Timo Mühlhaus

Mitochondria are endosymbiont-derived organelles that play a central role in cellular metabolism, energy production and stress responses. While single mitochondria represent functional units, they continuously exchange their contents through fusion and fission, facing stress conditions as a dynamic population. To date, it remains largely unknown how stress alters mitochondrial dynamics in plants and how altered dynamics affect mitochondrial properties and plant stress resilience. Here, we investigate mitochondrial dynamics in response to oxidative stress in the non-vascular model plant Physcomitrium patens . By creating mutants with impaired mitochondrial fission in different reporter lines for mitochondrial parameters, we additionally analyse effects of chronic changes to mitochondrial population dynamics. We found that Mito-Paraquat (MtPQ) treatment increased the glutathione redox potential E _GSH in mitochondria, the cytosol and chloroplasts, as monitored via roGFP2-based genetically encoded biosensors. Mitochondria elongated within hours and showed a concomitant and heterogenous increase of matrix EOS _red , that we propose as a marker for matrix protein damage. Mitochondrial fission mutants lacking PpELM1B (ELONGATED MITOCHONDRIA) displayed distinct changes of mitochondrial morphology parameters as determined by automated 3D-segmentation and feature mapping (MorphoMapper) of confocal z-stacks. Elongated mitochondria in Ppelm1b ^ge lines showed an oxidative matrix E _GSH shift and increased matrix EOS _red while matrix mixing still occurred, albeit at the same slow rate as in wildtype, within days. Macroscopically, Ppelm1b ^ge lines displayed reduced growth, decreased respiration, and a higher sensitivity to oxidative stress. Our results show that plant mitochondrial morphology and physiological parameters specifically shift in response to stress and impaired fission. Mitochondrial fission is vital to maintain a healthy mitochondrial population that sustains plant oxidative stress tolerance.

Periphytic biofilms are key mediators of freshwater ecosystem processes but are increasingly exposed to anthropogenic stressors. We compared biofilms sampled upstream and downstream of a wastewater treatment plant (WWTP) effluent, two source communities presumed to differ in prior exposure to wastewater‐associated press disturbance, to examine differences in tolerance to ciprofloxacin (CIP) exposure. We assessed biomass, photosynthetic pigments and fatty acid (FA) profiles as well as microbial community structure using qPCR and metabarcoding. Downstream‐sourced biofilms exhibited higher biomass stability, reduced pigment and FA losses and fewer bacterial and eukaryotic community shifts under CIP exposure than upstream‐sourced biofilms. Upstream‐sourced communities, despite higher baseline pigment and FA contents, showed pronounced sensitivity to CIP, including biomass declines and reduced prokaryotic diversity. Taxon‐specific responses revealed stress‐tolerant Chlorophyceae , cyanobacteria and heterotrophic taxa, which may have contributed to higher functional stability under the tested conditions. Our findings show that biofilm responses to CIP varied with source site and community composition. While not a direct test of environmental filtering, the observed patterns are consistent with its predictions. The results further suggest that legacy effects and spatial heterogeneity may influence biofilm vulnerability to antibiotic exposure, although this requires validation across multiple WWTP‐impacted and reference sites.

Artificial Intelligence (AI) has become a central analytical approach in plant science, supporting prediction, pattern discovery, and integration across molecular, phenotypic, and environmental data. While methodological advances have been rapid, progress is increasingly constrained less by algorithms than by the structure, semantics, and interoperability of the underlying data. In this review, we examine how the Findable, Accessible, Interoperable, and Reusable (FAIR) principles are currently realized in AI-driven plant research, and how different implementations shape which AI analyses are feasible in practice. We introduce a readiness level perspective that treats FAIR principles not as a binary property, but as a spectrum of operational implementations ranging from unstructured data to fully machine-actionable FAIR Digital Objects. Applying this framework to molecular, phenotypic, and integrative use cases, we show that FAIR-readiness is strongly method- and context-dependent. Task-focused AI applications can succeed at lower readiness levels, whereas integrative, transferable, and reproducible analyses consistently require richer metadata, explicit semantics, and traceable provenance. We also note that many advances depend on shared infrastructure and community-maintained practices that allow curation and reuse to accumulate across projects. We argue for a shift from FAIR compliance towards FAIR capability, to enable scalable and sustainable AI-driven discovery in plant science.

The amyloid precursor protein (APP), a key factor in Alzheimer's disease (AD) pathology, and its two mammalian homologs, amyloid precursor-like proteins 1 and 2 (APLP1 and APLP2), are considered as members of the synaptic adhesion molecule (SAM) family. They are localized to the pre- and postsynapse, form trans-cellular dimers, and have been shown to induce presynaptic differentiation in a heterologous synapse formation assay. We demonstrate that expression of all APP family members in non-neuronal cells also promotes dendritic excitatory postsynaptic differentiation in primary mouse neurons of either sex, similar to Neurexin1β and other SAMs. Synaptogenic activity was decreased by deletion of the E1 domain and increased upon inhibition of soluble APP (sAPP) generation, reinforcing that trans-cellular interaction of APP/APLPs can induce synaptogenesis. Consistent with this, the capacity of heterologously expressed APP to induce postsynaptic specializations in contacting dendrites was reduced by the absence of APP family members at the postsynaptic site and was lost in conditional triple knock-out (cTKO) neurons. Pharmacological analyses revealed that heterologous formation of pre- and postsynapses relies on proper microtubule and actin cytoskeleton dynamics, as well as the MAP kinase pathway, similar to what has been shown for Neurexin1β and Neuroligin1. However, inhibition of the PI3K/Akt pathway selectively impaired APP-induced postsynaptic differentiation, suggesting that distinct APP signaling pathways are required for pre- and postsynaptic differentiation. Collectively, our data highlight the role of all APP family members as SAMs in trans-synaptic signaling, providing key insights into their physiological function and advancing our understanding of AD-related synaptopathies.

Light intensities in nature fluctuate; however, our understanding of how plants acclimate to changing photon fluxes remains limited. Various chloroplast-located proteins contribute to the ability to tolerate fluctuating light (FL). Here, we show that an Arabidopsis (Arabidopsis thaliana) protein outside chloroplasts is involved in acclimation to FL. Early-Response to Dehydration 6-Like 14 (ERDL14) is a tonoplast-located sugar/H+ antiporter able to pump glucose and sucrose into the vacuole. The ERDL14 promoter contains several elements that enable its activation under high light (HL) and FL conditions, in which standard light intensities (3 mins at 120 µE) are periodically interrupted by short high-light phases (1 min at 1.000 µE). erdl14 loss-of-function mutants do not show phenotypic peculiarities under HL, which coincides with a marked down-regulation of ERDL14 gene expression after 2 days at HL. In contrast, FL treatment impairs erdl14 mutant development, most likely caused by down-regulation of the abundance of proteins involved in primary photosynthetic processes, leading to reduced photosynthetic quantum yield. This decreased abundance of proteins involved in primary photosynthetic processes is connected to increased cytosolic sugar levels in erdl14, indicating a “high-sugar repression”, which is prevented in the wild type by the pumping activity of ERDL14. Knock-out mutants lacking other vacuolar sugar importers, i.e. tst1 and tst2, are not as strongly impaired in their ability to cope with FL, indicating a specific function of ERDL14.

The DataPLANT consortium is driving a cultural shift toward reproducible Open Science in the plant research community by embedding Findable, Accessible, Interoperable, and Reusable (FAIR) principles into the research lifecycle from the outset. Central to this effort is the development of a science gateway that supports collaborative research, flexible workflow integration, and structured data publication. Built on a robust code and project management platform, this gateway offers tools for versioning, provenance tracking, and documentation, enabling transparent, iterative, and open research practices. At the heart of this infrastructure are DataHUBs, which serve as the technical complement of the Annotated Research Context (ARC)—a concept designed to organize, structure, and facilitate the management of collaborative research data. Aligned with the emerging paradigm of FAIR Digital Object (FDO), DataPLANT redefines data publication: no longer a static deposit, but the creation of rich, machine-actionable digital objects that integrate data, metadata and persistent identifiers. These FDOs are citable, immutable in their published state, yet support versioning and continued evolution. The DataPLANT infrastructure actively supports the creation, management, and dissemination of FDOs, ensuring that FAIRness is maintained throughout the entire data lifecycle. Treating FDOs as the ultimate result of a curated research process, rather than a post hoc archiving step, establishes a sustainable foundation for open and FAIR science in plant research and other fields.

Peroxiredoxins are thiol peroxidases, which detoxify peroxides, relay redox signals and act as chaperones. In eukaryotes, multiple peroxiredoxin-1 (Prx1)/AhpC-type isoforms frequently co-exist in the same subcellular compartment, yet have been assumed to assemble only as homo-oligomeric complexes. Here we show that hetero-oligomerization is a conserved and functionally relevant property of Prx1/AhpC-type peroxiredoxins. Using biochemical reconstitution, native mass photometry, electron microscopy and live-cell assays, we demonstrate formation of heterodimers and heterodecamers, with diverse subunit stoichiometries, in peroxiredoxin pairs from different eukaryotic kingdoms. In Saccharomyces cerevisiae , oxidative challenge induces Tsa1–Tsa2 heterodecamerization with substoichiometric Tsa2 incorporation sufficing to stabilize the decameric state. Functional hetero-oligomers are also observed forming among human, plant and Leishmania peroxiredoxins. Our findings provide new insights into peroxiredoxin structural plasticity with broad implications for redox biology, stress responses and cellular adaptation, and also challenge the long-held paradigm of peroxiredoxin homo-oligomerization.

Traditional search applications within Research Data Management (RDM) ecosystems are crucial in helping users discover and explore the structured metadata from the research datasets. Typically, text search engines require users to submit keyword-based queries rather than using natural language. However, using Large Language Models (LLMs) trained on domain-specific content for specialized natural language processing (NLP) tasks is becoming increasingly common. We present ArcBERT, an LLM-based system designed for integrated metadata exploration. ArcBERT understands natural language queries and relies on semantic matching, unlike traditional search applications. Notably, ArcBERT also understands the structure and hierarchies within the metadata, enabling it to handle diverse user querying patterns effectively.

Stress responses in biological systems arise from complex, dynamic interactions among genes, proteins, and metabolites. A thorough understanding of these responses requires examining not only changes in individual molecular components but also their organization into interconnected pathways and networks that collectively maintain cellular homeostasis. This review provides an overview of computational strategies designed to capture these multifaceted processes. First, we discuss the importance of data analysis in uncovering how stress adaptation unfolds, highlighting both classical approaches (e.g., ANOVA, t-tests) and more advanced methods (e.g., clustering, smoothing splines) that handle strong temporal dependencies. We then explore how enrichment analyses can contextualize these dynamic changes by linking regulated molecules to broader biological functions and processes. The latter half of the review focuses on network-based modeling techniques, emphasizing the construction and refinement of de novo networks to identify stress-specific regulatory networks. Pairwise approaches are discussed alongside advanced methods that include multi-omics data, literature knowledge, and machine learning. Finally, we address comparative network analyses, which facilitate cross-condition studies, revealing both conserved and distinct features that shape resilience. With continued advances in high-throughput experimentation and computational modeling, these methods will deepen our insights into how cells detect and counteract stress.

PSII assembly requires auxiliary factors, including Psb28. Although the absence of Psb28 in cyanobacteria has little effect on PSII assembly, we show here that the Chlamydomonas (Chlamydomonas reinhardtii) psb28 null mutant is severely impaired in PSII assembly, showing drastically reduced PSII supercomplexes, dimers, and monomers, while overaccumulating early PSII assembly intermediates reaction center II (RCII), CP43mod, and D1mod. The mutant had less PSI and more cytochrome b6f complex, its thylakoids were organized mainly as monolayers, and it had a distorted chloroplast morphology. Complexome profiling of the psb28 mutant revealed that THYLAKOID ENRICHED FRACTION 5 (TEF5), the homolog of Arabidopsis (Arabidopsis thaliana) PHOTOSYSTEM B PROTEIN 33/LIGHT HARVESTING-LIKE 8, comigrated particularly with RCII. TEF5 also interacted with PSI. A Chlamydomonas tef5 null mutant was severely impaired in PSII assembly and overaccumulated RCII and CP43mod. RC47 was not detectable in the light-grown tef5 mutant. Our data suggest a possible role for TEF5 in RCII photoprotection or maturation. Both the psb28 and tef5 mutants exhibited decreased synthesis of CP47 and PsbH, suggesting negative feedback regulation possibly exerted by the accumulation of RCII and/or CP43mod in both mutants. The strong effects of missing auxiliary factors on PSII assembly in Chlamydomonas suggest a more effective protein quality control system in this alga than in land plants and cyanobacteria.