IntroductionAging is a biological process marked by the progressive accumulation of cellular, tissue, and organ damage, including that of the brain1,2,3. Among the most vulnerable brain regions is the hippocampus, which undergoes functional decline with age4,5, leading to impairments in reasoning, processing speed, and memory4. A key contributor to this age-related hippocampal deterioration is oxidative stress and the accumulation of abnormal proteins within mitochondria, both of which promote mitochondrial dysfunction, a recognized hallmark of aging6,7. Mitochondria generate ATP through oxidative phosphorylation (OXPHOS), a crucial process for maintaining cognitive functions in the hippocampus8,9. However, increased oxidative stress and impaired quality control mechanisms render mitochondria particularly susceptible to dysfunction in the brain during aging8,10,11.Mitochondria contain several proteases responsible for maintaining protein homeostasis, including members of the AAA + (ATPases Associated with diverse cellular Activities) superfamily12,13. Among them, the Lonp1 protease plays a particularly crucial role in preserving the integrity of the mitochondrial proteome14. Lonp1 is localized in the mitochondrial matrix and is ubiquitously expressed across all cell types and tissues15. Remarkably, it is estimated that Lonp1 mediates the degradation of over 50% of mitochondrial proteins16,17. To that end, Lonp1 selectively targets a range of substrates, including subunits of complexes I and V of the electron transport chain, thus contributing to the remodeling and stabilization of OXPHOS18,19. In addition to its proteolytic function, Lonp1 is also involved in maintaining mitochondrial DNA (mtDNA)20,21, regulating the stability of mitochondrial transcription factor A (TFAM), and interacting with mitochondrial polymerase γ to support mtDNA replication and transcription22,23. These functions underscore the crucial role of Lonp1 in maintaining both mitochondrial structure and bioenergetic capacity. The critical nature of Lonp1 is further demonstrated by the fact that its complete deletion in embryos is lethal. At the same time, heterozygous mice exhibit significant mitochondrial abnormalities, including altered morphology and impaired bioenergetic function in colon enterocytes and embryonic fibroblasts15,24. It is well supported that Lonp1 also plays critical roles during adulthood and in aging. Indeed, expression and activity of mitochondrial Lonp1 decline with age in multiple tissues, contributing to mitochondrial proteostasis failure, oxidative stress, and functional defects25,26,27,28. In aged kidneys, for example, reduced Lonp1 correlates with organ dysfunction and mitochondrial abnormalities29. Moreover, in models of muscle disuse and neurodegeneration, loss or impairment of Lonp1 accelerates features of aging and mitochondrial dysfunction30. These observations indicate that Lonp1 plays a crucial role in maintaining mitochondrial function in adult and aging tissues. However, despite its well-established roles in peripheral tissues, the function of Lonp1 in brain cells, particularly within the hippocampus, a region crucial for learning and memory, and in the context of aging, remains unexplored.Additionally, the expression and regulation of the Lonp1 gene remain poorly understood. In humans, Lonp1 exists in three isoforms, generated through RNA alternative splicing31,32. Transcript variant 1 contains the mitochondrial-targeting sequence (MTS) and does not undergo splicing, whereas variants 2 and 3 result from differential splicing of exon 1, leading to potential differences in subcellular localization and function. However, the presence of these Lonp1 isoforms in mouse cells has not yet been confirmed. Additionally, several regulators of Lonp1 expression have been identified. These include transcription factors such as Nuclear Respiratory Factor 2 (NRF-2)33,34. Moreover, Sirtuin-1, although not a transcription factor itself, can modulate Lonp1 transcription indirectly through the deacetylation and activation of PGC1α, which in turn stimulates NRF-2 signaling35. Nonetheless, the full spectrum of transcriptional regulators of Lonp1 in the brain, particularly within the context of aging, remains largely unexplored. This highlights a critical unresolved question in understanding how Lonp1 expression is controlled in neuronal tissues and across different species.Epigenetic modifications of histones and DNA influence gene expression during aging36. In humans, the LONP1 gene undergoes methylation changes at CpG sites in response to environmental factors, such as nitrogen dioxide pollution37. Moreover, hypomethylation of a LONP1 enhancer region has been reported and linked to pathological conditions such as myalgic encephalomyelitis/chronic fatigue syndrome (ME/CFS)38. This evidence suggests that epigenetic regulation of LONP1 is responsive to genome–environment interactions. Mecp2 is the most abundantly expressed member of the methylated DNA-binding protein family in the brain36,39. Mutations in Mecp2 have been shown to alter the expression of genes involved in mitochondrial function40. Supporting this, deep RNA sequencing of hippocampal tissue from wild-type and Mecp2 knockdown rats revealed changes in the expression of mitochondrial proteases, including the serine protease Htra2 and the ATP-dependent protease Clpx41. These findings suggest that Mecp2 may regulate nuclear genes encoding mitochondrial proteins, such as Lonp1. However, whether Mecp2 modulates Lonp1 expression and whether this regulation is influenced by aging remains unknown, as this has not yet been investigated.In this study, we investigated the age-related regulation of Lonp1 expression and proteostasis in the hippocampus, with a focus on the transcriptional role of Mecp2. Our findings uncover previously unrecognized epigenetic mechanisms involved in Lonp1 regulation during aging. Specifically, we observed increased cytosine DNA methylation in the Lonp1 promoter region. Notably, we demonstrate for the first time that Mecp2 directly binds to the Lonp1 promoter, and that this interaction is significantly reduced in the aged hippocampus. This reduction contrasts with the observed increase in Lonp1 mRNA levels, suggesting that Mecp2 functions as a transcriptional repressor of Lonp1. Furthermore, we report an age-associated decrease in total Mecp2 protein and its phosphorylated form (pSer80) in the nuclear fraction of the hippocampus, potentially explaining Mecp2 dimished capacity to bind and regulate Lonp1. Additionally, in contrast to human cells, we found that only Lonp1 transcript variant 1, the isoform encoding the full-length mitochondrial-targeted protein, is expressed in the hippocampus and other energy-demanding tissues of SAMP8 mice. Finally, despite increased Lonp1 mRNA levels in aging, we observed a paradoxical decrease in Lonp1 protein levels, which we attribute, at least in part, to lysosome-mediated degradation in aged tissue. Collectively, our results reveal a novel methylation-dependent transcriptional regulation of Lonp1 by Mecp2 and identify a critical disconnect between Lonp1 mRNA and protein expression in aging. These findings provide new insight into the molecular mechanisms linking mitochondrial proteostasis and epigenetic regulation in age-related hippocampal dysfunction.Materials and methodsAnimalsSAMP8 mice (Senescence-Accelerated Mouse Prone 8) and SAMR1 (Senescence-Accelerated Mouse Resistant 1) 2- and 7-month-old (mo) females were handled according to National Institutes of Health (NIH, Baltimore, MD) guidelines. This study was reported under ARRIVE guidelines. Animals were housed in temperature-controlled cages (24 °C) on a 12-h light/dark cycle, with food and water available ad libitum. The Bioethics and Biosafety Committee of the Universidad San Sebastián, Chile, approved all experimental procedures described in this study (CEC Nº 23–2021-20 and 0001–04-04–22). Animals were anesthetized with isoflurane and euthanized via decapitation. At the time of euthanasia, SAMP8 and SAMR1 mice were 2-month-old (25-28 g) or 7-month-old (30-35 g). The hippocampus was dissected for biochemical analysis. Each group consisted of almost three animals for biochemical assays (n ≥ 3).Isolation of hippocampal subcellular fractionsMouse hippocampal subcellular fractions were obtained as previously described6,10. Hippocampus was extracted, suspended, and lysed in MSH buffer (230 mM mannitol, 70 mM sucrose, 5 mM Hepes, pH 7.4) supplemented with protease and phosphatase inhibitors in a glass homogenizer. The homogenates were centrifuged at 600 × g for 10 min at 4ºC. The pellet obtained, used for nuclear protein extraction, was resuspended in MSH buffer, homogenized, and then centrifuged at 600 × g for 10 min at 4 °C. The pellet was resuspended in RIPA buffer and shaken for 30 min at 4 °C. Protein samples were centrifuged at 14,000 rpm for 20 min at 4 °C, and the supernatant obtained corresponds to nuclear proteins. The supernatant fraction obtained at 600 × g was centrifuged at 8000 × g for 10 min; the pellet obtained corresponded to the mitochondrial-enriched fraction, and the supernatant obtained corresponded to mitochondria-free cytoplasm. Protein concentration was determined using the BCA kit (Thermo Fisher Scientific, USA).Epigenetic and methylation analysesIn silico prediction of CpG islands in the Lonp1 geneUsing the genomic sequence of chromosome 17 of Mus musculus, we extracted the nucleotide sequence for the genomic region comprising 3000 bp downstream and 1000 bp upstream of the Transcriptional Start Site (TSS) for the Lonp1 gene. The genomic sequence of Lonp1 was downloaded from GenBank (NC_000083.7; Mus musculus strain C57BL/6 J chromosome 17, GRCm39; region 56,932,873–56,936,873). From this, we predicted the presence of CpG islands using five different methods: 1) Gardiner-Garden (1987) in the Julia Programming Language (versión 1.9.4, JuliaLang, https://julialang.org/)42, 2) Methyl Primer Express (Version 1.0, https://www.thermofisher.com/order/catalog/product/cl/es/4376041)43, 3) MethPrimer44 (version 1.1, https://methprimer.com/cgi-bin/methprimer/methprimer.cgi), 4) EMBOSS Cpgplot (version EMBOSS:6.6.0.0, https://emboss.sourceforge.net/apps/cvs/emboss/apps/cpgplot.html)45, and 5) CpGProD (Version 1.1, Ponger & Mouchiroud, (https://doua.prabi.fr/software/cpgprod_query)46. Subsequently, a visual inspection was performed to identify consensus genomic regions predicted as CpG islands, facilitating the design of MSP-PCR primers for experimental validation.MSP-PCR primer designUsing MethPrimer, forward and reverse primers were designed for the two predicted CpG islands, 1 and 2, respectively, at 5- and 10-nucleotide lengths. Quality control of the design primer was done using MFEprimer (version 3.1 (https://mfeprimer3.igenetech.com/spec)47, where we evaluated the formation of homodimers, heterodimers, and hairpins and filtered out all primers that showed the formation of any of these structures. From this, the primers were manually selected based on PCR product size, GC content percentage, and Tm to maximize both the PCR product size and the coverage of the predicted CpG islands.Analysis of global methylation patterns by MSP-PCRThe methylation status of the Lonp1 gene promoter was evaluated using a methylation-specific PCR (MSP)47. Genomic DNA extraction was performed using the GeneJet Genomic DNA Purification Kit (Thermo Fisher Scientific, cat. K0721) according to the manufacturer’s instructions. Genomic DNA was then modified with sodium bisulfite using the Epijet bisulfite conversion kit (Thermo Fisher Scientific, cat K1461), followed by methylation-specific polymerase chain reaction (MSP-PCR) using a Gotaq G2 green master mix kit (Promega, cat M7823). For MSP-PCR, Methylation- or no-methylation-specific primers were designed using the Methprimer bioinformatics software44. PCR products were visualized using electrophoresis performed on 2.5% agarose gels. The methylation ratio was obtained by comparing the MSP-PCR product densitometry with the generated partitions to identify differences in methylation levels in the identified CpG islands.Dot blot 5-methylcytosine assay in hippocampal genomic DNADot Blot DNA assessed the 5-methylcytosine status of genomic DNA from the hippocampus following the manufacturer’s instructions48. Briefly, genomic DNA was fragmented by sonication using the Bioruptor@Pico (Diagenode) for ten cycles of 30-s each, with 30 s between cycles. 250 ng fragmented genomic DNA was denatured in DNA denaturation buffer (200 mM NaOH, 20 mM EDTA) and incubated at 95 °C for 10 min. Then, sodium citrate saline buffer (3.0 M NaCl, 0.3 M Sodium Citrate, pH 7.0) was added and incubated at 4 °C for 5 min. Samples were loaded onto PVDF membranes in a 32-well Slot Blot apparatus (Bio-Rad). The PVDF membrane was UV crosslinked at 1200 J/m2. The membrane was blocked in 5% milk in TBS-Tween and incubated with a 5-methylcytosine primary antibody, then with an anti-rabbit IgG peroxidase-conjugated secondary antibody, and finally visualized using an ECL Kit (Luminata Forte Western HRP substrate, Merck Millipore, USA).Mecp2 binding analysesIn silico prediction of Mecp2 bindingWe used the same sequence used for CpG island prediction in the Lonp1 gene. The position weight matrices (PWM) of two known mouse MeCP2 binding motifs were downloaded from the Cis-BP database (https://cisbp.ccbr.utoronto.ca/)49. The motifs (Cis-BP IDs: Mecp2_M00806_2.00, Mecp2_M09259_2.00) were transformed into MEME format ((MEME Suite Team, https://meme-suite.org/) and then used to scan the Lonp1 sequence with the FIMO program from the MEME Suite (version 5.5.5, included in MEME Suite Team, https://meme-suite.org/)50. Significant matches were defined as p-value