Rice agronomic traits and variability induced by mutagenesis

Alejandro Hernandez Soto (  alhernandez@itcr.ac.cr ) Instituto Tecnologico de Costa Rica: Tecnologico de Costa Rica https://orcid.org/0000-0001-9435-5117 Fabián Echeverría-Beirute Instituto Tecnologico de Costa Rica: Tecnologico de Costa Rica Ana Abdelnour-Esquivel Instituto Tecnologico de Costa Rica: Tecnologico de Costa Rica Andres Gatica-Arias Universidad de Costa Rica Escuela de Biología: Universidad de Costa Rica Escuela de Biologia Marta Valdez-Melara Universidad de Costa Rica Escuela de Biología: Universidad de Costa Rica Escuela de Biologia


Introduction
Induced mutagenesis is a valuable tool to support functional genomics studies and the development of new genotypes. Rice serves as an outstanding model not only because of its impact on the worldwide food supply chain but also because of the availability of technological resources to utilize. Rice was the rst crop sequenced in 2004 (Matsumoto et al., 2005), biotechnological techniques are available, and the genomic information is available to search for speci c target mutations, such as from the Rice Genome Annotation Project and Oryza Genome which can contribute to the precise engineering of the crop (Kawahara et al., 2013;Tanaka et al., 2020;Kajiya-Kanegae et al., 2021). Biological, chemical, and physical agents can induce mutagenesis, such as radiation, rst used on vegetables in 1928; ethyl methanesulfonate (EMS), which produces 2-10 mutations per Mb; and speci c mutations constructed with new breeding and genetic engineering techniques (Soriano, 1961;Serrat et al., 2014;Romero and Gatica-Arias, 2019;Viana et al., 2019;Yang et al., 2019a). In this review, we The mutation breeding principle is to generate heritable changes in the DNA by external agents. The changes result by exposing plant cells to physical (UV, Xray, gamma radiation) or chemical (sodium azide and ethyl methanesulfonate) agents (Mba et al., 2010). Induced mutagenesis offers a promising alternative for developing rice varieties resistant to biotic and abiotic stresses since it could accelerate the spontaneous mutation process and increase the pool of genes available for genetic improvement (Gressel and Levy, 2006;Oladosu et al., 2016;Viana et al., 2019).
Tissue culture Totipotency, a distinguishable characteristic of plant cells, allows each cell to regenerate an entire plant in principle. This process involves the culture on special growth media of tissue fragments or individual cells from a plant enabling the cells to grow and further division (Fehér, 2019). In this sense, tissue culture approaches are helpful to develop biotic or abiotic stress-tolerant plants. Among the techniques available, somaclonal variation enables changes in the DNA causing genetic and phenotypic variation among clonally propagated plants. The somaclonal variants obtained could be detected using in vitro selection by applying selective pressure in culture conditions (Larkin and Scowcroft, 1981;Bairu et al., 2011).

New breeding techniques CRISPR/Cas9
The clustered regularly interspaced short palindromic repeats (CRISPR)-associated endonuclease Cas9 (CRISPR/Cas9) system targets a speci c genomic sequence using an engineered 20 base pair (bp) RNA guide sequence that binds to its DNA and the Cas9, from S. pyogenes, recognizes the PAM sequence 5¢-NGG-3¢ generating double-stranded breaks in speci c genes at desired locations in the genome. This genome editing method allows the insertion, deletion, or modi cation of DNA with increased speci city and e ciency (Romero and Gatica-Arias, 2019).

CRISPR/Cpf1 system
The nuclease Cas12a requires a small crRNA for inducing double strand breaks with e ciencies similar to those of CRISPR/Cas9. Moreover, this nuclease uses a 22 nt spacer for its maximum e ciency and speci city and identi es a T-rich PAMs located upstream of the guide and generated staggered ends (Schindele et al., 2018).

Base editing
This system allows the conversion of nucleotides without inducing double-stranded DNA breaks or using donor templates. In this sense, it has been used for changing a C-G base pair into T-A, or A-T into G-C (Marx, 2018).

Prime editing
This system uses a catalytically impaired Cas9 endonuclease fused to a reverse transcriptase enzyme, and a prime editing guide RNA (pegRNA). This complex is capable of identifying the target site and replace the target DNA nucleotides without double-stranded DNA breaks or using donor templates (Anzalone et al., 2019;Lin et al., 2020).

Domestication genes
The Oryza genus is composed of species with a variety of genome structures, including six diploids (n = 12; named AA, BB, CC, ee, ff, gg) and ve polyploids (n = 24, named BBCC, CCDD, HHJJ, HHKK, and KKLL) (Kim et al., 2015;Nadir et al., 2017;Wing et al., 2018;Chen et al., 2019). Only two diploid (2n=24) species of rice have been domesticated and used for cultivation: Oryza sativa and African O. glaberrima. Rice domestication emerged because of the selection of speci c genes and the loss of function of speci c genes. Wild relatives have functional versions of genes such as sh4, waxy, BH4, qSH1, AN1, brown pericarp, PROG1, and OsG1, as described in the following text. The sh4 gene is related to reduced seed shattering (Os04g0670900). The waxy gene controls the amylose content (Os06g0133000). BH4 is related to the hull color of the seeds (Os04g0460200). The gene qSH1 is involved in seed shattering (Os01g0848400). The AN1 gene is related to seeds, morphology, and grain shape (Os04g0350700). RC Brown pericarp is involved in the seed coat (Os07g0211500). PROG1 is related to an erect plant structure (Os07g0153600). OsLG1 is related to a closed-panicle structure (Os04g0656500) . The importance of such genes is critical in understanding how de novo domestication and their further use in plant breeding can be achieved from wild Oryza varieties.
Osmoprotection by accumulating molecules such as trehalose is a possible pathway involved in salt tolerance, as proven currently in Arabidopsis (Li et al., 2011b;Nuñez-Muñoz et al., 2021). Other individual genes could be of interest, such as the Na+ transporter SKC1 (Os01g0307500) with a V395 that provides salt tolerance (Jayabalan et al., 2019). Knocking out an independent but closely related gene, OsEPFL9 (Os01g0824500), results in increased water use e ciency under stress because of the reduced stomatal count (Yin et al., 2017(Yin et al., , 2019. Other stress tolerance pathways have been shown to be effective. Low cadmium accumulation occurs after knocking out the metal transporter genes OsNramp5 (Os07g0257200) and OsNramp1 (Os07g0258400); the plants are able to resist heat stress when the gene OsNTL3 (Os01g0261200) is working correctly cold tolerance results from knockout of the OsMYB30 (Os02g0624300) gene and more cuticle wax is deposited when the gene DHS (Os02g0682300) is knocked out (Sasaki et al., 2012;Tang et al., 2017;Wang et al., 2018a;Chang et al., 2020;Liu et al., 2020a;Zeng et al., 2020).

Herbicide resistance monogenic traits
Rice is usually cultivated under two agronomical systems: paddy transplanted rice (PTR) and dry seeded rice (DSR). The rst is the conventional method, which requires water ooding and represents a sustainability issue because of water scarcity, methane production and the consumption of nonrenewable energy (Wang et al., 2017). DSR, on the other hand, represents opportunities for e cient water and nitrogen use, and a reduction of both greenhouse gas emissions and labor demand, especially in countries such as China, where 90% of rice is currently produced under PTR (Shekhawat et al., 2020). However, weed management is a challenge in DRS, speci cally during the rst 41 days after sowing (DAS). Another complication is weedy rice (O. sativa f. spontanea), which can result in yield losses of up to 50% (Nadir et al., 2017). Weedy rice usually involves increased seed longevity, seed shattering and stress tolerance (Durand-Morat et al., 2018). The use of chemical control represents a tool to manage weedy rice, but there are still challenges as described below.
The Herbicide Resistance Action Committee (HRAC) and the Weed Science Society of America (WSSA) classify herbicides into 34 groups and one unknown group based on their "mode of action" (MoA) at the biochemical level (Forouzesh et al., 2015;Dayan et al., 2019;Gaines et al., 2020). The discovery of a new mode of action has been rare in the last 30 years. A good example is leptospermone, and its analog inhibitors act as hydroxyphenylpyruvate inhibitors of dioxygenase (HPPD) (Dayan and Duke, 2020). Different modes of herbicide action, such as rotations, delay the emergence of herbicide-resistant weeds. However, weeds are evolving to resist multiple MoA types of herbicides. For example, Chloris radiata is found in Colombian rice elds with dual resistance to glyphosate (mode of action 9) and the acetolactate synthase (ALS) inhibitor imazomox (mode of action 2) (Hoyos et al., 2021). Weedy rice infestation in the USA resulted in 5.7 million tons lost and $457 million in environmental costs between 2002-2014 (Bzour et al., 2018). Mutations to provide herbicide tolerance were introduced into rice 20 years ago based on the Acetohydroxy acid synthase AHAS/ALS (Os02g0510200) gene mutation, providing tolerance to the mode of action 2 (Li et al., 2019). Rice herbicide tolerance varieties are used in the USA (700,000 Ha), Brazil (600,000 Ha), Uruguay (70,000 Ha), Argentina (32,000 Ha), Malaysia (95,000 Ha), and Italy (60,000 Ha), as well as in many Central America countries, such as Costa Rica, Honduras, Panamá, and the Dominican Republic (Singh et al., 2017). The incorrect use of this variety allowed introgression and outcrossing of the resistance into red rice, which means that weed herbicide control requires stricter farming practices, such as rotation (Liu et al., 2021). Alternatives such as aryloxphenoxy propionate-resistant rice (mode of action 1), which is the result of mutations in the ACCase2 (Os5g0295300) gene, already exist and will allow for herbicide rotation (de Andrade et al., 2018;Camacho et al., 2019).
Herbicide-resistant weeds to the inhibition of photosynthesis at PSII can also provide insights for rice models. The S264G mutation in psbA increases tolerance more than 50-fold in triazine herbicide-tolerant radish (MoA-5). However, it can also compromise tness because of less e cient photosynthesis (Lu et al., 2019). Other mutations, such as Val219Ile, Asn266Thr, Phe255Ile, and Ala251Val, can also provide tolerance (Gaines et al., 2020). It is important to note that the psbA mutation Val-219-Ile provides tolerance to the amide propanil MoA-5 on Cyperus difformis (Pedroso et al., 2016). Propanil is widely used in rice cultivation because the crop is naturally capable of degrading the molecule by a putative enzyme located in the mitochondria, and an additional resistance pathway could increase its e ciency (Matsunaka, 1967;Chen and Matsunaka, 1990). The described mutations could also result in herbicide tolerance in rice when targeting the homologous gene AAS46167, encoding protein P0C434, to address an additional MoA.
Rice is also known to be resistant to Bentazon (MoA-6), as it is degraded by cytochrome P450 CYP81A6 (Pan et al., 2006). Additionally, the P450 gene CYP72A31 is responsible for conferring tolerance to bispyribac sodium (BS) in Oryza sativa indica, while its absence in japonica rice varieties results in BSsensitive varieties (Zhang et al., 2002;Saika et al., 2014).

Bacteria, fungi and virus resistance
Rice breeding of pathogen resistance is possible by knockout of the Sweet 14,11,13 genes named Os11g0508600, Os08g0535200, Os12g0476200, respectively, since they act as a point of access for pathogens causing bacterial blight streaks (BLSs), such as Xanthomonas oryzae pv. oryzae (Xoo), and they reduce copper in the xylem Oliva et al., 2019;Varshney et al., 2019). The pathogen emerges by breaking the resistance pressure of varieties planted in approximately 80% of the total crop cultivation area carrying the resistance gene Xa4 on chromosome 11 introduced in the 60s in the variety IR20 (Quibod et al., 2020) Xanthomonas oryzae can also infect wild grasses and become an emergent microorganism that is di cult to control (Lang et al., 2019).
Another outstanding gene to target is the transcription factor IPA1 (Os08g0509600); higher expression levels of IPA1 result in increased yield and immunity when tested against Magnaporthe oryzae. Resistance relies on time-and pathogen-speci c phosphorylated activation of the transcription factor at Ser163 and subsequent WRKY45 promoter-resistant gene triggering within 48 hours after infection, while the yield of the nonphosphorylated protein binds to the DEP1 promoter (Jing Wang et al., 2018). A different way to achieve M. oryzae resistance is by knocking out OsERF922 ethylene response factor 922 (Os01g0752500) (Wang et al., 2016). For details, see Table 3 below.
5. Grain number, quality, weight and plant structure Rice quality traits are essential to achieve a better yield, consumer preference, and growth e ciency. The genes involved in grain number and size, plant density, structure, panicles, and owering are complex because of their interactions. However, new ndings and key mutations provide some insights into their regulatory mechanisms and greater predictability in achieving the desired phenotype, as described next (for details, see Figure 5 and Table 4 below).
Grain size. The GS3 Grain Size3 gene (Os03g0407400) is responsible for negatively controlling the grain length. Its mutation can result in better or worse weight and size that correlates with the composition of its domains: organ size regulation (OSR), a transmembrane necrosis factor receptor/nerve growth factor receptor (TNFR/NGFR), and a von Willebrand factor type C (VWFC) (Takano-Kai et al., 2013;Li et al., 2016;Shen et al., 2017;Yang et al., 2019b;Zeng et al., 2020). The wild type allele contains all of the domains and results in medium grains (Takano-Kai et al., 2013). Loss of function results in long-grain varieties; for example, Minghui 63 has a stop mutation C165A at the second exon, resulting in a loss of function and a long-grain phenotype. In contrast, a mutation or deletion in the fth exon creates a truncated protein with no VWFC domain and a short seed phenotype (Mao et al., 2010;Takano-Kai et al., 2013).
Grain number. Malfunction of the gene Os01g0197700 (GN1a) produces an increment of grain per panicle number and owering because of a lower degradation of cytokines produced by the corresponding cytokinin oxidation enzyme Shen et al., 2017;Huang et al., 2018). Another gene that correlates with increased production and downregulates cytokine level regulation is EP3 Erect Panicle 3 (Os02g0260200) (Li et al., 2011a;Shen et al., 2017).
Structure. Farmers prefer smaller plants with many panicles and fewer tillering traits. Knockout of the DEP1 (Os09g0441900) gene, as well as the loss of function of the HTD1 (Os04g0550600) gene coming from landraces produces short, dense, erect panicles (Zou et al., 2006;Li et al., 2016;Lacchini et al., 2020).
The transcription factor IPA1 Ideal Plant Architecture1 (Os08g0509600) -speci c mutations between bases 854 and 876 can increase the expression of transcription factor proteins because they interrupt OsmiR156 transcript cleavage. For example, C874A in the third exon (leucine to isoleucine) generates a rice plant with a reduced tiller number, increased lodging resistance, and an enhanced grain yield Wang et al., 2018b).
The number of panicles and consequently the yield can be increased by knocking out or indirectly blocking Pin1A and Pin15b. The indirect mechanism results in higher expression of DEP1 and LPA1, which interact to suppress PIN1a expression (Huang et al., 2018;Fu et al., 2019;Miao Liu et al., 2020). LPA1 is also important in the erect phenotype, and its knockout results in lamina inclination, while BAS1 seems to be important in stomata closing Mao et al., 2018).

. Other traits
Other rice traits provide value for breeding and for satisfying consumer preferences, such as nitrogen use, fragrancy, oleic acid content, and color. Regarding nitrogen, there is a better e ciency with a higher expression of the nitrate transporter OsNPF6.1 and the two transcription factors OsNAC42 and OsNLP4 (Tang et al., 2019;Yu et al., 2021b). Knockout of the FAD2 gene results in an oleic acid increment (Tiwari et al., 2016;Abe et al., 2018). Furthermore, a mutation in the Osor (Os02g0651300) gene results in potential orange-colored rice (Endo et al., 2019), and fragrancy can be increased or decreased by modulating the BADH2 gene, which prevents the formation of the aromatic compound 2AP (2-acetyl-1-pyrroline) (Shen et al., 2017). For details, check Table 5 below.

Conclusion
Induced mutations targeting speci c genes associated with known phenotypes, as described in this review, will allow for advances in more precise rice breeding to improve the varieties that farmers are already using. It can also result in new varieties and de novo domestication from wild relatives and extrapolate the results to other crops with homologous traits. Farmers urgently requires advances in this knowledge to respond to the challenges of climate change, consumer demands, water scarcity, nitrogen usage, and sustainable production.  T102I + P106S CRISPR In vitro resistance 1 mg l-1 glyphosate, 400x dilution Greenhouse. (Li et al., 2016b) (S)= Susceptible, genomic (*) Additional information at The Rice Annotation Project (RAP). Oliva et al., 2019;Varshney et al., 2019).
Encodes a protein with a UBA domain. Homozygous null mutant is embryonic lethal. (Li et al., 2012) heading and grain weight, heading dateand grain weightrelated protein The gene is upregulated by nitrogen starvation. OsNLP4 binds to the NRE motif and promotes the expression of OsNiR that encodes a critical nitrite nitrogen assimilation. Representation of salt tolerance traits mediated by three different methods: 1) overexpression, 2) knockout of speci c genes, and 3) particular sodium channels. Note that the rst corresponds to transcription factors that trigger adaptive responses labeled MSL37, NAC2, NAP, and P5CS. The second is a knockout of those that result in salt sensitivity: OsRR2, STL1, DST; and the sodium channel SKC1 in rice. The third is the sodium channel SKC1 containing amino acid V395. Created with BioRender.com.