Potential evidence for transgenerational epigenetic memory in Arabidopsis thaliana following spaceflight

Plants grown in spaceflight exhibited differential methylation responses and this is important because plants are sessile, they are constantly exposed to a variety of environmental pressures and respond to them in many ways. We previously showed that the Arabidopsis genome exhibited lower methylation level after spaceflight for 60 h in orbit. Here, using the offspring of the seedlings grown in microgravity environment in the SJ-10 satellite for 11 days and returned to Earth, we systematically studied the potential effects of spaceflight on DNA methylation, transcriptome, and phenotype in the offspring. Whole-genome methylation analysis in the first generation of offspring (F1) showed that, although there was no significant difference in methylation level as had previously been observed in the parent plants, some residual imprints of DNA methylation differences were detected. Combined DNA methylation and RNA-sequencing analysis indicated that expression of many pathways, such as the abscisic acid-activated pathway, protein phosphorylation, and nitrate signaling pathway, etc. were enriched in the F1 population. As some phenotypic differences still existed in the F2 generation, it was suggested that these epigenetic DNA methylation modifications were partially retained, resulting in phenotypic differences in the offspring. Furthermore, some of the spaceflight-induced heritable differentially methylated regions (DMRs) were retained. Changes in epigenetic modifications caused by spaceflight affected the growth of two future seed generations. Altogether, our research is helpful in better understanding the adaptation mechanism of plants to the spaceflight environment. In order to investigate whether the effects of spaceflight on plants persist in future generations, Xu et al studied the offspring of Arabidopsis thaliana seedlings that had been grown in a microgravity environment for 11 days. They found that epigenetic modifications caused by spaceflight potentially affected the growth of two future seed generations, shedding light on the longevity of the effects of spaceflight on plants.

D ifferences in epigenetic regulation of gene expression induced by environmental stress exposure can persist into the next generation of stressed plants 1 . In some cases, these intergenerational effects can even be passed down to at least two seed generations 2 . In plants, repeated exposure to stress triggers stress adaptation, in which previous stress increases the plant's ability to respond to future stresses 3 . This concept has been extended to many aspects of the formation of plant stress memories, in which the altered state of plant stress response is transmitted by mitosis or meiosis 1,4,5 . Stress memory in plants is of great interest, including its underlying molecular mechanisms and its potential effects on crop yields, especially in harsh environments. Stress also causes epigenetic changes in DNA methylation profiles and further genomic instability 2 . Previous studies on Arabidopsis thaliana and Scots pine (Pinus silestris) growing near the Chernobyl reactor have shown that increased genome methylation level is related to genome stability and stress tolerance 6 . Molinier and colleagues reported that exposure to UVC increased the rate of somatic recombination in the plant progeny, whereas increases in UVC-induced homologous recombination frequency (HRF) continued for five generations 1 .
DNA methylation is considered to be a possible mechanism for the formation of stress memories in plants 7 . DNA methylation occurs in three sequence environments in plant genomes, namely mCpG, mCHG, and mCHH, where H is not G, and is thought to play a role in transposable element (TE) silencing and maintenance of genomic stability, and to contribute to gene expression 8 . This potential effect on gene expression suggests that DNA methylation may complement genetic variation and promote phenotypic variation as a means of transmitting genetic information 9 . In fact, the DNA methylation state can be faithfully maintained during mitosis or meiosis by a number of pathways and enzymes. During gametogenesis and spermatogenesis, parental DNA methylation patterns are reconstructed and reproduced, although it is not clear to what extent the patterns of DNA methylation in plant genomes are reprogrammed. Because these processes occur during growth from plant embryos 10 , any cumulative variation in DNA methylation ("epi-alleles"), whether environmentally induced or spontaneous, can persist for generations. DNA methylation shows stable genetic ability and the frequency of epi-alleles in the literature is similar to that of genetic polymorphisms due to spontaneous gene mutations, which may occur at a higher-than-normal frequency under conditions of abiotic stress.
This potential has led to many studies of epigenetic alleles, mediated by intergenerational stress memory, which may open up exciting possibilities for crop (epi-) genomics. However, an alteration in plant phenotype caused by a change in intergenerational methylation is still of a rare occurrence 11 . Due to the lack of early reports on the inheritance of spaceflight effects on plants, in the current study, we systematically studied the potential effects of spaceflight on environmentally induced DNA methylation, using the model plant Arabidopsis. With an experiment designed to minimize genetic variation, the ability of plants to create transgenerational stress memories over three successive generations was observed and combined with in-depth DNA methylation profiling and RNA-sequencing analysis. Subsequently, we also tested the possibility of microgravity-induced heritable alleles being associated with adaptive gene expression in spaceflight.
DNA methylation is a major epigenetic modification that modifies gene expression and is passed down at least one generation 7,12 . Previous studies have demonstrated that significant differences in the modification of 5-methylcytosine (5-mC) within the genome of plants undergoing spaceflight, compared with on-ground controls 13,14 . The main purpose of the study was to analyze the transgenerational memory induced by the spaceflight environment in Arabidopsis thaliana plants. The experiment was part of the Chinese Academy of Sciences SJ-10 satellite mission and involved a follow-up experiment to the launch of the Arabidopsis seedlings which were flown in the SJ-10 recoverable satellite in April 13 . Detailed analysis of the growth and development phenotypes in three consecutive generations from the original spaceflight seedlings was carried out. To evaluate the dynamic DNA methylome alteration and the genetic effects on the offspring following spaceflight, we profiled the genome-wide DNA methylation patterns in the F 1 generation of Arabidopsis seedlings, using whole-genome bisulfite sequencing (WGBS) 15 . Systematic phenotypic observations, combined with genome-wide methylation analysis, revealed a plant adapting comprehensively to the stressful microgravity environment associated with spaceflight, from the perspective of epigenomic methylation level. We report here on how exposure of Arabidopsis seedlings to the microgravity environment during spaceflight contributed to the intergenerational changes in plant development and stress response. Some of these changes continued for generations, even in the absence of subsequent stress, and we discuss here how these heritable changes may be due to changes in genomic methylation involved in epigenetic regulation of gene expression.

Results
Experimental design of the space vs. ground control comparison treatments, and follow-up experiments. The Col-0 Arabidopsis seedlings were grown in a growth chamber on earth for 6 days before being transferred to one of the two treatments ( Fig. 1a, b): either on the Chinese recoverable scientific satellite SJ-10 (S0 population), under microgravity conditions, or on Earth ("ground control"), at 1 g gravity load, as the control (G0 population). After 60 h culture, the total DNA was extracted for genome-wide methylation analysis from the G0 and S0 populations of the parent generation. After completion of the spaceflight (11 days), a representative sample of G0 and S0 seedlings was transplanted into soil on Earth to grow them to maturity to measure growth and development traits, and to collect selffertilized seed from individual plants to breed the next generation (and subsequent generations). This seed production was continued for three more generations, and phenotype analysis was carried out on each generation.
In the F 1 generation (G1 and S1 populations), DNA was extracted from 6-day-old seedlings of each population and used for genome-wide methylation analysis to detect changes in methylation levels, and to evaluate the genetic and phenotypic effects of DNA methylation at the whole-genome level. Transcriptome sequencing was also performed on the F 1 generation at the same time. Similar offspring studies were subsequently carried out on later generations of the ground control (G2-G3) and spaceflight groups (S2-S3), all carried out under strict and consistent growth conditions (Fig. 1b).
The progeny of spaceflight-stressed plants (S1) exhibited different phenotypes, compared with ground controls (G1). It is known that heavy metal and nutrient deficiency stresses can cause plants to adapt to stress, and, in some cases, this adaptation can be passed on to the offspring of the stressed plants 5,9 . To identify whether intergenerational alterations induced by exposure to microgravity during spaceflight were reflected in plant development and stress tolerance, we systematically compared the offspring of the spaceflight and the ground control populations. Individual spaceflight (S0) and ground control (G0) plants were individually self-fertilized to generate populations of S1 and G1 seeds, and, subsequently S2, S3 and G2, G3 populations (Fig. 1b). The plant growth parameters of the three generations of the two populations, which were assessed included leaf area, root length, flowering time, silique length, root gravitropism, and abscisic acid (ABA)-sensitivity analysis.
As we reported earlier, although the leaf area of the S0 Arabidopsis seedlings increased significantly in response to spaceflight microgravity treatment, in comparison with the G0 population 13 , there were no significant differences between the offspring populations (Fig. 3c, d). Interestingly, we observed that seedling root length (Fig. 2a, b) and flowering time (Fig. 2c-e) were significantly higher in the spaceflight S1 generation compared with the corresponding ground control generation (G1), but only the effect on the seedling root length phenotype was subsequently preserved in the F 2 (S2) generation. Furthermore, none of these growth phenotypic differences observed between the S and G F 2 populations were observed in the F 3 (S3) generation (Figs. 2, 3).
In terms of ABA sensitivity, the spaceflight F 2 (S2) generation exhibited significantly higher ABA sensitivity than the control (G2) population, although this advantage could not be observed in the F 3 generation of the S and G populations ( Supplementary   Fig. 1). We also observed that the significant difference between S1 and G1 in the phenotype of root curvature after a 90°change in gravitropic orientation was not observed in the F 2 or F3 generations (Fig. 3a, b). All in all, through our observations of three generations of both populations, we concluded that the phenotypic effects on Arabidopsis seedlings of exposure to the stress of the spaceflight environment could be observed in the treated S0 population and their F 1 offspring, S1, but the heritable effects in subsequent generations varied, with the heritable effect declining with increasing numbers of generations. To our knowledge, our results represent the first systematic analysis of the heritability of microgravity effects of spaceflight on the growth and development of Arabidopsis.
Spaceflight did not cause large-scale changes in genomic DNA methylation patterns, but the distribution of methylation levels changed slightly during the F 1 generation. The phenotypic changes in the S0 generation, compared with the G0 generation, were associated with major changes in genomic DNA methylation patterns 13 . In order to study the cause of the intergenerational effect of spaceflight on phenotype, we compared the variation in DNA methylation in the F 1 generation of the two Arabidopsis seedlings were transferred from petri dishes 3 d before launching. Seedlings in culture chambers installed in growth chamber unit 1 were grown under microgravity conditions and fixed in space using RNAlater® as described in the previous report 13 . The seedlings in the culture chamber installed in growth chamber unit two in the satellite grew for 11 days under microgravity; after this time of spaceflight, they were still alive, were returned to Earth, and cultivated until seed harvest. b Multiple independent lineages began with individual inbred Arabidopsis plants, propagated by single-seed descent, with self-fertilization, for three generations, with half of the lineages exposed to the microgravity of space and the remainder maintained on Earth. Intergenerational stress memory tests were performed on the offspring plants. Whole-genome bisulfite sequencing (WGBS) and transcriptome analyses were performed on the F 1 generation. Various growth phenotypes were assessed in successive F 1 , F 2 , and F 3 generations. The length of primary roots of seven-day-old seedlings in the successive F 1 and F 3 generations of spaceflight and ground control samples (n > 40 for F1 generation, n > 150 for F3 generation), and the length of primary roots of six-day-old seedlings in the F 2 generation (n > 160); Scale bar indicates 1 cm. c-e The flowering time analysis of the three successive F 1 , F 2 , and F 3 generations of spaceflight and ground control samples (n > 30 for each generation). The error bar is the standard deviation. An asterisk *indicates a statistically significant difference between the spaceflight and control samples using Student's two-tailed t-test (p < 0.05). . c, d The rosette leaf area of the offspring of spaceflight and ground control samples. Leaf area was determined by using ImageJ software. The data shown are the mean ± standard deviation values obtained for eight leaves (n > 30). e The silique length analysis in the offspring of spaceflight and ground samples. The data shown are the mean ± standard deviation values (n > 10). An asterisk * indicates a statistically significant difference between the samples by using Student's two-tailed ttest (p < 0.05). Scale bar indicates 1 cm.
populations (S1 and G1) of Arabidopsis (Col-0) seedlings by using the bisulfite-sequencing (BS-Seq) analysis 16 . By assessing the transformation process, the correct cytosine sites for bisulfite transformation were found to be near 98%, indicating that there had been a good conversion and processing rate in the experimental process. Whole-genome bisulfite sequencing (WGBS) revealed genome-wide DNA methylation patterns in the offspring of plants grown in space, as well as their comparable ground controls. The cytosine methylation states of CpG, CHG, and CHH sequences in the F 1 generations are visualized in the Integrative Genomics Viewer (IGV) (Fig. 4a). Unlike our previous genome-wide DNA methylation data from the parental populations 13 , spaceflight did not cause large-scale changes in DNA methylation in the F 1 generation, but the distribution of methylation levels was slightly affected by spaceflight. The total methylation level and the distribution of 5-mC in the genome did not change significantly in the genome (Fig. 4a). Figure 4b-e shows the proportional distribution of CpG, CHG, and CHH methylation levels, comparing ground control values with the spaceflight values. In each methylation environment, total methylation decreased slightly during spaceflight, compared with the controls (Fig. 4b), although there was no significant difference in the mean genomic methylation levels between the S1 and G1 samples in each sequence context. However, the level of methylation, at below 10%, was slightly lower in each of the three methylation environments of the ground control population than in the spaceflight population in the F 1 generation (Fig. 4c-e).
Spaceflight-induced DNA methylation changes are associated with protein-coding genes in the F 1 generation. We further evaluated detailed methylation patterns within genes, including coding sequences (CDS) and noncoding regions. Figure 5 shows the average methylation levels in protein-coding genes in the CpG, CHG, and CHH environments. The results showed that the methylation level at CpG sites was higher than that at CHG and CHH sites in each gene region. In all three environments, the methylation level in the untranslated (UTR) region was lower than that in the coding region ( Fig. 5d-f). It was observed that the altered average gene DNA methylation level was largely uniform across the entire gene region. Only in the content of CHG was the methylation level different between the two populations, being slightly lower in the F 1 generation of the spaceflight population, S1, than in the ground control population, G1 (Fig. 5b, e).
Furthermore, no large-scale changes in the average methylation level were detected across the protein-coding region affected by CpG and CHH methylation, and the overall change in DNA methylation was independent of any particular region of the gene being expressed (Fig. 5a, c, d, f).
Differential methylation regions (DMRs) were studied to assess the genome-wide methylation content. The DMRs were determined by comparing the mean methylation levels within a 1000bp window between the spaceflight and ground control populations and detecting statistically significant methylation levels in the analysis. To further analyze the functional classification of genes with altered methylation patterns, we used Gene Ontology (GO) enrichment to classify these genes 17 . In the biological process category of CpG, altered methylation-related (AMR) genes were enriched for the ammonia assimilation cycle, glutamate biosynthetic processes, protein phosphorylation, protein ubiquitination, epigenetic regulation of gene expression, response to UVC and the tyrosine kinase signaling pathway ( Supplementary Fig. 2a). In the biological process category of CHH, the AMR genes were enriched with respect to mitochondrial transport, transmembrane transport processes, and translation processes (Supplementary Fig. 2b). In addition, Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway enrichment analysis was performed for these AMR genes 18 . In the biological process of CpG, AMR genes for mRNA transport, carbon metabolism, amino acid biosynthesis, glycerolipid metabolism, arginine and proline metabolism, and cysteine and methionine metabolism were enriched (Supplementary Fig. 2d). In the CHG context, AMR genes for amino acid biosynthesis, starch metabolism, nucleotide excision repair, the mRNA surveillance pathway and the Fanconi anemia pathway were enriched ( Supplementary Fig. 2e).
The methylation levels of transposable element (TE) genes were modified by spaceflight in the F 1 generation. Previous studies have shown that TE mobilization and silencing are often accompanied by the loss of DNA methylation 19 . TEs can affect genome size, produce insertions and other mutations, and affect gene expression patterns 20,21 . Under spaceflight conditions, many different methylated TEs were demethylated in the S0 population 13 . Figure 6a-c shows the methylation patterns of CpG, CHG and CHH contexts in TE regions, as well as the methylation levels in the upstream and downstream regions 2-kb from the CDS. In the spaceflight samples (S1), TEs exhibited a slightly lower methylation level in all three CHG, CHH, and CpG environments (Fig. 6a). These results show that, in these environments, most TEs have different degrees of methylation, and the somewhat hypomethylation of TE sites in the S1 population may be induced by active TEs.
Previous studies had demonstrated that different types of TEs affect methylation levels. In addition, TE methylation levels can affect the expression of transposons under environmental stress [22][23][24] . In order to further determine the relationship between the methylation state of A. thaliana and the various types of TE, TEs were divided into five groups: DNA, long interspersed nuclear elements (LINE), short interspersed nuclear elements (SINE), long terminal repeats (LTR), and rolling-circle (RC) TEs 25,26 . The analysis showed that, compared with other TE types, the methylation signal of RCs was relatively low, whereas SINEs showed significant hypermethylation levels in the CHH context, and all types of TEs in the spaceflight sample showed slightly lower global methylation levels than did those in the ground controls (Fig. 6d-f), indicating that the demethylation mechanism of different types of TEs in the spaceflight samples differed. In addition, we found that the SINEs, LINEs, and LTRs had significantly lower methylation levels in the CHH context than those in the CpG or CHG contexts. In conclusion, the findings suggest that TE demethylation plays a role in adaptation to microgravity.
Relationship between differential DNA methylation and gene expression. RNA sequencing (RNA-Seq) was used to establish the model of gene expression differences between F 1 generations of spaceflight (S1) and ground control plants (G1). In the comparison between the spaceflight and ground control populations, about 400 genes were differentially expressed. These genes were associated with responses to abiotic stresses, oxidative stimuli, and hormones, etc. Among all differentially expressed genes (DEGs), 81 differentially expressed genes were also differentially methylated in the CpG context of the S1 population (Fig. 7a). Figure 7a shows the number of DEGs mapped to DMR in the CpG, CHG, and CHH contexts in different gene structures (2-kb upstream and downstream of the transcriptional start site (TSS)). It also provides a complete heatmap of the list of DMR-DEG genes and summarizes the relevant characteristics of this group. For each gene, the ID number was consistent with the differential expression heatmap and indicated the mean differential methylation level of the CpG context in each region. The figure also shows a consistent or opposite trend between differential expression and DNA methylation. The opposite trend occurred for genes, which presented in the gene region as hypomethylated DMR upregulation or CpG-methylated hypermethylated DMR downregulation, while the consistent trend described the inverse relationship. Furthermore, expression of the ABA-activated signaling pathway, the response to abiotic stress stimuli and the cold stimulus, nitrogen compound metabolic process, etc. was examined by Gene Ontology analysis (Fig. 7b). The biosynthesis of amino acids, carbon metabolism, glycolipid metabolism, etc. were examined by KEGG analysis (Fig. 7c). The scatter plots of different gene expressions and the differences of CpG libraries in different gene regions also reflected the relationship between DNA methylation and gene expression of a certain type.
Isolation and verification of the spaceflight-induced heritable DMRs in the offspring. Due to the fact that root length, flowering time and silique length were obviously and significantly increased in the spaceflight F 1 generation (S1), compared with the ground control offspring (G1) (Figs. 2, 3), analysis of the nature of this heritable effect was carried out. The spaceflight-associated DEGs also participated in the photosynthesis pathway, nitrogen compound metabolism and signaling pathway, etc. To further test the relationship between the two aspects, methylation and gene expression, we selected one of the photosynthesis-associated genes, LHCA1, and the key nitrate signaling genes, TGA1/TGA4, for further analysis. LHCA1 encodes a component of the lightharvesting complex associated with photosystem I. LHCA1 interacts with WHIRLY1 to change the photochemical activity of photosystem I and participate in the light adaptation of Arabidopsis 27 . In accord with the F 1 generation result, the hypermethylation level of the LHCA1 gene body in the F 2 generation was verified by BSP (bisulfite-sequencing PCR) analysis, and this was associated with the corresponding expression level of LHCA1 being downregulated in the F 1 -F 3 generations of spaceflight plants, relative to the levels in the corresponding ground control populations (Fig. 8a).
In terms of the nitrogen compound metabolic process and the nitrate signaling pathway, NIN-like transcription factors and TGA1/TGA4 transcription factors play a central role in nitrate signaling [28][29][30][31][32] . Meanwhile, the spaceflight-associated hypomethylation of the nitrate signaling genes TGA1 and TGA4 were identified in the gene body region of the CpG context. We inferred that the increased gene expression level caused by hypomethylation may contribute to the increased nitrogen absorption in the F 1 generation of the spaceflight offspring (Fig. 8b, c). Furthermore, the hypomethylation levels of the inherited conserved methylation sites in TGA1 and TGA4 genes were also identified in the F 2 generation but not in the F 3 generation (Fig. 8b, c). To test whether nitrate assimilation was changed, we checked the Nitrate content and NR activity in the F 1 generation of the spaceflight offspring. We found that spaceflight environment increased Nitrate content and NR activity in the F 1 generation (Supplementary Fig. S7). Combined with the observation of the growth and development phenotypes (Figs. 2,  3), we inferred that the gene expression changes caused by spaceflight-induced inherited modification of methylation sites could, at least partly, explain the increased growth phenotypes in the spaceflight plant offspring.
Here, we also selected the DMR-DEG genes of the F 1 generation and showed the distribution of differential methylation patterns by using the IGV browser graphics (Fig. 8). Genes related to the ABA-activated signaling pathway, kinase signaling, and hormone signaling were represented in the DMR-DEG sets. Spaceflight-associated hypomethylation in the gene body region in the CpG context of the ABA-activated pathway genes CPK26 and KIN10 occurred in the F 1 generation. The ABA-activated signaling gene CPK26 is a member of the calcium-dependent protein kinases (CDPKs), and the Arabidopsis CDPKs are well known for their roles in plant growth regulation and abiotic stress responses 33 . Another gene, KIN10, encodes a sucrose nonfermenting 1 (SNF1)-related protein kinase, which is a central integrator of transcription networks in plant stress and energy signaling 34,35 . The methylation levels of the conserved inherited methylation sites of CPK26 and KIN10 were detected in the F 2 and F 3 generations of the spaceflight plants, but no differences in methylation levels were detected in the selected regions ( Supplementary Figs. S3, S4). Altogether, the lower gene expression and the increased ABA sensitivity of CPK26 and KIN10 in the F 2 generation of plants exposed to spaceflight, compared with the ground control plants, may not be due to DNA methylation changes. With respect to the JA-signaling pathway gene, JAR1 36 , the JAR1 gene body was found to be hypomethylated in the spaceflight offspring plants of the F 1 generation, though this hypomethylation level was not detected in the F 2 generation (Supplementary Fig. S5). Previous research had demonstrated that the NAC domain transcription factor (TF) VND-INTERACTING1 gene (VNI1) was a key regulator of xylem vessel differentiation. The VNI1 gene body was found to be hypomethylated in the spaceflight plant offspring in the F 1 generation, although this hypomethylation level was not detected in the F 2 generation (Supplementary Fig. S6).   LHCA1, b TGA4, and c TGA1 in the offspring of spaceflight and ground control plants. Visualization of methylation data was performed, using the Integrated Genome Viewer (IGV) (http://software.roadinstitute.org/software/igv). The inherited methylation level of CpG islands was verified by using BSP analysis. The error bar is the standard deviation. An asterisk * indicates a statistically significant difference between the spaceflight and control samples using Student's two-tailed t-test (p < 0.05).

Discussion
As sessile organisms, plants need to constantly adjust their responses to external stimuli in order to adapt to their environment. Due to the fact that seed dispersal is usually quite limited, offspring are likely to be affected by parental growth conditions. Since many environmental pressures last for several generations, other adaptation mechanisms may also exist. Epigenetic modification, which regulates gene expression without altering the DNA sequence, is an attractive alternative mechanism 37 . Epigenetic traits play a key role in controlling gene expression and subsequent responses to the environment 7 . Learning more about plant DNA methylation changes during spaceflight will provide a basic understanding of how plants adapt to stress and, more specifically, to the microgravity environment encountered during spaceflight.
Plants are affected by abiotic and biological environmental factors in many ways. In addition to changes in plant physiology and increased resistance/tolerance responses, genomic dynamics also change. For example, transposable elements are activated by abiotic and biotic stress conditions, mutations are induced by chemical agents, and homologous recombination is enhanced by stresses such as high temperature or UVB 26,38 . However, we know little about the evolutionary impact of these changes; because most of these changes occur in somatic tissues, they cannot be identified in future generations. In plants, germ cells are produced from somatic tissue late in development, so some of the genomic changes acquired during plant growth could be passed on to the next generation. In fact, the frequency of fixed gene mutations increases significantly in the offspring of UVB-or pathogen-exposed plants [38][39][40] . Genomic changes in the offspring of the stress group are expected to return to original levels. Similar to this general prediction, the reduction of methylation in the F 1 generation spaceflight samples (S1), compared to the control samples (G1), was obviously lower than that of the parental generation (S0) 13 (Fig. 4). We show in the current study that the reduction in DNA methylation levels decreased over generations from the original parent plants exposed to the spaceflight environment. The phenotypes observed in successive F 2 and F 3 generations are also consistent with this prediction.
Cross-generational effects in genetically uniform populations of plants and animals are usually associated with changes in genomic DNA methylation. This led us to compare the levels of 5-methylcytosine (5-mC) in genomic DNA isolated from the offspring of stressed S0 and non-stressed control G0 plants. The observation of various growth and development phenotypes and stress sensitivity of the offspring plants, cultured under nonstress conditions, was analyzed. Compared with the offspring of the same-generation control plants, the 5-mC content was slightly downregulated in S1 plants grown from seed from plants exposed to spaceflight stress (Fig. 4b), indicating that DNA methylation tended to decrease when stress was not maintained.
In a more detailed study, we measured the overall genomic methylation by WGBS analysis. The different methylation levels between the F 1 generations of spaceflight and ground samples were compared. Compared with the ground control (G1), the regions with lower methylation levels in S1 were regarded as hypomethylated regions, with these genes being involved in signal transduction, transcription, protein metabolism, and stress response. The general concept of spaceflight-induced DNA methylation changes supports the overall conclusions obtained in an earlier study from ecotype Col-0 seedlings growing on spaceflight 13 ; however, previous research by Zhou et al. (2019) did not observe wide-scale hypomethylation during spaceflight, in a study with the Arabidopsis ecotype Ws-0, that was germinated and lived on the International Space Station, ISS 14 . We speculate that the inconsistent results were mainly due to the ISS experiment and spaceflight experiment itself or the use of different Arabidopsis ecotypes. Recent transcriptome sequencing studies have shown that gene expression profiles and alternative gene splicing patterns in Arabidopsis ecotypes Col-0 and Ws-0 differed significantly in their adaptation to microgravity 41,42 , and these results suggest that the two ecotypes may have different genomic methylation patterns in response to the microgravity environment. Other possible reasons, such as the growth hardware, and the age of the seedlings on the spaceflight, may also play a role. This research experiment also highlights the importance of natural variation impacts when designing and interpreting changes related to the space microgravity environment. Of course, we should acknowledge the limitations of spaceflight experiments themselves. We cannot resolve gravity affects without performing a control grown on a centrifuge in space in the same hardware, and seedlings were only exposed to spaceflight for a short period of time, so it was not known how this translates to long-term growth of seedlings.
Similar to our result, previous research reported an increase in the plant root length in spaceflight and simulated microgravity (Random Positioning Machine, RPM) experiments 43 . They also found the red light was able to compensate for deleterious effects of microgravity on the root cell growth 43 . In addition, cell cycle acceleration and changes in essential nuclear functions induced by simulated microgravity in a synchronized Arabidopsis cell culture were also observed; as well as some observations on changes in chromatin-based regulation increased chromatin condensation, and the overall depletion of nuclear transcription 44 . It should be noted that this work shows that microgravity promotes genomic methylation levels, a conclusion that is inconsistent with our results. One reason, we speculate, is that spaceflight experiments and simulated microgravity experiments are different in themselves, or that the cultured cells and plant seedlings do not respond exactly the same way to microgravity stimulation. Another study showed that gravitational and magnetic field variations synergized to cause subtle variations in the global transcriptional state of Arabidopsis in vitro callus cultures 45 . These results expand our knowledge of how culture plant cells and plant seedlings are affected by microgravity.
Interestingly, we observed that root length, flowering time, and silique length were significantly increased in the spaceflight F 1 generation compared with the ground control F 1 offspring, but that only the root length phenotype was preserved in the F 2 generation (Figs. 2, 3). Here, we also detected transcriptome changes in the F 1 generation plants, using RNA-Seq analysis, and we identified many gene expression changes between the S1 generation of spaceflight-stressed plants and the G1 ground control plants. Three independent repeat sequences were used to compare the transcriptome of S1 and G1. Expression of 206 genes was upregulated at the RNA level and that of 188 genes was downregulated, when using 2-fold changes and p < 0.05 differences in expression as threshold criteria. This DEG group contained many genes involved in abiotic stress response and signaling. We identified spaceflight-associated hypomethylation of nitrate signaling genes TGA1 and TGA4 in the gene body region of the CpG context, which may contribute to the increased nitrogen absorption efficiency observed in the F 1 generation of spaceflight offspring (Fig. 8b, c). Furthermore, the hypomethylation level of TGA1 and TGA4 inherited conserved methylation sites were also identified in the F 2 (S2) generation but not in the F 3 (S3) generation (Fig. 8b, c). We inferred that the gene expression changes caused by microgravity-induced inherited modification of methylation sites could explain, at least partly, the increased growth phenotype in the spaceflight plant offspring.
In Arabidopsis, CpG methylation is mainly controlled by the methyltransferase, MET1, and the chromatin remodeling agent, decreased DNA methylation1 (DDM1). DNA methylation is also associated with genomic stability, especially with respect to transposable factors. We revealed here the relationship between DNA methylation and gene expression, which was involved in many types of abiotic stress genes previously associated with spaceflight responses, including stress response, cold response, cell wall remodeling, nitrate metabolism, plant hormone, Ca 2+ signaling, etc. In the offspring, we focused on the validation and analysis of ABA-responsive kinase signaling genes and nitrogen metabolism genes, the results showing that the spaceflightinduced genome methylation alterations, which were passed down to their offspring, contributed to the changes in gene expression level between the S and G samples, which result in the transgenerational adaptation by Arabidopsis to the spaceflight environment. At the same time, the spaceflight-induced DNA methylation difference between S0 and G0 tended to decrease between generations. All in all, our results are helpful in improving the strategies by which plants adapt to stress, and, more importantly, to the outer space environment.

Methods
Plant preparation and growth conditions. The experiment is part of the SJ-10 spaceflight project, which was launched from the Jiuquan satellite launch center in April 2016. Sample preparation prior to launch were consistent with launch conditions described in the previous article 13 . The scientific equipment used in our spaceflight study was obtained from Shanghai Institution of Technical Physics. The specifications of the equipment box, plant growth parameters, and SJ-10 satellite flight parameters are described in the previous reports 13,46,47 . According to our experimental design, in one part of the experiment, the seedlings continued to grow for 60 h under microgravity conditions. In another part, the seedlings grew in microgravity for 11 days in culture unit 2 and were still growing when they returned to Earth. Under spaceflight and ground conditions, all seedlings were grown in a growth chamber under 8000 lx, 22 ± 2°C and a 16-h photoperiod. The g-profile during launch was as follows: the first-level maximum static overload was 4.8 g in flight for 150 s; the second-level maximum static overload was 6.0 g in flight for 180 s 13 . Micro vibration level is superior to 1.5 mg for long-term, 7 mg for transient peak. Quasi steady state microgravity influenced by attitude adjustment thruster: 10 −6 -10 −5 g 48,49 . The ground controls were cultivated in a growth chamber, and grown for 11 days on Earth under the similar conditions as the spaceflight samples. It should be noted that the plant experimental unit is airtight, sealed on the ground and sent to the space. So the gas composition in the plant culture unit was the same as on-ground control.
Flowering time analysis. Flowering time was measured by observing the time from germination to when the first flower bud on each A. thaliana plant opened, and counting the number of rosette leaves.
ABA sensitivity assay. ABA treatment was performed at the germination stage by using the 1/2 Murashige & Skoog (MS) medium containing 0, 0.5, or 2 μM ABA. The plants were photographed after treatment for the times indicated.
Root gravitropism assay. Four-day-old seedlings grown on a Petri dish were incubated horizontally for a period of time to calculate the angle of root curvature, using the ImageJ software (https://imagej.net).
Total RNA and DNA extraction. Arabidopsis seedlings grow on plates. Each plate represents a biological replicate of 11-day-old seedlings, consisting of 15-18 individual seedlings for genomic DNA extraction and the remainder (9-12 seedlings) for RNA extraction. RNA was extracted using the RNeasy Plant Mini Kit (Qiagen, Hilden, Germany), and the quality and quantity of RNA samples were evaluated using an Agilent 2100 Bioanalyzer (Agilent Technologies, Inc.). Total DNA was extracted using the methodology described in previous reports 13 . Three biological replicates were used from the spaceflight and ground control samples for each of RNA-Seq and WGBS analysis.
RNA sequencing (RNA-Seq) and whole-genome bisulfite DNA sequencing. RNA was extracted and measured according to the manufacturer's guidelines. The RNA-Seq library and genome-wide bisulfite sequencing was carried out as described in the previous reports 13,14 . An Illumina sequencing library was constructed using NEBNext Ultra DNA Illumina Prep Kit (New England Biolabs, Ipswich, MA, USA), in combination with Illumina-specific methylation and barcode adapters. The EZ DNA Methylation™ Kit (Zymo Research, Irvine, CA, USA) was used to convert the library to bisulfite, according to the manufacturer's protocols. The nine barcoded bisulfite-Seq libraries and nine barcoded RNA-Seq libraries (with three biological replicates for each sample) were sequenced on a HiSeq3000 instrument. The primers used are shown in Supplementary Tables. S1, S2.
Read alignment and processing. FastQC was used to evaluate the quality of RNA-Seq sequence data. Low-quality sequences, such as adapters, were removed. Gene expression was performed by using the RSEM software (http://deweylab. github.io/RSEM/). The expected read counts per kilobase of transcripts per million map reads and fragments were extracted for further analysis. The threshold value of significantly differentially expressed genes was set at a false discovery rate (FDR) of 0.05, and a fold change was greater than 2. The input from Bisulfite-Seq was fine-tuned and quality control was carried out. Reads were compared with the reference genome of A. thaliana (Col-0) using BSmap 50 . The differential methylation region (DMR) was identified using the previously described method 13 . DMR was classified according to the characteristics of the gene body, promoter and downstream region. The ratio of cytosine transformed in CG, CHG, and CHH environments was calculated using reads labeled with the reference genome, indicating the conversion efficiency of bisulfite.
PCR program for bisulfite sequencing. Samples were pretreated with the QIA-GEN® EpiTect Bisulfite Kit, and the Genemark® Plant Genomic DNA Purification Kit was used to extract genomic DNA. The PCR program settings for cloning fragments were as described in the previous report 13 .
Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway enrichment analysis. KEGG (http://www.genome.jp/kegg/) analysis was used to determine the pathway enrichment of functional DMRs or differential expression genes with p-values less than 0.05.

Analysis of transposable elements (TEs).
A repeat masking program (http:// www.repeatmasker.org/) was used to separate and analyze TEs. We used Fisher's exact tests to study TEs for changes in methylation levels. The FDR method was used multiple times to test adjustments. A p-value less than 0.05 was defined as a methylated TEs with significant changes.