Expression, structural and functional characterization of the protein Proline Oxidase involved in different human diseases

Proline oxidase (PO) is a mitochondrial inner-membrane enzyme that catalyzes the first step in the proline degradation pathway. It is encoded by PRODH, a gene localized at 22q11.2: at least 16 PRODH missense mutations on the PRODH gene have been identified with moderate to severe effect on PO activity. Intracellular synthesis and degradation of free proline occur through a distinct set of enzymes with unique properties and regulatory mechanism: together they constitute a “proline cycle” which represents an important connection with other metabolic pathways. All organisms oxidize proline to glutamate in two enzymatic steps coupled by a non-enzymatic equilibrium: in the first step, proline oxidase (PO) catalyzes the oxidation of L-proline to Δ1-pyrroline-5-carboxilyc acid (P5C) a reaction that involves flavin adenine dinucleotide (FAD) as a cofactor. P5C is then alternatively subjected to three metabolic reactions: oxidation to glutamate, transamination to ornithine or reduction back to proline. Even if the biosynthetic pathways and biological role of proline have been well established, the role of proline oxidation in the cell remain largely unknown. The PRODH gene, maps in a genomic region (22q11) where several studies have localized genes that potentially confer increased susceptibility to schizophrenia. Schizophrenia is a common chronic and disabling brain disease of still largely unknown etiology and pathogenesis, that affects about 1% population worldwide. Schizophrenia occurs in all populations and is considered a complex disease with multiple factors contribution: the most widely accepted neurodevelopmental hypothesis of schizophrenia integrates environmental influences and causative genes. PRODH is considered a positional candidate gene for schizophrenia since delections of the 22q11 locus are found in 0.3-2% of patients with schizophrenia; moreover microdeletion of chromosome 22q11 is associated with velocardiofacial syndrome (VCFS) (or Di George syndrome) and patients with 22q11 deletion also exhibit cognitive deficits similar to those observed in schizophrenia. Interestingly, the localization of specific proline transporters within the glutamate synapses suggests that proline might serve as a modulator of glutamate transmission, hypothesis strongly supported by the observation that proline and glutamate are involved in the same metabolic pathways. Finally, association between plasma proline level and schizophrenia revealed that schizophrenia patients have significantly higher plasma proline levels than controls and that subjects with hyperprolinemia (HPI) have a 6-fold greater chance to develop schizophrenia than controls. HPI is an autosomal recessive disorder characterized by plasma proline concentration above normal, caused by absence or reduced activity of PO. This PhD project was aimed at understanding the role of PO under physiological and/or pathological processes and particularly as a candidate risk factor for schizophrenia. Despite the intriguing role of this protein in different human diseases, a correlation between genotype and phenotype of PO is still lacking mainly because the structural and functional characterization of the human protein has not yet been performed. For this reason we decided to characterize the recombinant PO by expressing the human protein in an heterologous system and to evaluate the effects of reported mutations on its catalytic properties. The final aim was to formulate a possible mechanism of PO regulation in the framework of glutamate metabolism and pathologies. E. coli is usually the most used organism for heterologous expression of recombinant proteins, but the expression of mammalian proteins, such as PO, results sometimes difficult in this host since E. coli does not perform most of the post-translational modifications. Moreover, human PO is localized in the mitochondrial inner membrane, so it could possess regions that negatively affect its expression in E. coli as a soluble protein. Therefore, by using an accurate bioinformatic analysis, several PO deletion variants were produced: all PO variants include a core region corresponding to the catalytic domain but they lack regions which could negatively affect the expression. Among the expressed PO variants, the PO-barrelN-His variant was expressed as a soluble protein and was subjected to a detailed biochemical characterization. A second aspect of the project focused on the expression of the wild-type and L441P inactive PO variant in a mammalian cell system. The U87 human glioblastoma cell line was selected since, despite representing the worst malignant form of astrocytic brain tumors, it is used as in vitro model of astroglia which is the responsible of the release and reuptake of signaling molecules, among which the key neurotransmitter glutamate. U87 clones stably expressing wild-type and L441P inactive variant PO were successfully produced, as confirmed by Western blot analysis. PO has been reported to be a mitochondrial protein thus, to verify the subcellular localization, PO variants were also overexpressed as chimeric proteins with EYFP fluorescent protein. Confocal images show that the fluorescent signal corresponding to PO-EYFP displays a peculiar “spaghetti like” pattern typical of mitochondria. Furthermore, PO-EYFP variants signal distribution largely overlapped with mitochondrial immunostain. Subsequently, the investigation of the functional differences between U87 cells transfected with the wild-type or the mutant PO was started. The cell growth rate was assayed over a period of 7 days: the growth curves (cell density/well vs time) display the typical pattern of exponential growth. No significant difference was evident between cells stably expressing the variant or the wild-type PO, while both grew slower than untransfected U87 cells. This result could indicate that expression of PO negatively affected the growth rate of transfected U87 cells, while the PO activity did not. The next step was the investigation of the cellular content of proline, since variation in the concentration of the amino acid is indicative of the cellular PO activity. Proline is reported to be directly involved in glutamatergic transmission and even though the biochemical mechanism by which this happens is still unknown, two hypotheses have been formulated: proline may accumulate in the synaptic terminals inhibiting glutamate release or act on a cell surface transporter inhibiting the release through a complex signaling cascade. For these reasons we measured the proline/glutamate and glutamine levels in cellular and extracellular space of U87 cells by HPLC analysis. At the same time we started the investigation of the expression level of the glutamate transporters GLT-1 and EAAT-1 by Western blot analysis to verify if expression of active or inactive PO variants could affect glutamatergic functionality. In conclusion, the purification of a soluble, active human PO variant (encompassing the whole catalytic domain of the enzyme), as obtained in the first part of the project, represents the fundamental step to investigate structure-function relationships in proline oxidase in analogy to the biochemical studies performed on the prokaryotic homologues of the protein. Further studies will be necessary for the identification of the effect of the reported polymorphisms of the enzyme on its structure and activity and to understand how they could affect the proline/P5C redox cycle. These studies represent a required step for the identification of compounds acting as positive or negative effectors of this enzyme and that, modulating its properties at the molecular level, could be potentially used as drugs.

Il progetto di dottorato dal titolo “Expression, structural and functional characterization of the protein Proline Oxidase involved in different human diseases” ha avuto l’obiettivo di studiare l’enzima prolina ossidasi (PO) umano e di chiarire un ipotetico meccanismo attraverso il quale l’enzima è coinvolto nella modulazione del metabolismo del glutammato e delle patologie ad esso correlate (ad esempio schizofrenia, iperprolinemia di tipo I, etc… ma anche in numerosi tumori). Nella prima parte del progetto è stato effettuato il clonaggio e l’espressione in forma ricombinante dell’enzima nel sistema eterologo E. coli. La variante di delezione PO-BarrelN-His (contenente l’intero dominio catalitico della proteina) è stata espressa in forma solubile: ciò ha permesso una accurata caratterizzazione biochimica della proteina ricombinante. Per la prima volta è stato così possibile dimostrare che l’enzima umano presenta proprietà biochimiche simili agli omologhi microbici. Nella seconda parte del progetto, il gene codificante per PO è stato espresso in cellule U87 di glioblastoma umano. Ciò ha permesso di valutare la vitalità cellulare e l’effetto sulla concentrazione cellulare di importanti metaboliti (prolina e glutammato). In particolare si è studiata la localizzazione subcellulare (mitocondriale), l’effetto sulla crescita cellulare e sull’espressione di importanti trasportatori di membrana. I risultati di questo studio rappresentano un importante punto di partenza per la progettazione di nuove molecole (inibitori della PO) in grado di modulare la concentrazione di L-prolina (e dei metaboliti correlati) quali nuovi farmaci per la cura di importanti patologie umane.