Plants grown in Apollo lunar regolith present stress-associated transcriptomes that inform prospects for lunar exploration

The NASA Artemis program renews interest in how the lunar environment affects terrestrial biology and the feasibility of using in-situ resources for sustained lunar habitation. Plants are central to life support concepts (food, oxygen, water recycling, CO2 scrubbing) and serve as model organisms for studying space-related biology. Despite decades of space plant research and reliance on hydroponic systems due to uncertainty about regolith substrates, plants had not previously been grown in true lunar regolith with modern molecular tools. This study asks whether Arabidopsis thaliana can successfully develop in Apollo lunar regolith and, if so, what metabolic strategies are engaged, as inferred from differential transcriptomes. Samples from Apollo 11 (10084), Apollo 12 (12070), and Apollo 17 (70051), representing mature/submature versus immature regoliths from basaltic terrains, were compared to the lunar simulant JSC-1A. The work informs both fundamental plant–lunar material interactions and the potential use of lunar regolith as a growth matrix in exploration habitats.

Prior plant space biology has elucidated responses to gravity, radiation, and other factors, advancing understanding of physiological adaptation in space. Most extraterrestrial plant growth system designs utilize hydroponics, partially due to limited knowledge of in-situ substrates like planetary regolith. During Apollo, biological quarantine and pathogen screening involved transient contact of lunar materials with organisms, including rubbing fines on leaves or sprinkling on seedlings and media; however, plants were not grown with lunar regolith as the supporting matrix. Regolith simulants such as JSC-1A have been used terrestrially but do not fully replicate lunar agglutinates or nanophase metallic Fe content and oxidation states. Lunar regolith maturity increases agglutinates and nanophase Fe, potentially influencing plant responses. This historical and materials context motivated direct growth experiments in actual lunar regolith using contemporary molecular analyses.

Plant material: Arabidopsis thaliana ecotype Columbia-0 (Col-0; TAIR CS70000) seeds from a single batch were used.

Lunar regolith materials and controls: The NASA AARB (formerly CAPTEM) provided 4 × 1 g regolith samples each from Apollo 11 (10084-2075, -2076, -2077, -2078), Apollo 12 (12070-105, -99, -106, -109), and Apollo 17 (70051-159, -160, -161, -162). All samples were sieved to <1 mm. JSC-1A lunar simulant (<1 mm) served as the terrestrial control.

Growth system: A small-scale system using four 48-well sterile culture plates was developed. Each well (12.5 mm diameter, 15 mm deep) contained 900 mg of regolith/simulant over a subsurface irrigation assembly: a Rockwool plug (10 × 15 mm cylinder compressed to ~7 mm) with a tuft extending through a drilled drainage hole, topped with a 13 mm, 0.45 µm nylon filter. Rockwool is an inert spun basalt fiber material.

Irrigation and nutrients: Plates were bottom-watered daily with 0.125× Murashige and Skoog (MS) nutrient solution, pH 5.7. Rockwool wicking prevented waterlogging. Apollo 17 samples and JSC-1A wetted by capillarity; Apollo 11 and 12 regoliths were initially hydrophobic and were actively stirred with nutrient solution to achieve uniform wetting, after which they behaved similarly to the other substrates.

Planting and growth conditions: 3–5 Col-0 seeds were pipetted onto each well surface. Plates were placed in ventilated terrarium boxes to reduce airflow while allowing exchange, and kept under growth lights in a secured plant growth room. Germination occurred within 48–60 h. Between days 6–8, seedlings were thinned to one per well. Plates were photographed daily. On day 20, aerial tissues (leaves and hypocotyl) were harvested at substrate level, snap-frozen in liquid nitrogen, and stored at −80°C for RNA-seq.

Experimental design and replication: Per plate, there were four JSC-1A control wells and three lunar wells (one each from A11, A12, A17). Across four plates: JSC-1A n=16; each Apollo site n=4. For transcriptomics, n=4 per group were targeted; Apollo 11 yielded n=3 due to one sample lacking viable RNA; Apollo 12 and 17 had n=4; JSC-1A controls used plant #4 from each plate for parity (n=4).

Calibration with simulants: Preliminary experiments using JSC-1A, JSC-Mars-1A, and sieved potting soil established the plate configuration and nutrient concentration. Serial dilutions identified 0.125× MS as optimal in JSC-1A; water alone did not support development beyond first true leaves. Given similar nutrient composition between JSC-1A and lunar regolith, supplementary nutrients were deemed necessary; however, insufficient lunar regolith was available to test nutrient necessity directly in lunar samples.

RNA isolation and sequencing: RNA was extracted with Qiagen RNeasy Plant Mini Kit; quality assessed via Qubit and Agilent Bioanalyzer. Poly(A)+ mRNA was isolated (NEBNext E7490) and strand-specific libraries prepared (NEBNext Ultra II Directional kit E7760). Libraries were barcoded, amplified (12 cycles), purified (AMPure), sized, and quantified. Libraries underwent Illumina FAB to minimize adaptor dimers and index hopping, diluted to 0.8 nM, and sequenced on one Illumina NovaSeq 6000 S4 lane (2×150 cycles) at 120 pM with 1% PhiX. Output: 2.5–3.0 billion paired-end reads (~950 Gb), average Q30 ≥ 92.5%, Cluster PF 85.4%; ~50 million paired reads per sample.

Bioinformatics and statistics: Reads were quality-checked (FastQC) and trimmed (Trimmomatic), mapped to TAIR10 (STAR), and quantified (RSEM). Differential expression used edgeR with FDR ≤ 0.05, |fold change| > 2, and average FPKM > 0 in at least one replicate of each comparison. Clustering and PCA assessed sample associations. DEG overlaps were visualized with BioVenn. Transcriptome analyses were conducted either grouped by Apollo site or by plant morphology categories (Severe, Small, Large) derived from growth phenotypes and sizes quantified from daily images.

• Germination and early development: Germination was near 100% in all Apollo regoliths and indistinguishable from JSC-1A, occurring 48–60 h after sowing. Cotyledon development and stems appeared normal across treatments.
• Growth and morphology: Roots from lunar wells were stunted relative to JSC-1A by day 6–8. After thinning, lunar regolith-grown plants developed more slowly, had smaller rosettes, and some showed severe stress phenotypes (stunting, deep pigmentation). JSC-1A replicates displayed uniform growth; lunar plants showed greater variability. Across sites, Apollo 11 plants performed worst, Apollo 12 intermediate, and Apollo 17 best.
• Site-based transcriptomes (day 20 aerial tissues vs JSC-1A): • Apollo 11: 465 DEGs (most among sites). • Apollo 12: 265 DEGs. • Apollo 17: 113 DEGs (fewest among sites). • Approximately 71% of DEGs across lunar samples were associated with ionic and oxidative stress responses (salt, metal, ROS). Coordinately expressed DEGs across all sites also included nutrient metabolism genes. The most repressed common gene was Protochlorophyllide Oxidoreductase-A (AT5G54190); a highly induced common gene of unknown function (AT5G26270) is associated with phosphate starvation. Other commonly induced defense/stress genes included Nitrilase-2 (AT3G44300; metal/cadmium), Jasmonate-regulated gene-21 (AT3G55970), Defensin-like (AT1G13609), and DUF677 family genes (AT3G28290, AT3G28300) linked to aluminum toxicity and jasmonate signaling.
• Morphology-based transcriptomes: • Phenotype groups: Severe (tiny, distorted, reddish-black), Small (small but proportionate and green), Large (closest to JSC-1A morphology but still smaller). • DEG counts vs JSC-1A: Severe = 1328 DEGs; Small = 130; Large = 150. Even Large and Small plants showed stress-response transcriptomes. • Unique signatures: Severe plants were dominated by ROS-associated DEGs and indicators of developmental stress (hormones, cell wall remodeling, calcium signaling). Small plants showed ROS signaling with minimal metal-stress signatures. Large plants had >60% of uniquely induced DEGs tied to salt/drought stress; many LEA proteins were 50–100-fold upregulated. The most downregulated unique DEG in Large plants was ATPC2 (AT1G15700), involved in ATP synthase regulation.
• Materials properties and performance: Apollo 11 and 12 regoliths were initially hydrophobic; Apollo 17 was not. Plants in the mature Apollo 11 regolith struggled most and showed the most DEGs; those in the less mature Apollo 17 regolith struggled least, suggesting regolith maturity negatively correlates with plant performance.
• Overall: Lunar regolith is usable as a primary plant growth substrate but is not benign; plants interpret it as a highly ionic, oxidative environment, eliciting stress-associated transcriptomes.

The study directly addressed whether plants can grow in actual lunar regolith and what metabolic pathways are engaged during adaptation. Arabidopsis germinated and developed in Apollo regoliths, demonstrating feasibility of using lunar regolith as a primary support matrix. However, growth delays, root stunting, and stress morphologies were common, and transcriptomes revealed strong induction of ionic (salt/metal) and ROS stress pathways across all sites, even in plants with near-normal morphology. The site-based differences (Apollo 11 > Apollo 12 > Apollo 17 in DEG counts and growth impairment) align with regolith maturity effects: mature soils have more nanophase metallic Fe and smaller grain sizes, potentially increasing surface reactivity and ion release, consistent with plants sensing a chemically reactive substrate. Morphology-based analyses further showed that physical establishment in regolith (root zone efficacy) contributes to success states, yet physiological stress remains a consistent feature. These findings imply that while lunar regolith can support plant life, significant mitigation—such as selecting less mature regolith, preconditioning substrates, managing ion availability, or engineering root zones—will be necessary to optimize plant performance for life support applications in lunar habitats.

This work provides the first demonstration of plant growth in true Apollo lunar regolith with concurrent transcriptomic profiling, showing that Arabidopsis can germinate and grow in diverse lunar soils but experiences substantial stress, predominantly ionic and oxidative. Performance varied by regolith source, with mature Apollo 11 regolith eliciting the most severe responses and immature Apollo 17 the least, indicating regolith maturity as a key determinant of substrate suitability. The results inform the prospects of using in-situ lunar regolith for plant production, highlighting the need for further characterization and mitigation strategies to reduce substrate reactivity and stress induction. Future research should investigate regolith preprocessing or amendments, nutrient and pH management, selection of less mature deposits, engineering of root-zone physical properties, and longer-duration growth to reproductive stages, alongside expanded multi-omics analyses to refine substrate optimization for lunar bioregenerative life support.

• Limited sample amounts constrained experimental scope; notably, there was insufficient lunar regolith to test directly whether supplemental nutrients were strictly necessary in lunar regolith (assumption based on JSC-1A calibrations).
• Transcriptome replication for Apollo 11 was reduced (n=3) due to one failed RNA sample; Apollo 12 and 17 had n=4; JSC-1A controls n=4.
• The experiment used a small-scale well-plate system with 900 mg substrate per well and assessed aerial tissues at day 20; results reflect early vegetative growth and aerial transcriptomes only.
• Apollo 11 and 12 regoliths required active wetting to overcome initial hydrophobicity prior to planting, a handling step that may differ from in-situ operations.