Alternative Transcripts Research Articles

Alternative transcripts recode human genes to express overlapping, frameshifted microproteins.

Overlapping genes were thought to be essentially absent from the human genome until the discovery of abundant, frameshifted internal open reading frames (iORFs) nested within annotated protein coding sequences. However, it is currently unclear how many functional human iORFs exist and how they are expressed. We demonstrate that, in hundreds of cases, alternative transcript variants that bypass the start codon of annotated coding sequences (CDSs) can recode a human gene to express the iORF-encoded microprotein. While many human genes generate such non- coding alternative transcripts, they are poorly annotated. Here we develope a new analysis pipeline enabling the assignment of translated human iORFs to alternative transcripts, and provide long- read sequencing and molecular validation of their expression in dozens of cases. Finally, we demonstrate that a conserved DEDD2 iORF switches the function of this gene from pro- to anti- apoptotic. This work thus demonstrates that alternative transcript variants can broadly reprogram human genes to express frameshifted iORFs, revealing new levels of complexity in the human transcriptome and proteome.

Re-appraising the evidence for the source, regulation and function of p53-family isoforms.

The p53 family of proteins evolved from a common ancestor into three separate genes encoding proteins that act as transcription factors with distinct cellular roles. Isoforms of each member that lack specific regions or domains are suggested to result from alternative transcription start sites, alternative splicing or alternative translation initiation, and have the potential to exponentially increase the functional repertoire of each gene. However, evidence supporting the presence of individual protein variants at functional levels is often limited and is inferred by mRNA detection using highly sensitive amplification techniques. We provide a critical appraisal of the current evidence for the origins, expression, functions and regulation of p53-family isoforms. We conclude that despite the wealth of publications, several putative isoforms remain poorly established. Future research with improved technical approaches and the generation of isoform-specific protein detection reagents is required to establish the physiological relevance of p53-family isoforms in health and disease. In addition, our analyses suggest that p53-family variants evolved partly through convergent rather than divergent evolution from the ancestral gene.

Alternative Splicing of the Last TKFC Intron Yields Transcripts Differentially Expressed in Human Tissues That Code In Vitro for a Protein Devoid of Triokinase and FMN Cyclase Activity.

The 18-exon human TKFC gene codes for dual-activity triokinase and FMN cyclase (TKFC) in an ORF, spanning from exon 2 to exon 18. In addition to TKFC-coding transcripts (classified as tkfc type by their intron-17 splice), databases contain evidence for alternative TKFC transcripts, but none of them has been expressed, studied, and reported in the literature. A novel full-ORF transcript was cloned from brain cDNA and sequenced (accession no. DQ344550). It results from an alternative 3' splice-site in intron 17. The cloned cDNA contains an ORF also spanning from exon 2 to exon 18 of the TKFC gene but with a 56-nt insertion between exons 17 and 18 (classified as tkfc_ins56 type). This insertion introduces an in-frame stop, and the resulting ORF codes for a shorter TKFC variant, which, after expression, is enzymatically inactive. TKFC intron-17 splicing was found to be differentially expressed in human tissues. In a multiple-tissue northern blot using oligonucleotide probes, the liver showed a strong expression of the tkfc-like splice of intron 17, and the heart preferentially expressed the tkfc_ins56-like splice. Through a comparison to global expression data from massive-expression studies of human tissues, it was inferred that the intestine preferentially expresses TKFC transcripts that contain neither of those splices. An analysis of transcript levels quantified by RNA-Seq in the GTEX database revealed an exception to this picture due to the occurrence of a non-coding short transcript with a tkfc-like splice. Altogether, the results support the occurrence of potentially relevant transcript variants of the TKFC gene, differentially expressed in human tissues. (This work is dedicated in memoriam to Professor Antonio Sillero, 1938-2024, for his lifelong mentoring and his pioneering work on triokinase).

Identification of candidate causal variants and target genes at 41 breast cancer risk loci through differential allelic expression analysis

Understanding breast cancer genetic risk relies on identifying causal variants and candidate target genes in risk loci identified by genome-wide association studies (GWAS), which remains challenging. Since most loci fall in active gene regulatory regions, we developed a novel approach facilitated by pinpointing the variants with greater regulatory potential in the disease’s tissue of origin. Through genome-wide differential allelic expression (DAE) analysis, using microarray data from 64 normal breast tissue samples, we mapped the variants associated with DAE (daeQTLs). Then, we intersected these with GWAS data to reveal candidate risk regulatory variants and analysed their cis-acting regulatory potential. Finally, we validated our approach by extensive functional analysis of the 5q14.1 breast cancer risk locus. We observed widespread gene expression regulation by cis-acting variants in breast tissue, with 65% of coding and noncoding expressed genes displaying DAE (daeGenes). We identified over 54 K daeQTLs for 6761 (26%) daeGenes, including 385 daeGenes harbouring variants previously associated with BC risk. We found 1431 daeQTLs mapped to 93 different loci in strong linkage disequilibrium with risk-associated variants (risk-daeQTLs), suggesting a link between risk-causing variants and cis-regulation. There were 122 risk-daeQTL with stronger cis-acting potential in active regulatory regions with protein binding evidence. These variants mapped to 41 risk loci, of which 29 had no previous report of target genes and were candidates for regulating the expression levels of 65 genes. As validation, we identified and functionally characterised five candidate causal variants at the 5q14.1 risk locus targeting the ATG10 and ATP6AP1L genes, likely acting via modulation of alternative transcription and transcription factor binding. Our study demonstrates the power of DAE analysis and daeQTL mapping to identify causal regulatory variants and target genes at breast cancer risk loci, including those with complex regulatory landscapes. It additionally provides a genome-wide resource of variants associated with DAE for future functional studies.

Colocalised Genetic Associations Reveal Alternative Splicing Variants as Candidate Causal Links for Breast Cancer Risk in 10 Loci.

Genome-wide association studies (GWASs) have revealed numerous loci associated with breast cancer risk, yet the precise causal variants, their impact on molecular mechanisms, and the affected genes often remain elusive. We hypothesised that specific variants exert their influence by affecting cis-regulatory alternative splice elements. An analysis of splicing quantitative trait loci (sQTL) in healthy breast tissue from female individuals identified multiple variants linked to alterations in splicing ratios. Through colocalisation analysis, we pinpointed 43 variants within twelve genes that serve as candidate causal links between sQTL and GWAS findings. In silico splice analysis highlighted a potential mechanism for three genes-FDPS, SGCE, and MRPL11-where variants in proximity to or on the splice site modulate usage, resulting in alternative splice transcripts. Further in vitro/vivo studies are imperative to fully understand how these identified changes contribute to breast oncogenesis. Moreover, investigating their potential as biomarkers for breast cancer risk could enhance screening strategies and early detection methods for breast cancer.

Impact of NSD1 Alternative Transcripts in Actin Filament Formation and Cellular Division Pathways in Fibroblasts.

Germline variants in the NSD1 gene are responsible for Sotos syndrome, while somatic variants promote neoplastic cell transformation. Our previous studies revealed three alternative RNA isoforms of NSD1 present in fibroblast cell lines (FBs): the canonical full transcript and 2 alternative transcripts, termed AT2 (NSD1 Δ5Δ7) and AT3 (NSD1 Δ19-23 at the 5' end). The precise molecular pathways affected by each specific isoform of NSD1 are uncharacterized to date. To elucidate the role of these isoforms, their expression was suppressed by siRNA knockdown in FBs and protein expression and transcriptome data was explored. We demonstrate that one gene target of NSD1 isoform AT2 is ARP3 actin-related protein 3 homolog B (ACTR3B). We show that loss of both canonical NSD1 and AT2 isoforms impaired the ability of fibroblasts to regulate the actin cytoskeleton, and we observed that this caused selective loss of stress fibers. Our findings provide novel insights into NSD1 function by distinguishing isoform function and demonstrating an essential role of NSD1 in regulating the actin cytoskeleton and stress fiber formation in fibroblasts.

Correlations of Long Noncoding RNA HNF4A-AS1 Alternative Transcripts with Liver Diseases and Drug Metabolism.

Hepatocyte nuclear factor 4 alpha antisense 1 (HNF4A-AS1) is a long noncoding RNA (lncRNA) gene physically located next to the transcription factor HNF4A gene in the human genome. Its transcription products have been reported to inhibit the progression of hepatocellular carcinoma (HCC) and negatively regulate the expression of cytochrome P450s (CYPs), including CYP1A2, 2B6, 2C9, 2C19, 2E1, and 3A4. By altering CYP expression, lncRNA HNF4A-AS1 also contributes to the susceptibility of drug-induced liver injury. Thus, HNF4A-AS1 lncRNA is a promising target for controlling HCC and modulating drug metabolism. However, HNF4A-AS1 has four annotated alternative transcripts in the human genome browsers, and it is unclear which transcripts the small interfering RNAs or small hairpin RNAs used in the previous studies are silenced and which transcripts should be used as the target. In this study, four annotated and two newly identified transcripts were confirmed. These six transcripts showed different expression levels in different liver disease conditions, including metabolic dysfunction-associated steatotic liver disease, alcohol-associated liver disease, and obesity. The expression patterns of all HNF4A-AS1 transcripts were further investigated in liver cell growth from human embryonic stem cells to matured hepatocyte-like cells, HepaRG differentiation, and exposure to rifampicin treatment. Several HNF4A-AS1 transcripts highly displayed correlations with these situations. In addition, some of the HNF4A-AS1 transcripts also showed a strong correlation with CYP3A4 during HepaRG maturation and rifampicin exposure. Our findings provide valuable insights into the specific roles of HNF4A-AS1 transcripts, paving the way for more targeted therapeutic strategies for liver diseases and drug metabolism. SIGNIFICANCE STATEMENT: This study explores the alternative transcripts of HNF4A-AS1, showing how their expression changes in different biological conditions, from various liver diseases to the growth and differentiation of hepatocytes and drug metabolism. The generated knowledge is essential for understanding the independent roles of different transcripts from the same lncRNA in different liver diseases and drug metabolism situations.

Targeted DNA-seq and RNA-seq of Reference Samples with Short-read and Long-read Sequencing

Next-generation sequencing (NGS) has revolutionized genomic research by enabling high-throughput, cost-effective genome and transcriptome sequencing accelerating personalized medicine for complex diseases, including cancer. Whole genome/transcriptome sequencing (WGS/WTS) provides comprehensive insights, while targeted sequencing is more cost-effective and sensitive. In comparison to short-read sequencing, which still dominates the field due to high speed and cost-effectiveness, long-read sequencing can overcome alignment limitations and better discriminate similar sequences from alternative transcripts or repetitive regions. Hybrid sequencing combines the best strengths of different technologies for a more comprehensive view of genomic/transcriptomic variations. Understanding each technology’s strengths and limitations is critical for translating cutting-edge technologies into clinical applications. In this study, we sequenced DNA and RNA libraries of reference samples using various targeted DNA and RNA panels and the whole transcriptome on both short-read and long-read platforms. This study design enables a comprehensive analysis of sequencing technologies, targeting protocols, and library preparation methods. Our expanded profiling landscape establishes a reference point for assessing current sequencing technologies, facilitating informed decision-making in genomic research and precision medicine.

Single-cell long-read targeted sequencing reveals transcriptional variation in ovarian cancer

Single-cell RNA sequencing predominantly employs short-read sequencing to characterize cell types, states and dynamics; however, it is inadequate for comprehensive characterization of RNA isoforms. Long-read sequencing technologies enable single-cell RNA isoform detection but are hampered by lower throughput and unintended sequencing of artifacts. Here we develop Single-cell Targeted Isoform Long-Read Sequencing (scTaILoR-seq), a hybridization capture method which targets over a thousand genes of interest, improving the median number of on-target transcripts per cell by 29-fold. We use scTaILoR-seq to identify and quantify RNA isoforms from ovarian cancer cell lines and primary tumors, yielding 10,796 single-cell transcriptomes. Using long-read variant calling we reveal associations of expressed single nucleotide variants (SNVs) with alternative transcript structures. Phasing of SNVs across transcripts enables the measurement of allelic imbalance within distinct cell populations. Overall, scTaILoR-seq is a long-read targeted RNA sequencing method and analytical framework for exploring transcriptional variation at single-cell resolution.

Hereditary sensory autonomic neuropathy type VI in the age of genetic testing

AbstractBackgroundHereditary sensory and autonomic neuropathy type VI (HSAN VI) is a rare recessive genetic disorder caused by mutations in the human dystonin (DST) gene. We report a novel homozygous alternate transcript mutation in the DST gene causing a severe neonatal form of HSAN VI.Patient DescriptionThis baby boy was born with severe hypotonia, respiratory distress, dysmorphic features, and bilateral club feet. Imaging, karyotyping, Prader–Willi assay, spinal muscular atrophy genetic panel and myotonic dystrophy genetic panel were all negative. A comprehensive neuropathy panel detected a homozygous pathogenic variant in the DST gene—alternate transcript NM_015546.4:c.1357G>A (p.Trp4525*). Nerve conduction studies revealed mixed axonal and demyelinating sensorimotor neuropathy, suggesting the possibility of motor involvement in severe forms of this rare condition. The infant ultimately developed sepsis and died from cardiorespiratory arrest. Neuropathological findings of focal and mild spinal nerve axonal degeneration were nonspecific.ConclusionCollective analysis of these patients would help to further characterize the spectrum of disease pathology and could provide insight into the neurophysiology and neuropathology of this rare condition.

R2Dtool: integration and visualization of isoform-resolved RNA features.

Long-read RNA sequencing enables the mapping of RNA modifications, structures, and protein-interaction sites at the resolution of individual transcript isoforms. To understand the functions of these RNA features, it is critical to analyze them in the context of transcriptomic and genomic annotations, such as open reading frames and splice junctions. We have developed R2Dtool, a bioinformatics tool that integrates transcript-mapped information with transcript and genome annotations, allowing for the isoform-resolved analytics and graphical representation of RNA features in their genomic context. We illustrate R2Dtool's capability to integrate and expedite RNA feature analysis using epitranscriptomics data. R2Dtool facilitates the comprehensive analysis and interpretation of alternative transcript isoforms. R2Dtool is freely available under the MIT license at github.com/comprna/R2Dtool.

TransTEx: novel tissue-specificity scoring method for grouping human transcriptome into different expression groups.

Although human tissues carry out common molecular processes, gene expression patterns can distinguish different tissues. Traditional informatics methods, primarily at the gene level, overlook the complexity of alternative transcript variants and protein isoforms produced by most genes, changes in which are linked to disease prognosis and drug resistance. We developed TransTEx (Transcript-level Tissue Expression), a novel tissue-specificity scoring method, for grouping transcripts into four expression groups. TransTEx applies sequential cut-offs to tissue-wise transcript probability estimates, subsampling-based P-values and fold-change estimates. Application of TransTEx on GTEx mRNA-seq data divided 199166 human transcripts into different groups as 17999 tissue-specific (TSp), 7436 tissue-enhanced, 36783 widely expressed (Wide), 79191 lowly expressed (Low), and 57757 no expression (Null) transcripts. Testis has the most (13466) TSp isoforms followed by liver (890), brain (701), pituitary (435), and muscle (420). We found that the tissue specificity of alternative transcripts of a gene is predominantly influenced by alternate promoter usage. By overlapping brain-specific transcripts with the cell-type gene-markers in scBrainMap database, we found that 63% of the brain-specific transcripts were enriched in nonneuronal cell types, predominantly astrocytes followed by endothelial cells and oligodendrocytes. In addition, we found 61 brain cell-type marker genes encoding a total of 176 alternative transcripts as brain-specific and 22 alternative transcripts as testis-specific, highlighting the complex TSp and cell-type specific gene regulation and expression at isoform-level. TransTEx can be adopted to the analysis of bulk RNA-seq or scRNA-seq datasets to find tissue- and/or cell-type specific isoform-level gene markers. TransTEx database: https://bmi.cewit.stonybrook.edu/transtexdb/ and the R package is available via GitHub: https://github.com/pallavisurana1/TransTEx.

μ-Opioid receptor transcriptional variants in the murine forebrain and spinal cord

Oprm1, the gene encoding the μ-opioid receptor, has multiple reported transcripts, with a variable 3′ region and many alternative sequences encoding the C-terminus of the protein. The functional implications of this variability remain mostly unexplored, though a recurring notion is that it could be exploited by developing selective ligands with improved clinical profiles. Here, we comprehensively examined Oprm1 transcriptional variants in the murine central nervous system, using long-read RNAseq as well as spatial and single-cell transcriptomics. The results were validated with RNAscope in situ hybridization. We found a mismatch between transcripts annotated in the mouse genome (GRCm38/mm10) and the RNA-seq results. Sequencing data indicated that the primary Oprm1 transcript has a 3′ terminus located on chr10:6,860,027, which is ∼ 9.5 kilobases downstream of the longest annotated exon 4 end. Long-read sequencing confirmed that the final Oprm1 exon included a 10.2 kilobase long 3′ untranslated region, and the presence of the long variant was unambiguously confirmed using RNAscope in situ hybridization in the thalamus, striatum, cortex and spinal cord. Conversely, expression of the Oprm1 reference transcript or alternative transcripts of the Oprm1 gene was absent or close to the detection limit. Thus, the primary transcript of the Oprm1 mouse gene is a variant with a long 3′ untranslated region, which is homologous to the human OPRM1 primary transcript and encodes the same conserved C-terminal amino acid sequence.

Alternative TSS use is widespread in Cryptococcus fungi in response to environmental cues and regulated genome-wide by the transcription factor Tur1.

Alternative transcription start site (TSS) usage regulation has been identified as a major means of gene expression regulation in metazoans. However, in fungi, its impact remains elusive as its study has thus far been restricted to model yeasts. Here, we first re-analyzed TSS-seq data to define genuine TSS clusters in 2 species of pathogenic Cryptococcus. We identified 2 types of TSS clusters associated with specific DNA sequence motifs. Our analysis also revealed that alternative TSS usage regulation in response to environmental cues is widespread in Cryptococcus, altering gene expression and protein targeting. Importantly, we performed a forward genetic screen to identify a unique transcription factor (TF) named Tur1, which regulates alternative TSS (altTSS) usage genome-wide when cells switch from exponential phase to stationary phase. ChiP-Seq and DamID-Seq analyses suggest that at some loci, the role of Tur1 might be direct. Tur1 has been previously shown to be essential for virulence in C. neoformans. We demonstrated here that a tur1Δ mutant strain is more sensitive to superoxide stress and phagocytosed more efficiently by macrophages than the wild-type (WT) strain.

Swi/Snf chromatin remodeling regulates transcriptional interference and gene repression

Alternative transcription start sites can affect transcript isoform diversity and translation levels. In a recently described form of gene regulation, coordinated transcriptional and translational interference results in transcript isoform-dependent changes in protein expression. Specifically, a long undecoded transcript isoform (LUTI) is transcribed from a gene-distal promoter, interfering with expression of the gene-proximal promoter. Although transcriptional and chromatin features associated with LUTI expression have been described, the mechanism underlying LUTI-based transcriptional interference is not well understood. Using an unbiased genetic approach followed by functional genomics, we uncovered that the Swi/Snf chromatin remodeling complex is required for co-transcriptional nucleosome remodeling that leads to LUTI-based repression. We identified genes with tandem promoters that rely on Swi/Snf function for transcriptional interference during protein folding stress, including LUTI-regulated genes. This study provides clear evidence for Swi/Snf playing a direct role in gene repression via a cis transcriptional interference mechanism.

Pan-cancer transcriptome analysis reveals widespread regulation through alternative tandem transcription initiation.

Abnormal transcription initiation from alternative first exon has been reported to promote tumorigenesis. However, the prevalence and impact of gene expression regulation mediated by alternative tandem transcription initiation were mostly unknown in cancer. Here, we developed a robust computational method to analyze alternative tandem transcription start site (TSS) usage from standard RNA sequencing data. Applying this method to pan-cancer RNA sequencing datasets, we observed widespread dysregulation of tandem TSS usage in tumors, many of which were independent of changes in overall expression level or alternative first exon usage. We showed that the dynamics of tandem TSS usage was associated with epigenomic modulation. We found that significant 5' untranslated region shortening of gene TIMM13 contributed to increased protein production, and up-regulation of TIMM13 by CRISPR-mediated transcriptional activation promoted proliferation and migration of lung cancer cells. Our findings suggest that dysregulated tandem TSS usage represents an addtional layer of cancer-associated transcriptome alterations.

The Effect of Alternative Splicing Sites on Mirtron Formation and Arm Selection of Precursor microRNAs.

Mirtrons represent a subclass of microRNAs (miRNAs) that rely on the splicing machinery for their maturation. However, the molecular details of this Drosha-independent processing are still not fully understood; as an example, the Microprocessor complex cannot process the mirtronic pre-miRNA from the transcript even if splice site mutations are present. To investigate the influence of alternative splicing sites on mirtron formation, we generated Enhanced Green Fluorescent Protein (EGFP) reporters containing artificial introns to compare the processing of canonical miRNAs and mirtrons. Although mutations of both splice sites generated a complex pattern of alternative transcripts, mirtron formation was always severely affected as opposed to the normal processing of the canonical hsa-mir-33b miRNA. However, we also detected that while its formation was also hindered, the mirtron-derived hsa-mir-877-3p miRNA was less affected by certain mutations than the hsa-mir-877-5p species. By knocking down Drosha, we showed that this phenomenon is not dependent on Microprocessor activity but rather points toward the potential stability difference between the miRNAs from the different arms. Our results indicate that when the major splice sites are mutated, mirtron formation cannot be rescued by nearby alternative splice sites, and stability differences between 5p and 3p species should also be considered for functional studies of mirtrons.

SimSpliceEvol2: alternative splicing-aware simulation of biological sequence evolution and transcript phylogenies

BackgroundSimSpliceEvol is a tool for simulating the evolution of eukaryotic gene sequences that integrates exon-intron structure evolution as well as the evolution of the sets of transcripts produced from genes. It takes a guide gene tree as input and generates a gene sequence with its transcripts for each node of the tree, from the root to the leaves. However, the sets of transcripts simulated at different nodes of the guide gene tree lack evolutionary connections. Consequently, SimSpliceEvol is not suitable for evaluating methods for transcript phylogeny inference or gene phylogeny inference that rely on transcript conservation.ResultsHere, we introduce SimSpliceEvol2, which, compared to the first version, incorporates an explicit model of transcript evolution for simulating alternative transcripts along the branches of a guide gene tree, as well as the transcript phylogenies inferred. We offer a comprehensive software with a graphical user interface and an updated version of the web server, ensuring easy and user-friendly access to the tool.ConclusionSimSpliceEvol2 generates synthetic datasets that are useful for evaluating methods and tools for spliced RNA sequence analysis, such as spliced alignment methods, methods for identifying conserved transcripts, and transcript phylogeny reconstruction methods. The web server is accessible at https://simspliceevol.cobius.usherbrooke.ca, where you can also download the standalone software. Comprehensive documentation for the software is available at the same address. For developers interested in the source code, which requires the installation of all prerequisites to run, it is provided at https://github.com/UdeS-CoBIUS/SimSpliceEvol.

Deciphering the cis-regulatory landscape of natural yeast Transcript Leaders.

Protein synthesis is a vital process that is highly regulated at the initiation step of translation. Eukaryotic 5' transcript leaders (TLs) contain a variety of cis-regulatory features that influence translation and mRNA stability. However, the relative influences of these features in natural TLs are poorly characterized. To address this, we used massively parallel reporter assays (MPRAs) to quantify RNA levels, ribosome loading, and protein levels from 11,027 natural yeast TLs in vivo and systematically compared the relative impacts of their sequence features on gene expression. We found that yeast TLs influence gene expression over two orders of magnitude. While a leaky scanning model using Kozak contexts and uAUGs explained half of the variance in expression across transcript leaders, the addition of other features explained ~70% of gene expression variation. Our analyses detected key cis-acting sequence features, quantified their effects in vivo, and compared their roles to motifs reported from an in vitro study of ribosome recruitment. In addition, our work quantitated the effects of alternative transcription start site usage on gene expression in yeast. Thus, our study provides new quantitative insights into the roles of TL cis-acting sequences in regulating gene expression.

Arabidopsis DNA glycosylase MBD4L improves recovery of aged seeds.

DNA glycosylases initiate the base excision repair (BER) pathway by catalyzing the removal of damaged or mismatched bases from DNA. The Arabidopsis DNA glycosylase methyl-CpG-binding domain protein 4 like (MBD4L) is a nuclear enzyme triggering BER in response to the genotoxic agents 5-fluorouracil and 5-bromouracil. To date, the involvement of MBD4L in plant physiological processes has not been analyzed. To address this, we studied the enzyme functions in seeds. We found that imbibition induced the MBD4L gene expression by generating two alternative transcripts, MBD4L.3 and MBD4L.4. Gene activation was stronger in aged than in non-aged seeds. Seeds from mbd4l-1 mutants displayed germination failures when maintained under control or ageing conditions, while 35S:MBD4L.3/mbd4l-1 and 35S:MBD4L.4/mbd4l-1 seeds reversed these phenotypes. Seed nuclear DNA repair, assessed by comet assays, was exacerbated in an MBD4L-dependent manner at 24 h post-imbibition. Under this condition, the BER genes ARP, APE1L, and LIG1 showed higher expression in 35S:MBD4L.3/mbd4l-1 and 35S:MBD4L.4/mbd4l-1 than in mbd4l-1 seeds, suggesting that these components could coordinate with MBD4L to repair damaged DNA bases in seeds. Interestingly, the ATM, ATR, BRCA1, RAD51, and WEE1 genes associated with the DNA damage response (DDR) pathway were activated in mbd4l-1, but not in 35S:MBD4L.3/mbd4l-1 or 35S:MBD4L.4/mbd4l-1 seeds. These results indicate that MBD4L is a key enzyme of a BER cascade that operates during seed imbibition, whose deficiency would cause genomic damage detected by DDR, generating a delay or reduction in germination.