Keywords

Sweetpotato (Ipomoea batatas (L.) Lam) is a globally important root, tuber, and banana (RTB) crop and cultivated in tropical and temperate zones of the world. As a hexaploidy crop with high heterogeneity, all the elite sweetpotato genotypes have to be propagated clonally, preventing further segregation of the genes and alleles combined in the elite genomes in the next generations. The asexual propagation of sweetpotatoes sometimes helps spread pathogens and pests, leading to dramatic yield loss. Sweetpotato breeders have to constantly develop new cultivars and lines with improved resistance to insects and pests that continuously evolve in the field. Sweetpotato breeders also need to put efforts to breed cultivars with improved yield, nutrition, and flavor due to market demands. Thus, sweetpotato breeding is of critical importance and innovative sweetpotato breeding methods need to be further developed in order to improve breeding efficacy and decrease breeding cycle and cost.

10.1 Methods Used for Sweetpotato Breeding

Three breeding methods are commonly employed in sweetpotato breeding programs, i.e., conventional breeding, molecular marker-assisted breeding, and transgenic breeding. As a hexaploid plant in the Convolvulaceae family, sweetpotato has the characteristics of self- and cross-incompatibility, high heterogeneity, and poor flowering and sterility in certain environments, which pose challenges for conventional sweetpotato breeding (Cervantes-Flores et al. 2011; Dhir et al. 1998; Yan et al. 2022). These challenges include the use of multiple parents for crossing, large (10,000–100,000) breeding populations for progeny evaluation, extended breeding cycles, genetic intricacies, unpredictable segregation, and labor-intensive procedures. Chromosomal linkage blocks and linkage-drag also prevent the generation of novel meiotic recombination of genes and alleles in conventional sweetpotato breeding. In the past few decades, both paired cross and polycross together with recurrent selection have been successfully implemented in conventional sweetpotato breeding for the development and release of elite cultivars with desirable traits. These traits include enhanced disease resistance, high yield, preferable taste, and improved nutrition such as high beta-carotene content, iron content, and dry matter, and low sweetness (Mwanga et al. 2016; Rolston et al. 1987).

Since conventional sweetpotato breeding is highly ineffective and time consuming, and conventional breeding for certain traits such as storage root yield and quality, resistance to root-knot nematode (RKN), sweetpotato virus disease (SPVD), and sweetpotato weevil has limited success (Collins et al. 2019; Ngailo et al. 2013; Oloka et al. 2021; Placide et al. 2015), molecular breeding has become a powerful and complementary means to conventional sweetpotato breeding for traits like these. Molecular breeding relies on the development of DNA-based molecular markers that are tightly linked to desirable traits to assist progeny selection. Markers including Simple Sequence Repeats (SSRs), Restriction Fragment Length Polymorphisms (RFLPs), Amplified Fragment Length Polymorphisms (AFLPs), and more recently Single Nucleotide Polymorphisms (SNPs) have been used in marker-assisted selection (MAS) in sweetpotato (Chap. 4). MAS works well for simple traits and major QTLs. However, complex traits may be controlled by many small-effect genes and alleles. Molecular markers associated with each trait may be corresponding to a large genomic region containing tens or hundreds of genes rather than the genes responsible for the trait. Selection of markers is based on statistical significance of individual markers and sometimes arbitrary. To further improve selection efficiency and prediction of progeny performance, genomic selection can be used to estimate the effects of all the markers for all the traits of each individual plant to calculate genome-estimated breeding values (GEBVs) so that GEBVs can be used to predict which progeny is good for further testing and release. Efforts are underway in genomic selection in sweetpotato as variously described in Chaps. 4, 6, 7, and 12.

As the key genes underlying certain agronomic traits have been cloned and characterized in various species including sweetpotato, overexpression and/or RNAi-induced silencing of these key genes have been used in transgenic sweetpotato to improve various single gene traits. Thus, genetic engineering has emerged as a valuable new tool in sweetpotato breeding to improve salinity and drought tolerance, disease and pest resistance, herbicide resistance, and starch, carotenoid, and anthocyanin biosynthesis (Liu 2017). Both biolistic bombardment- and Agrobacterium tumefaciens-mediated transformation have been successfully used in the generation of transgenic sweetpotato lines with limited transformation efficiency (Liu 2017). Most of the transgenic sweetpotato lines were obtained from the transformation of leaves, petioles, stems, storage roots, and embryogenic calli even though transgenic sweetpotato plants were also generated from embryogenic suspension cell cultures. The advantage of using genetic engineering for sweetpotato improvement is the manipulation of one or a few genes at a time, permitting further finetuning of the expression of the gene(s) of interest. The difficulties in transgenic sweetpotato breeding include the genotype-dependency of sweetpotato transformation and the public acceptance of the transgenic end products. Most elite sweetpotato breeding cultivars and lines are not transformable and the transformable genotypes are not elite. The permanent integration of the transgenes in the end products makes the sweetpotatoes GMO, which is a marketing hurdle.

In addition to the above-mentioned breeding methods, somatic hybridization (Guo et al. 2006; Yang et al. 2009; Zhang et al. 2001) and mutation breeding (Mansour et al., 2018; Moussa and Gomaa 2017; Luan et al. 2007; Shin et al. 2011; Wang et al. 2007; Yan et al. 2022) have been successfully used in sweetpotato breeding. Somatic hybrids were successfully obtained between sweetpotato and its wild relatives I. cairica (Guo et al. 2006), I. lacunose (Zhang et al. 2001), or I. triloba (Jia et al. 2022) even though many of these somatic hybrids contained significantly reduced chromosome numbers than the sum of the chromosome numbers in both parents. Using gamma irradiation, improved sweetpotato varieties have been developed from axillary buds for high yield and starch content (Shin et al. 2011), shoot apices for changed root flesh color and increased root yield (Wang et al. 2007), stems for modified yield and quality traits (Moussa and Gomaa, 2017), and calli for induced morphological changes (Lee et al. 2002). The use of gamma irradiation for sweetpotato breeding may suffer from the formation of chimera mutations, which can be avoided in the ethylmethanesulphonate (EMS)-mediated sweetpotato breeding that utilized sweetpotato leaf explant-derived calli to breed cultivars with enhanced salt tolerance (Luan et al. 2007). All of these mutation breeding methods were used in combination with tissue culture and may suffer from a low mutation rate.

10.2 Gene Editing Biotechnologies that Could Be Applied in Sweetpotato

Gene editing tools include meganucleases (Cohen-Tannoudji et al. 1998), zinc-finger nuclease (ZFNs) (Bibikova et al. 2002), transcription activator-like effector nucleases (TALENs) (Christian et al. 2010), and clustered regularly interspaced short palindromic repeat-associated nuclease (CRISPR) (Cong et al. 2013; Mali et al. 2013). These nucleases cut both DNA strands at the target sites to create double-strand breaks (DSBs). The DSBs are then repaired by the DNA repair mechanisms in plant cells. The dominant repair mechanism in plant cells is non-homologous end joining (NHEJ), which is prone to introducing random mutations. DSBs can also be repaired to a lesser extent by homology-directed repair (HDR) if a donor DNA template is provided for homologous recombination. Screening of the mutated genes can lead to the identification of the gene-edited plants.

Meganucleases have been employed as a gene editing tool since 1985. They work as a homodimer to bind their specific recognition sites ranging from 12 to 40 base pairs in length (Fig. 10.1a; Jurica et al. 1998). The rare presence of their recognition sites in plant genomes and the fusion of the DNA-binding domain with their catalytic domain limit their application in plants. In the 1990s, ZFNs emerged as a promising gene editing tool. ZFNs work as heterodimers with each ZFN containing a specific DNA-binding domain and a non-specific cleavage domain from the Fok I endonuclease (Fig. 10.1b; Pabo et al. 2001). Each DNA-binding domain contains 6–8 zinc fingers with each zinc finger recognizing and binding to 3-bp-long nucleotides. Design of each DNA-binding domain requires screening against expression libraries to confirm their binding specificity. Synthesis of each DNA-binding domain is tedious and costly. The prohibitive cost and intricate synthesis process have hampered the widespread adoption of ZFNs in plant applications. Subsequently, TALENs were developed, featuring another type of DNA-binding domain and a non-specific Fok I domain positioned at the carboxylic terminal (Fig. 10.1c; Christian et al. 2010). The DNA-binding domain of TALENs contains ~20 TALE repeats with each repeat identical in amino acid sequence except the Repeat Variable Diresidue (RVD) on positions 12 and 13 of each repeat that binds to a single nucleotide (Christian et al. 2010). Compared to ZFNs, TALENs offer greater user-friendliness, although their drawback lies in the necessity of constructing each repeat, which is costly.

Fig. 10.1
figure 1

Schematic representations of gene editing tools. a Meganuclease can bind 12–40-bp-long DNA sequences and precisely cleave both strands at its recognition site, resulting in sticky DSBs. b Zinc-finger nucleases (ZFNs) function as dimers, with each monomer comprising a DNA-binding domain and a nuclease domain. The DNA-binding domain comprises 3–6 zinc finger repeats, forming an array that identifies 9–18 nucleotides. The nuclease domain contains the type II restriction endonuclease Fok I. c Transcription activator-like nucleases (TALENs) operate as dimeric enzymes akin to ZFNs. Each subunit contains a DNA-binding domain—a highly conserved 33–34-amino acid-long sequence tailored for each nucleotide, and a Fok I nuclease domain. d In the CRISPR/Cas9 system, the Cas9 endonuclease is directed by the gRNA to achieve precise target cleavage. A 20-nucleotide-long recognition site precedes the protospacer adjacent motif (PAM) for this process

Full size image

More recently, CRISPR/Cas technologies have risen in prominence due to their efficiency and ease of use. Each CRISPR/Cas consists of two primary components, i.e., an endonuclease Cas protein and a single guide RNA (gRNA), which binds to the single-strand DNA through 17–20 nt at the 5′-end of each gRNA (Fig. 10.1d; Tsai et al. 2015). With the guidance of a gRNA, the Cas protein specifically binds to and cleaves the target DNA, triggering DNA repair. There are various types of Cas genes such as Cas9 (Chen et al. 2019; Zhu et al. 2020), Cas12a (Wada et al. 2022), and Cas13a (Abudayyeh et al. 2016, 2019; East-Seletsky et al. 2016; Konermann et al. 2018). Cas9/gRNA stands as a highly efficient tool that has been extensively used for precise modification of target genes in plants. It has been employed for gene disruption, gene insertion, gene replacement, base editing, and regulation of gene expression in crops. It has played a pivotal role in modifying agronomic traits, encompassing improvements in nutritional content, yield enhancement, and the development of stress and disease-resistant crops (Zhu et al. 2020; Gao 2021). It expedites crop breeding by facilitating the incorporation of desired traits while minimizing undesirable ones.

10.3 The Success of Gene Editing in Sweetpotato

To date, the CRISPR/Cas technology has been used in sweetpotato for gene editing while the other gene editing tools have never been reported for attempts in sweetpotato. CRISPR/Cas9 was used to knockout the granule-bound starch synthase I (GBSSI) gene controlling amylose biosynthesis and starch branching enzyme II (SBEII) gene responsible for amylopectin biosynthesis, and a mutation efficiency of 62–92% was achieved for multi-allelic mutations (Wang et al. 2019). Most of the detected mutations were point mutations and small insertions/deletions (indels), some of which caused amino acid changes or stop codons. The gbssI knockout showed reduced amylose content while the sbeII knockout had increased amylose content and decreased amylopectin content. Neither individual knockout caused significant changes in the total starch content. This pioneer example demonstrated the effectiveness of the use of CRISPR/Cas9 for gene editing and trait improvement in sweetpotato.

Moreover, Cao et al. (2022) successfully obtained the PDS mutant through the cut-dip-budding delivery system in several sweetpotato genotypes. The cut-dip-budding method was developed based on the “root suckering” ability of sweetpotatoes, which grows shoots from adventitious shoot primordia on roots. In this process, the shoot explants were cut off at the shoot–root junction and treated with A. rhizogenes containing Cas9/gRNA at the cut sites to induce transformed roots, followed by bud generation. Mei et al. (2024) also reported that they successfully obtained the PDS mutant through the gene editing with the injection delivery method in sweetpotatoes. The axillary buds were removed from sweetpotato shoots, and A. tumefaciens containing Cas9/gRNA was injected into the cut sites for the regeneration of gene-edited shoots. The use of gene editing for sweetpotato trait improvement provides precision in the modification of target genes. It also decreases the sweetpotato breeding cycle from 8 to 12 years for conventional breeding to just one generation, significantly accelerating sweetpotato breeding and trait improvement.

10.4 Challenges and Constraints in Sweetpotato Gene Editing

To conduct gene editing in sweetpotato, Cas/gRNA needs to be delivered into a sweetpotato genome, which is typically achieved through biolistic bombardment- or Agrobacterium-mediated stable sweetpotato transformation. However, sweetpotato transformation is highly genotype dependent. Most, if not all, of the elite sweetpotato cultivars and lines are not transformable, preventing the direct use of gene editing in these elite sweetpotato cultivars and lines for further trait improvement.

The transformation and permanent integration of Cas/gRNA into a sweetpotato genome makes the end product GMO. Unlike diploid crops such as tomato in which sexual propagation and genetic segregation can be used to segregate out the transgenes, segregating out the transgenes via sexual propagation cannot be applied in sweetpotato. As a highly heterogeneous hexaploid crop, elite sweetpotato breeding cultivars and lines can only be propagated asexually once all the favorable genes and alleles have been combined into an individual plant through crossing. This constraint prevents the removal of the transgenes from the end product via segregation following sweetpotato transformation and gene editing, making the end product GMO. GMO gene-edited sweetpotato suffers from public acceptance and marketing hurdles in some countries such as England and the E.U.

Challenges and constraints in sweetpotato gene editing also come from the incomplete whole genome sequences of limited sweetpotato cultivars and their hexaploid nature, making gRNA target design challenging for genes with unknown sequences or with many homologous sequences in a sweetpotato genome of interest. The first successful attempt to conduct whole genome sequencing in sweetpotato was reported in Yang et al. (2017) which published a half haplotype-resolved genome in a newly bred carotenoid-rich cultivar Taizhong6. Then, Yoon et al. (2022) published a haploid-resolved and chromosome-scale assembly of the whole genome sequence of sweetpotato cv. Xushu18, which identified 175,633 genes and suggested that cv. Xushu18 is an auto-hexaploid with an AAAAAB genome. In addition, investments by the Bill & Melinda Gates Foundation in the GT4SP (http://sweetpotato.uga.edu/gt4sp_download.shtml) and SweetGAINS (https://cipotato.org/cip_projects/sweetgains-africa/) projects have developed genomics tools for sweetpotato improvement and have significantly improved the sweetpotato genomic resources, respectively. In August 2022, the SweetGAINS project released the high-quality v1 genome assembly of cv. Beauregard with annotation that is now available for searches with the BLAST search tool (http://sweetpotato.uga.edu/blast.shtml). In addition, the complete chloroplast genome sequences have been reported for 16 sweetpotato cultivars (Zhou et al. 2018) and another 107 sweetpotato cultivars including cv. Xushu18 (Xiao et al. 2021; Yan et al. 2015; Yoon et al. 2022). The whole mitochondrial genomes have also been published for cv. Xushu18 (Yoon et al. 2022) and cv. JinShan 57 (Yang et al. 2022). The availability of these genomic sequences dramatically helps with gRNA design for gene editing in sweetpotato. However, there may exist various SNPs between the published cultivar genomes and cultivar genomes of interest, which may affect the accuracy of the designed gRNA target sites and decrease the editing efficiency with potential off-target effects and unintended consequences in cultivar genomes of interest. To improve gRNA design, PCR amplification followed by Sanger sequencing could be used to amplify a gene of interest from a cultivar genome of interest even though it may be challenging to amplify all the six alleles of a gene in a cultivar genome.

10.5 Future Perspectives of Gene Editing in Sweetpotato

To overcome the above-mentioned challenges and constraints of sweetpotato gene editing, innovative approaches for genotype-independent transgene-free gene editing in crops including sweetpotato need to be developed. A set of key growth and developmental regulatory genes have been recently found to make recalcitrant crop genotypes transformable, permitting genotype-independent transformation in certain groups of crops (for reviews, see Gordon-Kamm et al. 2019; Maren et al. 2022; Nagle et al. 2018; Nalapalli et al. 2021). The most well-known growth and developmental regulatory genes include the WUSCHEL (WUS), BABY BOOM (BBM), ISOPENTENYL TRANSFERASE (IPT), GROWTH-REGULATING FACTOR (GRF4), and GRF-INTERACTING FACTOR1 (GIF1) genes. A pioneering example came from the use of a low expression of the maize WUS2 gene and a constitutive expression of the maize BBM gene for genotype-independent crop transformation in four difficult-to-transform maize inbred lines as well as thirty-three commercial maize inbred lines (Lowe et al. 2016). Another pioneering example was published in Debernardi et al. (2020) for the use of the wheat GRF4-GIF1 chimeric gene for genotype-independent transformation in wheat, rice, citrus, and triticale. Moreover, Maher et al. (2020) reported the use of a low expression of the maize WUS2 gene and a high expression of the Agrobacterium IPT gene for Cas9-mediated gene editing and edited shoot regeneration from the mature plants of tobacco, potato, and grape, which avoided the use of plant tissue culture. All these examples offer a great promise for using growth and developmental regulatory genes to conduct genotype-independent transformation in sweetpotato for gene editing.

To date, several strategies have been developed for transgene-free gene editing in crops. These include transgene removal via genetic segregation, transient expression of the Cas/gRNA DNA without permanent integration of Cas/gRNA into crop genomes, and DNA-free (or protein) delivery of the pre-assembled Cas/gRNA ribonucleoproteins (RNPs) or pre-transcribed Cas/gRNA RNA. As previously mentioned, elite sweetpotato cultivars must be propagated asexually to maintain favorable alleles, thus preventing the removal of transgenes via genetic segregation.

Cas/gRNA, delivered by either particle bombardment- or Agrobacterium-mediated transformation, can be transiently expressed in the nucleus without stable integration into the host genome. Thus, transgene-free gene-edited plants can be regenerated and selected in the absence of a selection agent; these regenerated plants will contain the targeted edited gene(s) but will not contain the transgenes. Identifying the transgene-free gene-edited plants requires a highly efficient plant regeneration protocol and a large number of regenerated plants after transformation. Simultaneous editing of a gene of interest (GOI) and the acetolactate synthase gene (ALS) gene in tomato, tobacco, potato, and citrus provided a selection marker for edited cells as the mutated als gene conferred resistance to sulfonylurea herbicides (Huang et al. 2023). While this is an effective method for obtaining transgene-free gene-edited plants, it still relies upon crop transformation for DNA delivery, which remains an obstacle for sweetpotato.

Moreover, Cas/gRNA RNPs can be delivered into a crop genome in a DNA-free manner, bypassing crop transformation of the Cas9/gRNA DNA. Protein delivery of the Cas/gRNA RNP has been used to obtain transgene-free gene-edited plants in a variety of crop species using biolistic bombardment (Liang et al. 2018, 2019; Poddar et al. 2023; Svitashev et al. 2016), polyethyleneglycol (PEG)-mediated transfection (Andersson et al. 2018; Banakar et al. 2022; Brandt et al. 2020; Choi et al. 2021; Fan et al. 2020; Jiang et al. 2021; Kim et al. 2020a, b; Klimek-Chodacka et al. 2021; Lin et al. 2022; Malnoy et al. 2016; Murovec et al. 2018; Nicolia et al. 2021; Najafi et al. 2023; Park et al. 2019; Pavese et al. 2022; Sidorov et al. 2022; Subburaj et al. 2016; Yu et al. 2021; Woo et al. 2015; Wu et al. 2020), and lipofection-mediated transfection (Liu et al. 2020). Most methods utilize protoplasts, which have been stripped of their cell walls for an easy delivery of RNPs across the cell membranes. However, most crop species, including sweetpotato, do not have a well-established plant regeneration system from protoplasts. Although protoplast regeneration has been previously reported in sweetpotato, efficiency appears to be genotype dependent (Sihachakr and Ducreux 1987). Protoplast-based systems also tend to be more technically challenging than other methods, requiring specialized equipment and labor.

Delivery of gene editing RNA transcripts into plant cells has also been used to achieve transgene-free gene editing in plants. Biolistic bombardment and virus-mediated delivery methods have been used to deliver Cas9/gRNA transcripts into plant cells for transient editing (Ma et al. 2020; Zhang et al. 2016). A recently published grafting-based system delivered Cas9/gRNA transcripts from transgenic rootstocks to wild-type scions using tRNA-like sequence (TLS) motifs (Yang et al. 2023). The addition of TLS motifs to Cas9 and gRNA transcripts allows them to traverse graft unions, resulting in transgene-free gene editing of the wild-type scion tissues. Like the DNA-based transient transformation systems outlined earlier, cells that have been edited using these methods must also be regenerated in a selection-free environment, making it difficult to identify edited plants.

Taken together, gene editing is a promising and powerful bioengineering method and can be used together with other breeding techniques to improve sweetpotato traits as well as conduct gene functional analysis. Enabling tools such as genotype-independent transgene-free gene editing methods need to be developed to revolutionize sweetpotato breeding.