Declining metal availability in the Mesozoic seawater reflected in phytoplankton succession

The study addresses how seawater trace metal availability and oxygenation changes across Earth history influenced the evolution and ecological turnover of marine phytoplankton. While Precambrian trace metal chemistry is often inferred from indirect geochemical proxies and was tightly coupled to redox state, the Phanerozoic drivers are less clear. A pronounced phytoplankton succession occurred near the Palaeozoic/Mesozoic boundary (~252 Ma), with the open ocean shifting from green-algal (Archaeplastida) dominance to eukaryotes with red-algal-derived plastids (secondary endosymbionts, including haptophytes and heterokonts). The authors hypothesize that differences in metalloproteomes and metal transport strategies among major phytoplankton lineages reflect adaptation to changing trace metal availability, and that these biological patterns can serve as indicators of ocean metal chemistry changes, particularly a decline in open-ocean trace metals during the Mesozoic.

Prior work links trace element availability to ocean redox and biological evolution, with indirect records from iron formations, black shales, pyrite, and carbonates used largely for Precambrian reconstructions. Molecular clocks suggest primary endosymbiosis and the spread of red plastids predated ecological expansion by 0.5–1 Gyr, implying environmental selection later drove radiations. Studies show lineage-specific elemental stoichiometry and physiological metal demands, as well as the importance of metal transport systems (e.g., ABC transporters, P-type ATPases) in managing limitation and toxicity. Earlier analyses indicate green versus red lineage differences in elemental composition and speculate sulfate availability and oxygenation could have facilitated expansion of chlorophyll a+c phytoplankton. Geological evidence suggests changing oxygen levels across the Phanerozoic and potential variations in Zn and Cu through time, though Zn reconstructions remain debated. Together, these works motivate a comparative genomic approach to evaluate lineage-specific metalloproteomes and transporters as proxies for ancient seawater metal availability.

The authors used two complementary approaches. 1) Comparative genomics: They compiled genome-predicted proteomes from 26 phytoplankton species spanning cyanobacteria, primary endosymbionts (green lineage: Chlorophyceae, Trebouxiophyceae, Klebsormidiophyceae, Prasinophyceae; red lineage primary endosymbionts: Bangiophyceae, Florideophyceae, Cyanidiophyceae) and secondary endosymbionts (Cryptophyta, Haptophyta, Stramenopiles). Orthogroups were inferred using OrthoFinder with DIAMOND all-versus-all alignments, applying length-normalized bit-score transformations, MCL clustering (inflation 1.5), and species tree inference via STAG. Metal-binding proteins were identified from curated annotations and literature for metal-binding functions and transporters. Percentages of metalloproteins (by metal) relative to proteome size were calculated to reflect physiological metal requirements. Metal transporter families were enumerated, and their proportions relative to proteome size were computed. Principal component analysis of transporter family counts assessed lineage clustering. Amino acid composition (thiol/sulfur- and histidine-rich residues) of transporter families was compared across lineages to infer relative binding affinity/selectivity differences. Subcellular targeting of orthogroups (plastid, mitochondrion) was predicted to partition metalloproteins and transporters by compartment. Relationships were tested between metalloproteome fractions and published half-saturation coefficients (Ks or Km) for Fe, Mn, Zn across species. 2) Protein domain analysis: Complete proteomes (including additional species: Phaeodactylum tricornutum, Polarella glacialis, Symbiodinium microadriaticum, Symbiodinium pilosum) and 608 transcriptomes from MMETSP were scanned via DIAMOND against SCOP domains (E ≤ 1e-10). Identified domains were cross-referenced to MetalPDB to count metal-containing domains. Relative metal domain abundance was normalized by total domains and compared among groups. Additional data synthesis: Geological proxies of Zn and Cu from pyrite and shale, nutrient and O2 model outputs (COPSE reloaded, GEOCARBSULFOR), and modern phytoplankton abundance distributions from OBIS were incorporated to contextualize evolutionary and environmental patterns.

• Distinct metalloproteomes by lineage: Secondary endosymbiotic lineages have significantly lower percentages of Fe-, Zn-, and Cu-binding proteins than primary endosymbionts (P < 0.05), indicating reduced requirements for these trace metals. Results are consistent across complete proteomes and domain analyses (including MMETSP). For Mo, eukaryotes show higher abundance than prokaryotes.
• Physiological validation: Across species, the percentages of Fe-, Mn-, and Zn-binding proteins positively correlate with their respective half-saturation coefficients (Km), supporting encoded metalloproteome fractions as indicators of metal requirements.
• Transport system investment and composition: Cyanobacteria allocate ~2.8% of their proteomes to metal transporters versus 0.5–1% in eukaryotes (P < 0.05). ABC transporters are the most abundant transporter family across groups and comprise >90% of metal transport families in cyanobacteria. Red-lineage eukaryotes have a higher proportion of ABC transporters among metal transporters (~60%) than green lineages (~50%; P < 0.05). Green lineages possess significantly higher proportions of P-type ATPases dedicated to efflux (P < 0.001), indicating more specialized detoxification systems.
• Transporter diversity and strategy: Cyanobacteria exhibit streamlined, generalist uptake (e.g., multi-metal ABC families) and limited detoxification capacity (e.g., fewer P-type ATPases; some lack phytochelatin synthase), consistent with broad adaptability. Green algae have more complex, selective efflux systems suited to metal-rich environments. Red-lineage secondary endosymbionts emphasize generalist transporters, better fitting oligotrophic, metal-poor niches.
• Binding affinity signatures: Secondary endosymbionts show higher proportions of thiol-containing residues in key transporters (e.g., ZIP Zn transporters, ferric reductases) and elevated histidine content in high-affinity Ni transporters, multicopper oxidases, and Fe(III) transporters, suggesting greater selectivity and affinity under metal scarcity and competition with abundant Mg/Ca.
• Subcellular targeting patterns: Primary endosymbionts have larger plastid-targeted fractions of metalloproteins: Fe 30–53%, Cu 4–25%, Mn 14–80%, Co 30–67%, versus secondary endosymbionts: Fe 17–27%, Cu 0–8%, Mn 0–25%, Co 0–50%. Many specific metal transporters are plastid-targeted (Zn 30–70%, Fe 10–25%, Cu 17–100%, Co 30–90%). Mo-containing proteins are predominantly mitochondria-targeted (43–80% primary; 30–75% secondary), with no plastid- or mitochondria-specific Mo(VI) transporters detected.
• Fe and Zn strategy contrasts: An inverse relationship exists between the proportion of Fe-specific transporters and the percentage of Fe-binding proteins across proteomes, reflecting increased transporter investment as Fe becomes scarce. For Zn, the proportion of Zn transporters positively correlates with Zn-binding protein percentages; eukaryotes show higher Zn usage associated with regulatory domains (Zn fingers/RING), scaling with proteome size (r^2 = 0.91), consistent with increasing organismal complexity.
• Phylogenetic clustering: PCA of transporter families separates cyanobacteria, red, and green lineages, mirroring elemental stoichiometry patterns.
• Geological and ecological context: Pyrite/shale records indicate declines in Zn and Cu concentrations during the Mesozoic, coincident with radiations of secondary endosymbionts (diatoms, coccolithophores, dinoflagellates). Modern observations show secondary endosymbionts thrive in metal-poor open ocean, whereas primary endosymbionts are more coastal, aligning with inferred metal requirements and transport strategies.

The findings indicate that phytoplankton lineages carry distinct, evolutionarily inherited metalloproteomes and transporter architectures that align with past and present metal availability. Secondary endosymbionts evolved reduced proteomic reliance on Fe, Cu, and Zn, combined with higher-affinity, more selective transport systems, favoring survival in well-oxygenated, oligotrophic, and trace-metal-poor waters. In contrast, green-lineage primary endosymbionts retain more complex efflux and detoxification capacities that suit environments with higher and potentially toxic metal levels (e.g., coastal/terrestrial settings). Subcellular targeting patterns suggest that plastid-derived metal use (notably Fe, Co) diminished in secondary plastids, with increasing regulatory control by eukaryotic hosts, while Mo usage reflects eukaryotic innovations. The negative Fe and positive Zn associations between transport investment and proteome usage highlight divergent evolutionary pressures: Fe scarcity in oxygenated oceans drove transporter diversification and Fe economization, whereas Zn’s unique biochemical roles expanded with eukaryotic complexity. Geological records of declining Mesozoic Zn and Cu, together with modern biogeography, support the interpretation that reduced open-ocean trace metals were a key selective driver of Mesozoic phytoplankton succession. Thus, biological genomic signatures serve as a complementary proxy to geochemical archives for reconstructing seawater metal chemistry over time.

This study integrates comparative genomics, protein domain analyses, and subcellular targeting to demonstrate lineage-specific metal requirements and acquisition strategies in phytoplankton. Secondary endosymbionts are characterized by metal-lean proteomes and higher-affinity, generalist transport systems, consistent with adaptation to oligotrophic, metal-poor conditions that emerged prominently in the Mesozoic. Primary endosymbionts exhibit higher metal requirements and more specialized efflux capacities, aligning with metal-rich environments. The results imply a substantial decline in open-ocean trace metal availability at the onset of the Mesozoic, contributing to phytoplankton community shifts and altered ocean chemical buffering. Future work should expand taxonomic and genomic sampling, integrate expression and proteomic data to refine requirement estimates, and develop independent geochemical proxies to test the inferred Mesozoic decline in trace metals and its spatial structure.

• Metalloproteome inferences are based on genome-encoded potentials, not expression or post-translational regulation; actual cellular metal quotas can vary with environment.
• Domain-level analysis does not distinguish between transporters and other metalloproteins, potentially confounding inferences of metal use; data quality varies among MMETSP transcriptomes.
• Limited cyanobacterial sampling (n=2) and a non-exhaustive set of eukaryotic genomes may bias group averages.
• Subcellular targeting predictions and transporter functional annotations carry uncertainty; many ABC transporters are promiscuous and multi-substrate.
• Geological Zn and Cu reconstructions have uncertainties and potential diagenetic overprints; linking biological patterns to metal availability remains partly inferential and needs further proxy validation.