Enzymes involved in the metabolism of flavins in prokaryotic organisms: structure-function relationship
- Autores: Sonia Susana Arilla Luna
- Directores de la Tesis: Milagros Medina Trullenque (dir. tes.)
- Lectura: En la Universidad de Zaragoza ( España ) en 2013
- Idioma: español
- Tribunal Calificador de la Tesis: Carlos Gómez-Moreno Calera (presid.), Jose Antonio Ainsa Claver (secret.), José Ramón Peregrina Bonilla (voc.), Juan Antonio Hermoso Domínguez (voc.), Inmaculada Yruela (voc.)
- Materias:
- Texto completo no disponible (Saber más ...)
- Resumen
Flavin mononucleotide (FMN) and flavin adenine dinucleotide (FAD), the biologically active forms of riboflavin (Rf, vitamin B2), are essential cofactors for numerous enzymes (i.e., dehydrogenases, oxidases, reductases) that play critical roles in a large number of one- and two-electron oxido-reduction processes critical to the major metabolic routes in all organisms. Conversion of Rf into FMN and FAD is catalysed in all organisms by the sequential action of two enzymatic activities: an ATP:riboflavin kinase that transforms Rf into FMN, and an ATP:FMN adenylyltransferase that catalyses the subsequent adenylylation of FMN into FAD. However, whereas eukaryotes use two different enzymes for FMN (flavokinase) and FAD (adenylyltransferase) production, in most prokaryotes both activities are catalysed by a single bifunctional protein, the FAD synthetase (FADS). Functional analyses have shown the FMNAT process in FADS is reversible, while the RFK activity is essentially irreversible. The structure of some prokaryotic FADS has been determined : FADS folds in two modules, each one mainly related with one of the activities. The RFK-module of FADS shares considerably sequence and structural homology with monofunctional RFKs and it is able to catalyse the RFK reaction by itself. The FMNAT module does not present sequence or structural homology with eukaryotic FMNAT, rather belonging to the nucleotidyltransferase (NT) superfamily, and it does not appear to be self-sufficient to transform FMN into FAD. Additionally, recent studies based on sequence analysis suggested some microorganisms such as L.monocytogenes have an additional enzyme which shares its N-terminal module with FADS but differs in its C-terminal module, which is shorter and lacks the consensus motives for RFK activity.
The functional knowledge of this family of proteins is particularly based on the CaFADS one. At the beginning of this thesis, the putative presence of a flavin binding site for the adenylylation activity, independent of that related to the phosphorylation activity, was proven for the first time. This novel site was proposed to be in an open and preformed cavity in the FMNAT module. Models created for the interaction of the substrates in this cavity suggested the functional relevance of some residues putatively involved in the allocation of the isoalloxazine ring of flavins during the catalytic process. In this study, the roles for Pro56, Pro58, Phe62, Leu98, Tyr106 and Phe128 in allocation of the flavin in the FMNAT module were analysed by site-directed mutagenesis. The effects introduced by the mutations on the substrates and products binding parameters, as well as on the kinetic parameters for the RFK and FMNAT activities of CaFADS, indicated these residues are critical to maintain the required geometry in the FMNAT active site to keep both ligand binding and catalytic activity. A role was confirmed for all of them in the stabilisation of the isoalloxazine moiety of flavins and the transient state for the FMNAT/FADpp reactions by direct interaction with the isoalloxazine ring of flavins. Besides, Pro56, Pro58 and Phe128 are also implicated in the conformation of the binding site for adenine nucleotides. Thus, they were expected to provide additional implications in the spatial distribution of other residues to achieve a catalytical competent conformation for the FMNAT activity. On the other hand, Phe62 contributes to the binding affinity of flavins at the FMNAT site. However, although contributing to flavins stabilization, it does not seem relevant for catalytical orientation and rather might be implicated in the transition state stabilization. Kinetics observations also showed that the replacement of Pro56, Pro58 and Leu98 prevented the FMNAT reaction, while the aromatic character of Phe62, Tyr106 or Phe128 was required for this activity. In the case of Phe128, specific properties of Phe side-chain appear contributing to achieve the highest catalytical efficiency for the FMNAT reaction. All these results further support the proposed allocation of the flavin binding site in the big and preformed cavity at the FMNAT module of CaFADS. Moreover, they also show that achieving the catalytical competent allocation and the stabilization of the flavin ligands at this binding site are required for the FMNAT and FADpp reactions, suggesting the role of FADS in FMNAT activity may reside in the ability of the protein to stabilize the ligands rather than its direct participation in the catalytic process.
The substitution of some of these residues also affected the RFK activity, this effect being especially notable for Pro56 and Phe62. This observation supports a role of each of the CaFADS modules in modulating the efficiency of the catalytic activity taking place in the other module, as previously reported. On view of all of these experimental observations and the crystallographic structure of CaFADS, which suggests that the protein might get organized into a dimer of trimers where the three protomers of each trimer are connected in a head-tail disposition, in this project we also studied the functional relevance of quaternary organizations. Opposite to the RFK module, which is able to catalyse the phosphorylation process on its own, no FMNAT reaction at all was observed when the FMNAT module ((¿188-338)FADS) was individually expressed, even in presence of the RFK module ((¿1-182)FADS). In spite of the fact that (¿188-338)FADS was able to acquire a three-dimensional conformation, no flavin binding was detected. Thus, the absence of FMNAT activity was related to a no catalytically folding which prevents ligands to bind, supporting that a bifunctional enzyme is needed for the FMNAT module to acquire a catalytic competent conformation in the same protomer. On the other hand, in vitro and in vivo assays with CaFADS and C.ammoniagenes cells showed that despite CaFADS is mainly a monomer it is able to stabilize different quaternary organizations. In vitro AFM single molecule studies showed these organization are promoted upon the interaction with some of its ligands. Kinetics studies gave a functional relevance to these quaternary states and, together with the cooperativity observed between modules, suggested that a dynamic and regulated conversion between the monomer and the dimer of trimers must take place during the catalytic cycles of CaFADS. Site-directed mutagenesis studies on residues suggested to be involved in the stabilization of the dimer of trimers by the three-dimensional structure (Arg66, Lys202, Glu206, Phe206, Asp298, Val300 and Glu301), supported the functional relevance of quaternary assemblies and indicated each oligomeric assembly could be stabilized by no covalent interactions between residues located in different regions of protomers.
The significant differences between mammal FMNAT and the domain responsible for this activity in bacterial FADS, together with the need of these cofactors for flavoproteins and flavoenzymes required to the viability of microorganisms, make prokaryotic FADS particularly attractive targets for the treatment of diverse diseases as tuberculosis (caused by Mycobacterium tuberculosis) and listeriosis (caused by Listeria monocytogenes). Given that CaFADS is the best characterized member of this family of enzymes, one of the aims of this thesis was to analyze the possibility to extrapolate the knowledge about CaFADS to MtFADS and LmFADS. A sequence comparative analysis of MtFADS, Lm1FADS and Lm2FADS and the production of in silico structural models for each one of these proteins is here presented here. Sequence alignments indicated that FMNAT and RFK activities must be present in MtFADS and Lm1FADS, while Lm2FADS, despite having all motives related to the FMNAT activity, does not contain the consensus sequences for RFK activity. In fact, the C-terminal module of Lm2FADS rather shows homology with domain II of TtATP-PRT, particularly in the highly conserved region of ATP-PRT. Based on these observations, three-dimensional models were proposed for these proteins free, and in clomplex with ligands. As expected, these models showed MtFADS and Lm1FADS fold in two modules where residues involved in RFK an FMNAT activity are structurally conserved, supporting MtFADS and Lm1FADS are members of the prokaryotic FADS family. On the other hand, sequence differences between Lm2FADS and the RFK module of the FADS family prevented the production of a three-dimensional model based on these proteins. In consequence, each module was individually modelled, and a couple of relative orientations between them evaluated. The FMNAT module shows high structural homology and structural conservation of residues involved in FMNAT activity with prokaryotic FADSs whereas the C-terminal, showing structural homology with ATP-PRT, presents structurally conserved the residues involved in PRPP binding but it lacks residues involved in ATP stabilization. Given that the FMNAT module presents an ATP binding site for FMNAT activity and one of the two models shows the two active sites closer enough to share a single ATP binding site, it is proposed that Lm2FADS may contain an ATP-PRT-like transferase activity.
In an attempt to extend our studies on the prokariotic FADS family to pathogenic bacteria, protocols for the recombinant expression in E.coli of Lm1FADS and Lm2FADS and for their subsequent purification have been here developed. Additionally, we have initiated the biochemical characterization of these enzymes. Lm1FADS showed high similarities with CaFADS, being able to catalyze both the RFK and FMNAT activities. Lm2FADS only shows FMNAT activity, supporting the predictions from sequence analysis and production of three-dimensional models. Despite the FMNAT module seems to be redundant in these two enzymes, kinetics characterizations showed these two enzymes have different energetic requirements for activity. FMN synthesis resides only in Lm1FADS and it is enhanced in presence of reducing agents. Regarding the production of FAD, Lm1FADS is unable to catalyze it unless being in a reducing ambient, while Lm2FADS appears as the responsible of FAD synthesis in oxidizing environment and its FMNAT is negatively affected under reducing conditions. The existence of two enzymes with different energetic requirements involved in flavin metabolism could be consequence of the facultative anaerobic nature of L.monocytogenes. Preliminary interaction studies with flavins showed these enzymes are able to bind oxidized flavins. Thus, the effect the reducing environment has in the efficiency of these enzymes is not related to their inability to bind oxidized flavins, but rather related to conformational requirements of the catalytical competent binding of the substrates or catalysis. Given that the active sites of all these enzymes are rich in aromatic residues and the reducing environment could affect their ¿-electron organization, the sensitivity of these activities may reside in the modification of the aromatic networks involving in flavin allocation. Thus, the selectivity for reduced flavins observed in FMNAT activity of Lm1FADS but not in Lm2FADS and CaFADS could be related to Tyr129, since this is the only one residue involved in the FMNAT activity that it is not identical to the structurally equivalent residue in the other enzymes at the isoalloxazine binding site.
Finally, a protocol to express MtFADS as a fusion protein in E.coli has been developed and a protocol for its subsequent purification suggested. It should be taken in account MtFADS shows a very high tendency to be sent to inclusion bodies even when it is expressed in a phylogenetically closer system as M.smegmatis. This high insolubility of MtFADS could be related to an inability of the FMNAT module to fold correctly. However, while solubilisation and refolding of the enzyme from these inclusion bodies recovers RFK activity, it doesn¿t recover the FMNAT one. This suggests the reversibility and irreversibility of RFK and FMNAT unfolding, respectively, and dismisses this methodology to obtain MtFADS pure to homogeneity. Since several problems in heterologous expression of other proteins from M.tuberculosis have been previously described and, in a similar way, LmFADS also showed some difficulties to express in solution, it might be inferred that these pathogenic microorganisms might present some specific characteristics which make their proteins unstable out of these bacteria.
Taking in account all this structural and functional information, these enzymes could be good targets to envisage the rational design of selective antimicrobial drugs. Moreover, CaFADS could be a good template in preliminary steps through the design of structure-based inhibitors for MtFADS and LmFADS.
In summary, all together the results here presented considerably contribute to increase the knowledge of the prokaryotic FADS family. On one hand, they contribute to best understand the structural mechanism of substrate recognition and catalysis and, on the other hand, they provide a platform to deal with the search of specific inhibitors for the FMNAT activity in pathogenic microorganisms.
