48 well tissue culture plates (Genesee Cat #: 25-108) were used for initial growth. A 0.5-inch drill was used to bore a hole through the bottom of desired wells. A ~1cm X ~0.75cm piece of Grodan^® A-OK Stonewool starter cubes was cut and placed into the hole, leaving a small tuft of fibers on either side. Approximately 1gm of each soil type was placed into the wells and watered with watering solution through capillary action by placing the plate into a container with the liquid just covering the bottom of the container (Paul et al., 2022). Plates remained in the watering solution for ~60 sec or until the liquid had permeated the top of the soil.

An antioxidant solution was prepared by combining 0.75 mM glutathione, 0.75 mM ascorbic acid, and 0.75 mM proline, following the method outlined by El-Beltagi et al. (2020). For each gm of regolith simulant, 20 ml of washing fluid was added to achieve complete saturation, as per the procedure by Kim et al. (2021). A vessel equipped with a magnetic stir bar on a stir plate was used to maintain a full vortex column for 120 min at room temperature (Kim et al., 2021). The solution was transferred into a Nalgene Rapid-Flow Filter Unit and vacuum filtration was performed until the substrate was completely dry, either overnight or until no residual moisture remained. To recycle the LMS-1 regolith simulant, we took previously treated antioxidant regolith, which had already proven to support plant growth, and subjected it to another round of washing with the antioxidant cocktail, followed by vacuum filtration and regolith drying before replanting.

Seeds for Col-0 were obtained from The Arabidopsis Biological Resource Center (ABRC) at Ohio State University, stock CS28994. Seeds were surface sterilized in 2.7% sodium hypochlorite for 7 min, stratified in 4°C for three days, and transferred directly to assigned treatment of regolith simulant (LMS-1, Exolith Lab, Oviedo, FL, USA) or Earth soil (Sunshine, Mix 5, Sun Gro Horticulture, Agawam, MA, USA). Seeds were germinated under a 12-h photoperiod of an average illuminance of 5000 (+/− 250) lux attained with 3000:6500 Kelvin lights in a 1:1 ratio, and at a constant temperature of 22°C and allowed to grow under the same conditions for five weeks or until growth stopped, at which point whole plants were harvested for analyses.

To generate CAT2 (At4g35090) overexpression lines, pCB302-CAT2 fused with a C-terminal 2x HA epitope tag and constitutive 35S promoter was generously provided by Dr. Ping He (University of Michigan) and transformed into wild type Col-0 plants using Agrobacterium tumefaciens strain GV3101 via in planta vacuum infiltration (Tague and Mantis, 2006). T1 transformants with Basta resistance were selected and CAT2-HA expression was confirmed by western blotting. T2 transformants with Basta resistance were selected from transgenic lines that segregated as a single locus. T3 homozygous transgenic lines were used for further experiments.

Statistical analyses were executed utilizing GraphPad Prism 9 for MacOS. Each experimental setup considered individual plants, with growth stages specified in figure legends. A minimum of three biological samples were included in each experiment. PCR and ELISA-based assays were performed with at least three technical replicates. Significance was assessed through p-values, deeming differences significant when below 0.05.

Genomic DNA was prepared following the standard cetyltrimethylammonium bromide (CTAB) method (Doyle and Doyle, 1987) with minor modification. A total of 0.5–1 gm of frozen plant DNA tissue was ground to a powder and mixed in a 1:1 ratio (w/v) with the CTAB lysis buffer (100 mM Tris-HCl pH 8.0, 20 mM EDTA pH 8.0, 1.4 M NaCl, 2% CTAB, 2% mercaptoethanol) and incubated at 65°C for 1 h. Phenol:chlorophorm:isoamyl alcohol (25:24:1, v/v/v) was added to the plant extract in a 1:1 ratio, and total DNA was extracted into the aqueous phase by careful mixing. The phases were separated by centrifugation at 20,000 rpm for 20 min. The aqueous phase was transferred to a new tube and DNA was precipitated for 2 h at −20°C, by adding 3M sodium acetate (1/10 aqueous phase volume) and 100% 2-propanol (2% aqueous phase volume). DNA was pelleted by centrifugation at 17,000 rpm for 20 min, washed with 70% ethanol and treated with 10 mg/mL RNase A for 30 min at 37°C.

Single telomere analysis was performed using primer extension telomere repeat amplification (PETRA) as previously described (Heacock et al., 2004). Bulk telomere length was measure by TRF analysis using Southern blotting (Heacock et al., 2004). Telomere length was measured using the online tool, WALTER (Lyčka et al., 2021).

Three leaf discs with a diameter of 2.5 mm were excised from the 4-week-old Arabidopsis plants expressing CAT2 C-terminal tagged with 2xHA. Three discs were frozen and ground, and 80 μl of 2x SDS loading buffer (125 mM Tris-HCl, pH 6.8, 2% SDS, 0.05 Bromophenol blue, 20% glycerol, 200 mM β-Mercaptoethanol) was added. The mixture was vortexed and boiled at 95°C for 10 min. After boiling, the mixture was centrifuged at maximum speed for 5 min. Then, 8 μl of supernatant was loaded on a 10% SDS-PAGE gel. Following electrophoresis, proteins were transferred onto a PVDF membrane (Millipore) using a Bio-Rad Trans-Blot apparatus at 25 volts for 30 min. To detect CAT2-2xHA fusion protein, a-HA- antibodies conjugated to the horseradish peroxidase (HRP; 1:2,000) was used.

Plants were gently removed from soil to ensure that the roots were not damaged and excess soil was removed from the base of the plant. Plants were individually placed on a scale to record their biomass and each ample was wrapped in aluminum foil, flash-frozen in liquid nitrogen, and stored at -80°C. Stalk height, rosette area and leaf size was calculated using ImageJ software by first taking photos of the plants, ensuring the relevant area of the plant was in-frame. Photos were uploaded to a computer running ImageJ. A scale was set using a representative known size in the plant pot. This allowed ImageJ to convert pixels to cm accurately for each photo. Once the scale was set, a line was traced from the area of interest to measure.

Chromosome spreads were obtained from pistils and stained with DAPI (4’,6-diamidino-2-phenylindole) as described in (Armstrong et al., 2009) with minor modifications. After 24 h of incubation in fixation solution (3 parts 100% ethanol and 1 part acetic acid), pistils were separated from the rest of the flower using watchmaker’s forceps and a needle. Two to three pistils were rinsed twice in Milli-Q water, then twice in 1X citrate buffer with 5 min incubations at room temperature. A solution of 660 μl 1X citrate buffer and 330 μl enzyme mix (1% cellulase and 1% pectolyase in 1X citrate buffer) was used for digestion. Samples were incubated for 60-90 min at 37°C in the enzyme solution and then quickly transferred to cold 1X citrate buffer to stop the reaction and incubated for at least 15 min. Pistils were placed on a sterile microscope slide with 70 μl of 60% acetic acid for 2-5 min until pistils lost their white color. A coverslip was carefully placed on top and gentle taps were applied to spread the cells. The slides with spread pistil cells were placed in liquid nitrogen for 30 sec. The coverslip was removed using a blade, and the slides were rinsed with fixative. The slides were dried for 20 min at room temperature. For chromosome staining, Drop-n-Stain EverBrite™ Mounting Medium without DAPI (23009, Biotium) was used. 18 μl of Drop-n-Stain EverBrite™ was placed on the sample, covered with a coverslip, and sealed with clear nail polish.

Lunar regolith simulant is a commercially available substrate that mimics the soil composition of lunar regolith recovered from the lunar surface via prior space missions (genuine lunar regolith). Both lunar regolith simulant and more recently genuine lunar regolith have been characterized as media that support plant growth (Ming and Henninger, 1994; Wamelink et al., 2014; Paul et al., 2022). To further explore the growth and sustainability of A. thaliana, we performed experiments with the lunar regolith simulant LMS-1 (Long-Fox et al., 2023) supplemented with 0.125x MS watering using the conditions previously described by Paul et al. (2022) ( Supplementary Figures S1A-C ). Consistent with the former study, plants grown on LMS-1 germinated, but at a rate that was 10% of controls grown in standard Earth soil ( Supplementary Figures S1C, D ). In an attempt to improve germination efficiency, we employed a soil remediation technique to reduce the levels of heavy metals in the substrate by pre-washing with EDTA. The substrate was subsequently filtered through a Nalgene rapid-flow filter unit followed by vacuum filtration until dry (Xu et al., 2019; Kim et al., 2021) ( Supplementary Figure S1B ). This treatment did not support seed germination, and neither did supplementation of the LMS-1 with a human waste simulant ( Supplementary Figures S1C, D ).

We next sought to decrease oxidative damage associated with lunar regolith by pre-washing the LMS-1 with an antioxidant cocktail consisting of 0.75 mM glutathione, 0.75 mM ascorbic acid and 0.75 mM proline (El-Beltagi et al., 2020) ( Supplementary Figures S1A-C ). Compared to untreated LMS-1, plants germinated on antioxidant-treated substrate at a rate of nearly 60% of the Earth soil control ( Supplementary Figure S1D ) and grew to yield a rosette area that was much larger than plants grown in untreated LMS-1 (0.59 cm³ for antioxidant LMS-1 versus 0.014 cm³ for untreated LMS-1) (p = 0.0394 calculated as one-way ANOVA, n = 6) ( Figures 1A, B ). Plants grown in antioxidant-treated regolith simulant for 59 days displayed an increased biomass of 1.72 gm on average compared to untreated regolith plants (ave = 0.02 gm, p = 0.0337 calculated as one-way ANOVA) ( Figures 1C, D ). However, relative to Earth soil controls, biomass in plants grown in antioxidant washed LMS-1 was 2-fold lower (1.72 gm vs 3.7 gm) (p = 0.0164 calculated as one-way ANOVA) ( Figures 1C, D ).

Subsequent experiments indicated the individual components of the antioxidant cocktail had different effects on plant growth relative to the complete antioxidant cocktail and the untreated LMS-1 ( Figure 2 ). Compared to antioxidant-treated LMS-1, proline-treated substrate yielded plants with a similar rosette area (ave_proline = 0.98 cm³, ave_antioxidant = 0.58 cm³, p = 0.74 calculated as one-way ANOVA) ( Figure 2B ), biomass (ave_proline = 2.50 gm, ave_antioxidant = 3.66 gm, p = 0.90 calculated as two-way ANOVA) ( Figure 2C ) and height (ave_proline = 34.41 cm, ave_antioxidant = 34.40 cm, p > 0.99 calculated as one-way ANOVA) ( Figure 2D ). Likewise, plants treated with 0.75 mM glutathione displayed a similar growth-promoting ability was those treated with the complete antioxidant cocktail in terms of biomass (ave = 2.33 gm, p = 0.97 calculated as two-way ANOVA) ( Figure 2C ), rosette area (ave = 0.44 cm³, p = 0.99, calculated as one-way ANOVA) ( Figure 2B ), and plant height (ave = 26.91 cm, p = 0.11 calculated as one-way ANOVA) ( Figure 2D ). Conversely, 0.75 mM ascorbic acid failed to stimulate the same extent of plant growth compared to full antioxidant treatment ( Figure 2A ), leading to reduced plant height (ave = 12.88 cm, p < 0.0001 calculated as one-way ANOVA) ( Figure 2D ) and biomass (ave = 0.07 gm, p = 0.048 calculated as two-way ANOVA) ( Figure 2C ), but similar rosette area (ave = 0.18 cm³, p = 0.73 calculated as one-way ANOVA) ( Figure 2B ). These data support the conclusion that exogenous application of antioxidants can enhance plant growth on lunar regolith simulant.

In parallel with the chemical antioxidant treatment, we used a genetic approach to increase the level of antioxidant activity of A. thaliana in vivo by ectopically expressing the antioxidant enzyme Catalase 2 driven by the powerful CaMV 35S promoter (35S::CAT2). Ectopic Catalase 2 expression was verified using western blotting ( Supplementary Figure S2A ). 35S::CAT2 line 11 was chosen for further analysis. This line achieved an 89% germination rate on LMS-1 supplemented with the antioxidant cocktail, similar to the wild type Col-0 on this substrate ( Supplementary Figures S2B, C ). However, 35S::CAT2 showed a remarkable 77% germination rate on untreated LMS-1, compared to only 22% for Col-0 ( Supplementary Figures S2B, C ). Similar to Col-0 grown on untreated LMS-1 ( Figure 1 ), 35S::CAT2 plants that were germinated on untreated LMS-1 were not viable past 20-days of growth. We also analyzed plants completely devoid of CAT2, cat2 null mutants. These plants had a germination rate of only 15% on antioxidant-treated LMS-1 and completely failed to germinate on the untreated LMS-1 substrate ( Supplementary Figures S2B, C ). We conclude that enhancing antioxidant mechanisms through chemical or genetic means can substantially stimulate A. thaliana growth on lunar regolith simulant.

The establishment of interplanetary colonies will necessitate continuous plant growth across multiple generations. Our experiments and previous studies (Paul et al., 2022) indicate that while plants can germinate in untreated LMS-1, they arrest in vegetative growth without producing a germline. Strikingly, plants germinated in antioxidant-treated LMS-1 (generation 1 or G1) can successfully transition to the reproductive phase and produce viable seeds for generation 2 (G2) progeny ( Figures 3A, B ). A 60% germination rate was observed in both G1 LMS-1 and G2 LMS-1 (p = 0.5508 calculated as one-way ANOVA), although both sets of plants displayed a lower germination rate than the Earth soil control (p < 0.0005 calculated as one-way ANOVA) ( Figure 3A ). We collected seeds from G2 LMS-1 plants and determined that they successfully germinate, indicating that a third generation of viable plants may be possible. Analysis of G1 LMS-1-grown plants 23 days post germination revealed a rosette area that was reduced two-fold (ave_{G1 =} 0.51 cm³) compared to its Earth soil grown plants (ave_EC = 0.83 cm³, p = 0.0040 calculated by one-way ANOVA) ( Figure 3C ). The rosette area was four-fold less than for G2 antioxidant-washed LMS-1 plants compared to earth soil (ave_{G2 =} 0.28 cm³, p < 0.0001 calculated as one-way ANOVA) ( Figure 3C ). Biomass measurements taken 54 days post-germination also revealed a reduced net weight of both G1 (ave = 2.38 gm, p = 0.0347 calculates as one-way ANOVA) and G2 plants (ave= 1.72 gm, p = 0.0071 calculated as one-way ANOVA) relative to the Earth soil control (ave = 4.05 gm) ( Figure 3D ). Plant height of G2 antioxidant-washed LMS-1 plants was less (ave = 24.807 cm) than either G1 antioxidant-washed LMS-1 plants (ave = 31.311 cm, p = 0.01 calculated as one way ANOVA) or the Earth soil control (ave = 37.801 cm, p = 0.0001 calculated as one way ANOVA) ( Figure 3E ). These findings indicate that although plants can be grown for successive generations in antioxidant-treated lunar regolith simulant, generational growth leads to progressively reduced plant size and biomass, implying that inclusion of antioxidants in the lunar regolith simulant is insufficient to promote robust growth over successive generations.

Given previous studies showing that telomere length is maintained in Arabidopsis subjected to a variety of different environmental exposures (Lee et al., 2016; Campitelli et al., 2021; Barcenilla et al., 2023), we investigated whether telomere length would be similarly preserved in plants grown in lunar regolith simulant. Plants grown in untreated LMS-1 did not produce sufficient biomass for telomere length analysis, and therefore we measured telomeres in G1 and G2 plants grown in antioxidant-treated LMS-1. We used two strategies to assess telomere length: Terminal Restriction Fragmentation (TRF) analysis which determines bulk telomere length and Primer Extension Telomere Repeat Amplification (PETRA) which assesses telomere length on individual chromosome arms (Heacock et al., 2004). Average telomere length was quantitated by WALTER software (Lyčka et al., 2021). TRF analysis revealed telomeres of plants grown in Earth soil were an average of 3396 bp in length which, as expected, falls within the wild type size range of 2-5 kb for the Col-0 accession (Shakirov and Shippen, 2004) ( Figures 4A, B ; Supplementary Figure S3A ). Unexpectedly, however, telomeres in G1 plants grown on antioxidant-treated LMS-1 were shorter, measuring 3073 bp for a net decrease of 323 bp relative to the Earth soil control (ave = 3396 bp, p = 0.0021 calculated by one-way ANOVA) ( Figures 4A, B ). When the progeny of these plants were propagated to the next generation (G2), telomere length was even shorter (2560 bp) for a further loss of 513 bp relative to G1 plants (p = < 0.0001 calculated by one-way ANOVA) ( Figures 4A, B , Supplementary Figure S3A ). PETRA conducted on the telomeres from chromosome arms 5R, 3L, 2R and 1L likewise revealed telomere shortening. The average telomere length for Earth soil grown plants for these chromosome arms was 3317 bp ( Figures 4C, D ). In contrast, PETRA showed G1 antioxidant-washed LMS-1-grown plants had telomeres of 3063 bp on average (p = 0.02 calculated by one-way ANOVA), while for G2 antioxidant-washed LMS-1 plants the average telomere length was 2777 bp (p < 0.0001 calculated by one-way ANOVA) ( Figures 4C, D ). Thus, our PETRA data are consistent with the results of bulk telomere length analysis and indicate that relative to Earth soil grown plants, telomeres in plants grown in lunar regolith simulant for consecutive generations decline by approximately 300 bp in G1 and by an additional 500 bp in G2.

To confirm that telomere shortening was not an artifact of antioxidant treatment, we tested if plants grown on Earth soil that was pre-washed with the antioxidant cocktail undergo telomere shortening. Plants were grown for 10 days on this substrate and then telomere length was assessed by PETRA. We found no significant difference in telomere length between samples grown in untreated Earth soil (3449 bp) versus Earth soil treated with antioxidant cocktail (3617 bp, p = 0.08 calculated by unpaired t-test) ( Supplementary Figures S3B–D ). These data argue that some component(s) in the lunar regolith simulant are responsible for telomere shortening.

Telomere shortening can occur through a variety of mechanisms including loss of telomerase enzyme activity (Fitzgerald et al., 1999; Riha et al., 2001). To investigate whether telomerase was impaired by growth on lunar regolith simulant we measured telomerase enzyme activity using quantitative telomere repeat amplification protocol (Q-TRAP) (Kannan et al., 2008). We observed a significant reduction in relative telomerase activity, with G1 antioxidant-washed LMS1-grown plants registering only 0.39-fold the amount of telomerase activity of the Earth soil controls (p = 0.0047 calculated by one-way ANOVA) ( Figure 4E ). Similarly, the level of telomerase activity in G2 antioxidant-washed LMS-1 plants was half that of plants grown in Earth soil (0.51 fold, p = 0.035 calculated by one-way ANOVA), and not statistically different from the G1 plants (p = 0.79 calculated by one-way ANOVA) ( Figure 4E ).

Telomere shortening can also occur if chromosome ends become deprotected leading to nucleolytic digestion. However, we did not observe evidence of DNA degradation: PETRA products were represented by sharp bands ( Figure 4C ), indicating telomere tracts were largely intact. In addition, analysis of mitotically dividing cells in the pistils of antioxidant-washed LMS-1 grown A. thaliana did not reveal evidence of anaphase bridges that would be consistent with the end-to-end chromosome fusions (Heacock et al., 2004) ( Supplementary Figure S4 ).

Given the elevated oxidative stress response associated with plants grown in lunar regolith (Paul et al., 2022), we asked if the genome of plants grown in LMS-1 accumulates oxidative damage by assessing levels of genomic 8-oxoG. Compared to plants grown in Earth soil, A. thaliana grown in antioxidant washed LMS-1 displayed a 4-fold increase in genome oxidation for both G1 and G2 plants (p_{G1 lunar} = 0.0015, p _{G2 lunar} = 0.0027 calculated by one-way ANOVA) ( Figure 4F ). There was no difference in genome oxidation levels between G1 and G2 antioxidant-washed LMS-1 plants (p = 0.76 calculated by one-way ANOVA) ( Figure 4F ).

We also assessed the impact of individual components within the antioxidant cocktail on genome oxidation, since each component had a distinct effect on plant growth. LMS-1-grown plants treated with ascorbic acid (which exhibited smaller plant size, Figure 2 ) had the highest 8-oxoG levels compared to the Earth soil control (ave = 4.2 fold increase, p < 0.0001 calculated by two-way ANOVA) and the antioxidant cocktail (p = 0.043 calculated by two-way ANOVA) ( Figure 5A ). Conversely, for plants treated with proline, genome oxidation was similar to that of plants treated with the complete antioxidant cocktail (p = 0.55 calculated by two-way ANOVA). Interestingly in the case of glutathione treatment, relative genome oxidization was lower when compared to antioxidant cocktail full treatment (p = 0.0062 calculated by two-way ANOVA) and similar to Earth soil (ave = 1.82 fold change, p = 0.095 calculated as two-way ANOVA). These observations support the notion that antioxidants can ameliorate genome oxidation in lunar regolith simulant-grown plants. Notably, the extent of oxidative damage was inversely correlated with plant biomass (r = -0.95 calculated as Pearson correlation between plant biomass and 8-oxoG) ( Figure 5B ) and height (r = -0.72 calculated as Pearson correlation between plant height and 8-oxoG) ( Figure 5C ).

The addition of glutathione to the lunar regolith simulant was remarkably effective for maintaining telomere length. Telomere length in plants grown in this substrate (3917 bp) was the same as for Earth soil controls (3890 bp, p = 0.99 calculated as one-way ANOVA) ( Figures 5D, E ). Conversely, ascorbic acid treatment led to the most substantial reduction in telomere length compared to Earth soil (3353 bp, p < 0.0001 calculated as one-way ANOVA), as well as when compared to plants subjected to full antioxidant treatment (p = 0.0025 calculated as one-way ANOVA). Proline treatment resulted in telomeres that were shorter than plants grown in Earth soil (3438 bp, p < 0.0001 calculated as one-way ANOVA), but a similar length as in plants subjected to the full antioxidant cocktail (p = 0.088 calculated as one-way ANOVA) ( Figures 5D, E ). We conclude that inclusion of glutathione in the regolith simulant is sufficient to effectively preserve telomeres at the same length as in plants grown in Earth soil. By contrast, proline-treated LMS-1 yielded plants with telomeres of a length similar to those grown in the full antioxidant cocktail, whereas ascorbic acid treatment failed to replicate the effects of the complete antioxidant cocktail. Further analysis revealed a strong negative correlation between telomere length and 8-oxoG measured for each of the antioxidant treatments (r = -0.87 calculated as Pearson correlation) ( Figure 5F ). The variations in telomere length and overall size in LMS-1-grown plants, particularly the decrease in telomere length with ascorbic acid treatment and its recovery with glutathione, correlate with their impacts on the level of oxidative DNA damage. The relatively high 8-oxoG content in ascorbic acid-treated plants is accompanied by telomere shortening and smaller plant sizes. In contrast, glutathione appears to effectively counteract genomic oxidative stress, preserving telomere length and sustaining healthier plant growth. Taken together, these data indicate that genomic DNA oxidation is substantially increased in plants grown in lunar regolith simulant, consistent with the transcriptional activation of an oxidative stress response (Paul et al., 2022). However, antioxidant pre-wash, particularly with glutathione, can mitigate this damage.

Finally, since oxidative stress levels can affect organelle function (Gonzales-Ebsen et al., 2017), we monitored organellar function by assessing the abundance of chloroplasts and mitochondria in plants cultivated on LMS-1 using PCR to amplify genes encoded by mitochondria (cox1, rps4, and atp1), chloroplasts (psbA, clpP and ndhH) (Weihe, 2014) or nuclei (RpoTp and RpoTm) as a control. We detected no substantial difference in the amount of mitochondrial and chloroplast DNA in G1 plants grown in antioxidant treated LMS-1 (0.77 fold-change_mito, 0.63 fold-change_chloro), G2 plants grown in treated LMS-1 (1.06 fold-change_mito, 1.23 fold-change_chloro) and the Earth soil controls (1.01 fold-change_mito, 1.12 fold-change_chloro) ( Supplementary Figure S5 ). We conclude mitochondrial and chloroplast DNA levels are not impacted by cultivation in antioxidant-treated LMS-1 substrate.

Recycling soil can significantly enhance plant growth (Liu et al., 2008; Kasiviswanathan et al., 2022; Wang et al., 2022). To test if reused lunar regolith simulant provided a more hospitable environment, we compared the growth of wild type A. thaliana in untreated, antioxidant washed, and reused antioxidant washed LMS-1 ( Figure 6A ). Plants grown on reused LMS-1 displayed visually enhanced growth with respect to antioxidant-treated fresh LMS-1 ( Figures 6A, D ). At 18 days, plants grown in reused LMS-1 displayed greater rosette area (0.014 cm³) compared to fresh regolith simulant (0.008 cm³, p < 0.0023 calculated as one-way ANOVA) ( Figure 6B ). Leaf size was also increased in reused LMS-1 (3.02 cm) compared to fresh regolith simulant (1.35 cm, p < 0.0001 calculated as one-way ANOVA) ( Figure 6C ). After 34 days, plants grown in reused regolith simulant reached an average height of 17.8 cm compared to only 12.1 cm in fresh LMS-1 (p = 0.023, calculated as one-way ANOVA) ( Figures 6D, E ). Notably, plant height was similar for plants grown in parallel in Earth soil (ave = 21.06 cm, p = 0.38 calculated as one-way ANOVA) ( Figures 6D, E ). Total biomass at 34 days decreased in reused LMS-1 (ave = 0.4 gm) compared to the Earth soil controls (ave = 0.95 gm, p = 0.014 calculated by one-way ANOVA), but was similar to fresh LMS-1 (ave = 0.22 gm, p = 0.69 calculated as one-way ANOVA) ( Figure 6F ).

We next asked if the improvement in plant growth on reused regolith simulant correlated with other molecular phenotypes. Compared to Earth soil, both fresh LMS-1 (ave = 1.94-fold change, p < 0.0001 calculated as one-way ANOVA) and reused LMS-1 (ave = 1.42 fold-change, p = 0.008 calculated as one-way ANOVA) displayed increased genome oxidation ( Figure 6G ). However, plants grown in reused regolith simulant displayed decreased genome oxidation compared to fresh LMS-1 (p = 0.0036 calculated as one-way ANOVA) ( Figure 6G ). Moreover, telomere shortening was not as pronounced when plants were grown in reused regolith simulant (ave = 3423 bp) compared to fresh LMS-1 (ave = 2986 bp, p = 0.034 calculated as one-way ANOVA) ( Figures 6H, I ). Although telomere tracts were shorter than for plants grown in Earth soil, the difference was 394 bp for plants grown in reused LMS-1 (p = 0.06 calculated as one-way ANOVA) and 831 bp for fresh regolith simulant (p < 0.0001 calculated as one-way ANOVA), compared to Earth soil (ave = 3817 bp) ( Figure 6H, I ). Telomerase activity levels were indistinguishable in plants grown in recycled regolith versus fresh regolith with both sets of plants showing 40% reduction relative to plants grown in Earth soil (p < 0.001 calculated as ordinary one-way ANOVA) ( Figure 6J ). Because telomeres are substantially shorter in plants grown in fresh regolith compared to the recycled substrate, these findings support the conclusion that factors besides telomerase contribute to telomere shortening in plants grown in lunar regolith simulant.