Elevated CO2 concentration enhance Oryza sativa seed shattering and affects seed-shattering gene expression

Weedy rice (Oryza sativa L.) is widely recognized as a major constraint in cultivated rice systems globally. Seed shattering is related to the invasiveness and persistence of weedy ecotypes in eld and exacerbates its control in cultivated rice systems. Shattering traits are controlled genetically and by different environmental conditions. At present, a rapidly changing environment, including rising levels of carbon dioxide [CO2], could alter shattering frequency, with subsequent effects on weed seed input and competition. The objective of the current study was to evaluate the interaction between weedy rice seed shattering and the transcriptional seed shattering-regulation genes as affected by weedy rice genotypes and atmospheric CO2 concentrations. We examined seven biotypes and two atmospheric CO2 concentrations: ambient {a[CO2]} and enhanced {e[CO2]} concentration, 400 and 700 µmol mol-1 respectively. Our results indicate that e[CO2] increases weedy rice seed shattering. The gene expression analysis demonstrates an effect of [CO2] in the expression of all gene shattering-related genes (OsCPL1, qSH1, Sh4, SHAT1, OsXTH8, OSH15, and SH5), with high variability observed between genotypes. Here we showed that increased CO2 concentration affects greatly seed shattering in weedy rice and in minor effect cultivated rice, by modulation of seed shattering-related genes and weedy genotypes showed the highest upregulation level of this genes. Thus, increased CO2 concentration positively affect panicle number and grain yield mainly in cultivated rice.


Background
Among pernicious weeds in cultivated rice, weedy or red rice (Oryza sativa L.) is recognized as among the most competitive and signi cantly reduce rice yields even at small densities (Smith, 1988). Traits that can enhance weedy rice persistence and competitiveness include seed shattering and dormancy (Burgos Seed shattering has been widely studied to understand rice domestication (Cheng et al., 2016;Dong and Wang, 2015;Li et al., 2006;Thurber et al., 2010;Zhang et al., 2017). In rice, seed shattering is dependent on the proper formation and subsequent degradation of an abscission layer in the joint between lemma and pedicel (Dong and Wang, 2015). This layer's degradation begins in the grain ripening process, caused by ethylene's production, which inhibits auxin synthesis. Synthesis of enzymes that degrade and proteins that remodel the cell wall occurs, including β-1,4-glucanase, polygalacturonase, xyloglucanendotransglycosylase/hydrolase, and expansin (Taiz et al., 2017). As a consequence of hydrolytic enzymes' action, the middle lamella and the cell wall are degraded, causing the grain's fall (Roberts et al., 2002).
A number of genetic factors including several quantitative trait locus (QTLs) are associated with shattering in rice (Balanzà et al., 2016;Zhou et al., 2012). The gene qSH1 have been described in Oryza japonica subspecies (Konishi et al., 2006) and SH4 in the Oryza indica subspecies (Li et al., 2006); these genes together are responsible for almost 70% of the shattering in rice. The SH4 encodes a transcription factor with a Myb3 DNA binding domain and promotes abscission layer cells' hydrolyzing during the abscission process (Li et al., 2006;Lin et al., 2007). The SHAT1 gene encodes a transcription factor with an APETALA2 domain and acts on abscission layer cell differentiation (Zhou et al., 2012). There are also genes which can inhibit shattering. OsCPL1, which encodes a carboxy-terminal domain phosphatase-like protein, represses abscission layer development (Ji et al., 2010).
Although there are a number of descriptions regarding genetic regulation of abscission and seed shattering (Maity et al., 2021), there is little information about the effect of environmental factors in regulating these genes and the consequences for abscission and seed shattering (Patharkar and Walker, 2019). Among these factors it is worth investigating the role of rising CO 2 concentration [CO 2 ], which has increased by almost 30% since the 1960s and is expected to increase another 50% by century's end. It is widely recognized that CO 2 plays an essential role in plant morphology, fecundity, and development.
Recent and projected increases in [CO 2 ] can affect secondary metabolic processes, acting on different routes and interfering with metabolism at the cellular level (Kimball, 2016;Mhamdi and Noctor, 2016;Xu et al., 2015). During the abscission of tomato leaves, the transcriptional activation of multiple genes related to ethylene synthesis and heat shock proteins was induced by elevated CO 2 (800 µmol mol -1 ) (Pan et al., 2019). Previous studies demonstrate that in addition to the vital role as a substrate in photosynthesis, CO 2 plays an essential role in cell homeostasis and hormonal signaling (Shi et al., 2015). Druart et al., (2006) reported that with the increase in the environment [CO 2 ], there is an effect in the repression of genes related to cell wall formation and cell growth. As far as we know, this is the rst study to evaluate rice shattering in different [CO 2 ] and demonstrate the in uence of [CO 2 ] on breaking tensile strength (BTS).
The current study was undertaken to evaluate the effect of projected, future CO 2 concentration on seed shattering and in the transcriptional regulation of seven genes known to be related to seed shattering in rice, cell wall synthesis, tissue degradation of the abscission layer, and genes that in uence lignin biosynthesis. The study was conducted for both cultivated and weedy rice types to assess if differences to CO 2 were evident and if a CO 2 by genotype interaction was observed.

Gene expression
Taking into account the effect of the different [CO 2 ] on BTS value, we also tested the effect of different [CO 2 ] in the transcriptional regulation of seed shattering-related genes ( Table 2). The target genes are associated with the different abscission layer formation phases, from cell differentiation to lignin deposition regulation and abscission layer cell separation.
The genes qSH1, OSH15, SH4 and SHAT1 are involved with abscission layer formation therefore, displaying a role in seed shattering (Konishi et al., 2006). At a[CO 2 ] condition ( Fig. 4), qSH1, OSH15 and SHAT1 were upregulated in two of the three weedy rice biotypes (AVAR and AV53) which also showed high seed shattering compared to cultivated rice genotypes. Despite AV60 weedy rice also shown high seed shattering at a[CO 2 ] condition, qSH1, OSH15, SH4 and SHAT1 were downregulated. Batatais also showed a similar expression pattern observed in AVAR and AV53 respect to the upregulation of qSH1, OSH15 and SHAT1. However, Batatais showed the lowest seed shattering level among cultivated rice at a shattering capacity for rice cultivars, while for weedy rice biotypes, a high level of shattering was reported.
The low shattering observed for rice cultivars is related to the crop's domestication, which is linked to reducing the natural shattering of seeds (Li et al., 2011).
The mechanistic basis for the increase in shattering at the higher CO 2 concentration is not clear. The increase in [CO 2 ] has been reported to in uence several biochemical and physiological plant processes, affecting secondary metabolic processes, acting in different pathways, and interfering with cellular metabolism (Kimball, 2016;Noctor and Mhamdi, 2017;Xu et al., 2015). The shattering process involves the degradation of a speci c zone between the pedicel and the grain, coordinated by methods involving the action on hydrolytic enzymes, where the middle lamella and the cell wall are degraded (Roberts et al., 2002). Also, a hormonal balance between ethylene and auxin in uences the beginning of the shattering process, and environmental factors can affect this balance. Another critical point is the wide genetic and phenotypic diversity present in weedy rice biotypes (Singh, 2013). All these factors indicate that many processes can in uence shattering characteristics and that environmental by [CO 2 ] interactions will require additional study.

Page 6/19
The positive effect of high CO 2 in increasing cultivated rice yield while decreased in weedy rice Different pro les at e[CO 2 ] was observed between cultivated rice yield and weedy rice. Variability in panicle number and grain yield was observed for most of the weedy rice biotypes tested. Weedy rice biotypes are widely described to present high variability in several traits. For rice cultivars, the results were different, most responded with a greater panicles number and grain yield in e[ ]. In the present study, cultivars were more responsive to [CO 2 ] than weedy rice biotypes. Weedy rice results were also found in evaluating rice genotypes where the seed mass similar and the above-ground mass was increased in the CO 2 (Wang et al., 2020). In the present study, weedy rice biotypes tended to produce more biomass but were ine cient in the source-sink for grain production.
Increased CO 2 concentration affects seed shattering by altering gene expression pattern The molecular control of seed shattering has been unveiled and genes involved in abscission layer cell differentiation, lignin deposition and abscission layer cell separation were detected in rice. qSH1 is the major QTL for seed shattering in japonica rice and encodes for a BELL1-type homeobox protein and is involved with abscission layer cell differentiation (Konishi et al., 2006). qSH1 and OSH15, a KNOX protein, form a dimer that promote abscission layer cell differentiation (Yoon et al., 2017). The dimer qSH1 and OSH15 transcriptionally regulates SHAT1 (Zhou et al., 2012). SHAT1 encodes a transcription factor with an APETALA2 domain and regulate downstream genes involved in cell abscission layer differentiation. SHAT1 can also be activated by SH4 to activates qSH1 to maintain SH4 expression (Zhou et al., 2012). This ne-tuned network promotes abscission layer cell differentiation, the rst phase of seed shattering. Here, we show that the transcriptional regulation of these genes is affected by increasing in CO 2 concentration. We also show that in e[CO 2 ] condition all analyzed genotypes, weedy and cultivated rice, increase shattering level.
Speci cally related to qSH1 transcriptional regulation, no expression pattern was detected between weedy and cultivated rice (Fig. 4A). A previous study also evaluated the Batatais cultivar, AV60, and other weedy rice genotypes and found that the qSH1 gene was not differentially expressed among the genotypes (Nunes et al., 2014) and that the SNP associated with seed shattering was not present in the weedy rice (Nunes et al., 2015). Here, the qSH1 was expressed mainly in weedy rice biotypes in a[ CO 2 ] condition and all evaluated genotypes in e[CO 2 ] condition, including those with a low shattering level.
Thus, it seems that qSH1 is not a suitable marker gene for seed shattering in e[CO 2 ] condition. The high qSH1 expression in genotypes showing low shattering does not agree with the phenotype and may be an effect of other genes acting repressing the shattering (Htun et al., 2014).
As expected, OSH15 show a similar pro le as observed for qSH1 in a[CO 2 ] and e[CO 2 ] conditions (Fig. 4B).
However, it is interesting to note the higher transcript amount of OSH15 related to qSH1 in e[ CO 2 ] condition where all genotypes also showed increased shattering level. It can be explained by the fact that OSH15 displays two different roles in regulating seed shattering, in forming a dimer with qSH1 to promote abscission layer cell differentiation and also in forming a dimer with SH5, a BELL homeobox protein acting to inhibit lignin biosynthesis in abscission layer (Yoon et al., 2017). The interaction between OSH15 and SH5 may explain the upregulation of OSH15 in those genotypes showing low shattering degree (Batatais) (Fig. 1). SHAT1 is affected by e[CO 2 ] rice and weedy rice genotypes and its expression pattern at a[CO 2 ] condition in most of weedy rice genotypes highlight its role in seed shattering (Fig. 4D). The upregulation of SHAT1 in cultivated rice genotypes, mainly in Batatais, may be due the fact that SHAT1 can display other roles independent of SH4 as previously reported by Zhou  Homologous to qSH1, the SH5 gene encodes a BEL1-type homeobox protein that inhibit lignin biosynthesis at abscission layer and therefore promotes seed shattering (Yoon et al., 2014). Interestingly, SH5 is downregulated in almost all genotypes analyzed here at a[CO 2 ] condition, while SH5 is upregulated in all analyzed genotypes under e[CO 2 ] condition (Fig. 4F). Therefore, the transcriptional regulation of SH5 is sensitive to [CO 2 ] changes, with a vital effect on seed shattering in a future environment with increased [CO 2 ]. That effect was highlighted in Batatais genotype that have a high shattering level in e[CO 2 ] (Fig. 1).
In addition, the SH5 sensitivity to increased [CO 2 ] is also harmful since SH5 can also induce the expression of SHAT1 and Sh4, two important players for the proper abscission layer formation (Yoon et al., 2014). Also, considering that the abscission layer is a specialized cell layer located in the rachilla and is composed of small cells with thin nonligni ed walls surrounded by larger ligni ed cells (Yu et al., 2020), the expression of SH5 and OSH15 can be acting directly in inhibiting the deposition of lignin, increasing shattering. Moreover, these genes effects can stand out from the others, resulting in suffer repression due to the expression of others genes, since shattering has a complex and polygenic character (Yoon et al., 2017).
Finally, abscission layer cell separation is the last step for seed abscission. Cell wall degrading enzymes can be involved in this process. OsXTH8 is a cell wall remodeling enzyme that catalyzes cleavage of xyloglucan polymers (Jan et al., 2004). OsXTH8 was reported to be highly expressed in abscission layer, and it is proposed to facilitates the separation of the grain (Nunes et al., 2014). As observed for SH5, OsXTH8 expression was also highly sensitive to [CO 2 ] changes (Fig. 4G). The high expression of OsXTH8 in Batatais cultivar, which has low shattering, suggests that OsXTH8 maybe not the main factor acting in seed shattering and/or some post-transcriptional or post-translational mechanism may be acting avoiding OsXTH8 effect.

Seed shattering in a future atmosphere CO 2 concentration
The exact and detailed mechanism of how abscission layer formation and degradation occur is not fully understood. The genes identi ed in these processes may not act alone since other processes, such as synthesizing new cells and enzymes, can have a fundamental role. Also, a hormonal gradient between ethylene and auxin interconnects the process, acting precisely in the synthesis and secretion of enzymes that act on cell wall degradation and proteins that remodel the cell wall (Taiz et al., 2017). However, studies that underlie phytohormones' involvement in the speci cation of the abscission layer suggest that a precise balance between their biosynthesis and responses is of fundamental importance (Dong and Wang, 2015). Still, there is no complete elucidation of the molecular mechanisms and interactions of plant hormones underlying the abscission layer differentiation.
The plant's machinery works powered by the photosynthetic process, from where it obtains energy toward the formation of fundamental organic compounds toward the plant. When the plant is exposed to e[  (Ainsworth et al., 2006). In the latter, transcripts for ribosomal proteins, cell cycle, and cell wall loosening, necessary in cytoplasmic growth and cell proliferation, were highly expressed in growing soybean leaves grown in e[CO 2 ]. In rice plants, gene expression obtained from plants grown in e[CO 2 ] revealed to many gene expression, including senescence-associated protein 5, a gene associated with leaf senescence that triggers cell growth and structure (Fukayama et al., 2009). Thus, the expression of genes from the senescence-associated family can in uence the shattering process in rice. Still, in a study developed with tomato leaves (Pan et al., 2019) the transcription of multiple genes related to ethylene synthesis and heat shock proteins was induced by elevated CO 2 (800 µmol mol -1 ), acting in abscission in leaves. Therefore, there could be an effect of CO 2 on the hormonal balance in the abscission layer, affecting rice shattering.
Here, we identi ed that seed shattering-related genes are sensitive to changes in [ condition, whereas all analyzed genotypes showed increased shattering. Altogether, we demonstrate the harmful shattering pro le, mainly related to weedy rice, in a future increased [CO 2 ] environment (Fig. 5).
Atmospheric carbon dioxide directly determines the rate of photosynthesis in plants, affecting plant productivity and tness, and can act as a selective pressure, driving evolution. Changes in gene expression, linked to critical adaptive characteristics, represent phenotypic plasticity and offer clues to the selection targets during long-term (multigenerational) adaptation. Also, it is possible that, in addition to gene regulation, the mutation acts to give rise to new locally adapted alleles (Watson-Lazowski et al., 2016). The availability of new molecular tools, particularly high-throughput and inexpensive RNA and DNA sequencing, suggests that further studies can be developed, using previously impossible approaches that combine phenotyping, functional genomics, and population genetic analysis in nonmodel systems.

Conclusion
The present study demonstrated that e[CO 2 ] increases seed shattering in rice genotypes (estimated by the BTS measure), presenting a greater phenotypic effect for weedy rice biotypes. This was the rst study that evaluated the effect of [CO 2 ] on the shattering process. The mechanism associated with elevated CO 2 and shattering is unknown, as such, more approaches are needed to investigate other factors to elucidate this effect.

Plant material and growth conditions
Three weedy rice and four cultivated rice genotypes were analyzed (Table 1), previously phenotyped in other studies (Nunes et al., 2015(Nunes et al., , 2014. The experimental unit consisted of 8-L pots lled with soil (Albaqualf), fertilized at sowing, and it was seeded ten, and, after emergence, plants were trimmed to four plants per pot. The experiment was performed in a growth chamber maintained at 28/25°C day/night temperature. The cultural practices were based on rice production's technical recommendations for the south of Brazil (SOSBAI, 2018). The experimental design was completely randomized with three replications. The treatments included two [CO 2 ] levels: ambient (a[CO 2 ]) at 400 ± 20 µmol mol -1 and elevated (e[CO 2 ]) at 700 ± 20 µmol mol -1 . Table 1 Weedy rice and rice cultivars identi cation, species, and sub-species with origin in this study. Seed shattering phenotyping Seed shattering was evaluated according to (Nunes et al., 2015). Brie y, the quantitative evaluation of the BTS was obtained at seed maturity using force gauge equipment connected by a hook to the seed until the seed's release from the pedicel. The force direction was exerted longitudinally to the petiole and the grain. Four panicles per genotype and 10 grains from the panicle median part were evaluated. At the end of the experiment, the number of panicles, seed yield, and above-ground dry weight (ADW) per pot was determined. Data were analyzed by the two-tailed Student's t-test was used for statistical analyses (** indicates p-value<0.01 and * indicates p-value<0.05).
RNA extraction and cDNA synthesis RNA was extracted from rice and weedy rice genotypes described in Table 1. Four panicles were selected per pot and marked with plastic tags at anthesis. Ten days after anthesis, 30 ower-pedicel junctions were chosen in the middle of the panicle, consisting of approximately 30 mg of plant material and one replicate. This material was immediately placed in liquid nitrogen and kept at -80°C until the RNA extraction occurred. Each collected ower-pedicel structure consisted of a 1 mm region of the pedicel and 1.5 mm of the ower, corresponding to the abscission layer zone (Ji et al., 2006;Li et al., 2006).
The total RNA was extracted using PureLink™ Plant RNA Reagent (Invitrogen) according to the manufacturer's recommendations. RNA quantity and purity were veri ed by spectrophotometry in NanoVue (GE Healthcare) and integrity by agarose gel electrophoresis. The cDNA was synthesized with the reverse transcriptase SuperScript™ First-Strand Synthesis System III (Invitrogen) using oligo(dT). The quality of the cDNA was assessed using an RT-qPCR reaction in the LightCycler® 480 Instrument II thermocycler (Roche) using SYBR Green I (Invitrogen) and oligonucleotides for the reference gene Actin 1 ( Table 2). Test reactions were performed before conversion into cDNA (to con rm digestion) and after conversion to cDNA. In both tests, genomic DNA was used as a positive control since the oligonucleotide for Actin 1 was designed in exon junctions and had a greater amplicon in the presence of an intron serving as good control of the absence of genomic DNA.

Quantitative reverse transcription PCR (RT-qPCR)
The quanti cation of gene expression in RT-qPCR was performed according to the MIQE Guidelines (Bustin et al., 2009) using oligonucleotides for target and reference genes ( The quanti cation of gene expression was calculated using the comparative ΔΔCT method (Livak and Schmittgen, 2001), using a baseline the expression of the low seed shattering genotype Nipponbare in each treatment and normalized for the reference genes OsACT1, OsEF1α, and OsUBQ5.    Effect of CO 2 concentrations (a[CO 2 ] = 400 μmol mol -1 and e[CO 2 ] = 700 μmol mol -1 ) in above-ground dry weight (g pot -1 ) (ADW) of the weedy rice biotypes AVAR, AV53 and AV60 and rice cultivars Batatais, IRGA 417, IRGA 424 RI and Nipponbare. Error bars correspond to standard error. The two-tailed Student's t-test was used for statistical analyses (** indicates p-value<0.01 and * indicates p-value<0.05).