QDPR homologues in Danio rerio regulate melanin synthesis, early gliogenesis, and glutamine homeostasis

Dihydropteridine reductase (QDPR) catalyzes the recycling of tetrahydrobiopterin (BH4), a cofactor in dopamine, serotonin, and phenylalanine metabolism. QDPR-deficient patients develop neurological symptoms including hypokinesia, truncal hypotonia, intellectual disability and seizures. The underlying pathomechanisms are poorly understood. We established a zebrafish model for QDPR deficiency and analyzed the expression as well as function of all zebrafish QDPR homologues during embryonic development. The homologues qdpra is essential for pigmentation and phenylalanine metabolism. Qdprb1 is expressed in the proliferative zones of the optic tectum and eye. Knockdown of qdprb1 leads to up-regulation of pro-proliferative genes and increased number of phospho-histone3 positive mitotic cells. Expression of neuronal and astroglial marker genes is concomitantly decreased. Qdprb1 hypomorphic embryos develop microcephaly and reduced eye size indicating a role for qdprb1 in the transition from cell proliferation to differentiation. Glutamine accumulation biochemically accompanies the developmental changes. Our findings provide novel insights into the neuropathogenesis of QDPR deficiency.


Introduction
Dihydropteridine reductase (human: DHPR; mouse / zebrafish: Qdpr) is the key recycling enzyme of the cofactor tetrahydrobiopterin (BH 4 ). The homodimer uses NADH to supply two hydrogen atoms to BH 2 to recover BH 4 [1]. BH 4 is initially formed in a three-step pathway denovo from GTP and is then salvaged in a recycling pathway via pterin-4a-carbinolamine dehydratase (PCBD) and DHPR [2,3]. This pathway is highly conserved among species [4][5][6]. The a1111111111 a1111111111 a1111111111 a1111111111 a1111111111 zebrafish genome contains three DHPR homologs, Qdpra, Qdprb1 and Qdprb2, the function of which has remained largely unknown. BH 4 is cofactor in the enzymatic reaction of phenylalanine hydroxylase (PAH), tyrosine hydroxylase (TH) and tryptophan hydroxylase (TPH). PAH catalyzes the formation of tyrosine from phenylalanine and is therewith essential for phenylalanine degradation and synthesis of the neurotransmitter precursors L-dopa through TH, whereas TPH is responsible for serotonin biosynthesis. The BH 4 pathway is also needed for the activity of nitric oxide synthases (NOS) [2] and has been linked various clinical entities including autism [7], pain regulation [8] and cardiovascular diseases [9].
Hyperphenylalaninemia (HPA) due to inborn defects of PAH is the biochemical hallmark of phenylketonuria (PKU). In the majority of BH 4 deficiencies HPA also arises reflecting impaired PAH function due to shortage of BH 4 and is treated by diet or with BH 4 (sapropterin dihydrochloride) [10]. In contrast to PKU, BH 4 deficiencies are complicated by depletion of dopamine and serotonin, which can be overcome by oral supplementation with the neurotransmitter precursors L-Dopa and 5-HT [11].
In comparison to other BH 4 deficiencies, DHPR-deficient patients show a higher frequency of severe neurological symptoms including hypotonia, dystonia, microcephaly, epilepsy and brain atrophy [11,12]. Due to deficient DHPR activity 7,8-dihydrobiopterin (BH 2 ) accumulates that has been suggested to play a significant role in the neuropathogenesis of affected patients by inhibiting nitric oxide synthase and aromatic acid hydroxylases [13]. Oral BH 4 supplementation as recommended in other BH 4 deficiencies is likely to aggravate BH 2 accumulation and, therefore, dietary restriction of phenylalanine intake to prevent HPA is the treatment of choice in DHPR-deficient patients. Qdpr-knockout mice have been generated to better understand the pathophysiology of DHPR-deficiency. Indeed, these mice show similar biochemical alterations including HPA and BH 2 accumulation and, additionally, have a higher sensitivity to oxidative stress [5]. However, these animals do not develop a neurological phenotype. Unlike in humans, mice have a high cerebral dihydrofolate reductase (Dhfr) expression. This protein is essential for homeostasis of folates and can also recycle BH 2 [14]. High cerebral Dhfr activity in Qdpr-deficient mice may consequently explain the apparent lack of a neurological phenotype [5].
Since Qdpr knockout mice do not develop a clinical phenotype similar to DHPR-deficient patients, we aimed to establish a novel animal model to study the pathophysiology of DHPR deficiency. For the study of genetic diseases including inborn, metabolic disorders, zebrafish have become increasingly valuable as many proteins are conserved in their function, developmental changes can be observed ex utero, and their genome and gene expression can easily be manipulated. In fact, an increasing number of inborn errors of metabolism are identified and characterized in vivo via gene manipulations in zebrafish [15,16]. BH 4 metabolism has previously been studied in zebrafish, yet only regarding its role in pigment synthesis [6]. The de novo synthesis pathway of BH 4 generates pterin precursors for the yellow pigments in zebrafish eyes [6,17]. The dark pigment melanin is generated from tyrosine via L-dopa and eumelanin. Since this pathway is initiated by Th, melanin formation depends on de novo synthesis and recycling of BH 4 . Accordingly, down regulation of Pcd catalyzing the first step of BH 4 recycling lead to reduced melanin production in zebrafish [18,19].
The aim of this study was to analyze the DHPR zebrafish homologues (Qdpra, Qdprb1, Qdprb2) and their role in zebrafish development. We show that Qdpra is involved in the regulation of phenylalanine degradation and pigment synthesis. Qdprb1 on the other hand is essential for early gliogenesis and regulation of the glutamine homeostasis shining new light onto the pathophysiology of DHPR deficiency.

Fish husbandry
Zebrafish were maintained at 28˚C at a 14/10 h light/dark cycle and in accordance to [20]. Zebrafish were held in accordance with all international and national laws and obligations as registered at the Regierungspräsidium Karlsruhe (Az. .81/G-85/16). Embryos were raised in E3 medium at 28.5˚C until desired [21]. To inhibit pigment synthesis, embryos were kept in 0.003% 1-phenyl-2-thiourea (PTU) (Sigma) in E3 medium after gastrulation. AB wildtype fish and the following transgenic lines were used in this study: tg(HUC/D:GFP; Park [22] and tg (NBT/lyn:GFP).
The regional council (Regierungspräsidium Karlsruhe, Baden-Württemberg, Germany) prospectively approved this zebrafish research (Approval number .81/G-85/16). Data from one patient with DHPR deficiency was collected within the iNTD study, approved by the local ethic board of the Medical Faculty in Heidelberg (Approval number S-471/2014) [23].

Whole mount fluorescent immunohistochemistry
Whole mount fluorescent immunohistochemistry was performed as described previously [26]. Embryos were incubated overnight with primary antibody (pAb Rabbit Anti-phospho-Histone H3 (Ser10) (Millipore, Cat# 06-570) and, after a washing step, incubated overnight with secondary antibody (Cy2 AffiniPure Goat Anti-Rabbit IgG (H+L), Cat# 111-225-144) or DAPI. Secondary antibody was washed off and embryos were used for imaging. For fluorescence confocal microscopy, embryos were mounted in 1% low-melt agarose in glass-bottom dishes (MatTek or LabTek). Embryos were imaged using A TCS SP5 MP (Leica) inverse laser scanning microscope. Subsequent image analysis was performed with Fiji software.

Real time-quantitative PCR
RNA was isolated from 25-30 embryos of each condition from the same experiment using Trizol according to manufacturer's protocol. 500 ng isolated RNA was reverse transcribed via First Strand Maxima Synthesis kit (ThermoFisher). RT-qPCR was performed using SYBR Green Mix (Bioline) with previously tested primers on a BioRAD CFX Connect. Tests were run in duplicates of at least three independent experiments. Expression was normalized to Elongation factor 1 alpha expression. S1 Table shows the list of studied genes and used primer pairs.

Melanin quantification via spectrophotometry
Melanin content was determined as shown by [27,28]. Zebrafish were lysed in 1.0M NaOH with 10% DMSO. Lysate was incubated at 80˚C for 1 hour and absorbance measured at 475nm. Wildtype were set as 100%.

Biochemical measurements and exposures
Lysates from zebrafish embryos in water were used for analysis of amino acids. Embryos were lysed and homogenized on ice and in water by repeated passing through a 27G hypodermic needle (Terumo) and sonification using a "Branson 450 Sonifier" at 50% duty for 10 pulses with three repeats. Samples were stored at -20˚C until measurement. Deproteinization was done by addition of 20% sulfosalicylic acid and precipitates were removed by centrifuging samples at 13,000rpm for 30 minutes. Amino acids were analyzed by Biochrom 30 + Cation exchange chromatography. For glutamine exposures, embryos were exposed to 1 or 20mM L-Glutamine in E3 medium. Medium was changed daily. Exposure from sphere stage until 72 hpf was considered an early developmental hyperglutaminemia and exposure from 48 hpf until 72 hpf was considered a late developmental hyperglutaminemia.

Glutamine measurement in CSF of a patient with DHPR deficiency
Glutamine in CSF was measured during diagnostic work-up in a patient with DHPR deficiency as described previously [29]. At the time of lumbar punction the patient was not on medical treatment.

Statistics
All experiments were repeated for a minimum of three biological replicates. Single comparisons were analyzed via student's t-test and multiple comparisons using one-way ANOVA with post hoc Bonferroni and Holm evaluation. Values of p<0.05 were considered to be significant � .

Results
The zebrafish genome contains three annotated QDPR homologues, Qdpra, Qdprb1 and Qdprb2. Therefore, we tested their function individually.

Qdprb2 analysis
Qdprb2 is the homologue with the lowest homology to human DHPR with 62% amino acid conservation. Qdprb2 expression was not detected in whole mount in situ hybridization (WISH). RT-qPCR analysis showed very transient maternal qdprb2 expression (Figure A in S1 File). Consistently, functional knockdown with a qdprb2 specific Morpholino (MO) resulted in no morphological phenotype. We concluded that qdprb2 function has diverged during evolution and did not retain a functional role similar to its mammalian counterpart.

Qdpra regulates BH 4 -related phenylalanine and melanin metabolism
The homologue qdpra has 72% homology to the human DHPR protein. As shown by whole mount WISH, it is first visible at 18 hours post fertilization (hpf) in the eye and neural crest derived pigment cells, which is more apparent at 24 hpf when qdpra is also expressed in the retinal pigment epithelium of the eye (Fig 1A). At 48 hpf the expression is detected in developing melanophores along the axis of the body, as well as in the retinal pigment epithelium and choroid fissure of the eye. At 72 hpf, qdpra expression in the eye is decreasing, while expression in the liver arises. At 120 hpf, only the liver retains qdpra expression ( Fig 1A). RT-qPCR levels show a rise in qdpra expression at 20 hpf coinciding with the early production of melanin in melanophores [30], and, later on, expression levels increase at 72 hpf, co-inciding with the expression of Pah in the liver ( Figure B in S1 File, Figure A in S2 File).
Qdpra knockdown using a splice blocking MO did not result in visible morphological changes of the embryo. RT-PCR confirmed the predicted excision of exon 3 ( Figure B in S File). During the rapid increase of melanophore pigmentation at around 26 hpf treated fish appear pale (Fig 1B), which becomes fully apparent at 72 hpf, when pigmentation and melanin content is strongly reduced compared to controls (Fig 1C and 1D). Injection of qdpr mRNA from mouse restored melanin pigmentation confirming the specificity of the effect (Figure C in S2 File).
In summary, our findings indicate that Qdpra is necessary for melanin synthesis in melanophores. Biochemical analyses as well as region-specific and temporal expression profiles of qdpra suggest a role in hepatic phenylalanine metabolism.

Qdprb1 influences brain development
We next characterized the homologue qdprb1, which shows also 72% homology to the human homologue, yet only 73% homology to Qdpra. After weak expression during gastrulation, qdprb1 becomes mainly confined to the optic tectum (OT), eye and mid-hindbrain boundary (MHB) at 24 hpf (Fig 2A). At 48 hpf, qdprb1 expression in OT and MHB is retained, while the eye localizes along the developing ciliary marginal zone (CMZ), which is the main area for proliferating cells in the retina. This CMZ expression is retained at 72 hpf, by which time qdprb1 is also more confined to an area of proliferative cells in the OT [31] (Fig 2A). Additionally, qdprb1 expression at 72 hpf is observed in the inner retinal layer. mRNA expression levels show a continuous expression throughout development (Figure C in S1 File).
Knockdown of qdprb1 using splice blocking MO resulted in morphological abnormalities of the developing brain, first seen at around 18 hpf and affecting mid-and hindbrain. Morphants showed decreased brain size at 30 hpf and microcephaly at 72 hpf with aberrant midand hindbrain development (Fig 2B). Trunk and tail structures of qdprb1 hypomorphic embryos remained unaffected.
Co-injection of p53 MO [32] together with qdprb1 MO did not rescue the observed phenotypes highlighting that they did not result from p53 mediated apoptosis due to qdprb1 MO toxicity ( Figure

Metabolic alterations upon qdprb1 knockdown
Next, we assessed the biochemical consequences of qdprb1 reduction. Hypomorphic conditions did not affect phenylalanine levels at 72 hpf ( Fig 3A). Intriguingly, the most prominent metabolic alteration we found in these embryos was a strong increase of glutamine ( Fig 3B). Glutamine levels started rising at 18 hpf and steadily increased until 72 hpf ( Fig 3C). Glutamine accumulation was not rescued by the co-injection with p53 MO and replicated upon ATG MO injection ( Figure A, B in S4 File). This finding suggested that increased glutamine levels due to by qdprb1 hypomorphic conditions may be the cause underlying the observed brain abnormalities in fish. Contrary to the "Trojan Horse" [33][34][35] implying toxicity by glutamate and ammonia in mitochondria, we did not observe any changes in glutamate or NH 3 in qdprb1 morphants (Figure C in S4 File) suggesting a different link between increased glutamine levels and brain size, or that these two alterations are caused independently from each other.
Due to the unexpected metabolic phenotype of qdprb1 morphants, we next enzymatically characterized protein. We expressed and purified zebrafish Qdprb1 and human QDPR using the Champion pET SUMO Protein and Peptide Expression System (Thermo Fisher Scientific). In the subsequent enzyme assay both proteins reduced BH 2 to BH 4 confirming that zebrafish Qdprb1 has indeed in vitro properties of a dihydropteridine reductase (Qdprb1, Vmax 21 mU/ mg, Km 50μM; QDPR, Vmax 15mU/mg, Km 30μM).

Qdprb1 depletion increases proliferative markers
In line with its brain specific mRNA expression, qdprb1 hypomorphic embryos exhibit a brain specific morphological phenotype. Indeed, for instance spinal cord and lateral line neurons Early marker genes for the midbrain and mid-hindbrain boundary such as otx2 and Wnt1 only showed a reduction in the size of their expression domains suggesting that early patterning of qdprb1 expressing brain areas is unaffected ( Figure B, C in S5 File). Consistently, qdprb1 knockdown in tg(HuC/D:GFP) transgenic embryos expressing GFP in developing neurons confirmed that brain patterning and neuronal development occurred (Fig 4A and 4B). In contrast, sizes of the brain, most prominently of the optic tectum, and of the eye at 72 hpf were reduced, which could be rescued by co-injection of Qdprb1 mRNA (Fig 4A and 4B).
As early brain patterning appeared to be unaffected, we next investigated subsequent cellular processes to explain the observed microcephaly in qdprb1 hypomorphic embryos. Since we already excluded p53 mediated apoptosis, we focused our attention on cell proliferation. Indeed, we found an increased number of cells positive for the mitotic marker phospho- histone H3 (pH3) (72hpf) in the eye (Fig 3C and 3D) and the optic tectum ( Figure D in S5  File), i.e. qdprb1 expressing brain areas. Noteworthy, pH3 positive retinal cells were not restricted to the CMZ but were found throughout the retina. This is unlike wildtype retinal cells, which start to differentiate once leaving the CMZ, indicating that qdprb1 influences the transition from precursor cell proliferation and differentiation. DAPI staining of the retina of Qdprb1 hypomorphic embryos at 3 dpf showed that individual layers remained intact but the overall size was decreased and the CMZ was broadened (Fig 4E). The characteristic pH3 expression and localization in brain and eye as well as retina size and CMZ width was restored by qdprb1 mRNA injection (Fig 4C, 4B and 4D). In conclusion, qdprb1 knockdown does not influence major early brain patterning processes but instead causes a 4-fold increase of pH3 positive cells in the eye and optic tectum suggesting a function of Qdprb1 in promoting cell differentiation proliferation. The insets show dorsal views of the left eye, which is reduced in size but still layered upon qdprb1 suppression. This phenotype is rescued upon co-injection of qdprb1 mRNA. The overall GFP signal remains unchanged. (B) The brain area of Qdprb1 hypormorphic embryos is reduced by about 15% compared to wildtypes and rescued by co-injection of qdprb1 mRNA. (C) Z-stack overlays of DAPI (pink) and pH3 (green) staining of dorsally imaged retinas reveal an increased number of pH3-positive retinal cells in Qdprb1 hypormorphic embryos (3 dpf), which are not restricted to the CMZ like found in wildtype retinas. Number and distribution of pH3-positive cells is restored by co-injection of qdprb1 mRNA. (D) Statistical analysis shows a significant increase in pH3 positive cells upon qdprb1 knockdown (n = 6) in comparison to wildtype zebrafish (n = 7), which could be rescued by addition of qdprb1 mRNA (n = 3). (E) Z-confocal image of DAPI staining (pink) of the retina at 3 dpf shows retinal layers although decreased overall size and broadened CMZ in qdprb1 hypomorphic embryos. Due to the apparent microcephaly upon qdprb1 knockdown, an increased number of pH3 positive cells was rather surprising. Nevertheless, defects of neural stem cell differentiation can also result in microcephaly [36]. By RT-qPCR expression analysis in 26 hpf qdprb1 morphants we indeed found significant down regulation of genes critical for neurogenesis in brain and eye (elavl3, fold change 0.60 +/-0.09; neurod1, fold change 0.63 +/-0.02; notch1a, fold change 0.64 +/-0.09; rx2, fold change 0.38 +/-0.04), whereas genes of cell proliferation were up-regulated (myca, fold change 1.80 +/-0.32; ccnd1, fold change 1.40 +/-0.19). Reduction of gfap expression (fold change 0.63 +/-0.07) suggested that glial cell differentiation was also impaired. At 72 hpf neuronal markers suggested that dopaminergic neurons developed normally. This notion was corroborated by WISH staining of gch1 giving a similar expression pattern as found in wild-type zebrafish (Fig 5A). Expression of the neuronal differentiation genes

Qdprb1 knockdown depletes glial cell markers in the developing brain and eye
In mammals, astrocytes are the keepers of brain glutamine levels, removing glutamate from the synaptic cleft and shuttling it back to neurons in the form of glutamine. While teleosts lack typical star-shaped astrocytes, they contain radial glia that express specific astrocytic markers, such as glutamine synthetase a (glula), and gfap [37,38], suggesting that these cells have a similar function to astrocytes. Radial glial cells positive for astrocytic markers will further collectively be considered as astroglia in this manuscript.
To better understand the link between qdprb1 knockdown and the rise of glutamine levels, we studied the expression of genes critical for cerebral glutamine/glutamate cycle in 72hpf qdprb1 morphants, i.e. the stage with highest glutamine accumulation. Expression of the neuronal glutaminase a (glsa) was not affected (fold change 0.83 +/-0.12) in qdprb1 morphants, whereas its astrocytic counterpart, glula, was strongly down regulated ( Fig 5B). This finding was confirmed in WISH experiments, showing decreased expression of glula in the inner retinal layer of the eye and the proliferative region of the midbrain (Fig 5C). Furthermore, the expression of second astrocytic marker Gfap depends on Qdprb1 function suggesting a functional role for qdprb1 in astroglia development (Fig 5D).
In mammals, SLC1A3 (EAAT1) and SLC1A2 (EAAT2) mediate the astrocytic glutamate uptake. The zebrafish orthologues of SLC1A2 display a distinct expression pattern with slc1a2a being present in the ONL and slc1a2b being expressed in the astroglia of the eye and brain [39]. RT-qPCR analysis and WISH staining showed an almost complete loss of slc1a2a (Fig 5E  and 5F). Relative mRNA levels of slc1a2b were unchanged but WISH experiments showed a reduced expression in eye and midbrain (Fig 5G). In analogy to Slc1a2a/b, the two zebrafish orthologues of SLC1A3 show a distinct expression pattern with only Slc1a3b being expressed in brain tissue. Slc1a3b expression was unaltered in Qdprb1 hypomorphic embryos ( Figure A, B in S6 File). Of note, the Genbank ID HM138691.1 used by [39] refers to slc1a2a mRNA, yet the annotated nucleotide sequence is identical to the one of the paralogue slc1a2b in the RefSeq database (used in our study). The same discrepancy occurs for slc1a3a.
Next, we analyzed expression of cerebral glutamine transporters. In mammals, system A (SLC38A1, 2) and system L (SLC7A5, 8) are responsible for the uptake of glutamine in neurons [40] but their role in the zebrafish brain is virtually unknown. We analyzed expression of slc7a5 and slc38a2 in qdprb1 morphants by RT-qPCR showing a mildly reduced expression of the latter transporter ( Figure C in S6 File).
In summary, our data indicate a function for qdprb1 in the regulation of genes involved in brain glutamine metabolism and astroglia development.

Effects of chronic exposure to high glutamine
To shed more light on the mechanism of glutamine accumulation in Qdprb1 hypomorphic embryos, we assessed whether exposing zebrafish to high glutamine levels may induce a similar phenotype. Early developmental exposure (i.e. 5hpf to 72hpf) to high concentrations of glutamine (20 mM) resulted in developmental defects (Fig 6A), including small size of the head. RT-qPCR analysis revealed down regulation of gfap and slc1a2a expression as well as of qdprb1, though the latter effect was less pronounced (Fig 6B). Beside of highly increased cellular glutamine levels (wt, 55.93 +/-1.62 μmol/mg; exposed, 469.7 +/-283.53 μmol/mg) we also found ammonia accumulation (wildtype, 64,5 +/-3,17 μmol/mg; exposed 255.7 +/-147.72 μmol/mg) in treated zebrafish, which is not found in qdprb1 morphants. Later developmental exposure (i.e. 48hpf to 72 hpf) to 20 mM glutamine (Fig 6C) or exposure to lower glutamine levels (1 mM) did not induce a phenotype. Finally, we tested the glutamine synthase inhibitor L-methionine sulfoximine qdprb1 morphants. It indeed reduced glutamine concentrations providing more evidence for a link between Qdprb1 activity and glutamine homeostasis, but failed to rescue the observed morphological phenotype.
Overall these experiments suggest a possible regulatory feedback between qdprb1 and glutamine metabolism in early phases of development as well as a role for glutamine in astroglia development.

Glutamine in patients with DHPR deficiency
In case of a 1.5-year-old so far untreated female patient with DHPR deficiency, we documented an increase of glutamine in CSF (760 μmol/l (control range: 373-556 μmol/l)), while plasma levels remained normal. Clinically the patient presented a severe neurological picture with infantile Parkinsonism, epilepsy and brain atrophy. In follow-up analyses after treatment initiation with neurotransmitter precursors, folinic acid and phenylalanine reduced diet glutamine levels had normalized in CSF.

Discussion
QDPR is required for the regeneration of BH 4 that is a critical co-factor for the generation of dopamine and serotonin. Defects in the synthesis and recycling of BH 4 lead to severe infantile Parkinsonism. In comparison to defects in BH 4 biosynthesis, patients with inborn variants of QDPR present a higher frequency of severe neurological symptoms, including muscular hypotonia, dystonia, microcephaly, epilepsy and brain atrophy [11,12]. The aim of our study was to establish a suitable model to investigate the pathophysiology of QDPR deficiency. To this end, we characterized expression and function of the three zebrafish QDPR homologues Qdpra, Qdprb1, and Qdprb2.
Amongst these three genes, only Qdpra and Qdprb1 appear to have a zygotic function during embryonic development, while Qdprb2 may have a maternal contribution as judged by its expression.
At early stages (up to 24hpf), qdpra is expressed in the eye and neural crest cells. Later on (120hpf) its expression is mainly defined to the liver. We find that Qdpra regulates melanin content and pigment size, overall linking this paralogue to melanin synthesis. The temporal expression pattern of qdpra suggests that the requirement of BH 4 recycling for melanin synthesis is restricted to the early stages of melanophore development from neural crest cells, which has been shown for a number of genes [17]. Overall, qdpra expression correlates with gch2, the second paralogue of the de novo synthesis pathway initiator GCH, that is also expressed in pterin synthesizing xantophores [41] and BH 4 dependent melanophores [42]. The lack of qdpra expression in dopaminergic or serotonergic neurons, as labeled by gch1 and th respectively, makes it unlikely that BH 4 recycling in these cells involves this qdpr paralogue. The observed hyperphenylalaninemia in qdpra hypomorphs is in line with the biochemical phenotype of patients and likely reflects the requirement of BH 4 recycling for Pah activity. Accordingly, we find coexpression of qdpra and pah in the zebrafish liver.
Unlike the other two qdpr homologues, qdprb1 is localized in the zebrafish brain, more specifically in the proliferative regions of the eye (CMZ) and optic tectum as well as in the midhindbrain boundary. This qdprb1 expression in proliferative cells appears conserved and is also seen in proliferative brain regions of Medakafish [43] and embryonic mice [44]. Noteworthy, the proliferative regions of the eye and OT are genetically highly similar [45]. They harbor stem cells that differentiate into all glia and neurons required in the neuronal network [46,47] and stay proliferatively active throughout development. Suppression of qdprb1 results in microcephaly. While early brain patterning appears not affected, we observed an increased number of undifferentiated pH3-positive cells in the eye and OT. P53 mediated apoptosis was not enhanced. Further, we show up-regulation of pro-proliferative genes and down-regulation of genes associated with neural and glial differentiation. Our data suggest that qdprb1 influences the generation of neurons and astroglia from neuronal stem cells in the eye and OT. If qdprb1 function is suppressed, precursor cells fail to exit the cell cycle on time and remain undifferentiated. Indeed, the expression of proliferation promoting genes myca and ccnd1 stayed up-regulated at 72 hpf. This suggests that the block of neural stem cell differentiation persisted during the observed time window of development. Of note myca has been reported to play also a role in the maintenance of the neural stem cell pool in the optic tectum [45]. The role of zebrafish QDPR homologues in the pathophysiology of DHRP deficiency Upon qdprb1 knockdown the astroglia markers gfap and glula remained strongly reduced during the studied time window of development, while expression of some neuronal markers were unchanged indicating a more pronounced influence on gliogenesis. Also in the murine cerebellar cortex, Qdpr may play a role in glia cell development and/or function as it is strongly expressed glial cells [48].
Neuronal and glial differentiation starts with the switch of neural progenitor cells from symmetric to asymmetric cell division under the influence of pro-neuronal or pro-glial genes. Our data suggests that qdprb1 knockdown interferes with the underlying signaling network. Although the underlying mechanism remains to be elucidated, the observed changes in gene expression can explain the microcephaly of qdprb1 hypomorphs. Indeed, there is more evidence linking BH 4 to cellular proliferation and differentiation. Proliferation of erythroid cells has been shown to be dependent on BH 4 availability [49], while BH 2 application blocked differentiation of these cells [50]. Using the pheochromocytoma cell line PC12 Anastasiadis et al. found that cellular BH 4 levels influence c-Myc expression [51]. Further, tetrahydro-4-aminobiopterin, a structure analog of BH 4 and NOS inhibitor, blocks dentritic cell-mediated T cell differentiation independent of NO activity suggesting an alternative and novel BH 4 -dependent signaling pathway [52]. Though these studies could not identify the underlying mechanism, they found no evidence that the observed effects are based on BH 4 functioning as co-factor for PAH, TH, TPH or NOS. Of note, BH 2 exposure of zebrafish tested in our study did not affect embryonic development (up to 1 mM in tank water) suggesting that the lack of qdprb1 but not of BH 2 recycling induces the underlying pathomechanism. It needs to be addressed in future studies how and whether BH 4 / BH 2 directly or, as suggested by our study, via QDPR influence cell proliferation and differentiation.
The most prominent biochemical phenotype upon qdprb1 knockdown was a strong increase of glutamine levels. Considering the brain-specific expression of qdprb1, this finding indicated impairment of cerebral glutamine homeostasis, which, at least in mammals, is regulated by astrocytes. In line with this notion several genes involved in the astrocytic glutamate uptake and glutamine synthesis were strongly affected in qdprb1 hypomorphs. Microglia and oligodendrocytes have been identified and characterized in zebrafish [53,54], whereas teleosts in general lack typical star-shaped astrocytes [53]. Studies indicate that astroglia of zebrafish mediate axon guidance in early embryo development [55], have regenerative capacity, and can replace damaged or lost neurons [56][57][58]. Expression of typical astrocytic markers (e.g. glul, gfap and aqp4) suggests that further functions of mammalian astrocytes including glutamate recycling are conserved [37,53]. To get first insights into whether the reduction of glial markers was due to the observed increase of glutamine, we exposed wildtype zebrafish to different glutamine concentrations. This treatment resulted in reduced expression of qdprb1, gfap, and slc1a2a partially mimicking the qdprb1 knockdown phenotype. However, unlike observed in qdprb1 hypomorphic conditions, treated zebrafish also displayed elevated ammonia levels. One may speculate that this finding supports the "Trojan horse" theory [33][34][35] underlying the observed phenotype, i.e. induced brain damage by glutamine being shuttled to mitochondria and broken down into glutamate and ammonia, which will result in free radical production and swelling of the mitochondria. It must be taken into account that the glutamine treatment affects the entire embryo rather than only a few cell populations in qdprb1 knockdown embryos. Therefore, additional, more precise approaches are needed to explore the potential link between qdprb1, glutamine and glial development. If for instance astroglia cells in teleosts are found to be the sole source for cerebral glutamine production, then the effects of qdprb1 knockdown (glutamine increase and glia development) are two separate events. Alternatively, glutamine generation may derive from elevated pterin deaminase activity due to disturbed BH 4 recycling. This protein has mainly been studied in prokaryotes but it was also found in rat liver and zebrafish, and is thought to regulate BH 4 metabolism [59].
Cohort studies on BH4 deficiencies do not report increased glutamine levels in CSF or blood [11]. Retrospective evaluation of CSF samples of BH 4 deficient patients in our centre documented high glutamine levels in one young, untreated, and severely affected female DHPR-deficient patient with infantile Parkinsonism, epilepsy and brain atrophy. The fact that glutamine levels normalized after onset of treatment may suggest that glutamine levels correlate with disease severity and are normal in treated or less severely affected patients. Thus, there may well be a link between DHPR-deficiency and glutamine levels, which is reproducible in our fish model but had so far been overlooked in patients.

Conclusion
This study is the first to characterize the expression and function of the three zebrafish QDPR homologues qdpra, qdprb1, and qdprb2. While we could not identify a zygotic function for qdprb2 in embryonic development, qdpra executes the expected BH 4 recycling function in melanin producing cells and in the liver. Our results further suggest that qdprb1 is required for neuronal and glial differentiation of neural progenitor cells emerging from the CMZ and OT.  Table. List of studied genes and used primer pairs. (TIF)