1.1 Approaches for achieving “gene stacking” in plants

In plants, responses to abiotic stresses such as drought, require alteration in numerous genes and their interconnected pathways. Several strategies have been in practice to stack or pyramid the desired transgenes in different crops. Broadly speaking, two types of gene stacking approaches are used: the conventional approach which is the marker-assisted selection (MAS), and the advanced molecular stacking. However, in the conventional approach, nonlinking and segregation of transgenes as well as development of homozygous lines is still a challenge (Que et al., 2010). A recent system for flexible and in vivo stacking of multiple genes within a T-DNA of Agrobacterium, termed as GAANTRY, has also been reported (Collier et al., 2018). Figure 1 presents a bird's eye view of these approaches being used for gene pyramiding in crops and the following text briefly describes the same.

1.2 Gene pyramiding via MAS

Pyramiding of key genes into a single line for enhanced stress tolerance is a powerful strategy in crop breeding. To develop multi-stack hybrids, gene stacking is a strategy to successfully introduce one or more transgenes by crossing the desired parents having different transgenes. The conventional crop breeding strategies utilize traditional techniques and routine natural processes that require extensive field trials (Dormatey et al., 2020; Janila et al., 2016). These strategies include recurrent selection, iterative hybridization, backcross breeding, and pedigree crossing (Figure 1). However, it is difficult to trace the presence of targeted genes, limiting the selection of desired progeny (Malav et al., 2016). Pyramiding of genes/Quantitative Trait Loci (QTLs) to confer tolerance to drought stress is now feasible due to advancements in molecular markers (Das & Rao, 2015). Plant breeders use MAS to transfer genes from pyramided lines into the required crop to enhance the proportion of the desired population of improved genotypes (Das & Rao, 2015; Magar et al., 2014; Pradhan et al., 2015; Shamsudin et al., 2016). By using MAS, plant scientists successfully utilize genes/QTL allele-linked markers to improve qualitative as well as quantitative traits against abiotic stress such as drought, besides monitoring these polygenic traits at a time (Dormatey et al., 2020). There are several recent reviews published on this approach (for reference see Gouda et al., 2020) and therefore, presenting the details of this approach is beyond the scope of this review.

1.3 Gene pyramiding via molecular stacking

Molecular gene stacking or simultaneous gene pyramiding is another strategy where a few genes are introduced at the same time in a plant using transgenic approaches such as Agrobacterium-mediated transformation or by biolistic methods (Keshavareddy & Kumar, 2018; Srivastava et al., 2017). The molecular gene stacking can be achieved in plants via two routes: co-transformation or re-transformation, for which the details are presented below.

1.4 Co-transformation mediated gene pyramiding

Co-transformation is one of the most efficient strategies for the introduction of multi genes in plants. It can be done either by multiple plasmid co-transformation of unlinked transgenes consisting of various plasmids, (or discrete DNA fragments) or by single plasmid co-transformation of the linked transgene, where each transgene is linked with its own promoter and they are transformed together into the plant via Agrobacterium-mediated transformation or the biolistic method (Gupta et al., 2018, Segolela et al., 2019; Figure 1). A major advantage of this method is that the co-introduced transgenes tend to co-integrate at the same chromosomal position in a high proportion of transgenics. In contrast, it is difficult to assemble complex plasmids with multiple gene cassettes using similar promoters. Moreover, the integration of the transgene in high copy number and undesirable incorporation of T-DNA segments has also been reported in many cases (Douglas & Halpin, 2010; Salim et al., 2020).

Recent advancements in molecular biology have enabled the development of more useful and highly refined tools such as GAANTRY (Gene Assembly in Agrobacterium by Nucleic acid Transfer using Recombinase technologY) which can overcome some of the above-mentioned limitations of gene pyramiding. GAANTRY is an excellent platform for genetic engineering of plants by stably stacking multiple genes in the T-DNA of Agrobacterium plasmids (Collier et al., 2018). This system was designed to efficiently utilize cloning vectors and sequential site-specific recombination events to insert transgenes of interest in controlled orientation within a T-DNA and swapping of selection markers carried on an Agrobacterium virulence plasmid, eliminating the need for a binary vector. GAANTRY is a simple, efficient, inexpensive, high-fidelity procedure, and demonstrated to assemble and stably maintain extremely large T-DNAs by sequentially stacking genes to produce high-quality transgenic events of low copy number and free of backbone contaminations. Using this system, researchers have generated high-quality transgenic events by stacking 10 cargo sequences together (Collier et al., 2018).

1.5 Re-transformation mediated gene pyramiding

Re-transformation is a multi-trait or combined trait event, where a plant harboring a transgene is transformed several times with multiple genes. In this method of gene pyramiding, there are fair chances for the transgenes to get inserted at multiple loci and each event critically requires a separate selectable marker (Figure 1). This method is primarily used in crops that are uneasy to propagate through sexual cultivation. However, the induction of transgenes and the requirement of a range of selectable markers limits its utilization. Further, it is difficult to retain transgenes in the offspring after genetic segregation. In this regard, marker-free transformation provides an alternative approach which is also considered to be more environment-friendly (Joshi et al., 2020).

1.6 Gene pyramiding for enhanced tolerance to drought

In terms of plant responses to abiotic stresses, salinity and drought are strongly interconnected as similar cellular destruction and related signaling pathways have been reported to be activated (Urao et al., 2000; Zhu, 2001; Zhu, 2002). Plants face critical osmotic and mechanical injuries during drought stress, leading to the production of reactive oxygen species (ROS) that denature cellular proteins (Ahmad et al., 2010; Begum et al., 2019; Kosar et al., 2020; Thomashow, 1999; Tripathy et al., 2012). Although single gene engineering has been found to be useful towards improving drought tolerance in crops, future improvements of complex traits will likely require and benefit from the introduction of multiple genes. The transformation of the desired crops with the genes involved in drought and salt stress in a predictable and precise manner has been achieved by genetic engineering (Lawlor, 2013). A set of stress-responsive genes have been reported getting activated in response to drought and salinity stress which in turn leads to the activation of several stress-related proteins and the accumulation of protective metabolites. These stress-responsive genes are the primary targets for genetic manipulation to enhance drought tolerance in crops. Regarding this, initial studies focused on the single gene transgenics, including the manipulation of genes that targeted only metabolite accumulation to confer tolerance against salinity or drought stress (Bajaj et al., 1999; Blumwald, 2003). However, attempts have also been made to target pyramiding of multi-genes that can enhance endurance several folds. Some of the representative success stories of attaining improved tolerance against drought and salinity stress by using the gene pyramiding approach are listed in Table 1.

1.7 Gene pyramiding for engineering metabolic pathways

Methylglyoxal (MG) is a cytotoxic metabolite produced in response to abiotic stress in plants, including salinity and drought (Kaur et al., 2014). As a byproduct of the glycolysis pathway (Sousa et al., 2013), MG leads to ROS generation in cells (Singla-Pareek et al., 2003). The glyoxalase pathway involves two enzymes, that is, glyoxalase І (GLY I) and glyoxalase II (GLY II) that detoxify MG (Singla-Pareek et al., 2003, Mustafiz et al., 2011; Figure 1). Initial attempts for gene pyramiding for the two key genes of this metabolic pathway were carried out in the model plant tobacco followed by rice (Gupta et al., 2018; Singla-Pareek et al., 2003). The GlyI gene was cloned from Brassica juncea and the GlyII gene was cloned from Oryza sativa into plant transformation vectors under the control of the 35S cauliflower mosaic virus (CaMV) promoter. The construct was transformed into tobacco through Agrobacterium-mediated transformation. In the process, both single and double-transgenic plants were generated. The transgenic plants carrying overexpression of both the Gly genes were found to show a synergistic effect for enhanced stress tolerance than the plants carrying overexpression for either of the single genes of the pathway (Singla-Pareek et al., 2003). This proof of concept in tobacco led to the translation of this technology to rice. Attempts were made to engineer the glyoxalase pathway in a high yielding rice IR64 (Gupta et al., 2018). The GlyI gene from Brassica juncea and the GlyII gene from Oryza sativa under the 35S CaMV promoter were cloned together into the pCAMBIA1304 vector (Gupta et al., 2018). The construct was transformed into rice through Agrobacterium-mediated transformation, and overexpression lines were generated. Higher amounts of GlyI and GlyII proteins and low accumulation of MG in the transgenic plants were observed compared to wild-type (WT) plants, even under nonstress conditions, thus confirming that both the transgenes were stable and functionally active (Gupta et al., 2018). Surprisingly, 80% of transgenic seeds showed germination in the presence of 200 mM NaCl and also were able to maintain their greenness at the vegetative stage. The double transgenic plants, overexpressing both GlyI and GlyII genes together, exhibited less chloroplast damage and 70% chlorophyll retention under salinity stress. ROS accumulation was significantly lower in transgenic plants as compared to WT under salinity conditions. At the same time, the glyoxalase-overexpressing transgenic plants showed high shoot and root biomass as compared to WT plants under water-deprived conditions. However, typical drought stress symptoms like leaf rolling and chlorophyll bleaching were observed under 25 days of drought conditions in the transgenic lines. Upon rewatering, these symptoms disappeared and healthy transgenic plants were observed (Gupta et al., 2018). A similar attempt was made in tomato where lower lipid peroxidation and ROS production were observed, representing enhanced salt tolerance in the transgenic plants (Viveros et al., 2013). These reports clearly demonstrated the advantage of gene pyramiding in tackling drought and salinity stress as compared to single gene manipulation. Another important finding of the study was the fact that even though the two transgenes were regulated under the control of a constitutive 35S promoter, it did not result in any yield penalty or silencing of the transgenes. In contrast, the use of the same promoter to derive the transgene constitutively has been shown to have a yield penalty in rice (Xiao et al., 2007) and/or gene silencing, when present in more copies in several plants species (Matsunaga et al., 2019). Further investigations need to be done in this area to conclude the usefulness of the same promoter for gene pyramiding programs.

In another set of experiments, the glutamine synthetase (GS) enzymes which catalyze critical nitrogen metabolic reactions in plants, were attempted for gene pyramiding. GS has two isoforms, the cytosolic GS1 which is responsible for the primary assimilation of inorganic nitrogen available from the soil in the form of nitrate or ammonia, and the re-assimilation of NH₄⁺ released by protein degradation (Bernard & Habash, 2009). On the other hand, the chloroplastic GS2 is responsible for re-assimilation of NH₄⁺ released during photorespiration and nitrate reduction (Lam et al., 1996; Leegood et al., 1995). In a recent report by James et al. (2018), the cytosolic OsGS1 and chloroplastic OsGS2 genes were co-expressed in rice (Figure 2). The genes were cloned in the pMDC99 vector next to the actin promoter and then transformed into rice using agrobacterium. The dual-overexpression lines showed higher chlorophyll content, fresh weight, and relative water content compared to WT under drought stress. The transgenic plants showed very high proline levels while the malondialdehyde (MDA) content and electrolyte leakage were reported to be lower in pyramid lines as compared to the single transgenic plants (James et al., 2018).

1.8 Gene pyramiding for engineering antioxidant pathways

Ascorbic acid (AsA), also known as vitamin C, is an antioxidant involved in ROS scavenging, programmed cell death, premature senescence, cell division, and elongation, and oxidative stress tolerance mechanisms (Akram et al., 2017). Li et al. (2019) engineered the ascorbic acid metabolic pathway by pyramiding guanosine diphosphate (GDP)-mannose 3,5-epimerase (GME), GDP-D-mannose pyrophosphorylase (GMP), GDP-L-Gal phosphorylase (GGP), and L-Gal-1-P phosphatase (GPP) genes into tomato plants. These genes were cloned under the control of CaMV35S into the pMV vector and transformed into Agrobacterium tumefaciens by electroporation. They observed that multi-gene overexpression lines had increase in ascorbic acid content compared to the single-gene transgenic lines. Furthermore, they discovered that with an increase in AsA, the tolerance to oxidative stress was also enhanced in pyramided lines (Li et al., 2019). Recently, it has also been demonstrated that exogenously applied ascorbic acid has a positive effect on the growth of safflower plants under drought stress (Farooq et al., 2020).

Superoxide dismutase (SOD), a ROS scavenging enzyme detoxifies hydrogen peroxide (H₂O₂) into water and oxygen with the help of the ascorbate peroxidase (APX; Asada, 1999). Shafi, Gill, et al. (2019a) developed transgenic Arabidopsis lines overexpressing cytosolic Cu/Zn-SOD and APX genes isolated from Potentilla atrosanguinea and Rheum australe, respectively. The genes were individually overexpressed in Arabidopsis thaliana and later crossed to obtain dual transgenic lines. The transgenic lines showed enhanced antioxidant enzyme activity, callus induction, and shoot regeneration (Shafi, Gill, et al., 2019a) besides higher cellulose biosynthesis under salinity stress (Shafi et al., 2015; Shafi, Zahoor, et al., 2019b). Similarly, potato plants overexpressing PaSOD and RaAPX showed higher activities of SOD and APX, increased starch accumulation, improved growth, enhanced expression of genes and transcription factors directly involved in lignin biosynthesis, and reduced accumulation of ROS under salt stress (Shafi et al., 2017). Further, transgenic Festuca arundinacea (Lee et al., 2007) and potato plants (Tang et al., 2004) overexpressing Cu/Zn-SOD and APX in chloroplasts under the control of the oxidative stress-inducible promoter SWPA2 showed reduced thiobarbituric acid reactive substances (TBARS), ion leakage, and chlorophyll degradation under methyl viologen, H₂O₂, and heavy metal stress. In addition, enhanced root growth under salt stress was observed in sweet potato plants overexpressing Cu/Zn-SOD and APX under the control of the SWPA2 promoter (Yan et al., 2016).

1.9 Gene pyramiding for engineering ion sequestration in cells

One of the most important components of high salinity is the ionic stress which is due to accumulation of excess of Na⁺ ions in the cell. Excessive Na⁺ disrupts the K⁺ homeostasis and interferes with numerous metabolic processes that affect plant growth and development (Assaha et al., 2017; Kronzucker et al., 2013; Nieves-Cordones et al., 2016). Therefore, maintaining a balanced cytosolic Na⁺/K⁺ ratio has become a crucial target to improve salinity tolerance. One way to achieve this target is to overexpress the genes involved in ion homeostasis, for example, salt overly sensitive (SOS) genes (Shi et al., 2003), AVP1 (Gaxiola et al., 2001), and H⁺-ATPases and vacuolar Na⁺/H⁺ exchanger (AtNHX1) genes (Apse et al., 1999; Zhang et al., 2001). The ion homeostasis pathway includes SOS1, SOS2, and SOS3 genes in Arabidopsis (Zhu, 2001). It has been reported that the transformation of SOS1, SOS2, and SOS3 genes together in yeast results in higher tolerance against salt stress in contrast to single SOS gene transformation (Quintero et al., 2002). AtNHX1 is a vacuolar protein involved in Na⁺ transport into vacuoles (Apse et al., 1999), and its activity is controlled by SOS genes (Qiu et al., 2004). As SOS1 and AtNHX1 activity depend on SOS2 and SOS3 expression, scientists developed overexpression lines of AtNHX1 + SOS3, SOS2 + SOS3, and SOS1 + SOS2 + SOS3 in A. thaliana and observed lower accumulation of Na⁺ in the transgenic plants under salinity conditions (Yang et al., 2009; Figure 2). Unexpectedly, the AtNHX1 + SOS3 and SOS1 + SOS2 + SOS3 lines did not show greatly improved salt tolerance compared to single transgenic plants. Furthermore, the same attempt to generate transgenic lines co-expressing SOS1 + SOS2 + SOS3 was made in tall fescue (Ma et al., 2014). The three SOS genes, under the control of the stress-inducible rd29A promoter were cloned together into a single vector. Enhanced tolerance against salt stress was observed in the transgenic plants with pyramided genes. Additionally, SOS1 + SOS2 + SOS3 overexpression lines showed less growth retardation, fewer yellow leaves with decreased death rate, indicating a healthy phenotype under salinity conditions as compared to the WT plants (Ma et al., 2014). Likewise, Pehlivan et al. (2016) developed transgenic lines co-expressing AtNHX1 + SOS1 and observed a highly enhanced salinity tolerance (250 mM NaCl) in Arabidopsis. A similar approach was used by Fan et al. (2019) where they co-expressed SOS1 and H⁺-ATPase (AHA1) genes from Sesuvium portulacastrum into Arabidopsis to enhance salinity tolerance (Fan et al., 2019).

Sushma et al. (2014) co-expressed PgNHX1 and AtAVP1 in rice where enhanced salt and drought stress tolerance was observed. The PgNHX1-AtAVP1 overexpressing plants showed root growth enhancement, a lower percent reduction in chlorophyll content, higher cell viability, higher cell membrane stability, and stabilized higher K⁺/Na⁺ ratios under NaCl and PEG treatment. These results clearly state that the PgNHX1-AtAVP1 overexpressing lines had higher tolerance against salinity and drought stress as compared to the WT plants (Sushma et al., 2014). Recently, Biradar et al. (2018) attempted pyramiding of the salt responsive protein 3–1 (SaSRP3-1) gene and the vacuolar H⁺-ATPase subunit c1 (SaVHAc1) from Spartina alterniflora (halophyte grass) into rice. SaSRP3-1 and SaVHAc1 were independently transformed into pCAMBIA 1305.2 vectors and later used to develop a single transgenic plant. After that, a double transgenic plant harboring SaSRP3-1 and SaVHAc1 genes was developed by crossing individual transgenic lines of SaSRP3-1 and SaVHAc. They observed improved salinity tolerance in the double transgenic plants at the seedling, vegetative, and reproductive stages. Increased K⁺/Na⁺, chlorophyll content, RWC, and shoot-root growth was also reported in double transgenic plants as compared to the WT under salt stress (Biradar et al., 2018).

1.10 Gene pyramiding for engineering osmolyte accumulation

Betaine is an osmolyte known to have a role in salinity tolerance. Zhou et al. (2008) co-expressed the betaine synthesis gene (BADH) from the halophyte Atriplex hortensis with the NHX1 gene from the halophyte Salicornia europaea in tobacco. Tobacco transformation was carried out using the agrobacterium vector pBinBS having different cassettes of BADH and SeNHX1 under the CaMV35S promoter. They observed higher accumulation of Na⁺ and betaine in leaves and vacuoles in the double transgenic plant as compared to single transgenics under salinity stress. Also, the osmotic pressure and biomass were high in co-transformed plants (Zhou et al., 2008). Similarly, Duan et al. (2009) generated overexpression lines of tobacco to enhance their salinity tolerance, harboring the betA (betaine) gene from Escherichia coli encoding choline dehydrogenase and the AtNHX1 gene from A. thaliana. Firstly, single transgenic lines were generated using Agrobacterium-mediated transformation, followed by the double transgenic plant through the cross-breeding of the betA and AtNHX1 transgenic plants. They observed highly improved tolerance in the double transgenic plants since the two genes belong to different salt tolerance pathways (Duan et al., 2009). Similarly, a transgenic maize plant co-expressing betA from Escherichia coli and TsVP from Thellungiella halophila was also developed and tested for its improved drought tolerance (Wei et al., 2011). The authors observed higher glycine betaine content in pyramided lines as compared to the WT and single transgenic lines under drought stress conditions. Moreover, the solute potential of pyramided lines was found to be more negative than the WT and single transgenic lines after drought treatment, indicating that pyramided lines had great potential for osmotic adjustment to facilitate water uptake (Wei et al., 2011). Another group attempted the pyramiding of five genes, including the Vitreoscilla globin gene (vgb), the levansucrase gene (SacB), jasmonic acid ethylene-responsive element (JERF36), the Bacillus thuringiensis crystal protein (BtCry3A), and oryzacystatin I (OC-I) into a hybrid poplar plant and observed enhanced tolerance against drought and salinity conditions (Su et al., 2011). Populus × euramericana “Guariento” hybrid obtained from crossing Populus nigra and Populus deltoids was used in the study. All the five genes were regulated by the CaMV promoter and were transformed into hybrid poplar using particle bombardment. The authors reported that the combination of five genes in hybrid poplar conferred tolerance to waterlogging, drought, and salinity conditions and also was resistant to insects (Su et al., 2011).

Nguyen et al. (2013) cloned the abscisic acid-inducible protein PHV A1-Hordeum vulgare gene (HVA1) and the mannitol-1-phosphate dehydrogenase gene (mtlD) in maize and reported enhanced drought and salt stress tolerance. The HVA1 and mtlD genes were cloned in pBY520 and JS101 vectors, respectively, under the control of the actin promoter of rice (Act1) and the potato protease inhibitor II terminator (pin II). These constructs also contained a selectable marker under 35S CaMV promoter and nopaline synthase (Nos) terminator. Transformation into maize calli was done by bombardment technique using tungsten particles coated with plasmids containing the desired genes. WT plants and single/double transformants were subjected to drought (15 days no watering) and salinity stress (upto 300 mM NaCl). They observed 67% survival for the double transgenic plants, that is highest compared to single transgenics (52 and 45% of HVA1 and mtlD) and WT (35%) under drought stress. Also, the relative water content (RWC) was 85% in HVA1 + mtlD, 81% in HVA1, 77.6% in mtlD transgenic plants in contrast to the RWC (57.1%) reported in the WT plants. The percentage of reduction in shoot fresh weight and shoot and root dry weights of the HVA1 + mtlD plants were 13, 18.4, and 21.4%, respectively. These reductions were 32.1, 14.8, and 25.0% for HVA1 transgenic plants and 25.9, 27.0, and 21.9% for the mtlD transgenic plants, respectively. This study showed that the double transgenic plants were surviving better than the WT and accumulated higher mannitol under drought conditions as compared to single transgenics (Nguyen et al., 2013). He et al. (2011) enhanced glycinebetaine synthesis in transgenic tobacco by co-overexpressing two bacterial (Aphanothece halophytica) genes, that is glycine sarcosine methyltransferase gene (ApGSMT2) and dimethylglycine methyltransferase gene (ApDMT2). Both lower the genes were cloned into the pCUE-GSD vector under the control of the CaMV35S promoter, and the selection marker was for 5-enolpyruvylshikimate-3-phosphate synthase (EPSPs) gene, which provides tolerance to the herbicide glyphosate. Agrobacterium-mediated transformation was used to generate double-transgenic lines. The dual-overexpression lines showed lower electrolyte leakage and lower MDA content. Higher glycine betaine content and chlorophyll content under drought stress in transgenic plants indicated enhanced osmoprotection as compared to the WT plants (He et al., 2011). Later, the same team performed the pyramiding of these genes in maize which again resulted in enhanced drought tolerance (He et al., 2013).