Adenine Base Editor Creates Novel Substitution Mutations in eIF4G Gene of Rice

Two single nucleotide polymorphic mutations and deletion affecting Y 1059 V 1060 V 1061 amino acid residues in a host translation initiation factor four gamma (eIF4G) gene in rice are reported to confer resistance to rice tungrospherical virus in resistant genotypes. A CRISPR-based adenine base editing vector was used to target these residues in a susceptible indica cultivar, ASD16.Agrobacterium-mediated transformation of ASD16 generated 16 missense mutants and two deletion mutants. Substitution mutations occurred at A 5 > G 5 and A 4 > G 4 , where 5.5 % and 3.37 % of adenosines got converted to guanosines, respectively. The mutantsgenerated had missense mutations affecting the YVV residues and the residues immediately adjacent to YVV.Thus,these novel mutationsare promising candidates in imparting resistance against rice tungro disease.


INTRODUCTION
Rice tungro disease (RTD) causes severe yield losses in rice-growing endemic regions of South and Southeast Asia (Herdt, 1988;Azzam and Chancellor, 2002;Muralidharan et al., 2003). Rice tungro spherical virus (RTSV), along with its counterpart RTBV (Rice tungro bacilliform virus, a dsDNA virus), is responsible for RTD (Bunawan et al., 2014). Management of tungro disease via the development of broad-spectrum resistance through suppression of RTSV has been the preferred choice since rice plants infected with RTBV exclusively are incapable of spreading the disease. Thus, RTSV resistant cultivars rather than RTBV could successfully reduce the incidence of tungro disease in the field (Hibino, 1996;Anjaneyulu et al., 1995;Lee et al., 2010). Host-pathogen interaction studies of diseasecausing plant viruses reveal that most RNA viruses exploit the host machinery to perform their life cycle (Dreher and Miller, 2006;Pyott et al., 2016;Li, 2019). RTSV is one such RNA virus that leverages a host translation initiation factor four gamma (eIF4G) gene of rice to replicate and establish within rice plant (Lee et al., 2010). Lee et al. (2010) identified that naturally available cultivars resistant to RTSV had nucleotide polymorphisms and deletions affecting Y 1059 V 1060 V 1061 amino acid residues in eIF4G gene in japonica genotypes. The mutations resulted from substitutions at nucleotide positions 4387 (A > G) and 4390 (T > C). This suggested that mimicking such naturally occurring mutations in susceptible genotypes would successfully impart tungro disease resistance in target cultivars.
The latest genome editing tool via base editors has unravelled the possibilities of creating highly specific targeted substitution mutations in the host genome (Komor et al., 2016). Adenine base editors (ABEs) convert an A•T base pair to a G•C base pair, while Cytosine base editors (CBEs) convert a C•G base pair into a T•A base pair (Gaudelli et al., 2017;Komor et al., 2016). Unlike the Cas9 of CRISPR/ Cas9 system, which has two nuclease domains, the Cas9 in base editors have only one active cleavage domain and hence, the Cas9 in base editors is referred to as nCas9/Cas9n (nickase Cas9) or dCas9 (dead Cas9). This nCas9 or dCas9 is fused to cytosine deaminase (in case of CBEs) or adenosine deaminase (in case of ABEs) that characterize a base editor pair (Gaudelli et al., 2017;Komor et al., 2016). Of the two base editors, ABE has been widely accepted for base editing in rice crop as CBEs were found to induce unintended off-target mutations and have low editing efficiency (Jin et al., 2019;Hao et al., 2019). A series of adenine base editing vectors have been developed within a short span of four years, to yield maximum A > G substitution with negligible non-canonical substitutions (Gaudelli et al., 2017;Li et al., 2018;Kim et al., 2019). An adenine base editor, ABE7.10 has been reported to create high substitution mutations, up to 59.1 % in a japonica rice variety, Zhonghua 11 (Lee et al., 2018). Considering the efficacy and specificity of ABE, the plasmid harboring ABE7.10, pH-PABE-7-esgRNA (Lee et al., 2018) was used in the present study to create substitution mutations in the YVV residues of eIF4G gene of indica cultivar, ASD16, to impart resistance against rice tungro disease.

Agrobacterium-mediated rice transformation
Agrobacterium culture harboring the recombinant plasmid was used to transform an RTD susceptible indica cultivar ASD16, a cross derivative of ADT 39 and CO 51. Immature seeds (14-16 days after flowering) of ASD16 were collected from Paddy Breeding Station, Tamil Nadu Agricultural UNIVERSITY. Embryos were isolated from these seeds and used as explants for Agrobacteriummediated transformation following the protocol of Hiei and Komari (2008). Well proliferated and friable yellow calli were subjected to two rounds of stringent selection in 50 mgL -1 of hygromycin B. The calli that survived on hygromycin selection were subcultured onto pre-regeneration, regeneration and rooting media. Regenerated plants with well-developed roots were hardened and maintained in transgenic greenhouse.

On-target mutation analysis of putative T 0 mutants by Sanger sequencing
Plant genomic DNA was isolated from T 0 putative mutants and ASD16 wild type following CTAB method (Porebski et al., 1997). Molecular analyses by PCR for hpt (hygromycin phosphotransferase) and cas9 genes were performed using sequence-specific primers to confirm that the putative mutants developed had T-DNA with genes required for editing ( Table 1). The target region encompassing the sgRNA sequence was amplified with eIF4G primers (eIF4G F and eIF4G R; Table 1) in PCR positive mutants. PCR amplicons were purified (NucleoSpin Gel and PCR Clean-up Kit, Machery Nagel) and Sanger sequenced (Eurofins, Bengaluru). Analysis of results obtained from Sanger sequencing was performed using web-based tools, DSDecodeM (http://skl.scau.edu.cn/dsdecode/) (Xie et al., 2017;Liu et al., 2015) and CRISPR-ID (http://crispid.gbiomed.kuleuven.be/) (Dehairs et al., 2016) to identify the position of substitution mutations in the sgRNA sequence. The percentage of substitution contributed by a base at a specific position in the sgRNA sequence was predicted using web-based tool, EditR (http://baseeditr.com) (Kluesner et al., 2018). Base substituted mutants with missense mutations were identified based on these results.

Agrobacterium-mediated transformation of ASD16
Agrobacterium-mediated transformation of elite rice cultivar ASD16, with the strain LBA4404, harboring the recombinant plasmid pH-PABE7-esgRNA+sgRNA was successful in generating 139 independent events from 22 batches of cocultivation, comprising of 2220 immature embryos. This gave an average transformation efficiency of 6.26 % (Table 2).

Identification of on-target mutations harboring missense mutations
Molecular characterization of T 0 putative mutants by PCR for the presence of hpt and cas9 genes confirmed successful integration of T-DNA in all the 139 independent events generated ( Fig. 3a and Fig.  3b). Amplification of the target region in these T-DNA positive events gave expected amplification of 577 bp (Fig. 3c). The sgRNA sequence has 4 adenine (A) residues in the editing window at positions A 4, A 5, A 7 and A 8 . Substitution mutations at A 5 and A 8 lead to missense mutations as A 5 > G 5 and A 8 > G 8 would result in V > A missense mutations at both the positions, while A 4 > G 4 and A 7 > G 7 would result in silent mutations. Sanger sequencing analysis of the T-DNA positive mutants identified 16 events harboring missense mutations and two harboring deletion mutations (Table 2). All the substitution mutations observed were in monoallelic form. Out of the 16 missense mutants, 14 had one missense mutation, while two had two missense mutations (YK-ASD16-141 and YK-ASD16-150) ( Table 3). The majority of the mutants (13 mutants) had substitution mutations at A 4 and A 5. No substitution mutations were detected at A 7 and A 8 positions. At A 5 , 5.5 % of the adenosines were converted to guanosine, while at A 4 , 3.37 % of adenosines were converted to guanosine (Fig. 4). No silent mutation, arising from A 4 > G 4 alone, was observed and all the 5 mutants which had A 4 > G 4 substitution were observed along with A 5 > G 5 . Four mutants (YK-ASD16-141, YK-ASD16-150, YK-ASD16-147 and YK-ASD16-151B) had substitutions immediately upstream of the YVV residues, resulting in S > F. In contrast, one mutant (YK-ASD16-234) had substitution downstream, resulting in D > H. These novel mutations are in close proximity to the YVV residues, which were reported earlier by Lee et al. (2010) in naturally available resistant genotypes. Macovei et al. (2018) reported that CRISPR/ Cas9-mediated genome-edited rice mutants targeting a stretch of 14 amino acid residues 'SVLFPNLAGKSYVV', could successfully confer resistance against rice tungro disease. Thus, the 16 missense mutants with substitutions affecting YVV residues and residues immediately adjacent to YVV residues could serve as promising candidates for imparting resistance against rice tungro disease. The two deletion mutants (YK-ASD16-246 and YK-ASD16-354) had similar homozygous deletion mutations of three nucleotides 'GTT', encoding valine in the target GKSYVVD residues (Table 3). Observations of such deletion mutants with adenine base editors have also been reported earlier with low frequency (Li et al., 2018;Li et al., 2021). This mutation observed in the two deletion mutants was similar to that of the naturally available resistant genotype, TKM 6, which had deletion of the V residue (Lee et al., 2010). They serves as promising candidates in imparting resistance against tungro disease. Thus, 18 mutants were identified, giving a mutation efficiency of 12.95 % (Table 2). (Arrows indicate expected base substitution of A > G. Frequency of expected nucleotides are highlighted in blue and those of substituted nucleotides are highlighted in yellow. As deletion mutation was observed at these nucleotide positions, total percentage value is less than 100.) Besides the canonical A > G substitution, we have also observed non-canonical substitution of G > A at 14 th position of the sgRNA sequence. At G 14 , 1.78 % of the guanosines got converted to adenosine (Fig.  4). Similar observations on non-canonical editing using adenine base editors and, more precisely, ABE7.10 have also been reported earlier (Lee et al., 2018;Kim et al., 2019, Jeong et al., 2020. A possible explanation for this observation is the role of the adenosine deaminase enzyme. Unlike cytosine base editing, adenine base editing does not occur spontaneously in vivo as no enzymes are known to deaminate adenine in DNA (Gaudelli et al., 2017). Thus, the deaminase enzyme used in the construction of ABE 7.10 is sourced from E.coli (ecTadA, E.coli tRNA specific adenosine deaminase). The ecTadA enzyme harbors a common catalytic site for deamination of cytosine and adenine residues (Jeong et al., 2020). This explains the non-canonical substitutions of cytosine to adenine/thymine/ guanine while using an ABE7.10 in human and mouse cells (Lee et al., 2018;Kim et al., 2019, Jeong et al., 2020. Such cytosine substitutions were favored when C is present in a TC*N fashion and the editing window was limited between 5-7 bp (Lee et al., 2018;Kim et al., 2019). Table 3. Nucleotide traces and predicted protein sequences in T0 mutants. The PAM sequence is underlined in the wild type allele. Substitutions in the nucleotide sequences and predicted protein sequences are denoted in red.A1: Allele 1, A2: Allele 2, WT: Wild type.

CONCLUSION
The present study is a report on the application of adenine base editing vector systems in creating targeted base substitution in indica rice. The successful generation of eIF4G mutants in the local elite cultivar ASD16 harboring mutations similar to that of naturally available tungro resistant genotypes can impart tungro disease resistance. The inheritance of mutation needs to be studied in subsequent T 1 and T 2 generations. Their performance needs to be assessed in homozygous T 2 progeny by conducting bioassay against rice tungro virus. Besides, their agronomic performance also needs to be compared with the ASD16 wild type in the T 2 population. These promising mutants in ASD16 background can be used directly for cultivation or as a parent to introgress the trait to other elite genotypes, once characterized for RTD resistance.

FUNDING AND ACKNOWLEDGMENT
The authors would like to thank ICAR-NASF (ICAR/CRISPR-Cas-7003/2017-18) for the funding and Department of Plant Biotechnology, Centre for Plant Molecular Biology and Biotechnology, Tamil Nadu Agricultural University, Coimbatore for the facilities. YK also thank ICAR-NASF for the fellowship.