Introduction

A number of plant viruses, are the deadliest pathogens of crops and have been extensively investigated (Scholthof et al. 2011). The top ten plant viruses were listed based upon their respective scientific importance and significant economic losses in a wide range of crops (Scholthof et al. 2011). Among these ten viruses, mostly enlisted viruses have positive-sense and single-stranded RNA as genetic material. However, in Cauliflower mosaic virus, enlisted as a devastating virus, double-stranded DNA, act as an alternate form of the nucleic acid genome with unusual translation strategies. In contrast, viruses containing two single-stranded DNA as genetic material are also enlisted named as Tomato yellow leaf curl virus (TYLCV) and African cassava mosaic virus (ACMV), which are transmitted by whitefly vectors, and caused loss of billions of US dollars. Despite having different genetic materials (single or double stranded nucleic acid), all these enlisted viruses are significant contributors to severe crop losses worldwide. Huge economic loss caused by these viruses attracted considerable worldwide attention to defect viral infection for the development of promising tools for crop improvement.

These viral pathogens have complex interactions with the host plants as well as vectors and show sophisticated integration of their respective genome in the molecular mechanisms of infected living plant cells. Thus, several control measures (selective and non-selective) have been exploited against plant virus disease epidemics. The non-selective control measures cover a wide range of viral pathogens or vector species by targeting both the source of inoculum and rate of virus spread and also include cultural and legislative measures. On the other hand, strategies including host resistance and biological controls are considered as selective control measures. The selective control measures decrease the rate of virus spread but show inactivity against the initial virus source (Jones and Dangl 2006). Besides, utilization of attenuated/mild virus strains in a biological control approach, have been documented for providing protection against disease caused by the same or related virus species using transgenic approach (Fuchs and Gonsalves 2007; Lin et al. 2007; Prins et al. 2008; Nishiguchi and Kobayashi 2011). Furthermore, in transgenic lines, the mechanism of resistance was explained either by protein derived or RNA derived model as documented by Baulcombe (1996). In protein derived model, the protein part (coat protein) of the attenuated viral strain has been exploited to affect the process of virus infection (Bazzini et al. 2006). In contrast, host plants develop resistance against secondary virus infection via RNA mediated cross-protection, which is identical to post-transcriptional gene silencing (Ratcliff et al. 1999).

The RNA silencing approach has been documented as a novel gene regulatory mechanism by limiting the transcript level of the designated gene(s) in providing the resistance against virus attacks. However, viruses are evolved with a counter-attacking host defense mechanism by encoding suppressor(s) to save viral RNA genomes from RNA silencing machinery (Agrawal et al. 2003). Also, during this approach, there are chances of recombination with other potential viruses and subsequently, it leads to the generation of new pathogenic variant and resistance breaking virus strains (Varsani et al. 2018). Very recently, genome editing has emerged as a promising technique for improvement of viral resistance in crop plants by either targeting viral sequence or host factors. Genome editing employed nucleases such as ZFNs, TALENs and CRISPR-Cas9 to engineer plants for resistance against viruses. ZFNs and TALENs are costly and time-consuming and these limitations can be overcome by CRISPR Cas9 system. In this system, the gRNA-cas9 complex is employed for targeting and cleaving the viral genome (Uniyal et al. 2019a; b).

Studies have suggested CRISPR as an efficient technique to confer resistance in plants especially to geminiviruses (Ali et al. 2015, 2016; Ji et al. 2015; Uniyal et al 2019a). Ali et al. (2016) has analysed the CRISPR Cas9 machinery and found that targeting coding sequence using the same machinery generates viral variants that are capable of evading the CRISPR/Cas9 system. Although, no escape is detectable in a case when the target is in non-coding sequence. CRISPR approach can be easily used for creating RNA virus-resistant plants. But instead of targeting the viral sequence as a guide RNA target DNA sequence only, the CRISPR-Cas 9 system is allowed to target plant genes responsible for viral infection. The CRISPR-Cas9 system has been seen as a durable strategy for conferring viral resistance but there are chances of off-targets and mutation rate (Uniyal et al. 2019b). The repair of these off-targets cleavages also give rise to unwanted rearrangements as inversions, deletions and translocations. Every method for resisting plants against viruses has certain limitations and so day by day new technologies or improvement in previous technologies are being made as strategies for the decrement in off-targets (Zhang et al. 2015; Uniyal et al. 2019b).

However, the development of in-vivo attenuated virus(es) using the novel genome recoding approach are insensitive to gene transfer and subsequently reduces the risk of recombination as documented by Kuo et al. (2018). Secondly, this strategy has also not an issue of off-target effects. Thus, rewriting of codon or genome recoding could be a promising alternative strategy for the development of tolerance against plant virus in crop plants. In the present review article, the basic introduction and distinguish features of genome recoding are comprehensively explained. The details of recoding strategies adopted in various organisms are discussed and future scopes of rewriting genome in plants for viral attenuation are also outlined to understand the usage, impact and contribution in term of developing virus tolerant crops.

Synonymous codons: basics of genome recoding

In all living organisms, every cell has almost universal genetic code with 61 coding codons (comprising of four bases) for 20 amino acids along with three stop codons. The genetic code is defined as a set of rules that establish the correspondence between nucleotide triplet (as codon in genetic material) along with specific amino acids in any protein. The earlier belief that genetic code is universal is now known to have many exceptions, and so the genetic code exhibits flexibility (Bacher et al. 2007). This flexibility is the cause of formation of genetic code variants by four different natural mechanisms, named biased codon usage, codon reassignment, ambiguous decoding and natural genetic code expansion. However, the genetic code concept shows degeneracy which means that each codon usually codes for one specific amino acid but a single given amino acid might be coded by one, two, three, four, or six different codons (Tuorto and Lyko 2016). During the translation process, the corresponding codon in mRNA is paired with its complementary anticodon on specific tRNA that is charged with specific amino acid and once the correct pairing occurs, the respective amino acid is incorporated into the growing polypeptide chain. This concept of degeneracy can be explained by wobble base pairing (Crick 1966), which emphasized the importance of first and second bases in the codon by interacting with the anticodon of tRNA for the encoding of protein. Thus, a set of codons that encode for a single amino acid is termed as synonymous codons. Earlier, it was reported that all synonymous codons for a specific amino acid would be distributed randomly along the sequence of genetic information in DNA. Later, it was documented that synonymous codons are not used equally and randomly to code for a particular amino acid (Cannarozzi et al. 2010). In routine, some codons are preferred over others and this phenomenon is defined as codon usage bias. According to the proposed codon biased theory, the preferred codons correlate with the abundances of isoacceptor tRNAs and the degree of this correlation might be related to protein production levels of an individual specific gene (Rocha 2004).

The choice of synonymous codons for the encoding purpose has been found to affect the rate and fidelity of the translation process of protein of interest (Brule and Grayhack 2017; Gingold and Pilpel 2011). The replacement of synonymous codons with each other (termed as synonymous mutation) for a particular given amino acid still maintains the same unique amino acid sequence in particular protein but affects the translation efficiency (Li et al. 2015). The synonymous mutation or silent mutation influences many cellular processes and can also assign several valuable features including resistance to viruses, bio-containment, genetic isolation or resistance to horizontal gene transfer (Kuo et al. 2018).

Redesigning of synonymous codon: generation of recoded organisms including attenuated viruses

Genome recoding involves the engineering of endogenous translational components, orthogonal translational components and the synthesis of a non-canonical amino acid (ncAA). The incorporation of ncAA can be done by various methods including selective pressure incorporation method, stop or sense codon suppression method and frameshift method. The outline of steps involved in the generation of genome recoded organisms (GRO) is described in Fig. 1. Liu and Schultz (2010), reported the addition of approximately 70 unnatural amino acids to the genetic codes of E. coli, yeast and mammalian cells by the development of new orthogonal aminoacyl-tRNA synthetase/tRNA pairs. This report offered opportunities for the generation of improved or novel proteins as well as probes of structure and functions of proteins (Liu and Schultz 2010). Similarly, Lajoie et al. (2013a), documented the genome recoding of E. coli MG1655 by replacing all UAG stop codons with synonymous UAA codons that subsequently permitted the reassignment of translation function of UAG. In addition, this report also exhibited improved efficiency for in-corporation of ncAA and showed the expansion of chemical diversity of protein in GMO. Interestingly, the recoded E. coli was also documented with increased resistance to T7 bacteriophage. Overall, this work demonstrated that new genetic codes could be utilized for the development of viral resistance in genetically modified organisms. In the second report of Lajoie et al. (2013b), all instances of removal of 13 rare codons from a panel of 42 highly expressed essential genes were carried out and suggested that genome-wide removal of 13 codons is a feasible process. This work proposed a well-suited strategy for designing recoded genomes with reduced fitness. Ma and Isaacs (2016) documented a genetically recoded E. coli strain (lacking both UAG stop codon and the recognition protein release factor 1) with the utilization of alternative code that conferred resistance to multiple viruses. Thus, altering the genetic code obstructs the multiple viruses and encourages the use of recoded organisms as a strategy to reduce viral population fitness. Collectively, all these attempts provide new opportunities to generate proteins with enhanced or novel properties in targeted organisms.

Fig. 1
figure 1

The diagrammatic representation of the steps involved in the generation of genomically recoded virus. The resulted recoded virus upon transfection into living cells shows attenuation

Full size image

The synonymous redesigning of virus genome based upon alteration in codon usage, codon pair usage, and di-nucleotide content, could lead to attenuated viruses. In-fact, synonymously recoded viruses subsequently lead to the development of live attenuated vaccines. The recoding strategy involves hundred/thousands of nucleotide substitutions and shows lesser chances of risk of virus phenotypic reversion with point mutation. Thus, synonymous recoding of virus can lead to exploring numerous aspects of virus biology and the development of novel therapeutic strategies (Martínez et al. 2016). A number of similar reports, reflecting the utilization of synonymous recoding of viral genome of numerous viruses (poliovirus, influenza virus, chikungunya virus, human respiratory syncytial virus) has also been documented (Burns et al. 2006; Martrus et al. 2013; Nougairede et al. 2013).

Concept of codon bias variants and viral attenuation

Each organism has a unique preference of different set of codons over others and this phenomenon is termed as codon bias that can tune expression at multiple levels (genome/operon/gene) (Sharp and Li 1987; Quax et al. 2015). The codon biased frequency of synonymous codons is also seen at the genomic level in almost all organisms. In-fact, the codon bias is the phenomena that affect the translation efficiency of the viral genome after replacing with synonymous codons although does not change the amino acid sequence. The metrics that is used for the frequency of the optimized codons is termed as CAI (codon adaptation index). The CAI for some specific organisms is based upon the codon usage frequency in a reference set of highly expressed genes, while CAI for a unique gene is proposed to be determined by comparing its codon frequency with this reference set. The codon usage frequency plays a significant role in controlling the multitude of cellular processes including protein expression levels and protein folding (Pechmann and Frydman 2013; Shen et al. 2015). Several variations in codon bias have been documented including the occurrence of rarely translated codons at the 5′ end of the coding sequence and co-occurrence of certain codons (Cannarozzi et al. 2010). It has been proved that not only the overall frequency of synonymous codons but also the order of codons in a given gene is also biased, called codon co-occurrence bias. Once a codon has been used in coding sequences of highly expressed genes subsequently, it favours the codons that use the same tRNA (Cannarozzi et al. 2010). This co-occurrence bias phenomena have been documented in archaea, bacteria and eukaryotes (Botzman and Margalit 2011; Zhang et al. 2013).

Apart from codon biased frequency and co-occurrence, how a codon resides is also under selective constraint and referred to as codon pair bias (CPB). The CPB is the non-random fashion of distribution of nucleotides neighbouring a specific codon at the genomic level in organisms. Any adjacent pair of unique amino acids can be encoded by approximately 36 different pairs of synonymous codons and some synonymous codon pairs are used more frequently than others in specific species. For encoding adjacent specific amino acids (for example, alanine and glutamate), some codon pairs are usually underrepresented (nnUAnn), whereas, some are most preferred (nnGCnn). Some codon pairs are found more or less frequently depending upon the overall frequencies of two specific codons that form a particular codon pair in the open reading frames (Mueller et al. 2006). The preferences of codon pair bias are dissimilar in phylogenetically distant species. Coleman et al. (2008) developed an algorithm for quantification of CPB. For all possible codon pairs (except stop codon pair) in a particular organism/species, a score is calculated for every codon pair, termed as codon pair score (CPS) that is the natural logarithm of the ratio of the observed frequency of codon pair over expected frequency of all human coding regions (Coleman et al. 2008). CPS value for a particular codon pair can be either positive or negative based on its frequency in the human genome. Positive CPS value represents overrepresented codon pair whereas negative CPS value represents underrepresented codon pair. The codon pair bias for a particular coding region can be calculated by the average of all the possible values of CPSs of all the codon pairs (Coleman et al. 2008; Martínez et al. 2016).

The concept of CPB has also been utilized for virus attenuation (reduction in virulence of a pathogen, so that it can be used for making vaccine). Coleman et al. (2008) also documented the synthesis of novel attenuated polioviruses by using rare codons to encode same amino acid sequence as present in wild-type poliovirus and proposed a strategy for investigating the consequences of genomic level manipulation of CPB. This work documented that underrepresented codon pairs are responsible for poor translation and also act as a promising reason for reduction of the reproductive fitness of the virus and attenuation in polioviruses. The mechanism of viral genome hijacking the cellular machinery of a living cell depends upon the host protein synthesis and chaperone mechanisms. Thus, utilization of underrepresented codon pairs reduces viral fitness by interfering the protein synthesis. A new strain of poliovirus labelled as “PV-min” has been developed by recoding the poliovirus genome. This new strain was recoded with underrepresented codon pairs with respect to the human genome and subsequently reported for lower CPB score than human and worked as an attenuated form. In conclusion, this proposed strategy named as “synthetic attenuated virus engineering” (SAVE) allows re-engineering of the genetic material of viruses for diverse functions including live attenuated vaccine candidates. In addition, many similar reports also documented same strategies for attenuating different viruses by recoding the viral genome with the introduction of a large amount of underrepresented synonymous substitutions (Takata et al. 2017; Eschke et al. 2018).

However, it has been reported that the selection of disfavoured codon pairing leads to an increase in frequencies of di-nucleotide (CpG and UpA) which also attenuate the replication (Tulloch et al. 2014). In this report, the echovirus (human gut virus) was used to create two different types of mutant viral strain by changing two parameters (codon pair and frequencies of di-nucleotide). One set of viral strain had altered codon pair frequencies with a constant number of CG or UA nucleotide pairs. In another set of the viral strain, codon pair frequencies were kept constant and di-nucleotide frequencies (for CG and UA) were changed. Among these two different sets of mutants, only the set with increased di-nucleotide frequencies (for CG and UA) was found to be weakened in terms of replication ability. Similarly, Kunec and Osterrieder (2016), analysed codon pair preference in vertebrates (Five: human, pig, mouse, chicken and zebrafish) and arthropods (four: Aedes aegypti, Anopheles gambiae, Culex quinquefasciatus, and Ixodes scapularis) and documented that host codon pair preferences were found to marginally influence the encoding of viruses and attenuation was mainly carried out by an increase in CpG/TpA di-nucleotides and they emphasized that codon pair bias is a direct consequence of di-nucleotide bias.

CpG and TpA are the two di-nucleotides that deviates most from the expected distribution in various organisms. TpA di-nucleotides are found in regions such as TATA boxes or origin sites that require binding of proteins as in unwinding of DNA double helix while CpG di-nucleotides are mostly found at or near transcription start site as in promoter region (Karlin and Ladunga 1994). Codon pairs containing these two nucleotides are the most underrepresented pairs (Kunec and Osterrieder 2016). So, rewriting the viral genome for attenuation not only increase the underrepresented codon pairs but in doing so also inadvertently increases CpG and also TpA but in less amount (Kunec and Osterrieder 2016). The strategy of virus attenuation by di-nucleotide bias suggested the role of an increased level of the CpG and TpA di-nucleotides in improved innate response against recoded viruses as an increment in these two di-nucleotides are recognised by the host cell as non-self (Greenbaum et al. 2009; Atkinson et al. 2014). Mechanism of attenuation by synonymous replacement is controversial. It was stated that CpG/TpA di-nucleotide frequencies are higher in recoded virus, and more recognizable by the immune system that can prevent the viral replication in contrast to the RNA viruses which have lower frequencies. (Kumagai et al. 2008). So, overall the attenuation is dependent on the host immune system machinery. But in primates or the higher organism, these CpG motifs are methylated at cytosine and deaminated to form thymine which makes it difficult to understand the mechanistic reason of attenuation by a change in these dinucleotides (Hodgkinson and Eyre-Walker 2011).

The main genomic target is the capsid coding region in these viruses for genome recoding. Burns et al. (2009) stated the possibility of attenuation of the replication capacity of poliovirus by synonymous replacement of codons that are used for coding capsid. In this report, 542 non-preferred synonymous substitutions in the nucleotides of the virus capsid region were introduced. Virus containing all 542 substitutions was reported with reduced yield. Mueller et al. (2006) also introduced 680 substitutions in the same region of poliovirus which resulted in a yield of the non-viable virus. Several viruses attenuated by genome recoding have been listed in Table 1.

Table 1 Enlisting of viruses, genetic materials, crops, and attenuation type used for attenuation purposes
Full size table
Table 2 List of viruses used for the genome recoding in various organisms
Full size table

In view of all these observations, the concept of recoding genome can be systematically utilized for attenuation of virus and control viral diseases (Table 2). Certain features distinguish it from other strategies of viral attenuation (Coleman et al. 2008). These characteristics are: (a) no change in amino acid sequence compared to wild type (b) no phenotypic reversion and (c) can be combined with other synthetic approaches (Flanagan et al. 2001). However detailed knowledge of virology is necessarily required for genome recoding. The recoding strategy is nowadays used for creating vaccines against various types of bacteria.

Possible recoding strategies in organisms

The first and foremost thing for recoding is the computational designing of a recoded fragment by the reshuffling of synonymous codons in a way that amino acid sequence remains the same but there is a change in CPB and dinucleotide frequency of CpG and TpA. After that recoding of the genome can be carried out in three basic ways. The first way is changing or editing the existing genome by site-specific mutations in the target codon (Isaacs et al. 2011). The second strategy is rebuilding the native genome with the new recoded fragment (Lau et al. 2017). The third is a cumbersome process of complete de novo synthesis from recoded or synthesised fragments (Hutchison et al. 2016). Editing in the existing genome involves short oligonucleotides to change the specific target codon by recombination. Isaacs et al. (2011) applied this method and used multiplex automated genome engineering (MAGE) and conjugative assembly genome engineering (CAGE) to change 321 TAG codon to TAA in E. coli. In the second method that is rebuilding by segments, recoded DNA fragments are designed and synthesised de novo and then assembled into plasmid or vector. Plasmid or vector contains site-specific recombination sites corresponding to specific non-recoded fragment. Once plasmid is transfected inside a living cell, by recombinational events recoded fragment is integrated into the targeted host genome. Lau et al. (2017) documented the genome rewriting of Salmonella typhimurium by a process called SIRCAS that is stepwise integration of rolling circle amplified genome. This process involved in-silico designing of the recoded genome of S. typhimurium by replacing TTA/TTG leucine codon with CTA/CTG codons respectively. Then 10–20 kb recoded segments containing antibiotic (kanamycin or chloramphenicol) resistance cassettes were created and assembled into yeast artificial chromosome (YAC). The recoded segments also contain homology regions that allow its integration into the main chromosome. The third method of recoding by de novo synthesis involves de novo synthesis of smaller fragments to be assembled into larger ones and eventually bypasses the need of native organism. Hutchison et al. (2016) documented the building of minimal genome (synthesised chemically 531 kb genome having 473 genes involved in translation and transcription) of Mycoplasma mycoides containing only essential genes for life apart from native organism genome size, 1097 kb. A brief outline of virus recoding strategy has been shown in Fig. 2.

Fig. 2
figure 2

Creation of virus-resistant organism using genome recoding

Full size image

Scope of rewriting genome for generating virus tolerant crop plants

Phyto-viruses have affected many economically important crops including cotton, wheat, maize and others. For protecting these crops from viruses, they are being engineered by introducing the mild and attenuated form(s) of the virus affecting crops that cross-protect the engineered plants from that same virus. This incited role of RNA interference has been reported in this cross-protection mechanism. This strategy was rationally the most effective one but also limited as mentioned above because of the chance of recombination and generation of a new rewired viral genome. Moreover, this approach is not transgene-free and public acceptance can also be a major problem. Before this method, some conventional methods were also used that includes UV absorbing sheets and reflective mulches, and these methods are cost-effective and successful in minimizing the incidence of vectors. Continuous molecular research is being explored in the field of virology to understand viral control measures and resistance.

The new emerging recoding strategies although yet not tested in plants, can open new hope for the development of resistance against virus-mediated disease(s). Multiple plant viruses target the plastid and as its size is much smaller than the nuclear genome, so production and installation of synthetic chloroplast genome for virus resistance is under progressive research. Theoretically, a recoded plastome can result in resistance against all plastid-targeted virus as it will not be able to replicate or express its genes in plasmid. Although till now, no report of the level of difficulty in installation and maintenance of a synplastome has been reported in plants (Pixley et al. 2019). Generally taking the idea of cross-protection in plants, if the virus is attenuated by synonymously replacing its codons as done in the above-mentioned studies and injected into native plant, it can offer protection against the same invaded virus. Genome recoded virus are unable to transfer the gene to the same virus as mentioned by Ostrov et al. (2016), so there is no risk of generation of new genetic virus and recombination. The method can overcome the previous attenuation process but yet to be tested in plants. Recoding has established attenuation of human viruses and is a new strategy of developing an animal vaccine and could have successful chances for plant viruses too.

Possible challenges for developing virus tolerant plant using genome recoding strategy

The investigation of codon usage bias source is mainly carried out in silico and depends upon many approaches that can be sub-grouped into host-dependent and host-independent. The host-dependent methods depend mainly upon total codon statistics for an organism (plants/plant viruses) and further its comparison with respect to either specific gene(s) or upon tRNA abundance in same organisms (Sharp and Li 1987). These approaches have very limited available information in plants and plant-specific viruses ( Xu et al. 2008; Wang et al. 2018). A few reports of tRNA statistic has been only found in some model viruses (Wright 1990). The host independent subgroup measures only for uneven synonymous codons and expressed as an effective number of codons. At present, two models (neutralist and selective models) have been proposed to explain the primary reason of codon usage bias. These are neutralist and selective models. According to the neutralist model, the major reason for codon bias is due to the differences in the mutational processes. While the selective model proposes the major reason for the same to be abundances of tRNAs required for the optimization of an efficient and accurate translation process. In multicellular organisms, the observed codon usage is an exquisite balance between both selective and mutational pressures. However, in the case of plant viruses, the base composition and translational selection was found to be inadequate to explain codon usage bias as described by Cardinale et al. (2013). The codon preferences of plant viruses (luteoviruses and potyviruses; ssRNA and geminiviruses (ssDNA) were analysed and found to be irrelevant to base composition or translational selection. Very few reports are available regarding codon bias in plant viruses including begomoviruses. Xu et al. (2008) reported that the synonymous usage variations in the protein-encoded gene of begomoviruses are mainly influenced by mutation bias. These mechanisms need to be explored in the respective organisms before researchers can move ahead with genome recoding and this is the major challenge that needs to be considered before drafting strategy for the development of virus tolerance crops.

Apart from viruses, the codon bias usage in multicellular organisms is much more complex than unicellular organisms. However, the first comprehensive report of species-specific codon choice in plant genes was documented by Murray et al. (1989). The relative use of synonymous codons has also been found different in monocots and dicots. This report also offered a few highly biased plant genes as suitable candidates for experimentally altering codon bias in a plant. Besides, several factors namely length of the gene, the composition of GC, expression level, gene translation initiation signal, amino acids composition in protein, the abundance of tRNA and protein structure have been documented to impart influence on the codon usage bias within or among species (Prabha et al. 2017; Zhao et al. 2016). The codon usage bias is also found to be associated with the phylogenetic relationship among the given species. A greater variation in codon usage bias has been found in distant phylogenetic species (Zhao et al. 2016). The first report about the differences in the usage of synonymous codons in different genes that are expressed in different tissues of Arabidopsis was documented in 2012. Camiolo et al. (2012) proposed that the possible cause of the evolution of codon bias might be an adaptive response to different tRNAs in different tissues. Genes expressed in various tissues were found to be a tissue-specific compositional signature for codon usage in Arabidopsis. In the latest report, the composition bias with selection and mutation pressure affecting the codon usage pattern of the protein-coding genes was documented (Paul et al. 2018). With the advent of whole-genome sequencing of numerous plant species, the genome-wide pattern of codon bias has also emerged in many plant species including cotton species (Wang et al. 2018). The better understanding of the codon bias by using genome sequencing in different plants/viruses is also another major challenges. However, in most of the cases, the recoded genes showed reduced fitness and testing multiple designs in a single genome could lead to unacceptable fitness impairment (Lajoie et al. 2013b). The limits of genetic recoding in essential genes were analysed by Lajoie et al (2013a). The future design should only change codons of interest whose variants are associated with normal growth.

The non-availability of potential species-specific in-silico codon optimizer program could also be considered as another bottleneck challenge for drafting genome recoding strategy in plants and viruses. The in-silico analysis could help in translation efficiencies of native and codon-optimized genes and can be useful for comparative analysis of translation responses of different defence genes/optimized against plant viruses from a different host and species-specific specificity. Apart from this, many other techniques including ribosome profiling also offer to explore diagnose limiting steps for recoding strategy.

Broader implications and associated issues

Genome recoding strategy has great potential for enhancement of genome and imparting new properties including virus resistance, biocontainment and incorporation of non-natural amino acid. The resistance to horizontal gene transfer, improved functions and genome reduction, and usage to address biological questions are some other diverse significant applications of genome recoding.

In industrial fermentation, virus contamination is a serious issue. The recorded cells designed by genome recoding could be resistant to decoding nucleic acid signals from infected viruses and eventually it will prevent the bacterial microorganism, which is useful in dairy sector from viral contamination (Samson and Moineau 2013). Another promising application is the addition of new amino acid that could be responsible for improving and even expanding protein functions (Xiao et al. 2015). In addition, it also provides a new opportunity in the field of protein chemistry. The incorporation of ncAA can be easily achieved and eventually the building blocks for protein will be increased using the same strategy. This strategy has also a positive impact on protein therapeutics including the bispecific antibodies, immunotoxins, and vaccines development (Kuo et al. 2018).

Genome recoding might also be useful for blocking functional horizontal gene transfer from engineered microbes into the wild. The concept of reassigning stop codon, as sense codon could be a potential factor for making recoded host genes untraceable by most of the other surrounding microbes. The usage of toxins to prevent DNA transfer from recoded organisms to the surrounding environment could also be manageable. Besides all the diverse applications, the establishment of new or improved functions in the organism of interest can be achievable by involving entire genome synthesis or insertion of a new gene cluster.

Except for microbes and yeast, genome recoding at higher organisms and plants has many technical leaps and involve many societal and ethical questions. There are also risks for almost every strategy and similar risks are possible in genome recoding strategy. However, it varies from organism to organism and plant to plant and offers a pretty remarkable moment from a regulatory standpoint for a new strategy for genome rewriting. However, this strategy would have a huge breakthrough in the prevention of plant diseases mediated by viruses.

Conclusion and future direction

Recoding genome with the replacement of synonymous codons with specific one in a way that it changes the codon pair bias and dinucleotide bias has emerged as the novel strategy of developing attenuated virus(es). These genomically recoded viruses are employed to develop vaccines in animals and also confer viral resistance. This strategy is successfully tested in animal viruses to attenuate them. But this recoding strategy is yet to be tested for plant viruses for conferring viral resistance in plants and can open up a new avenue in controlling diseases mediated by plant viruses. It is transgene-free editing and hence will be less regulated than genetically engineered crops and can be accepted by the mass public quickly. Besides, more attention must be paid to get validated data of genome recoding concerning different plant viral attenuation and subsequently can be used to prevent a lot of agricultural losses.