Hidden peptides in flowering control: sORF-encoded regulators in Arabidopsis thaliana and Lupinus albus

The floral transition marks the commitment of the shoot apical meristem (SAM) to reproductive development and is governed by a complex gene regulatory network integrating photoperiod, temperature and endogenous developmental signals. The contribution of non-canonical gene products, specifically peptides encoded by small open reading frames (sORFs) and their associated long non-coding RNA (lncRNA) loci, to the control of flowering has not been systematically characterised. This thesis addresses this gap through two complementary experimental approaches: a phloem proteomics strategy in Arabidopsis thaliana (hereafter Arabidopsis) and an integrated morphological, transcriptomic, translatomic and functional investigation of the floral transition in Lupinus albus, a legume crop species for which the mechanisms controlling reproductive development had not previously been characterised at this level of resolution. In the first chapter, phloem exudates were collected from Arabidopsis plants at four developmental time points spanning the vegetative-to-reproductive transition, size-fractionated and analysed by reversed-phase liquid chromatography-electrospray ionisation mass spectrometry (LC-ESI-MS). A total of 725 unique proteins were identified, of which approximately 52 % matched previously reported phloem proteomic datasets, validating sample quality. Orthogonal Partial Least Squares (OPLS) analysis demonstrated stage-dependent clustering of samples, with 7-day long-day (LD) samples forming the most distinct group, coinciding with confirmed floral meristem (FM) identity. Re-interrogation of the mass spectrometry data against a curated sORF database identified 77 peptides with differential abundance across developmental stages. Among 12 candidates selected for detailed study, two (ATSO2_000592 and sORF1511) were of particular interest because their coding sequences (CDSs) are embedded within lncRNA loci on chromosome 1 (AT1G09053 and AT1G67365, respectively) and both showed progressive accumulation across the time course. Constitutive overexpression of sORF1511 in Arabidopsis produced a consistent early-bolting phenotype in three independent transgenic lines relative to wild-type (WT) Columbia-0 (Col-0), without a reduction in rosette leaf number. Overexpression of ATSO2_000592 had no detectable phenotypic effect. Critically, sORF1511 overexpression did not alter transcript levels of the host lncRNA AT1G67365, supporting the conclusion that the observed phenotype is attributable to the peptide rather than to altered lncRNA activity. Expression profiling showed that APETALA1 (AP1) transcripts were detectable 14 days after sowing (das) in overexpressor plants versus 20 das in WT, without changes in SUPRESSOR OF OVEREXPRESSION OF CONSTANS 1 (SOC1) or TERMINAL FLOWER 1 (TFL1) levels. Genome-wide RNA-Seq further revealed a 14-fold-enriched abscisic acid (ABA)-responsive transcriptional signature and strong upregulation of FLOWERING PROMOTING FACTOR 1 (FPF1), a gibberellin (GA)-sensitivity regulator, consistent with a possible transcriptional reprogramming linking sORF1511 action to GA-mediated stem elongation and stress-related signalling pathways. In the second chapter, scanning electron microscopy (SEM) of the L. albus SAM at 12, 15 and 17 das defined a morphological and molecular framework for the floral transition, validated by the progressive transcriptional induction of three AP1-like genes as FM identity markers. An improved genome annotation (LualbAMIv1BRv5) was generated by integrating ab initio gene prediction with RNA-Seq and proteomic evidence, achieving 98.1 % BUSCO completeness, 96.5 % OMArk consistency and a PSAURON score of 97.3; without accurate 5′ untranslated region (UTR) annotation, upstream Open Reading Frames (uORFs) within these regions would remain undetectable. Parallel RNA-Seq and ribosome profiling (Ribo-Seq) on shoot apices at 12, 15 and 17 das revealed that the transcriptional upregulation of the Arabidopsis homologues genes AP1 (LaAP1), LEAFY (LaLFY) and FRUITFULL (LaFUL) show coordinated increases in both transcript abundance and ribosome occupancy and a substantial number of genes display changes in translational efficiency independently of transcript abundance. Application of the RiboTaper pipeline identified 1,788 uORFs, 80 downstream ORFs (dORFs), 416 non-coding ORFs (ncORFs) and RibORF 1,573 uORFs and 172 dORFs. By de novo transcriptome assembly 1,071 intergenic sORFs have been identified. Mass spectrometry on shoot apices confirmed the in vivo production of hundreds of sORF-encoded peptides at each developmental stage, establishing that these elements are genuine translatome components rather than computational artefacts. Five uORFs located within the 5′ UTRs of flowering-related transcription factors (designated with the prefix La- to denote L. albus origin, with gene names reflecting the Arabidopsis homologue of highest sequence similarity without implying functional equivalence), ARABIDOPSIS THALIANA HOMEOBOX 16 (LaATHB-16), SHORT VEGETATIVE PHASE (LaSVP), LATE ELONGATED HYPOCOTYL (LaLHY), ASYMMETRIC LEAVES 1 (LaAS1) and SQUAMOSA PROMOTER BINDING PROTEIN-LIKE 1/12 (LaSPL) were selected for detailed characterisation. Hereafter, the uORF associated with each gene is referred to as uORFLaATHB-16, uORFLaSVP, uORFLaLHY, uORFLaAS1 and uORFLaSPL, respectively. Transcription start site sequencing (TSS-Seq) confirmed co-transcription of each uORF with its downstream main ORF (mORF) at all three stages. Ribosome occupancy profiling showed stage-dependent dynamics: LaATHB-16, LaAS1 and LaSPL showed higher uORF occupancy in the reproductive phase; LaSVP showed stronger uORF translation in the vegetative stage; and LaLHY shifted from dual uORF-mORF occupancy at 12 das to predominant uORF occupancy at 17 das. The LaLHY-associated uORF displays the hallmarks of a conserved peptide uORF (CPuORF), with extended stretches of conserved residues across legumes and into Brassicaceae. Transient dual-luciferase reporter assays in L. albus demonstrated that uORFLaATHB-16 enhances downstream mORF translation with no change in reporter mRNA levels, confirming a purely translational effect, whereas uORFLaSPL represses mORF translation and simultaneously reduces reporter transcript abundance. Exogenous application of synthetic peptides encoded by uORFLaSVP and uORFLaAS1 produced statistically significant effects on specific growth parameters in L. albus seedlings at 10-50 nM, indicating biological activity outside the translational regulatory context. In Arabidopsis, sORF1511 demonstrates that lncRNA-embedded coding elements can function as developmental regulators with systemic signalling potential. In L. albus, genome reannotation combined with Ribo-Seq and proteomic validation uncovered thousands of actively translated sORFs, including uORFs that exert mechanistically diverse effects on the translation of key flowering regulators. These findings demonstrate that translational regulation by sORF-encoded peptides constitutes a regulatory layer that is invisible to transcriptomics alone and that operates in both model and crop species.

The research presented in this thesis was supported by grants PID2020-113035RB-I00 and PID2023-147023OB-I00 from MICIU/AEI/10.13039/501100011033 and by the European Regional Development Fund (ERDF/EU) awarded to Reyes Benlloch Ortiz. The PhD candidate, Irene Moreno Sanguino, was supported by a predoctoral contract (ref. PRE2021-097035) funded by MICIU/AEI/10.13039/501100011033 and co-funded by the European Social Fund Plus (ESF+)