi-
cally reacts with the α-PDXP antibody in WT, but not in PDXP-KO
A
37 -
25 -
50 -
37 -
α-a
ctin
α-
PDXP
brain sk. mus. sm. intest. liver
WT KO WT KO WT KO WT KO
brain kidneyheart liver
Ponceau
kDa
37
25
37
WT KO WT KO WT KO WT KO
WT KO WT KO WT KO WT KO
20
15
20
15
kDa
37
25
brain
α-
PDXP
B
C
α-
PDXP
α-
P-cofilin
α-
cofilin
hippocampus
WT KO WT KO WT KO WT KO
α-
PDXP
α-
P-cofilin
α-
cofilin
WT KO
0.0
0.5
1.0
1.5
2.0
relative intensity
P-cofilin/cofilin
p = 0.71
p = 0.34
WT
0
1
2
3
4
20
15
20
15
37
25
relative intensity
P-cofilin/cofilin
kDakDa
>
>
KO
Fig. 2. Characterization of con-
ditionally PDXP-deficient mice. (A)
Western blot analysis of PDXP expres-
sion in different wildtype (WT) mouse
organs, and loss of detectable PDXP
expression in Pdxp
flx/flx
; EIIa-Cre mice
(KO). Data are representative of n≥3
assays performed with tissues dissected
from at least three different WT and
three different KO mice. The position
of the ~32 kDa PDXP protein is in-
dicated by an arrowhead. The identity
of the ~25 kDa band is currently un-
known. (B, C) Effect of PDXP deletion
on Ser3-phosphocofilin (P-cofilin) le-
vels in whole brain (B) or hippocampal
lysates (C); n= 4 mice per genotype
were analyzed. P-cofilin signals were
normalized to the respective cofilin
signals; the densitometric quantifica-
tion is shown on the right. The whis-
kers indicate minimum and maximum
values. Samples in (A) were obtained
from 7 to 8 months-old male mice;
samples in (B) and (C) were from
4 months-old male mice. Whole-brain
lysates in (B) and hippocampal lysates
(C) were generated from two separate
mouse cohorts. Sk. mus., skeletal
muscle; sm. intest., small intestine.
E. Jeanclos et al. BBA - Molecular Basis of Disease xxx (xxxx) xxx–xxx
4
brains is currently unclear. Because PDXP isoforms of a corresponding
molecular mass are not known to exist, this band might represent a
proteolytic PDXP degradation product.
After the molecular cloning of PDXP [8], an enzyme termed
chronophin was isolated from bovine brain fractions and shown to
dephosphorylate the actin regulatory protein cofilin on serine-3 [31];
surprisingly, chronophin and PDXP turned out to be identical. PDXP/
chronophin can regulate cofilin-dependent actin re-organization at the
leading edge of immune and cancer cells [32–35], and mediate an ATP-
sensing mechanism for cofilin dephosphorylation in neuronal cells ex-
posed to energy-stress [36]. In addition, it has recently been reported
that the constitutive genetic ablation of Pdxp (referred to in that work
as PLPP/CIN) leads to a moderate, ~45% increase in cofilin phos-
phorylation in the mouse hippocampus; conversely, substantial trans-
genic overexpression of the phosphatase resulted in a ~30% decrease in
cofilin phosphorylation levels [37]. We analyzed the phosphocofilin/
cofilin ratio in lysates of whole brain (Fig. 2B) and isolated hippocampi
(Fig. 2C) in our mouse model. Phosphocofilin levels were more variable
in PDXP-KO than in WT samples, but we did not detect significant
PDXP-dependent changes.
UPLC-MS/MS-based quantification of PLP, the primary known in
vitro metabolite substrate of PDXP [7,8,38], revealed threefold in-
creases in PLP levels in whole brain extracts of adult PDXP-KO com-
pared to WT mice. PLP levels were also significantly elevated in skeletal
muscle and in red blood cells. We did not find PLP changes in liver, in
the small intestine or in plasma (Fig. 3A, Table 1). These findings in-
dicate that PDXP is not involved in the regulation of extracellular PLP
levels, and suggest that PDXP is not essential for PLP hydrolysis in liver
and intestine. However, in contrast to the brain, liver can release PLP
into the plasma, and secreted PLP can subsequently be hydrolyzed to PL
by alkaline phosphatase, a membrane-bound ectoenzyme. It is therefore
possible that PDXP loss in the liver impacts intracellular PLP levels, but
that this effect is undetectable due to PLP secretion and extracellular
PLP hydrolysis [39]. Likewise, the loss of PDXP in the small intestine
[6] might be compensated for by direct PLP transfer from the intestinal
lumen to the circulation, or by alkaline phosphatase activity [40]. The
lack of increased PLP levels in the plasma of PDXP-KO mice is con-
sistent with a critical role of the extracellular, tissue non-specific al-
kaline phosphatase TNAP for PLP hydrolysis in plasma, as revealed by
the markedly elevated plasma PLP levels measured in patients with
hypophosphatasia, an inborn error characterized by impaired TNAP
activity [41], and by the increased plasma PLP concentrations found in
TNAP-targeted mice (ref. [42] and see below).
Further analysis of the adult brain demonstrated that PDXP loss led
to a comparable PLP increase in all investigated areas (cortex, mid-
brain, hindbrain and cerebellum). We also observed a ~twofold in-
crease in PLP levels in the brains of 6 day-old (postnatal day 6, P6)
PDXP-KO pups (Fig. 3B, Table 2). In brain, PL levels also increased in a
PDXP-dependent manner (Tables 1, 2). However, because the relative
increases in PL and PLP levels were similar in PDXP-KO brain extracts,
we speculate that extracellular TNAP partially hydrolyzed PLP to PL
during cell lysis [42,43]. PN levels in whole-brain extracts of adult
PDXP-KO mice appeared to be increased (Table 1), but the analysis of
different adult brain regions or of P6 brains did not confirm this result
(Table 2). By comparison, mice lacking TNAP have reduced PLP levels
in the brain, but strongly elevated PLP levels in plasma (brain: WT,
4.5 ± 1.0 nmol/g wet tissue weight; TNAP-KO, 1.6 ± 0.3 nmol/g wet
tissue weight; serum: WT, 183 ± 104 nM; TNAP-KO, 3166 ±
1288 nM, ref. [42]). Taken together, these data clearly demonstrate
that PDXP is a key determinant in the control of cellular PLP con-
centrations in vivo.
A presumed primary function of PDXP is to keep the levels of free
intracellular PLP low. This is important for cellular homeostasis because
PLP is an electrophilic small molecule that can bind and inactivate
cellular nucleophiles. In addition, elevated PLP levels may impinge
upon cellular PLP biosynthesis. In vitro, PLP functions as an effective
product inhibitor of PNPO, and it additionally binds PNPO tightly at a
non-catalytic site [35,36]. We therefore investigated whether elevated
levels of PLP influenced PNPO. Western blot analysis of PNPO expres-
sion revealed ~2.4-fold higher PNPO levels in PDXP-KO compared to
WT brain lysates. In contrast, PDXK expression was not significantly
altered (Fig. 3C). It is possible that higher intracellular PLP levels en-
hance PNPO stability and half-life. Alternatively, elevated PLP levels
might inhibit PNPO activity, and trigger a compensatory increase in
PNPO protein expression.
3.2. Effect of PDXP loss on amino acid levels in the brain
Because PLP is an essential cofactor of many enzymes involved in
amino acid metabolism [1], we next asked whether elevated PLP levels
affected steady-state amino acid levels in the adult brain. As shown in
the Supplementary Table S1, changes were restricted to a subset of
amino acids in particular brain regions. Serine and alanine levels were
significantly increased in the cortex (+14% or +21%, respectively),
and citrulline levels were higher in the basal ganglia region (+14%).
Glutamine and glycine concentrations decreased in PDXP-KO hind-
brains (−20% or −13%, respectively). The levels of pipecolic acid, a
catabolite of lysine, were higher in all parts of the brain, with increases
ranging from +23% in the midbrain to +33% in the cortex. In de-
veloping PDXP-deficient brains at P6, alanine and threonine levels were
elevated (+13% or +35%, respectively), and methionine and pheny-
lalanine levels were reduced (−35% or −43%, respectively). Together,
these data indicate that higher intracellular PLP concentrations affect
amino acid metabolism, but do not lead to a general increase in the
activities of PLP-dependent enzymes.
3.3. Elevated GABA levels and improved spatial learning and memory in
PDXP-KO mice
PLP is an obligatory cofactor for decarboxylation reactions in the
biosynthesis of dopamine, (nor)epinephrine, serotonin and GABA.
Specifically, PLP is required for the aromatic L-amino acid decarbox-
ylase, which catalyzes the decarboxylation of L-3,4-dihydrox-
yphenylalanine (DOPA) to the (nor)epinephrine precursor dopamine,
and the decarboxylation of 5-hydroxy-L-tryptophan to 5-hydro-
xytryptamine (serotonin) [44]. Quantification of these neuro-
transmitters and their catabolites in brain extracts of 9 months-old male
mice did not show PDXP-dependent changes (Fig. 4A). Thus, elevated
cellular PLP concentrations did not increase the physiological activity
of DOPA decarboxylase. Interestingly though, we found significantly
higher amounts of GABA in whole brain extracts of 8 months-old PDXP-
KO mice, as well as in the cerebellar and midbrain region of 4 months-
old PDXP-KO mice (Fig. 4B). GABA concentrations were also sig-
nificantly higher in whole brains and in dissected cerebellae of P6
PDXP-KO compared to WT pups (Fig. 4B). The rate-limiting step in
GABA biosynthesis consists of the PLP-dependent decarboxylation of
glutamate [45,46]; glutamate itself is formed by glutamate transami-
nase, which also uses PLP as a cofactor. Yet, similar to dopamine,
epinephrine and serotonin, the amounts of glutamate were comparable
between 4 months-old PDXP-KO and WT mice (Fig. 4C). We conclude
that loss of PDXP expression results in a selective increase in cerebral
GABA levels.
One distinguishing feature of GABA biosynthesis is that it is medi-
ated by two glutamate decarboxylases, which are encoded by separate
genes [47]. The majority of GABA in the brain is synthesized by the
67 kDa glutamate decarboxylase (GAD67), a constitutively active, cy-
tosolic enzyme that is typically saturated with PLP. In contrast, the
65 kDa glutamate decarboxylase (GAD65) is enriched in synapses and
bound to the membranes of synaptic vesicles. GAD65 predominantly
exists as an inactive apo-enzyme in cells, and PLP-binding transiently
increases holo-GAD65 formation in response to the demand for extra
GABA in neurotransmission [46,48,49]. Although overall GABA
E. Jeanclos et al. BBA - Molecular Basis of Disease xxx (xxxx) xxx–xxx
5
p < 0.0001
p = 0.01
4 mo
A
C
0
100
200
300
400
500
0
100
200
300
400
500
cere-
bell.
hind-
brain
cortex mid-
brain
P6
whole
brain
WT
KO
PLP (nmol/g protein)
n.s.
n.s.
p < 0.0001
p < 0.0001
p < 0.0001
p < 0.0001
p = 0.0001
whole
brain
skel.
muscle
liver
sm.
intest.
PLP (nmol/g protein)
WT KO WT KO WT KO
WT KO WT KO WT KO
PDXK PNPO
B
α-PDXK
α-PDXP
α-actin
α-PNPO
α-PDXP
α-actin
0
1
2
3
WT KO WT KO
rel. signal intensity
p = 0.04
n.s.
37 -
37 -
25 -
50 -
37 -
37 -
50 -
25 -
25 -
37 -
kDa
kDa
7-8 mo
Fig. 3. Effect of whole-body PDXP deletion on PLP levels in vivo. PLP levels were quantified by UPLC-MS/MS. (A) Organs were isolated from 7 to 8 months-old male
mice, n=4–8 per genotype; the indicated brain regions were dissected from 4 months-old male mice, n= 4 per genotype (B). The whiskers indicate minimum and
maximum values. For further details and data on other B6 vitamers, see Tables 1 and 2. (C) Western blot analysis of PDXK and PNPO levels in whole brain lysates of
7–8 months-old male mice; lysates of n= 3 mice per genotype were analyzed. Right panel, blots were analyzed densitometrically after normalization to actin. WT,
wildtype mice; KO, Pdxp
flx/flx
; EIIa-Cre mice. Data in (C) are mean values ± S.E.M.; n.s., non-significant. Skel. muscle, skeletal muscle; sm. intestine, small intestine;
cerebell., cerebellum; P6, postnatal day 6.
Table 1
UPLC-MS/MS-based quantification of B
6
vitamers in WT and PDXP-KO mice.
nmol/g protein nPLP PL PA PMP PM PN
Brain WT 8 91.29 ± 6.30 25.54 ± 1.56 0.27 ± 0.09 156.1 ± 13.6 0.79 ± 0.20 0.06 ± 0.01
PDXP-KO 7 298.8 ± 7.45 74.95 ± 5.95 0.35 ± 0.12 173.6 ± 10.2 1.3 ± 0.24 0.33 ± 0.05
p-Value < 0.0001 < 0.0001 0.60 0.33 0.12 < 0.0001
Skeletal muscle WT 8 235.0 ± 12.0 5.10 ± 1.62 0.05 ± 0.02 75.65 ± 13.17 0.66 ± 0.26 0.01 ± 0.01
PDXP-KO 7 331.4 ± 31.9 8.83 ± 3.40 0.13 ± 0.05 98.69 ± 28.22 0.79 ± 0.35 0.02 ± 0.02
p-Value 0.01 0.32 0.12 0.45 0.77 0.74
Small intestine WT 4 42.75 ± 3.77 7.80 ± 0.66 0.42 ± 0.23 97.25 ± 15.73 0.46 ± 0.12 0.01 ± 0.01
PDXP-KO 4 50.77 ± 3.56 6.62 ± 0.05 0.69 ± 0.12 116.3 ± 3.6 0.71 ± 0.08 0.41 ± 0.17
p-Value 0.17 0.12 0.34 0.28 0.14 0.06
Liver WT 8 152.5 ± 10.9 4.32 ± 0.84 0.68 ± 0.13 222.9 ± 14.3 0.28 ± 0.09 0.024 ± 0.01
PDXP-KO 7 140.9 ± 6.2 3.86 ± 0.56 0.60 ± 0.13 223.7 ± 9.6 0.24 ± 0.06 0.017 ± 0.01
p-Value 0.39 0.66 0.64 0.96 0.70 0.60
nmol/L nPLP PL PA PMP PM PN
Plasma WT 8 146.9 ± 14.9 528.0 ± 45.2 85.88 ± 12.92 246.9 ± 41.2 n.d. n.d.
PDXP-KO 7 143.4 ± 9.3 483.7 ± 33.3 65.14 ± 6.86 142.4 ± 23.2 n.d. n.d.
p-Value 0.85 0.45 0.2 0.05
Red blood cells WT 8 3776 ± 468 393.4 ± 31.4 14.91 ± 1.81 347.4 ± 23.5 n.d. n.d.
PDXP-KO 7 10,321 ± 1402 332.9 ± 23.0 10.20 ± 2.55 389.5 ± 53.2 n.d. n.d.
p-Value < 0.001 0.15 0.15 0.46 ––
Shown are mean values ± S.E.M. Nindicates the number of analyzed mice. For comparison with the other B
6
vitamers, results of the PLP measurements depicted in
Fig. 3A and B are shown again here. Pyridoxine 5′-phosphate (PNP) could not be detected. Statistically significant differences between the two genotypes were
analyzed using unpaired t-tests. All mice were 7–8 months-old males; n.d., not determined.
E. Jeanclos et al. BBA - Molecular Basis of Disease xxx (xxxx) xxx–xxx
6
concentrations were increased, we found that GAD67 expression was
lowered by ~30% in PDXP-KO compared to WT brains, possibly as a
result of a negative feedback response (Fig. 4D). GAD65 monomer or
dimer levels were not detectably affected by the absence of PDXP ex-
pression. Interestingly, we observed a prominent high molecular weight
(~220 kDa), GAD65-positive band in immunoblots of PDXP-KO brain
lysates, which was only faintly visible in the controls (Fig. 4D). PLP-
binding to GAD65 has profound effects on protein conformation
[50,51], and elevated PLP levels might thereby favor the formation of
SDS-resistant GAD65 complexes. Taken together, these data suggest
that increased cellular PLP concentrations trigger holo-GAD65 forma-
tion, which augments synaptic GABA biosynthesis in PDXP-KO mouse
brains.
GABA is involved in spatial learning behavior, for example by
regulating adult hippocampal neurogenesis [52,53]. We therefore in-
vestigated the cognitive abilities of PDXP-KO mice in the Barnes maze.
This task reports on hippocampus-dependent spatial reference
memory by evaluating the ability to learn and remember the location
of a hidden escape zone using a set of visual cues [54]. Fig. 5 de-
monstrates that PDXP-KO mice tended to require less time to locate
the escape chamber (reduced target latency), and to search fewer
holes before locating the escape chamber (less primary errors), but
these trends did not reach statistical significance. However, PDXP-KO
mice required significantly reduced path lengths (distances) to find
theescapechambercomparedtoWTmice.PDXP-KOdeletionalsoled
to shorter escape latency, i.e. mice needed less time to enter the escape
chamber. The analysis of reversal learning showed that PDXP-KO mice
performed significantly better in terms of path length required to find
the relocated escape chamber and escape latency (Fig. 5). Together,
these results indicate that PDXP loss improved spatial learning and
memory.
3.4. Anxiety-like behavior and impaired motor performance in PDXP-KO
mice
Because of the important roles of GABA for neuronal activities in-
volved in emotion, learning and memory, and the pivotal role of GABA
for muscle tone and the regulation and execution of movements [55],
we further examined PDXP-dependent mouse behavior and motor
performance in a series of non-invasive tests. In the open field arena (a
sensorimotor test used to assess general activity levels, locomotor ac-
tivity and exploration), PDXP-KO mice spent less time in the center of
the box. This difference was observed in 4–5 months old (young) and in
9–10 months old (middle-aged) mice (Fig. 6A). In the elevated plus
maze (a test used to evaluate anxiety-related behavior), young PDXP-
deficient mice behaved similar to WT controls, whereas middle-aged
PDXP-KO mice entered the open arms less frequently and also stayed
there for shorter times (Fig. 6B). Analysis of dark/light exploration
showed a tendency towards anxiety-like behavior in PDXP-KO mice (as
indicated by an increased latency to enter the lit compartment). How-
ever, this trend did not reach statistical significance; likewise, the time
spent in the lit compartment was not different between WT and KO
mice (Fig. 6C). Marble burying (a test that reports on anxiety-like,
obsessive-compulsive, and repetitive behavior) increased in older
PDXP-KO compared to age-matched WT mice (Fig. 6D). General loco-
motor activity (measured for all mice in parallel with the parameters
shown in Fig. 6A–C) was not different between the two genotypes in
these tasks. A sucrose preference test was conducted in 9 months-old
male mice to evaluate anhedonia as an indicator of depressive-like
behavior, yet no PDXP-dependent differences were observed (% sucrose
intake relative to total fluid intake in mL: WT, 93.70 ± 2.08 mL, n=3;
KO, 88.70 ± 5.15 mL, n=3;p= 0.42. Data are means ± S.E.M.). We
conclude from these results that PDXP loss resulted in a mild, anxiety-
like phenotype that was more prominent in middle-aged than in
younger mice.
We next analyzed PDXP-dependent motor function. To exclude
potential body weight differences as a confounding factor, we first
compared the body weights of 4–5 months-old and 8–9 months-old
male mice. There was a trend towards higher body weights in middle-
aged, whole-body PDXP-KO mice, but this tendency did not reach sta-
tistical significance (4–5 months-old: WT, 29.64 ± 0.30 g, n=8;
PDXP-KO; 30.18 ± 1.06 g, n=8, p= 0.63; 8–9 months-old: WT,
31.23 ± 0.63 g, n= 12; PDXP-KO, 33.36 ± 0.86 g, n= 12, p= 0.06.
Data are means ± S.E.M.). The inverted screen test showed a sig-
nificantly decreased performance of older PDXP-KO compared to WT
mice (Fig. 7A). In the weights test, performance was already impaired
in younger PDXP-KO mice (Fig. 7B; see Experimental Procedures for
details). To evaluate muscle strength directly, we conducted grip
strength measurements. Fore paw grip strength measurements in
8–9 months-old male mice did not show PDXP-dependent differences
(WT 1.46 ± 0.04 N, n= 12; KO 1.42 ± 0.07 N, n= 11, p= 0.61). In
contrast, hind paw grip strength measurements revealed poorer per-
formance of PDXP-KO mice (WT 0.92 ± 0.02 N, n=8; KO
0.77 ± 0.02 N, n= 11; p< 0.0001. Data for front and hind paw
measurements are means ± S.E.M.).
PLP is a cofactor of glycogen phosphorylase, the enzyme that cat-
alyzes the rate-limiting step in glycogenolysis. Because muscle glycogen
is an important energy substrate during exercise, and PLP levels in
skeletal muscle were significantly increased in PDXP-KO mice (see
Fig. 3A), we asked whether PDXP loss in skeletal muscle or in the
Table 2
UPLC-MS/MS-based quantification of B
6
vitamers in the brains of adult and postnatal WT and PDXP-KO mice.
nmol/g protein nPLP PL PA PMP PM PN
Cerebellum WT 4 85.10 ± 7.05 9.18 ± 0.69 0.032 ± 0.03 162.1 ± 22.46 0.14 ± 0.02 0.01 ± 0.01
PDXP-KO 4 266.9 ± 11.31 19.91 ± 3.82 0.12 ± 0.03 193.2 ± 18.26 0.14 ± 0.03 0.04 ± 0.02
p-Value < 0.0001 0.03 0.1 0.32 0.95 0.22
Hindbrain WT 4 89.08 ± 8.08 18.17 ± 1.99 n.d. 225.0 ± 7.200 0.55 ± 0.22 0.04 ± 0.02
PDXP-KO 4 234.8 ± 14.02 44.23 ± 1.97 n.d. 234.5 ± 14.82 0.63 ± 0.08 0.07 ± 0.05
p-Value 0.0001 < 0.0001 –0.59 0.75 0.58
Cortex WT 4 64.33 ± 3.89 13.16 ± 1.15 0.09 ± 0.03 145.1 ± 13.59 0.37 ± 0.09 0.007 ± 0.01
PDXP-KO 4 250.8 ± 18.15 34.53 ± 3.07 0.16 ± 0.07 158.7 ± 10.24 0.38 ± 0.10 0.06 ± 0.03
p-Value < 0.0001 < 0.001 0.41 0.46 0.97 0.10
Midbrain WT 4 85.41 ± 0.17 15.35 ± 1.25 0.06 ± 0.04 155.8 ± 15.03 0.39 ± 0.11 0.010 ± 0.004
PDXP-KO 4 277.8 ± 6.02 43.18 ± 2.94 0.04 ± 0.04 191.1 ± 6.6 0.74 ± 0.17 0.11 ± 0.05
p-Value < 0.0001 0.0001 0.73 0.07 0.14 0.09
P6 brain WT 4 95.25 ± 6.82 28.91 ± 1.21 0.07 ± 0.05 51.39 ± 4.98 0.17 ± 0.05 n.d.
PDXP-KO 4 201.8 ± 3.97 48.00 ± 1.63 0.04 ± 0.04 53.53 ± 8.86 0.26 ± 0.03 0.01 ± 0.01
p-Value < 0.0001 < 0.0001 0.65 0.84 0.18 0.36
Shown are mean values ± S.E.M. Nindicates the number of analyzed mice. For comparison with the other B
6
vitamers, results of the PLP measurements depicted in
Fig. 3A and B are shown again here. Pyridoxine 5′-phosphate (PNP) could not be detected. Statistically significant differences between the two genotypes were
analyzed using unpaired t-tests. Adult mice were 4 months-old males. P6, postnatal day 6; n.d., not detected.
E. Jeanclos et al. BBA - Molecular Basis of Disease xxx (xxxx) xxx–xxx
7
A
whole brain
p < 0.001
p < 0.01 p < 0.01
p = 0.02
p < 0.001
B
0
10
20
30
40
50
60
70
0
50
100
150
200
250
adult P6
whole brain
cerebellum
cerebellum
hindbrain
cortex
midbrain
glutamate (µmol/g protein)
Cadult P6
whole brain
cerebellum
cerebellum
hindbrain
cortex
midbrain
n.s.
n.s.
GABA (µmol/g protein)
n.s.
n.s.
n.s.
n.s.
n.s. n.s.
DA
DOPAC
epinephr.
HVA
5-HT
5-HIAA
0.0
0.5
1.0
1.5
3
6
9n.s.
n.s.
n.s. n.s.
n.s.
n.s.
neurotransmitter/metabolite
(nmol/g wet brain)
WT
KO
α-
GAD 67
WT KO WT KO WT KO
150 -
100 -
kDa
250 -
150 -
100 -
75 -
50 -
WT KO WT KO WT KO WT KOWT KO
-
-
-
-
-
-
-
α-
GAD 65
α-
FAK
α-
FAK
p = 0.001 n.s.
p < 0.001
rel. signal intensity
WT KO WT KO
WT KO
<<
<<
GAD67 GAD65
GAD65 HMW
rel. signal intensity
0
5
10
15
0.00
0.25
0.50
0.75
1.00
1.25
D
Fig. 4. Effect of whole-body PDXP deletion on neurotransmitter levels in the brain. (A) HPLC-based quantification of the indicated neurotransmitters and their
catabolites in whole-brain extracts of 9 months-old male mice (n= 5 per genotype). DA, dopamine; DOPAC, 3,4-dihydroxyphenyl acetic acid; HVA, homovanillic
acid; epinephr., epinephrine; 5-HT, 5-hydroxytryptamine (serotonin); 5-HIAA, 5-hydroxyindole acetic acid. (B) UPLC-MS/MS-based quantification of GABA levels in
the brain. Whole brains were isolated from 8 months-old male mice (n= 4 per genotype). The indicated brain regions were dissected from 4 months-old male mice
(n= 4 per genotype). P6, analysis of whole brains or dissected cerebellae from pups at postnatal day 6 (n= 4 per genotype). (C) UPLC-MS/MS-based quantification
of glutamate levels in the brain. Samples were the same as those analyzed in (B). The whiskers indicate minimum and maximum values. (D) PDXP deficiency affects
GAD67 and GAD65. Left and middle panels, GAD67 and GAD65 immunoblots. Brain lysates were generated from 8 months-old male mice, n= 4 per genotype. Focal
adhesion kinase (FAK) served as a loading control. Blots were analyzed densitometrically, and GAD65/67 signals were normalized to FAK signals. The top right panel
shows the quantification of GAD monomers (indicated by “<”in the blots), and the bottom right panel of the GAD65 high-molecular weight (HMW) complexes
(indicated by “<<”in the blots). Note that all bands corresponding to GAD monomers and HMW complexes were analyzed densitometrically and are included in
the figure, but that some symbols are overlapping and thus hidden. Data are mean values ± S.E.M.; n.s., non-significant. WT, wildtype mice; KO, Pdxp
flx/flx
; EIIa-Cre
mice.
E. Jeanclos et al. BBA - Molecular Basis of Disease xxx (xxxx) xxx–xxx
8
nervous system was primarily responsible for the apparent motor im-
pairments. To this end, mice deficient for PDXP either in skeletal
muscle (Pdxp
flx/flx
; Acta1-Cre
+/−
; mKO) or in neural cells (Pdxp
flx/flx
;
Nestin-Cre
+/−
; nKO) were generated. We verified efficient PDXP dele-
tion in skeletal muscle and whole brain extracts in these two lines
(Supplementary Fig. S2A, B). Because of the known metabolic issues
related to knockouts driven by Nestin-Cre [56], mice with a neural cell-
directed PDXP deletion were only characterized at 4–5 months of age.
PDXP deletion in neural cells led to an increase in whole brain PLP
levels (Supplementary Fig. S2C). Although there was a tendency to-
wards increased GABA concentrations in nKO mice, these changes were
statistically non-significant (Supplementary Fig. S2D). Similar to brain
extracts generated from whole-body PDXP-KO mice (see Fig. 4D), we
detected a high molecular weight, GAD65-positive band in im-
munoblots of nKO brain lysates (Supplementary Fig. S2E). These results
suggest that the appearance of high molecular weight GAD65 com-
plexes is a sensitive indicator of increased PLP levels, and that cell types
other than neurons, for example glial and endothelial cells [57] con-
tribute to the overall regulation of GABA levels in whole-body PDXP-KO
mice.
Grip strength measurements in young mice showed that PDXP ab-
lation in neural cells significantly reduced hind paw strength, whereas
PDXP loss in skeletal muscle increased hind paw strength in older mice
(Fig. 7C). Hence, PDXP deficiency in skeletal muscle does not account
for the decreased grip strength of globally PDXP-deficient mice. We
conclude that a dominant loss of PDXP function in another organ, most
likely the brain, causes the decrease in neuromuscular strength ob-
served in whole-body PDXP-KO mice.
4. Discussion
We demonstrate that PDXP (also known as PLPP, chronophin or
CIN) is a major regulator of PLP concentrations in vivo, and that whole-
body PDXP ablation elevates PLP and GABA levels in the brain.
Spatial learning and memory was improved in PDXP-deficient
compared to WT mice, and GABA may be involved in this behavior.
Many neuropsychiatric disorders and neurodegenerative diseases en-
compass dysfunctional GABA system components, and the disruption of
GABAergic inhibitory circuits with a resulting excitatory/inhibitory
imbalance is likely to contribute to some of the clinical features of these
disorders [58–60]. GABA signaling is also impaired in the aged hip-
pocampus, suggesting that pharmacological strategies might improve
memory function and spatial navigation abilities in the elderly popu-
lation [61,62]. A recent neuroimaging study that investigated the as-
sociation between vitamin B6 blood levels and cognition, brain struc-
ture and functional connectivity in healthy older humans (N > 600;
age range 55–85 years) has shown that vitamin B6 supplementation
was positively related to cortical folding and thus to the preservation of
brain structure. However, neuropsychological testing did not detect an
improvement in cognitive performance, and only slight changes in
functional connectivity were found [63]. Based on these data, the au-
thors hypothesized that vitamin B6 might counteract the gray matter
loss that frequently occurs in older age, and stabilize cognitive abilities.
A meta-analysis of brain GABA levels in different neuropsychiatric
diseases has revealed a GABAergic deficit in the brains of subjects with
autism spectrum disorders and major depression [64], and subtype-
selective GABA-A receptor modulators show potential as novel anti-
depressants [65,66]. GABA concentrations below a certain threshold
lead to seizures, and a number of anti-epileptic drugs target the GA-
BAergic system [67]. GABA levels are determined by GABA biosynth-
esis via PLP-dependent glutamic acid decarboxylases, and GABA cata-
bolism through the PLP-dependent GABA transaminase. The selective
GABA transaminase suicide inhibitor vigabatrin [68] increases GABA
levels, and is used for the treatment of infantile spasms and as an ad-
junctive therapy for refractory and complex partial seizures in adults
[69,70]. Despite its clinical efficacy, vigabatrin therapy is associated
with harmful side effects, such as the dose-dependent risk for pro-
gressive and permanent visual field defects [69,70]. It is tempting to
speculate that a combination of a pharmacological PDXP inhibitor to-
gether with vigabatrin may further increase GABA levels in the brain
and allow for a vigabatrin dose reduction.
The anxiety-like phenotype of PDXP-KO mice was unexpected,
given the physiological role of GABA in counteracting anxiety via
GABA-A receptors [58,65,66]. However, PLP and GABA levels were
already elevated postnatally in our mouse model. GABA levels affect
brain development [71] and synaptogenesis [72], and can thereby
impact stress responses later in life. For example, postnatal adminis-
tration of the GABA-A receptor agonist muscimol results in anxiety-like
behavior in 3-months old mice [73]. The EIIa-Cre driven, embryonic
ablation of PDXP might affect brain development (through GABA and/
or other mechanisms), and influence stress susceptibility and anxiety in
that way. PDXP loss might also alter the balance between excitatory and
inhibitory transmission in specific synapses in the adult organism. For
example, GABA co-released with glutamate controls the activity of the
lateral habenula, a region in the dorsal thalamus that regulates mood
and anxiety [74]. Finally, we found elevated D/L-serine levels (stereo-
isomers were not resolved in our analysis) in the cortex, and decreased
glycine concentrations in the hindbrains of PDXP-KO mice. Serine ra-
cemase is a PLP-dependent enzyme, and D-serine is a co-agonist at N-
methyl-D-aspartic acid (NMDA) glutamate receptors in the brain. Gly-
cine acts as an inhibitory neurotransmitter in the spinal cord and
brainstem, and has excitatory effects in the cerebral cortex due to its
agonistic activity at glutamatergic NMDA receptors. These changes may
contribute to the observed behavioral phenotypes. Further work is ne-
cessary to investigate the electrophysiological consequences of PDXP
1234567
0
5
10
15
three-trial blocks
primary errors (#)
1234567
0
20
40
60
80
100
three-trial blocks
1234567
0
50
100
150
three-trial blocks
1234567
0
200
400
600
800
1000
three-trial blocks
distance (cm)
KO
WT
target latency (sec)escape latency (sec)
p=0.026
p<0.001
p=0.006
p=0.007
p=0.008
p=0.007
p=0.004
p=0.013
p=0.027
p=0.030
Fig. 5. Whole-body PDXP loss improves performance in the Barnes maze.
Spatial learning and memory was assessed over five consecutive days, and three
trials were performed per day (‘three-trial blocks’, shown as mean
values ± S.E.M.). Afterwards, reversal learning was tested by moving the es-
cape chamber to the opposite side of the maze (indicated by the vertical dotted
line), and mice were analyzed for another two days in six additional trials.
Scored parameters were target latency (time to locate the hole with the escape
chamber), primary errors (number of holes searched before finding the hole
equipped with the escape chamber), escape latency (time needed to enter the
escape chamber) and distance (overall path length required before reaching the
hole with the escape chamber). Tests were performed with n= 12 mice per
genotype. All mice were 7 months-old males. Data are mean values ± S.E.M.,
p-values indicate statistically significant differences between WT and KO mice
in each trial block. WT, wildtype mice; KO, Pdxp
flx/flx
; EIIa-Cre mice.
E. Jeanclos et al. BBA - Molecular Basis of Disease xxx (xxxx) xxx–xxx
9
loss in specific brain regions and synapses, and to characterize the in-
volved receptors and signaling pathways in our mouse model [75].
In addition to a GABA-dependent decrease in muscle tone [76], the
apparent motor deficits of PDXP-KO mice might be influenced by in-
creased anxiety. In addition, age-dependent processes related to con-
tinuously elevated cellular PLP levels may contribute to the observed
phenotype. An increased anxiety-like behavior could add to the retarda-
tion of spatial learning that was observed when PLPP/CIN (PDXP)-defi-
cient mice were tested in the Morris water maze in another study [37].
The water maze imposes a much stronger aversive stimulus and more
stressful conditions on mice than the dry-land Barnes maze used in our
study, and increased stress and anxiety may confound the assay outcome
[54]. The discrepant results obtained in the two PDXP-KO mouse models
might also be linked to differences in the analyzed mouse strains, as well
as in animal age, -sex or -housing conditions. The mice investigated in our
present study were conditionally PDXP-deficient males of the indicated
ages; animals were on a C57BL/6J background and maintained under
specific pathogen-free conditions. Kim et al. have described mice with a
constitutive PDXP knockout on a mixed 129/SvEv-C57BL/6J albino
background; age, sex and housing conditions of the animals were not
A
C
D
0.0
0.5
1.0
1.5
2.0
0
10
20
30
40
50
B
0
10
20
30
40
p=0.002
0
20
40
60
80
p<0.0001
p=0.034
0
10
20
30 p=0.010
p=0.015
overall center time (%)
WT KO KO
WT
4-5 mo 9-10 mo
time on open arms (%)
WT KO KOWT
4-5 mo 9-10 mo
open arm entries (%)
WT KO KOWT
4-5 mo 9-10 mo
time in lit compartment (%)
latency (min)
WT KO KO
WT
4-5 mo 9-10 mo
WT KO KO
WT
4-5 mo 9-10 mo
0
5
10
15
20
marbles buried (#)
WT KO KOWT
4-5 mo 9-10 mo
p=0.016
n.s.
p=0.001
p=0.042
812128
812128
812128812128
8121288 12128
n.s.
n.s. n.s.
n.s.
0
500
1000
1500
2000
2500
distance (cm)
0
1500
3000
4500
6000
distance (cm) distance (cm)
0
500
1000
1500
2000
2500
WT KO KOWT
4-5 mo 9-10 mo
812128
WT KO KO
WT
4-5 mo 9-10 mo
812128
WT KO KOWT
4-5 mo 9-10 mo
812128
Fig. 6. Whole-body PDXP loss results in a
mild anxiety-like behavior. (A) Open field
test. Scored parameters were time in the
center of the open field arena and overall
locomotion (horizontal distance covered in
the 30 min test period). (B) Elevated plus
maze. Open arm entries, time on open arms
and overall locomotion were scored during
the 10 min test period. (C) Dark/light box.
The latency to enter and the time spent in
the lit compartment was measured together
with the overall locomotion in the 10 min
test period. (D) Marble burying. The
number of marbles buried within the 30 min
test period was counted. Mo, months. Data
are mean values + S.E.M. The number of
analyzed mice is indicated in the bars. WT,
wildtype mice; KO, Pdxp
flx/flx
; EIIa-Cre mice.
E. Jeanclos et al. BBA - Molecular Basis of Disease xxx (xxxx) xxx–xxx
10
reported [37]. The inbred strain background can impact behavioral
measures, and learning and memory is an age-dependent process [77–79].
Previous work by us and others has investigated the cofilin reg-
ulatory function of PDXP [31–36], and a recent study has reported ef-
fects of PDXP on hippocampal cofilin phosphoregulation in vivo [37].
Yet, a direct cofilin phosphatase activity of PDXP is difficult to reconcile
with crystallographic work from different laboratories, including our
own [21,38,80] (Protein Data Bank entries 2OYC, 2P69, 2P27, 2CFR,
2CFS, 2CFT, 5GYN, 4BX3, 4BXO, 4BX2, 4BKM, 5AES). PLP is deeply
buried within the catalytic cleft of PDXP [80], and the active site is
occluded by a large capping domain typical of small molecule-directed,
haloacid dehalogenase (HAD)-type hydrolases [81,82]. Given these
structural features, extensive conformational changes would be re-
quired to accommodate a protein substrate. Without structural in-
formation on the mechanism of cofilin dephosphorylation by PDXP, it
currently appears more likely that the effects of PDXP on phosphoco-
filin that are found in some studies are caused by indirect mechanisms.
Supplementary data to this article can be found online at https://
doi.org/10.1016/j.bbadis.2018.08.018.
Transparency document
The Transparency document associated this article can be found, in
online version.
Acknowledgements
We thank Angelika Keller and Kerstin Hadamek for excellent tech-
nical assistance, and Heinrich Blazyca and Drs. Dennis Klein and Rudolf
Martini for generous help with grip strength measurements. Marjolein
Bosma is thanked for the analysis of B6 vitamers. We acknowledge the
assistance of Charlotte Auth and Gabriel Christmann with immunoblots.
This work was supported by the Deutsche Forschungsgemeinschaft
SFB728 and SFB688 (to AG) and the IZKF Würzburg (to LH). CW was
supported by a predoctoral fellowship from the Medical Faculty of the
University of Würzburg (Graduate School of Life Sciences) and by a
Kaltenbach predoctoral fellowship from the German Heart Foundation.
References
[1] R. Percudani, A. Peracchi, A genomic overview of pyridoxal-phosphate-dependent
enzymes, EMBO Rep. 4 (2003) 850–854.
[2] J.P. Richard, T.L. Amyes, J. Crugeiras, A. Rios, Pyridoxal 5′-phosphate: electrophilic
catalyst extraordinaire, Curr. Opin. Chem. Biol. 13 (2009) 475–483.
[3] M.L. di Salvo, R. Contestabile, M.K. Safo, Vitamin B(6) salvage enzymes: me-
chanism, structure and regulation, Biochim. Biophys. Acta 1814 (2011) 1597–1608.
[4] N. Darin, E. Reid, L. Prunetti, L. Samuelsson, R.A. Husain, M. Wilson, B. El Yacoubi,
E. Footitt, W.K. Chong, L.C. Wilson, H. Prunty, S. Pope, S. Heales, K. Lascelles,
M. Champion, E. Wassmer, P. Veggiotti, V. de Crecy-Lagard, P.B. Mills,
P.T. Clayton, Mutations in PROSC disrupt cellular pyridoxal phosphate homeostasis
and cause vitamin-B6-dependent epilepsy, Am. J. Hum. Genet. 99 (2016)
1325–1337.
[5] J.W. Whittaker, Intracellular trafficking of the pyridoxal cofactor. Implications for
health and metabolic disease, Arch. Biochem. Biophys. 592 (2016) 20–26.
[6] M. Albersen, M. Bosma, N.V. Knoers, B.H. de Ruiter, E.F. Diekman, J. de Ruijter,
W.F. Visser, T.J. de Koning, N.M. Verhoeven-Duif, The intestine plays a substantial
role in human vitamin B6 metabolism: a Caco-2 cell model, PLoS One 8 (2013)
e54113.
[7] M.L. Fonda, Purification and characterization of vitamin B6-phosphate phosphatase
from human erythrocytes, J. Biol. Chem. 267 (1992) 15978–15983.
[8] Y.M. Jang, D.W. Kim, T.C. Kang, M.H. Won, N.I. Baek, B.J. Moon, S.Y. Choi,
O.S. Kwon, Human pyridoxal phosphatase. Molecular cloning, functional expres-
sion, and tissue distribution, J. Biol. Chem. 278 (2003) 50040–50046.
[9] M.L. di Salvo, M.K. Safo, R. Contestabile, Biomedical aspects of pyridoxal 5′-
phosphate availability, Front. Biosci. (Elite Ed.) 4 (2012) 897–913.
[10] P. Zhang, K. Tsuchiya, T. Kinoshita, H. Kushiyama, S. Suidasari, M. Hatakeyama,
H. Imura, N. Kato, T. Suda, Vitamin B6 prevents IL-1beta protein production by
inhibiting NLRP3 inflammasome activation, J. Biol. Chem. 291 (2016)
24517–24527.
[11] P.M. Ueland, A. McCann, O. Midttun, A. Ulvik, Inflammation, vitamin B6 and re-
lated pathways, Mol. Asp. Med. 53 (2017) 10–27.
[12] P.T. Clayton, B6-responsive disorders: a model of vitamin dependency, J. Inherit.
Metab. Dis. 29 (2006) 317–326.
[13] P.B. Mills, E. Struys, C. Jakobs, B. Plecko, P. Baxter, M. Baumgartner,
M.A. Willemsen, H. Omran, U. Tacke, B. Uhlenberg, B. Weschke, P.T. Clayton,
Mutations in antiquitin in individuals with pyridoxine-dependent seizures, Nat.
Med. 12 (2006) 307–309.
[14] P.B. Mills, R.A. Surtees, M.P. Champion, C.E. Beesley, N. Dalton, P.J. Scambler,
S.J. Heales, A. Briddon, I. Scheimberg, G.F. Hoffmann, J. Zschocke, P.T. Clayton,
Neonatal epileptic encephalopathy caused by mutations in the PNPO gene encoding
pyridox(am)ine 5′-phosphate oxidase, Hum. Mol. Genet. 14 (2005) 1077–1086.
[15] G.F. Hoffmann, B. Schmitt, M. Windfuhr, N. Wagner, H. Strehl, S. Bagci, A.R. Franz,
P.B. Mills, P.T. Clayton, M.R. Baumgartner, B. Steinmann, T. Bast, N.I. Wolf,
J. Zschocke, Pyridoxal 5′-phosphate may be curative in early-onset epileptic en-
cephalopathy, J. Inherit. Metab. Dis. 30 (2007) 96–99.
[16] M. Albersen, M. Bosma, J.J. Jans, F.C. Hofstede, P.M. van Hasselt, M.G. de Sain-van
der Velden, G. Visser, N.M. Verhoeven-Duif, Vitamin B6 in plasma and cere-
brospinal fluid of children, PLoS One 10 (2015) e0120972.
[17] B. Plecko, M. Zweier, A. Begemann, D. Mathis, B. Schmitt, P. Striano,
M. Baethmann, M.S. Vari, F. Beccaria, F. Zara, L.M. Crowther, P. Joset, H. Sticht,
S.M. Papuc, A. Rauch, Confirmation of mutations in PROSC as a novel cause of
vitamin B6-dependent epilepsy, J. Med. Genet. 54 (2017) 809–814.
[18] H. Schaumburg, J. Kaplan, A. Windebank, N. Vick, S. Rasmus, D. Pleasure,
M.J. Brown, Sensory neuropathy from pyridoxine abuse. A new megavitamin syn-
drome, N. Engl. J. Med. 309 (1983) 445–448.
[19] K. Dalton, M.J. Dalton, Characteristics of pyridoxine overdose neuropathy syn-
drome, Acta Neurol. Scand. 76 (1987) 8–11.
[20] S.M. Gospe Jr., M.P. Adam, H.H. Ardinger, R.A. Pagon, S.E. Wallace, L.J.H. Bean,
K. Stephens, A. Amemiya (Eds.), Pyridoxine-dependent Epilepsy, GeneReviews®,
Seattle (WA), 1993.
[21] G. Knobloch, N. Jabari, S. Stadlbauer, H. Schindelin, M. Kohn, A. Gohla, Synthesis
of hydrolysis-resistant pyridoxal 5′-phosphate analogs and their biochemical and X-
ray crystallographic characterization with the pyridoxal phosphatase chronophin,
Bioorg. Med. Chem. 23 (2015) 2819–2827.
[22] H. Prinsen, B.G.M. Schiebergen-Bronkhorst, M.W. Roeleveld, J.J.M. Jans,
M.G.M. de Sain-van der Velden, G. Visser, P.M. van Hasselt, N.M. Verhoeven-Duif,
AB
0
5
10
15
20
25 p=0.001
p=0.012
0
1
2
3
4
WT KO WT
4-5 mo 9-10 mo
p<0.0001
p<0.0001
test score
time (min)
0.6
0.8
1.0
1.2
1.4
hindpaw strength
(normalized to mean of controls)
p=0.004
p=0.039p=0.007
n.s.
KO WT KO WT
4-5 mo 9-10 mo
KO
24 24 36 36 88 12 12
C
4-5 mo 9-10 mo
WT KO WT KO WT mKOnKOWT
89 97 1211 66
Fig. 7. PDXP loss impairs motor performance and reduces muscle strength. (A)
Inverted screen test. Latency to fall offthe grid into the cage was measured. (B)
Weights test. The test score represents the number of chain links the mice lifted
for 3 s, multiplied by the time the weight was held in sec. Data in (A) and (B) are
mean values + S.E.M. (C) Hind paw grip strength measurements. Grip strengths
were normalized to the mean values of the respective control cohorts. The
whiskers show minimum and maximum values; the median value is indicated in
the box. The number of analyzed mice are given in the bars or underneath the
box plots. KO, whole-body PDXP-KO; nKO, neural cell-specific PDXP-KO; mKO,
skeletal muscle-specific PDXP-KO. All animals were males; WT signifies floxed
littermate controls (for nKO and mKO mice) or WT mice (for whole-body KO).
Mo, months; n.s., non-significant.
E. Jeanclos et al. BBA - Molecular Basis of Disease xxx (xxxx) xxx–xxx
11
Rapid quantification of underivatized amino acids in plasma by hydrophilic inter-
action liquid chromatography (HILIC) coupled with tandem mass-spectrometry, J.
Inherit. Metab. Dis. 39 (2016) 651–660.
[23] R.R. Schur, M.P. Boks, E. Geuze, H.C. Prinsen, N.M. Verhoeven-Duif, M. Joels,
R.S. Kahn, E. Vermetten, C.H. Vinkers, Development of psychopathology in de-
ployed armed forces in relation to plasma GABA levels, Psychoneuroendocrinology
73 (2016) 263–270.
[24] M. van der Ham, M. Albersen, T.J. de Koning, G. Visser, A. Middendorp, M. Bosma,
N.M. Verhoeven-Duif, M.G. de Sain-van der Velden, Quantification of vitamin B6
vitamers in human cerebrospinal fluid by ultra performance liquid chromato-
graphy-tandem mass spectrometry, Anal. Chim. Acta 712 (2012) 108–114.
[25] A. Raab, S. Popp, K.P. Lesch, M.J. Lohse, M. Fischer, J. Deckert, L. Hommers,
Increased fear learning, spatial learning as well as neophobia in Rgs2(−/−) mice,
Genes Brain Behav. 17 (2018) e12420.
[26] R.M. Deacon, Digging and marble burying in mice: simple methods for in vivo
identification of biological impacts, Nat. Protoc. 1 (2006) 122–124.
[27] T. Strekalova, R. Spanagel, D. Bartsch, F.A. Henn, P. Gass, Stress-induced anhedonia
in mice is associated with deficits in forced swimming and exploration,
Neuropsychopharmacology 29 (2004) 2007–2017.
[28] M.Y. Liu, C.Y. Yin, L.J. Zhu, X.H. Zhu, C. Xu, C.X. Luo, H. Chen, D.Y. Zhu,
Q.G. Zhou, Sucrose preference test for measurement of stress-induced anhedonia in
mice, Nat. Protoc. 13 (2018) 1686–1698.
[29] R.M. Deacon, Measuring the strength of mice, J. Vis. Exp. 76 (2013), https://doi.
org/10.3791/2610.
[30] B. Kohl, S. Fischer, J. Groh, C. Wessig, R. Martini, MCP-1/CCL2 modifies axon
properties in a PMP22-overexpressing mouse model for Charcot-Marie-tooth 1A
neuropathy, Am. J. Pathol. 176 (2010) 1390–1399.
[31] A. Gohla, J. Birkenfeld, G.M. Bokoch, Chronophin, a novel HAD-type serine protein
phosphatase, regulates cofilin-dependent actin dynamics,Nat.CellBiol.7(2005)21–29.
[32] M. Schulze, O. Fedorchenko, T.G. Zink, C.B. Knobbe-Thomsen, S. Kraus, S. Schwinn,
A. Beilhack, G. Reifenberger, C.M. Monoranu, A.L. Siren, E. Jeanclos, A. Gohla,
Chronophin is a glial tumor modifier involved in the regulation of glioblastoma
growth and invasiveness, Oncogene 35 (2016) 3163–3177.
[33] V. Delorme-Walker, J.Y. Seo, A. Gohla, B. Fowler, B. Bohl, C. DerMardirossian,
Chronophin coordinates cell leading edge dynamics by controlling active cofilin
levels, Proc. Natl. Acad. Sci. U. S. A. 112 (2015) E5150–E5159.
[34] C.X. Sun, M.A. Magalhaes, M. Glogauer, Rac1 and Rac2 differentially regulate actin
free barbed end formation downstream of the fMLP receptor, J. Cell Biol. 179
(2007) 239–245.
[35] M. Zoudilova, P. Kumar, L. Ge, P. Wang, G.M. Bokoch, K.A. DeFea, Beta-arrestin-
dependent regulation of the cofilin pathway downstream of protease-activated re-
ceptor-2, J. Biol. Chem. 282 (2007) 20634–20646.
[36] T.Y. Huang, L.S. Minamide, J.R. Bamburg, G.M. Bokoch, Chronophin mediates an
ATP-sensing mechanism for cofilin dephosphorylation and neuronal cofilin-actin
rod formation, Dev. Cell 15 (2008) 691–703.
[37] J.E. Kim, Y.J. Kim, D.S. Lee, J.Y. Kim, A.R. Ko, H.W. Hyun, M.J. Kim, T.C. Kang,
PLPP/CIN regulates bidirectional synaptic plasticity via GluN2A interaction with
postsynaptic proteins, Sci. Rep. 6 (2016) 26576.
[38] C. Kestler, G. Knobloch, I. Tessmer, E. Jeanclos, H. Schindelin, A. Gohla,
Chronophin dimerization is required for proper positioning of its substrate speci-
ficity loop, J. Biol. Chem. 289 (2014) 3094–3103.
[39] S.P. Coburn, Vitamin B-6 metabolism and interactions with TNAP, Subcell.
Biochem. 76 (2015) 207–238.
[40] H. Mehansho, M.W. Hamm, L.M. Henderson, Transport and metabolism of pyr-
idoxal and pyridoxal phosphate in the small intestine of the rat, J. Nutr. 109 (1979)
1542–1551.
[41] M.P. Whyte, J.D. Mahuren, L.A. Vrabel, S.P. Coburn, Markedly increased circulating
pyridoxal-5′-phosphate levels in hypophosphatasia. Alkaline phosphatase acts in
vitamin B6 metabolism, J. Clin. Invest. 76 (1985) 752–756.
[42] K.G. Waymire, J.D. Mahuren, J.M. Jaje, T.R. Guilarte, S.P. Coburn, G.R. MacGregor,
Mice lacking tissue non-specific alkaline phosphatase die from seizures due to de-
fective metabolism of vitamin B-6, Nat. Genet. 11 (1995) 45–51.
[43] H. Zimmermann, D. Langer, Tissue-nonspecific alkaline phosphatase in the devel-
oping brain and in adult neurogenesis, Subcell. Biochem. 76 (2015) 61–84.
[44] M.K. Rahman, T. Nagatsu, T. Sakurai, S. Hori, M. Abe, M. Matsuda, Effect of pyr-
idoxal phosphate deficiency on aromatic L-amino acid decarboxylase activity with
L-DOPA and L-5-hydroxytryptophan as substrates in rats, Jpn. J. Pharmacol. 32
(1982) 803–811.
[45] E. Roberts, K. Kuriyama, Biochemical-physiological correlations in studies of the
gamma-aminobutyric acid system, Brain Res. 8 (1968) 1–35.
[46] D.L. Martin, K. Rimvall, Regulation of gamma-aminobutyric acid synthesis in the
brain, J. Neurochem. 60 (1993) 395–407.
[47] D.F. Bu, M.G. Erlander, B.C. Hitz, N.J. Tillakaratne, D.L. Kaufman, C.B. Wagner-
McPherson, G.A. Evans, A.J. Tobin, Two human glutamate decarboxylases, 65-kDa
GAD and 67-kDa GAD, are each encoded by a single gene, Proc. Natl. Acad. Sci. U.
S. A. 89 (1992) 2115–2119.
[48] J.J. Soghomonian, D.L. Martin, Two isoforms of glutamate decarboxylase: why?
Trends Pharmacol. Sci. 19 (1998) 500–505.
[49] C.S. Pinal, A.J. Tobin, Uniqueness and redundancy in GABA production, Perspect.
Dev. Neurobiol. 5 (1998) 109–118.
[50] I. Kass, D.E. Hoke, M.G. Costa, C.F. Reboul, B.T. Porebski, N.P. Cowieson, H. Leh,
E. Pennacchietti, J. McCoey, O. Kleifeld, C. Borri Voltattorni, D. Langley, B. Roome,
I.R. Mackay, D. Christ, D. Perahia, M. Buckle, A. Paiardini, D. De Biase, A.M. Buckle,
Cofactor-dependent conformational heterogeneity of GAD65 and its role in auto-
immunity and neurotransmitter homeostasis, Proc. Natl. Acad. Sci. U. S. A. 111
(2014) E2524–E2529.
[51] G. Giardina, R. Montioli, S. Gianni, B. Cellini, A. Paiardini, C.B. Voltattorni,
F. Cutruzzola, Open conformation of human DOPA decarboxylase reveals the me-
chanism of PLP addition to Group II decarboxylases, Proc. Natl. Acad. Sci. U. S. A.
108 (2011) 20514–20519.
[52] C. Lieberwirth, Y. Pan, Y. Liu, Z. Zhang, Z. Wang, Hippocampal adult neurogenesis:
its regulation and potential role in spatial learning and memory, Brain Res. 1644
(2016) 127–140.
[53] S. Ge, D.A. Pradhan, G.L. Ming, H. Song, GABA sets the tempo for activity-depen-
dent adult neurogenesis, Trends Neurosci. 30 (2007) 1–8.
[54] F.E. Harrison, A.H. Hosseini, M.P. McDonald, Endogenous anxiety and stress re-
sponses in water maze and Barnes maze spatial memory tasks, Behav. Brain Res.
198 (2009) 247–251.
[55] U. Rudolph, H. Mohler, Analysis of GABAA receptor function and dissection of the
pharmacology of benzodiazepines and general anesthetics through mouse genetics,
Annu. Rev. Pharmacol. Toxicol. 44 (2004) 475–498.
[56] J. Declercq, B. Brouwers, V.P. Pruniau, P. Stijnen, G. de Faudeur, K. Tuand,
S. Meulemans, L. Serneels, A. Schraenen, F. Schuit, J.W. Creemers, Metabolic and
behavioural phenotypes in Nestin-Cre mice are caused by hypothalamic expression
of human growth hormone, PLoS One 10 (2015) e0135502.
[57] S. Li, T.P. Kumar, S. Joshee, T. Kirschstein, S. Subburaju, J.S. Khalili, J. Kloepper,
C. Du, A. Elkhal, G. Szabo, R.K. Jain, R. Kohling, A. Vasudevan, Endothelial cell-
derived GABA signaling modulates neuronal migration and postnatal behavior, Cell
Res. 28 (2018) 221–248.
[58] U. Rudolph, H. Mohler, GABAA receptor subtypes: therapeutic potential in Down
syndrome, affective disorders, schizophrenia, and autism, Annu. Rev. Pharmacol.
Toxicol. 54 (2014) 483–507.
[59] O. Marin, Interneuron dysfunction in psychiatric disorders, Nat. Rev. Neurosci. 13
(2012) 107–120.
[60] B. Luscher, Q. Shen, N. Sahir, The GABAergic deficit hypothesis of major depressive
disorder, Mol. Psychiatry 16 (2011) 383–406.
[61] J.A. McQuail, C.J. Frazier, J.L. Bizon, Molecular aspects of age-related cognitive
decline: the role of GABA signaling, Trends Mol. Med. 21 (2015) 450–460.
[62] A. Rozycka, M. Liguz-Lecznar, The space where aging acts: focus on the GABAergic
synapse, Aging Cell 16 (2017) 634–643.
[63] K. Jannusch, C. Jockwitz, H.J. Bidmon, S. Moebus, K. Amunts, S. Caspers, A com-
plex interplay of vitamin B1 and B6 metabolism with cognition, brain structure, and
functional connectivity in older adults, Front. Neurosci. 11 (2017) 596.
[64] R.R. Schur, L.W. Draisma, J.P. Wijnen, M.P. Boks, M.G. Koevoets, M. Joels,
D.W. Klomp, R.S. Kahn, C.H. Vinkers, Brain GABA levels across psychiatric dis-
orders: a systematic literature review and meta-analysis of (1) H-MRS studies, Hum.
Brain Mapp. 37 (2016) 3337–3352.
[65] H. Mohler, The GABA system in anxiety and depression and its therapeutic po-
tential, Neuropharmacology 62 (2012) 42–53.
[66] K.S. Smith, U. Rudolph, Anxiety and depression: mouse genetics and pharmacolo-
gical approaches to the role of GABA(A) receptor subtypes, Neuropharmacology 62
(2012) 54–62.
[67] M. Bialer, H.S. White, Key factors in the discovery and development of new anti-
epileptic drugs, Nat. Rev. Drug Discov. 9 (2010) 68–82.
[68] B. Lippert, B.W. Metcalf, M.J. Jung, P. Casara, 4-Amino-hex-5-enoic acid, a selective
catalytic inhibitor of 4-aminobutyric-acid aminotransferase in mammalian brain,
Eur. J. Biochem. 74 (1977) 441–445.
[69] W.D. Shields, J.M. Pellock, Vigabatrin 35 years later - from mechanism of action to
benefit-risk considerations, Acta Neurol. Scand. Suppl. (2011) 1–4.
[70] E. Ben-Menachem, Mechanism of action of vigabatrin: correcting misperceptions,
Acta Neurol. Scand. Suppl. (2011) 5–15.
[71] A. Represa, Y. Ben-Ari, Trophic actions of GABA on neuronal development, Trends
Neurosci. 28 (2005) 278–283.
[72] W.C. Oh, S. Lutzu, P.E. Castillo, H.B. Kwon, De novo synaptogenesis induced by
GABA in the developing mouse cortex, Science 353 (2016) 1037–1040.
[73] A.A. Salari, A. Bakhtiari, J.R. Homberg, Activation of GABA-A receptors during
postnatal brain development increases anxiety- and depression-related behaviors in
a time- and dose-dependent manner in adult mice, Eur. Neuropsychopharmacol. 25
(2015) 1260–1274.
[74] S.J. Shabel, C.D. Proulx, J. Piriz, R. Malinow, Mood regulation. GABA/glutamate
co-release controls habenula output and is modified by antidepressant treatment,
Science 345 (2014) 1494–1498.
[75] M. Farrant, Z. Nusser, Variations on an inhibitory theme: phasic and tonic activa-
tion of GABA(A) receptors, Nat. Rev. Neurosci. 6 (2005) 215–229.
[76] F. Crestani, K. L ow, R. Keist, M. Mandelli, H. Mohler , U. Rudolph, Mol ecular
targets for the my orelaxant acti on of diazepam, Mol. Pharmacol . 59 (2001)
442–445.
[77] J.N. Crawley, J.K. Belknap, A. Collins, J.C. Crabbe, W. Frankel, N. Henderson,
R.J. Hitzemann, S.C. Maxson, L.L. Miner, A.J. Silva, J.M. Wehner, A. Wynshaw-
Boris, R. Paylor, Behavioral phenotypes of inbred mouse strains: implications and
recommendations for molecular studies, Psychopharmacology 132 (1997) 107–124.
[78] K.R. Bailey, N.R. Rustay, J.N. Crawley, Behavioral phenotyping of transgenic and
knockout mice: practical concerns and potential pitfalls, ILAR J. 47 (2006)
124–131.
E. Jeanclos et al. BBA - Molecular Basis of Disease xxx (xxxx) xxx–xxx
12
[79] I.Schneider,W.S.Tirsch,T.Faus-Kessler,L.Becker,E.Kling,R.L.Busse,A.Bender,
B. Feddersen, J. Tritschler, H. Fuchs, V. Gailus-Durner, K.H. Englmeier, M.H. de Angelis,
T. Klopstock, Systematic, standardized and comprehensive neurological phenotyping of
inbred mice strains in the German Mouse Clinic, J. Neurosci. Methods 157 (2006) 82–90.
[80] S.C. Almo, J.B. Bonanno, J.M. Sauder, S. Emtage, T.P. Dilorenzo, V. Malashkevich,
S.R. Wasserman, S. Swaminathan, S. Eswaramoorthy, R. Agarwal, D. Kumaran,
M. Madegowda, S. Ragumani, Y. Patskovsky, J. Alvarado, U.A. Ramagopal,
J. Faber-Barata, M.R. Chance, A. Sali, A. Fiser, Z.Y. Zhang, D.S. Lawrence,
S.K. Burley, Structural genomics of protein phosphatases, J. Struct. Funct. Genom. 8
(2007) 121–140.
[81] A. Seifried, J. Schultz, A. Gohla, Human HAD phosphatases: structure, mechanism,
and roles in health and disease, FEBS J. 280 (2013) 549–571.
[82] A.M. Burroughs, K.N. Allen, D. Dunaway-Mariano, L. Aravind, Evolutionary geno-
mics of the HAD superfamily: understanding the structural adaptations and cata-
lytic diversity in a superfamily of phosphoesterases and allied enzymes, J. Mol. Biol.
361 (2006) 1003–1034.
E. Jeanclos et al. BBA - Molecular Basis of Disease xxx (xxxx) xxx–xxx
13