Estructura de proteínas mitocondriales y complejos DNA-proteína

Abstract

[spa] El desarrollo de mi tesis doctoral abarca el análisis estructural y en solución de dos proteínas mitocondriales: la isoforma 1 del Factor de Transcripción A Mitocondrial (TFAM) y el dominio N-terminal de la isoforma 2 de la proteína 3A portadora de un dominio de la familia ATPasa AAA+ (ATAD3A). La capacidad de TFAM de doblar el ADN mitocondrial (ADNmt) ha sido descrita a nivel atómico, pero muy poco se conoce del mecanismo de acción de ATAD3A. ATAD3A se encuentra insertada en la membrana mitocondrial interna con el dominio C- terminal localizado en la matriz mitocondrial, pero la posición del dominio N-terminal aún no ha quedado totalmente acordada. Al contrario que ATAD3A, TFAM se encuentra soluble en la matriz mitocondrial y asociada al ADNmt. La expresión heteróloga de ATAD3A en bacteria no era viable por la presencia de una región transmembrana, así que decidimos expresar los dominios N- y C-terminal (NTD y CTD, respectivamente) por separado sin incluir tal región. Sin embargo, los constructos diseñados para el CTD no fueron solubles pese a los diversos intentos en diferentes cepas bacterianas, protocolos de plegamiento, cribado masivo de constructos, y hasta expresión en sistemas eucariotas como la cepa KM71H de la levadura Pichia pastoris y la línea celular humana Expi293FTM. Aún así, para el NTD obtuvimos un fragmento soluble que incluía la secuencia D51-S225 con una cola de seis histidinas en el extremo C-terminal. El análisis de este fragmento mostró una alta flexibilidad y bajo nivel de plegamiento, por lo que analizamos su estructura secundaria y dinámica mediante Resonancia Magnética Nuclear. En octubre de 2022 se identificó un nuevo promotor de la transcripción en la cadena ligera del ADNmt, denominado LSP2, así que intentamos cristalizar el complejo TFAM/LSP2 con el objetivo de ver si había diferencias respecto a la estructura cristalográfica del complejo TFAM/LSP. En el caso de TFAM se recurrió a un constructo que abarca la secuencia S43- C246 con una cola de seis histidinas en el extremo C-terminal. Para LSP2, se diseñó una molécula de ADN de 22 pares de bases. Tras comprobar por EMSA que la estequiometría ideal de proteína:ADN era 2:1, se llevaron a cabo ensayos de cristalización y mejora de las condiciones prometedoras hasta que se obtuvieron cristales del complejo aptos para difracción de rayos X. Finalmente, construimos la estructura cristalográfica del complejo TFAM/LSP2 y pudimos observar las diferencias con TFAM/LSP.

[eng] My PhD thesis focuses on the structure of two nuclear-encoded mitochondrial proteins, the isoform 1 of the Transcription Factor A from Mitochondria (TFAM) and the N-terminal domain of the isoform 2 of the ATPase family AAA+ domain-containing 3A protein (ATAD3A). While TFAM presumable compacts the mitochondrial DNA (mtDNA) by imposing U-turns to the nucleic acids, which has described at the atomic level, very little is known about the biochemical function of ATAD3A. However, it has been proved that ATAD3A is involved in subcellular processes such as cholesterol trafficking between the endoplasmic reticulum and mitochondria, maintenance of mitochondrial membranes and regulation of the mtDNA. TFAM has an N-terminal mitochondrial targeting sequence, directing its migration to the organelle, where it is found soluble and associated to the mtDNA. On the contrary, ATAD3A lacks a standardized targeting sequence, still unknown, and is located embedded in the mitochondrial inner membrane with its C-terminal ATPase domain facing the mitochondrial matrix; its N-terminal domains is thought to be at the intermembrane space, has a predicted helicoidal structure and probably interacts with protein partners. Regarding ATAD3A, since a transmembrane region is present in its sequence, which may hamper protein solubility when performing heterologous expression in Escherichia coli, different protein constructs of both N- and C-terminal domains (NTD and CTD, respectively) were designed excluding this predictably insoluble transmembrane region. Unfortunately, none of the CTD constructs succeeded in solubility despite trying protein expression in several bacterial strains, on-column refolding protocols, coexpression with chaperones, massive screening of constructs at the Expression of Soluble Proteins by Random Incremental Truncation (ESPRIT) facility in Grenoble (France), and expression in eukaryotic systems such as the yeast strain KM71H of Pichia pastoris and the human cell line Expi293FTM. In the end, we decided to stop such research on the CTD of ATAD3A because no positive result was obtained in our hands. Nevertheless, we were able to produce a soluble and stable construct of the NTD which spanned from D51 to S255, with a six-histidine tail positioned at its C-terminus. This protein fragment was very reluctant to crystallize, so we decided to study it in-solution. By Dynamic Light Scattering (DLS) and Small Angle X-ray Scattering (SAXS) we could confirm it is an elongated and flexible particle, Size Exclusion Chromatography couple to Multi Angle Laser Light Scattering (SEC-MALLS) analysis proved it is in a monomeric state and Circular Dichroism (CD) showed a predominant alpha-helix profile. All these results led us to perform structural studies by Nuclear Magnetic Resonance (NMR), which let us identify the alpha-helix and flexible regions, but no definitive folding has been fully determined because the 2D NMR spectra showed high overlapping and low signal/noise ratio in several peaks that should correspond to the central part of the predicted alpha-helix structure, suggesting a potential rigid subdomain. Regarding TFAM, the discovery of a new transcription promoter in the light strand of the mtDNA, named LSP2, which binds to the transcription factor, was a great opportunity to try to crystallize the complex and check if the compacting mechanism was the same as described for the crystallographic structure of TFAM/LSP. To achieve such goal, a TFAM construct spanning from S43 to C246, encompassing the full-length sequence upon mitochondrial targeting sequence processing, was produced. In addition, we designed, from the described sequence of LSP2, a DNA of 22 base pairs to verify the ability of TFAM to bind to LSP2 sequence by Electrophoretic Mobility Shift Assay (EMSA). Once the optimal protein:DNA ratio was established to be 2:1, several crystallization screenings and optimization of the crystallization conditions were performed until protein/DNA crystals suitable for X-ray diffraction were obtained. We reconstructed the crystallographic structure of the new complex TFAM/LSP2 and analyzed its differences with TFAM/LSP.