Environmental gradients reveal stress hubs pre-dating plant terrestrialization

Land plant terrestrialization transformed Earth’s biosphere, establishing embryophytes as dominant contributors to terrestrial biomass and atmospheric oxygen. Phylogenomic work has clarified that land plants evolved from streptophyte algae, with unicellular Zygnematophyceae emerging as the closest algal relatives. With genomes now available for several zygnematophytes (including Mesotaenium endlicherianum), it is possible to redefine the shared molecular chassis underpinning environmental responses and to identify genes that facilitated terrestrialization. Land plants deploy multilayered sensing, signaling and response systems (for example, ABA, LEA proteins, phenylpropanoids) and many stress-relevant routes have homologues in streptophyte algae, though their functionality and deployment under stress remain unclear (e.g., ABA receptor homologues acting in ABA-independent ways). This study probes how a close algal relative responds to key terrestrial stressors by systematically varying irradiance and temperature, linking physiological performance with genome-wide expression to uncover conserved hubs governing stress responses.

The paper situates Mesotaenium within Streptophyta, referencing phylogenomic studies that identify unicellular Zygnematophyceae as the closest relatives of land plants. It reviews conserved stress-response components in land plants (ABA signaling, LEA proteins, specialized metabolites) and notes homologues in streptophyte algae, though some pathways (e.g., PYL ABA receptor) may function differently (ABA-independent). Prior comparative transcriptomics across streptophyte algae have highlighted conserved stress signaling circuits, including roles for chloroplast/photosynthesis-related responses and calcium/kinase signaling in Zygnematophyceae. These studies motivate testing whether algal homologues are engaged under environmental stress combinations relevant to terrestrialization.

• Organism and environmental gradients: Mesotaenium endlicherianum SAG 12.97 was cultivated and distributed into 504 wells (42 twelve-well plates). A custom gradient table provided a two-dimensional gradient: temperature 8.6 ± 0.5 °C to 29.2 ± 0.5 °C (x-axis) and irradiance 21.0 ± 2.0 to 527.9 ± 14.0 µmol photons m⁻² s⁻¹ (y-axis). Cultures were exposed for 65 h (RNA-seq and physiology) and 89 h (microscopy). Pre-experiments established viable and stress ranges.
• Physiological measurements: Maximum quantum yield of PSII (Fv/Fm) measured by IMAGING-PAM; in vivo absorbance at 480, 680, and 750 nm via microplate reader. Triplicate runs provided 504 Fv/Fm and 4,536 absorbance measurements per replicate. Statistical analyses used Kruskal–Wallis with Fisher’s LSD and Bonferroni correction; PCA, clustering, and regression assessed effects of light and temperature.
• RNA extraction and sequencing: For each of the 42 conditions, 0.4 ml from each of 12 wells was pooled, pelleted, and RNA extracted (Spectrum Plant Total RNA Kit), DNase-treated, and libraries prepared (strand-specific, polyA selection). Sequencing on Illumina NovaSeq 6000 generated 128 transcriptomes totaling 9,892,511,114 reads (~1.5 Tbp).
• Read processing and quantification: QC by FastQC/MultiQC; trimming with Trimmomatic. A new genome annotation (V2) was generated using REAT integrating RNA-seq assemblies (HISAT2, Scallop, StringTie, Portcullis, Mikado), homology (SPALN), and ab initio predictors (Augustus, SNAP, Glimmer, CodingQuarry) consolidated with EvidenceModeler and Minos; completeness assessed by BUSCO and AED. Expression quantified via Kallisto pseudoalignment; counts imported with tximport; filtering and TMM normalization with edgeR; mean–variance modeling via limma-voom.
• Exploratory and differential expression analyses: PCA, distance and Spearman correlation heat maps. The gradient table was partitioned into nine sectors plus an Fv/Fm < 0.5 stress cohort. limma with duplicateCorrelation performed 37 contrasts (absolute FC ≥ 2; BH-adjusted P ≤ 0.01). GO enrichment via clusterProfiler using expressed genes as background.
• Co-expression network analysis: WGCNA on limma-voom expression (signed network, bicor, soft-threshold β=16; module merge cut height 0.25; min module size 30) produced 26 modules. Module–trait correlations computed for light, temperature, absorption, and Fv/Fm; hubs defined as top 20 intramodularly connected genes per module. GO enrichment per module.
• Comparative analyses: 212 public RNA-seq datasets from Zygnema circumcarinatum, Marchantia polymorpha, Physcomitrium patens, and Arabidopsis thaliana processed with the same WGCNA pipeline. Orthofinder used to identify orthogroups; module similarity assessed by Jaccard indices; hub connectivity compared across species; maximum-likelihood phylogenies inferred for 160 Mesotaenium hubs to assess ancestry.
• Lipid droplet (LD) biology: Microscopy with BODIPY 493/503 staining quantified LDs under gradient conditions (Mann–Whitney U tests). Subcellular fractionation isolated lipid-rich phases; LC–MS proteomics (MaxQuant/Perseus) identified proteins enriched in LD fraction. LD lipid composition profiled by TLC and GC. Transient expression of mCherry-tagged Mesotaenium LD proteins (OLE, CLO, HSD/steroleosin, LDAP) in Nicotiana tabacum pollen tubes confirmed LD localization.
• Transposable elements: InterProScan identified transposon-related domains; expression response of TE-related genes assessed across conditions.

• Physiological response across gradients:
  • Fv/Fm and absorbance decreased with increasing irradiance; low temperature exerted a stronger negative effect on physiology than light. At 20.5 ± 1.0 °C, Fv/Fm dropped from 0.66 ± 0.02 (21 µmol photons m⁻² s⁻¹) to 0.042 ± 0.04 (≈535 µmol photons m⁻² s⁻¹). Lowest Fv/Fm (down to zero) occurred at highest light and lowest temperature.
  • Physiology was better explained by combined effects of temperature and light (bifactorial model adjusted R² ≈ 0.776) than by single factors (light R² ≈ 0.095; temperature R² ≈ 0.652).
  • Eurythermy appears foundational for euryphoty: high light tolerated at moderate temperatures (20.5–25.3 °C) but not at thermal extremes.
• Transcriptome landscape:
  • 128 transcriptomes (≈9.9 billion reads, ≈1.5 Tbp) revealed clear separation by high temperature and light (PCA: PC1 35%, PC2 18.1%). Clusters: (1) high light and/or high temperature, (2) low temperature (8–17 °C), (3) moderate conditions.
  • Comparing low light + low temperature (LLI_LT) vs high light + high temperature (HLI_HT) yielded 6,578 DEGs (|FC| ≥ 2, adj. P ≤ 0.01), the largest contrast. Across 37 comparisons, 8,157 genes were significantly regulated; 30 genes were consistently regulated across all comparisons, enriched for light-harvesting and ROS-related genes (e.g., ELIP) and fatty acid metabolism.
  • Cross-species DEG orthogroup sharing with Mesotaenium: 3,107–6,458 shared HOGs, with 46.6–73.0% shared within Zygnematophyceae and 15.8–30.4% outside; 4.6–59.8% of shared HOGs exhibited shared regulation depending on treatment. Most conserved responses involved chloroplasts/photosynthesis; within Zygnematophyceae, kinase and calcium-dependent signaling were prominent.
• Co-expression modules and conserved hubs:
  • WGCNA identified 26 modules among 17,905 expressed genes. Notable module–trait correlations:
    • Green: positively correlated with light (r = 0.88, P = 6×10⁻⁴³) and negatively with Fv/Fm (r = −0.79, P = 6×10⁻²⁹); enriched for ROS biology.
    • Purple: negatively correlated with light (r = −0.94, P = 3×10⁻⁶⁰), positively with absorption and Fv/Fm (r = 0.71 and 0.79), enriched for cell division hubs (cyclins, kinesins, Tesmin, TSO1).
    • Brown: negatively correlated with temperature (r = −0.95, P = 7×10⁻⁶⁵), enriched for plastid translation/transcription; 12 of top 20 hubs ribosome-related.
    • Light cyan: positively with light (r = 0.93, P = 1×10⁻⁵⁶) and negatively with Fv/Fm (r = −0.67, P = 5×10⁻¹⁸), featuring thioredoxins, light-induced proteins, pigment/apocarotenoid metabolism.
    • Blue: negatively with light (r ≈ −0.76) and positively with Fv/Fm (r ≈ 0.67); enriched for plastid signaling and photomorphogenesis (GLK1, COP1 targets, COL3, BROTHER OF LUX ARRHYTHMO), ethylene-related TFs (EIN3-like, ERFs), ABA-related factors (ABF2).
    • Yellow: positively with light (r = 0.62, P = 10⁻¹⁴), negatively with absorption and Fv/Fm (r = −0.79 and −0.81), enriched for chloroplast proteostasis (Clp, FtsH) and plastid–nucleus coordination (pTAC6, GUN2) and HY5.
  • Comparative WGCNA across Zygnema, Marchantia, Physcomitrium, and Arabidopsis revealed conserved module similarities (blue, brown, turquoise, yellow) by Jaccard indices and conserved hub connectivity.
  • Phylogenies for 160 Mesotaenium hubs (135 retrieved): 107 hub families present in or before the last common ancestor of Zygnematophyceae and land plants, indicating pre-terrestrial origins (~600 Mya).
• Ancient signaling and cell wall integrity:
  • MAPKs and phototropins appeared as hubs (blue module); brassinosteroid and cell wall integrity signaling components enriched (pink module) including EXORDIUM-like and COBRA family proteins, suggesting conserved cell wall perturbance signaling loops.
• Lipid droplets (LDs) as stress response:
  • LD-like inclusions formed under stress; BODIPY staining confirmed LDs. LD counts per cell differed significantly across conditions.
  • LD-associated genes were induced under high temperature and/or high light (e.g., HSD1/steroleosin, OLE7/oleosin, LDAP, PUX10). A CGI-58 homologue was a top hub (green module).
  • Proteomics of lipid-rich fractions identified 739 proteins (LD fraction) and 1,574 (total extract); 14 proteins significantly enriched in LD fraction, including hallmark LD proteins (OLE, CLO, HSD/steroleosin, LDAP). Transient expression confirmed LD localization.
  • LDs were rich in TAGs; lipid profiles varied with culture age.
• Transposable elements:
  • Of 1,748 TE-related genes (6,186 domain entries), 96 passed expression thresholds; high temperature (29 °C) most strongly affected TE mobilization.
• Data resources: All RNA-seq (PRJNA832564), proteomics (PRIDE PXD037847), images and phylogenies (Zenodo) made publicly available.

By integrating continuous gradients of temperature and light with physiological and transcriptomic profiling, the study reveals that Mesotaenium orchestrates stress responses through conserved genetic programmes centred on plastid function, ROS management, photomorphogenesis, and cell wall integrity—paralleling land plants. The strong dependence of physiology (Fv/Fm) on the combined effects of temperature and irradiance underscores the ecological context of early streptophyte evolution, where eurythermy likely enabled tolerance to fluctuating light (euryphoty). Co-expression modules and their hubs (GLK1, HY5, ABF2, COP1 network, Clp/FtsH proteases, EXL/COBRA, MAPKs) map onto ancient, conserved circuits that integrate chloroplast status with nuclear gene expression and environmental signaling (including ethylene and ABA-associated components), supporting a shared chassis predating terrestrialization. The cross-species analyses (shared orthogroups, conserved module similarities, and hub connectivity) demonstrate that stress regulatory networks were already established in algal ancestors, with treatment- rather than phylogeny-driven similarities for some responses. The formation of lipid droplets with hallmark proteins and TAG-rich content under stress, together with LD gene induction and conserved hub status of CGI-58, indicates that LD-mediated stress buffering is an ancient response predating land plants. Collectively, these findings address the central question by pinpointing recurrent hubs and modules—plastid/cell wall-derived signals and LD formation—as stress response centres that arose before plant terrestrialization.

This work establishes a high-resolution, combinatorial environmental gradient framework to uncover conserved stress-response hubs in Mesotaenium, a close algal relative of land plants. The study identifies co-expression modules tied to plastid operation, photomorphogenesis, ROS homeostasis, cell division, and cell wall integrity, with many hubs demonstrably predating terrestrialization. It further shows that lipid droplet formation with canonical protein markers is a conserved stress response in streptophyte algae. These results redefine core environmental response circuits shared across 600 million years of streptophyte evolution, illuminating the molecular toolkit that facilitated land plant emergence. Future work should experimentally dissect retrograde signaling components and hormone-associated factors in streptophyte algae, extend gradient-based multi-stressor analyses to additional species and environmental axes (e.g., desiccation, osmotic stress), and functionally validate conserved hubs through perturbation genetics and reverse engineering of module interactions.

• The study focuses on a single zygnematophyte species (Mesotaenium endlicherianum); responses may vary across streptophyte diversity.
• Environmental variables were limited to irradiance and temperature under laboratory conditions; other terrestrial stressors (e.g., desiccation, osmotic, UV, nutrient limitations) were not concurrently tested.
• Pooling wells per condition for RNA-seq captures average responses and may mask cell-to-cell variability.
• Genome re-annotation and orthology inferences, while comprehensive, may still miss lineage-specific or lowly expressed genes, affecting module membership and hub assignment.
• Correlative co-expression and module–trait associations do not alone establish causality; functional validation is needed.