Grapevine is one of the most widespread fruit crops in the world, with a production of about 77.1 million tons (FAOSTAT, 2019). It is cultivated both for the manufacture of wine and for its consumption as fresh fruit, and to a lesser extent to produce raisins, juices, and spirit drinks, the first use being the one for which it has more hectares allocated. Actual and future environmental conditions impose the decisions and choices regarding the management of the crop, following a path toward more sustainable alternatives. The problems related to climate change and the spread of numerous diseases require addressing solutions that can be found in the natural genetic variation of the genus Vitis (Vezzulli et al., 2019). Although grapevine can be improved through conventional breeding, it is a difficult and time-consuming process, due to the 2–3 years generation cycle and the long period of time required for the selection and testing of reliable progeny. Also, for grapevine, whose varieties are highly heterozygous, this way is even more difficult (Gray and Meredith, 1992).

Several years ago, genetic engineering emerged as an outstanding tool for the improvement of plants (Anderson et al., 2019). Genetic transformation offers the possibility of genetically modifying plants to improve agronomic traits of interest without altering the varietal identity using recombinant DNA technology, such as the transfer of resistance to diseases or herbicides to established crops. Grapevine was considered recalcitrant to genetic transformation, since one of its problems is the regeneration of plants from the tissues used for genetic transformation (Mullins et al., 1990; Nakano et al., 1994). The regeneration rate of grapevine plants after transformation and selection by antibiotics ranges between 10 and 30% of the total transformed material, and the transformation efficiency, which varies according to the genotype, down to 1%, although a 33% has been described (Torregrosa et al., 2015).

Accordingly, overcoming the problems related to tissue regeneration is one of the most essential challenges in the generation of transgenic and edited grapevine plants. In this sense, several factors have been identified that should be reviewed: the grapevine genotype; the type of tissue used to obtain the explants; the transformation methodology; the availability of regenerative transformable material, and the selection process/procedure based on antibiotics. Grapevine genetic transformation was mainly performed by infection with Agrobacterium tumefaciens (Torregrosa et al., 2015), and to a lesser extent through the biolistic techniques (Kikkert et al., 2001; Vidal et al., 2010).

The first critical factor in grapevine transformation is the production of highly regenerative transformable material, where the regeneration efficiency greatly depends on the different genotypes (Gray and Meredith, 1992). Somatic embryogenesis is the preferred regeneration procedure for the genetic transformation of grapevine. The source of starting material, the type, and the quality of the embryogenic cultures are key factors for a successful transformation (Martinelli and Gribaudo, 2001). However, despite these limitations, it has been possible to produce grapevine varieties resistant to fungal, viral, and bacterial diseases (Mauro et al., 1995; Scorza et al., 1996; Agüero et al., 2006; Nirala et al., 2010; Capriotti et al., 2020).

Beyond grapevine, the production of transgenic and edited plants requires, for most crops, the ability to regenerate plants from transformed tissues. This step is another critical factor, often reported as the biggest bottleneck in the process (Altpeter et al., 2016). After transgenic integration into recipient cells, tissue culture protocols are required to regenerate transgenic seedlings from transformed cells. This process is time consuming, usually several months, and generally requires the addition of expensive and environmentally harmful reagents (antibiotics and herbicides) to the growing medium to select the putative transgenic plants. The overexpression of certain transcription factors has been recently reported (Debernardi et al., 2020; Maher et al., 2020) as an interesting alternative to improve the transformation and regeneration processes of plants (including grapevine).

In recent years, new breeding technologies (NBT), such as gene editing via CRISPR-Cas9, have emerged as innovative genetic improvement tools for various crops of agronomic importance (Dalla Costa et al., 2017, 2019). The key elements in this system are Cas nucleases and CRISPR RNAs. The Cas9 endonuclease can cut at specific DNA target sites with the help of two small RNA molecules called CRISPR RNA (crRNA) and “trans-encoded” CRISPR RNA (tracrRNA). These two molecules can be fused artificially to form a chimeric RNA molecule called “single guide RNA” (sgRNA) (Mali et al., 2013; Qi et al., 2013). In conjunction with Cas9, sgRNA can form an “RNA-guided endonuclease,” a high precision tool capable of strategically introducing targeted mutations in the host genome (Jinek et al., 2012; Samanta et al., 2016).

On the other hand, alternative technologies are also emerging for the delivery of engineered genes into the plant cells. Carbon dots (CDs) were described as almost spherical water-soluble nanoparticles (NP) consisting of crystalline carbon domains synthesized from cheap starting materials such as peptides, carbohydrates, and, in general, a wide range of carbon sources (Li et al., 2012; Zheng et al., 2015; Peng et al., 2017). They have multiple advantages, such as being easy to obtain and for displaying an efficient plant cell uptake. It has been shown that CDs can act as a vehicle to transport plasmids to plant cells in transient transformation in several important crop species without negative impacts on photosynthesis or growth (Doyle et al., 2019; Schwartz et al., 2020).

Along with a thorough report of commonly known limitations related to grapevine transformation and regeneration, we primarily present in this review a detailed description of new alternative technologies that could provide solutions to overcome these drawbacks.

Since the genotype is one of the most influential factors in the success of a transformation protocol, the effects of the genetic background on the efficiency of plant regeneration and the corresponding culture conditions have been extensively studied (Dalla Costa et al., 2019). Thus, it has been reported that each genotype shows specific sensitivity to the infection with Agrobacterium, as well as to the antibiotics used to eliminate the bacteria, and/or to those used to select transgenic events (Zhou et al., 2014). The grapevine genetic background also influences somatic embryogenesis, the regeneration method most used in genetic engineering protocols in this crop. Several procedures have been developed for somatic embryogenesis from grapevine genotypes, and much research has been carried out using several plant tissues/organs as starting explants. The list of the plant parts widely used as suitable material for obtaining somatic embryos includes: ovaries (Yamamoto et al., 2000; Vidal et al., 2003, 2006; Gambino et al., 2005; Kikkert et al., 2005; López-Pérez et al., 2005); anthers (Franks et al., 1998; Gambino et al., 2005; Agüero et al., 2006; Prado et al., 2010); leaves (Martinelli et al., 1993; Nakano et al., 1994; Das et al., 2002); and, less frequently, stigmas and styles (Morgana et al., 2004; Carimi et al., 2005); stamen filaments (Nakajima and Matsuta, 2003; Acanda et al., 2013); nodal sections (Maillot et al., 2006, 2016); whole flowers (Gambino et al., 2007); mature seeds (Peiró et al., 2015); and tissues derived from vegetative structures, such as leaves and petioles (Martinelli et al., 1993; Das et al., 2002). The fact that most of the frequently used protocols are carried out using floral organs is indeed a strong limitation for obtaining somatic embryos, since the experiments can just be started at the flowering time. During the regeneration process, the germination of aberrant embryos and early germination can occur, which can also limit the obtaining of transformed grapevine plants since these embryos do not develop into normal plants. Among the most common aberrations, it is possible to find embryos without cotyledons, with different numbers of cotyledons, or with fused cotyledons, and trumpet-shaped or cauliflower-like cotyledons (Goebel-Tourand et al., 1993; Bornhoff et al., 2005; Li et al., 2006; Bharathy and Agrawal, 2008; Martinelli and Gribaudo, 2009; Peiró et al., 2015).

It has been proven that the synthesis and accumulation of reserve proteins during zygotic embryogenesis is regulated by abscisic acid and/or water stress (Dodeman et al., 1997). Therefore, to achieve a correct maturation of somatic embryos, two methodologies have been used: the exogenous application of abscisic acid (ABA) (Morris et al., 1990; Goebel-Tourand et al., 1993; Sholi et al., 2009) and the use of culture media with reduced water potential (Klimaszewska et al., 2000; Walker and Parrott, 2001; Kong et al., 2009; Buendía-González et al., 2012). Another strategy used has been the modulation of polyamine metabolism (Faure et al., 1991). The use of semipermeable cellulose acetate membranes has emerged as an effective alternative to improve the maturation of somatic embryos, and this is due to its ability to limit the availability of water for the embryo. In grapevines, it has been shown that the water stress produced by the membrane triggered an increase in endogenous levels of ABA, a fact that improved the maturation of somatic embryos (Acanda, 2015).

The choice of the Agrobacterium strain shows a significant effect on a successful transformation procedure of plant tissues. This factor also includes the corresponding bacterial culture conditions like density at the time of infection, coculture times, and the culture media used (Le Gall et al., 1994; Franks et al., 1998). One of the first Agrobacterium strains used for grapevine transformation was LBA4404 (Bouquet et al., 2008), showing low transformation efficiencies. To improve those transformation rates, hypervirulent Agrobacterium strains were later developed. Among them, EHA105 is currently the strain most used in grapevine transformation (Scorza et al., 1996; Franks et al., 1998; Iocco et al., 2001; Wang et al., 2005; Dutt et al., 2008; Dhekney et al., 2011; Dabauza et al., 2014; Li et al., 2015). AGL1, a hypervirulent Agrobacterium strain (Lazo et al., 1991) has also been used, verifying that agroinfiltration with this strain transformed all the tested genotypes (Urso et al., 2013).

Grapevine tissue necrosis is a regular problem that occurs during or after Agrobacterium transformation. This browning is triggered as a response to the bacteria and it was reported to be cultivar-specific (Bouquet et al., 2008). Polyphenols can be oxidized by air, peroxidases, or polyphenoloxidases. Peroxidases and polyphenoloxidases have been associated with mechanical injury and response to environmental stress. The stress response may involve the release of polyphenols from vacuoles and the de novo synthesis of phenol (Perl et al., 1996). These authors assume that the hypersensitive response could be due to the oxidation caused by high levels of peroxidase activity. Perl et al. (1996) improved plant viability and inhibited tissue necrosis in cv. “Superior Seedless” by using a combination of polyvinylpolypyrrolidone and dithiothreitol. A still unexplored alternative that could reduce this browning is the use of Agrobacterium vitis strains (A. tumefaciens biovar 3), which is a natural grapevine pathogen (Kikkert et al., 2001).

The differentiation of the positive transformation events among all the non-transformed plant tissues after infection with the Agrobacterium strain is a crucial step of the procedure. This is possible due to the action of marker genes, which are usually integrated together with the genes engineered to be expressed in the plant. There are two main types of marker genes: selection marker genes and reporter genes.

The use of marker genes allows the cells or plants carrying them to be selected in the presence of a selective agent, such as an herbicide or an antibiotic. The neomycin phosphotransferase (nptII) gene has been the most widely used in grapevine. This gene confers resistance to aminoglycoside antibiotics, such as kanamycin (Kan), paramomycin, or neomycin (Nakano et al., 1994; Scorza et al., 1995, 1996; Yamamoto et al., 2000; Iocco et al., 2001; Vidal et al., 2003, 2006; Bornhoff et al., 2005; Gambino et al., 2005; Wang et al., 2005; Agüero et al., 2006; Li et al., 2006, 2015; Dhekney et al., 2008, 2011; López-Pérez et al., 2008; Jin et al., 2009; Gago et al., 2011; Dabauza et al., 2014). Although kanamycin is the most widely selective agent used in grapevine transformation, this crop shows a high sensitivity to this antibiotic, and it is generally difficult to find a balance between the appropriate concentration for selection, without losing the viability of the embryos and plants. Another gene frequently used in grapevine transformation is hygromycin phosphotransferase (hptI), which confers resistance to the antibiotic hygromycin (Hyg) (Franks et al., 1998; Torregrosa et al., 2000; Fan et al., 2008; Nirala et al., 2010; Nookaraju and Agrawal, 2012; Dai et al., 2015). Different strategies have been assayed regarding the application timing and the optimal concentration of kanamycin (Franks et al., 1998; Iocco et al., 2001; Wang et al., 2005) and hygromycin (Franks et al., 1998; Fan et al., 2008; Nirala et al., 2010; Dai et al., 2015). Moreover, Saporta et al. (2014) made a comparison of Kan and Hyg in “Albariño” cell suspensions. Since grapevine antibiotic sensitivity showed a high genotype-specific behavior (Gray and Meredith, 1992; Torregrosa et al., 2000), sensitivity-specific assays are required for every new genetic transformation platform.

On the other hand, reporter genes give transformed plants an easily recognizable and measurable selection characteristic or phenotype. Once these genes are integrated, they allow us to know where they are expressed, in what quantity, when, and in which tissues they are transcribed. The most widely used reporter gene is the uidA gene that encodes β-glucuronidase (GUS) (Jefferson et al., 1987). This protein hydrolyzes substrates such as X-Gluc (5-bromo-4-chloro-3-indoxyl-beta-D-glucuronide) and causes a blue precipitate. The disadvantage of this method is that it is usually destructive. Alternatively, the green fluorescent protein gene (GFP), which encodes a protein that generates a chromophore emitting green fluorescence when excited by blue light or ultraviolet light, is commonly utilized as a non-destructive reporter system (Chiu et al., 1996). Both genes have been extensively used in grapevine genetic transformation (Baribault et al., 1989; Nakano et al., 1994; Scorza et al., 1995; Franks et al., 1998; Iocco et al., 2001; Das et al., 2002; Wang et al., 2005; Dutt et al., 2007, 2008; López-Pérez et al., 2008; Gago et al., 2011). More recently, He et al. (2020) reported RUBY as a non-invasive reporter that could be especially useful for monitoring gene expression in tissue culture experiments under sterile conditions in large crop plants such as fruit trees.

Alternatively to the plant transformation mediated by A. tumefaciens, several works reported the use of biolistic methodologies (Torregrosa et al., 2002; Vidal et al., 2003, 2006; Joubert et al., 2013). Biolistic is a physical transformation method, which consists of the high-speed projection of microparticles, usually gold and tungsten, impregnated with DNA (Sanford et al., 1987). The first grapevine work using biolistics was reported by Hébert et al. (1993), in which they transformed “Chancellor,” a Vitis complex interspecific hybrid. On the other hand, Scorza et al. (1995) carried out the first transformation of seedless table grapes using this technique. Nowadays, this technique is a useful transient expression tool for functional analysis in various plant materials, such as, cell suspension culture, leaf sections, and somatic embryos (Vidal et al., 2010; Jelly et al., 2014; Dalla Costa et al., 2019). A biolistic protocol based on the transient genetic transformation of “Cabernet Sauvignon” cell suspensions was developed by Torregrosa et al. (2002) to analyze the effect of anaerobiosis on the regulation of the expression of the VvAdh gene in response to anaerobiosis. Finally, biolistics transformation was also used to study the regulation of the defense gene VvPGIP in leaves sections of “Chardonnay” and ‘Thompson Seedless” somatic embryos (Joubert et al., 2013). One of the most important advantage of biolistic techniques with respect to the infection with A. tumefaciens is probably to skip the treatment with antibiotics to eliminate the bacteria, and in the case of grapevine, to avoid the before mentioned hypersensitivity reaction caused by infection. As disadvantages, low penetration depth, random integration, and putative damage to target tissue were also reported (Cunningham et al., 2018). Particle bombardment is still quite difficult to perform and requires the fine tuning of a series of critical variables such as helium pressure, particle diameter, cartridge preparation, or distance from target plant material. Additionally, purchasing a biolistic device and consumables can be expensive (Jelly et al., 2014).

The technical and biological problems mentioned above, together with the strong rejection of the consumers and the regulation of the appellations of origin, have prevented the wide development of grapevine genetic transformation. Furthermore, due to the scarce natural genetic resistance/tolerance of V. vinifera genotypes, stable transformation has been mainly oriented to the improvement of resistance to pathogens and insects (Vivier and Pretorius, 2002; Table 1). Most of the reports on fungal resistance have focused on the use of pathogenesis related proteins (PR), among which, glucanases and chitinases stand out. On the other hand, the accumulation of phytoalexins and stilbenes has been a proven strategy used to obtain resistance to fungi (Fan et al., 2008; Dabauza et al., 2014; Dai et al., 2015). The improvement of plant resistance to bacteria has been targeted by using antimicrobial genes like lytic peptides (Scorza et al., 1996; Vidal et al., 2003, 2006; Dandekar et al., 2012; Li et al., 2015). The insertion of virus capsid proteins (virus coat proteins; CP) has been used to increase the resistance to viruses such as GFLV, GVA, or GVB (Gölles et al., 1997; Tsvetkov et al., 2000; Gambino et al., 2005). Finally, the introduction of resistance to insects like root-knot nematodes or the grapevine phylloxera have been attempted by means of hairy roots transformation (Franks et al., 2006; Yang et al., 2013). In addition, the improvement of tolerance to abiotic stress has also been studied. Some of the most troublesome stress problems that have been addressed are resistance to cold (Jin et al., 2009; Tillett et al., 2012) or different sources of oxidative damage (Zok et al., 2009).

A compilation of most of the works carried out on stable transformation of V. vinifera, including the transformation methods, marker genes, and antibiotics used for selection are shown in Table 1. It is worth mentioning that most of the grapevine genetic transformation studies have used varieties such as “Thompson Seedless” or “Chardonnay,” highlighting the genotype-specificity of this procedure.

Plants are sessile organisms that are dependent on the living conditions of the environment around them, and for this reason plants have developed great plasticity to accommodate environmental effects by altering metabolism or development (Fehér et al., 2003). Pluripotency and totipotency, exceptional properties for tissue culture techniques, have contributed to several biotechnological applications. Pluripotency refers to the ability of one cell type to form another cell type, tissue, or organ, while totipotency refers to the ability of a single cell to develop, through embryogenesis, into a complete organism (Jha and Kumar, 2018). The totipotency theory was first proposed by Guttenberg (1943), but regeneration protocols were established after Skoog and Miller (1957) introduced changes in the concentration of auxins and cytokinin in the culture media. Since then, it has been possible to establish shoot regeneration for many plants (Lardon and Geelen, 2020).

The most common types of regeneration in plants are somatic embryogenesis and de novo organogenesis (Pulianmackal et al., 2014; Kareem et al., 2016). During somatic embryogenesis, dedifferentiated cells generate bipolar structures where it is possible to differentiate root and shoot meristems (Pulianmackal et al., 2014; Xu and Huang, 2014; Horstman et al., 2017; Méndez-Hernández et al., 2019). This process is achieved through abiotic stress induction or by the addition of auxins. Consequently, zygotic embryogenesis-like structures are formed, due to the action of transcription factors such as LEAFY COTYLEDON (LEC) 1 and 2, AGAMOUS- LIKE 15 (AGL15), FUSCA 3 (FUS3), BABYBOOM (BBM), and EMBRYOMAKER (EMK) (Horstman et al., 2017; Méndez-Hernández et al., 2019). On the other hand, de novo organogenesis consists of the formation of new meristems from pluripotent stem cells to build organs (Xu and Huang, 2014). This process is governed by the plant hormones auxin and cytokinin and a transcriptional cascade involving WUSCHEL-RELATED HOMEOBOX (WOX) 11 and 12, WOX5 and 7, and LATERAL ORGAN BOUNDARY DOMAIN (LBD) 16 and 29, and SHOOT MERISTEMLESS (STM) (Liu et al., 2014; Xu, 2018).

On the other hand, GROWTH-REGULATING FACTOR (GRF) genes were reported as plant-specific transcription factors involved in the establishment and maintenance of meristems and in the cellular proliferation of developing primary organs (Liebsch and Palatnik, 2020). GRF proteins interact with a transcription cofactor, GRF INTERACTION FACTOR (GIF), forming a functional transcriptional complex (Kim and Kende, 2004). GIFs can act as transcriptional coregulators enhancing the activity of GRFs. MicroRNA396 (miR396) is a conserved miRNA that recognizes a complementary sequence in GRF mRNA from seed plants. Finally, GRF expression is regulated by posttranscriptional repression by mir396 (Rodriguez et al., 2016). The combinatory action of the miR396-GRF/GIF system in regulating plant growth makes it a very valuable tool for improving crops of agronomic interest.

Up to now, most of the work carried out in grapevine genetic transformation has had relative success due to the problems related to the transformation processes and the difficulty in plant regeneration. This process is generally time-consuming and involves several laborious steps (Figure 1A).

As shown in the present review, the preferred transformation method for stable grapevine transformation is Agrobacterium, and that most of the works have focused mainly on the use of few grapevine cultivars like “Chardonnay” and “Thompson Seedless.” Regarding the use of antibiotics for the selection of transformants, the determination of the optimal concentration for each genotype emerged as a necessary step since antibiotic sensitivity showed to be genotype-dependent. Accordingly, the possibility of using reporter genes such as GFP or RUBY as a control of the transformation appears as an attractive alternative. Therefore, the use of new technologies and their combination is required to facilitate the recovery of many plants in a large number of grapevine cultivars. In this sense, the application of the technologies proposed by Debernardi et al. (2020) and Maher et al. (2020) would be very useful to increase regeneration rates (Figures 1B,C). As previously mentioned, the tissues used for transformation are sensitive to infection with A. tumefaciens. For this reason, the use of nanoparticles-derived delivery systems, such as CDs, emerges as an alternative to overcome this problem with the advantage that it would allow to extend the range of hosts. This work proposes the possibility of combining the technologies developed by Debernardi et al. (2020) and Maher et al. (2020) together with the delivery of vectors mediated by nanoparticles, with the aim of overcoming problems and limitations related to the classical methodology of grapevine transformation and plant regeneration (Figure 1D).

GC and CC wrote the manuscript. SM provided critical feedback. DL conceived and performed the final edition of the manuscript. All the authors approved the final version of the manuscript.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.