Formidable challenges facing humankind include a burgeoning global human population (Godfray et al., 2010; Gerland et al., 2014) and the increasing frequency and intensity of droughts related to global warming (Cook et al., 2014; Singh et al., 2015). In addition, abiotic stresses including high salinity, temperature extremes, increased UV radiation, heavy metals, and high light intensities are and will continue to be major constraints for global crop production and food security (Lesk et al., 2016). Among these abiotic stresses, drought is of major concern as it has dire effects on crop productivity (Fahad et al., 2017), plant growth, and development (Yordanov et al., 2000). By the end of the 21st century, rapid changes in the global climate will likely increase the frequencies of drought by more than 20% compared to current rates (Lobell et al., 2011; Cook et al., 2014). Indeed, Daryanto et al. (2016) showed that an approximately 40% decrease in water availability can decrease wheat (Triticum aestivum L.) and maize (Zea mays L.) yields by 20.6% and 39.3%, respectively. High salinity, another abiotic stress that is harmful to crop production, affects 20% of total cultivated and 33% of irrigated agricultural lands worldwide (Shrivastava and Kumar, 2015). Similarly, low temperatures and heatwaves cause significant reduction in crop yields across the world (Sanghera et al., 2011; Challinor et al., 2014; Hatfield and Prueger, 2015). Thus, novel approaches to mitigate the negative impacts of abiotic stresses on crop yields must be explored and developed to avoid socio-economic collapse due to climate change.

Approaches to enhance sustainable bioenergy production by engineering the CAM pathway into C₃ crops to enhance their water-use efficiency (WUE) on marginal lands are already underway (Borland et al., 2014, 2015; Yang et al., 2015; Liu et al., 2018). Another approach to enhance abiotic stress tolerance is to modulate the expression of transcription factors (TFs) or the functions of abiotic stress-adaptive genes that might already be present, but that are not normally expressed in unstressed or C₃ plants (Hussain et al., 2011; Rabara et al., 2014; Joshi et al., 2016; Wang et al., 2016; Bechtold, 2018). This approach would involve bioengineering a small number of regulatory genes with potentially global effects made possible by the role of TFs in gene regulation (Rabara et al., 2014; Joshi et al., 2016).

Transcription factor-based genetic engineering could direct such regulatory TFs to modulate a large number of downstream abiotic stress-responsive genes (Rabara et al., 2014). Stress tolerance in plants is generally under polygenic control (Tran et al., 2010) and some of the genes regulating stress-tolerance responses happen to code for TFs (Villalobos et al., 2004; Qiu and Yu, 2009; Zhang et al., 2011; Yang et al., 2012; Cai et al., 2014; Chen et al., 2014; Swain et al., 2017). Therefore, TFs might be ideal candidate regulators for improving abiotic stress tolerance in crop plants. To date, TF-based genetic engineering has mainly repurposed TFs from Arabidopsis thaliana, Glycine max, Oryza sativa, and T. aestivum (Rabara et al., 2014; Joshi et al., 2016; Wang et al., 2016). As far as we know, no reports have analyzed the effects of overexpressing TFs from crassulacean acid metabolism (CAM) plants which have greater abiotic stress tolerances than mesophytes. Most CAM plants are naturally adapted to low-water environments and many other abiotic stresses compared to the agronomically important C₃ plants (Borland et al., 2009). Here, we consider using a “next-generation TF-based” approach to exploit abiotic stress-responsive TFs from CAM plants to improve abiotic stress tolerance in crop plants. The current technologies for TF-based approaches to improve plant abiotic stress tolerance have been extensively discussed and reviewed by Rabara et al. (2014), Joshi et al. (2016), and Wang et al. (2016), and those details are therefore only briefly summarized below.

Plants are sessile organisms that exhibit various biochemical, physiological, and molecular adaptations to extreme environments (Joshi et al., 2016). For instance, water-deficit stress activates the expression of stress-responsive genes encoding enzymes that synthesize compatible protective sugars, antioxidants, and proteins, including heat shock proteins and some classes of late embryogenesis abundant (LEA) proteins (Tran et al., 2010; Joshi et al., 2016; Wang et al., 2016). In addition to the stress-induced upregulation of the above proteins, the expression of various regulatory proteins including TFs, protein kinases, and protein phosphatases is also activated (Wang et al., 2016).

Transcription factors are master regulators of many cellular processes and can also interact with other transcriptional regulators (Joshi et al., 2016). Importantly, they play a pivotal role in different abiotic stress responses by binding to the upstream cis-regions of promoters in many stress-responsive genes (Yamaguchi-Shinozaki and Shinozaki, 2006). Many studies have been conducted to identify and characterize families of TFs including AP2/ERFBP, MYB, WRKY, NAC, NF-Y, and bZIP that are involved in abiotic stress responses (Umezawa et al., 2006; Golldack et al., 2011; Leyva-González et al., 2012; Wang et al., 2016; Swain et al., 2017; Zanetti et al., 2017). Several TFs have already been overexpressed in crop plants and A. thaliana to improve abiotic stress tolerance (Qiu and Yu, 2009; Zhang et al., 2011; Yang et al., 2012; Cai et al., 2014; Chen et al., 2014). For example, the NAC family is one of the largest TF families in plants and is involved not only in plant growth and development, but also in transcriptional reprogramming associated with plant stress responses (Tran et al., 2010; Nakashima et al., 2012). Mao et al. (2012) reported that overexpression of the TaNAC2 gene from wheat can enhance tolerance to drought, salt, and freezing stresses in A. thaliana. In addition, functional characterization of the NAC045 (Zheng et al., 2009) and SNAC1 genes from O. sativa enhanced drought and salt tolerance in rice (Hu et al., 2006). Furthermore, overexpression of either GmMYB76 or GmMYB177 from soybean significantly enhanced salt and freezing tolerance in A. thaliana (Liao et al., 2008).

A relatively less explored yet high-potential approach is to discover novel abiotic stress-adaptive regulatory genes in extremophytes (i.e., CAM xerophytes and halophytes, desiccation-tolerant plants, or resurrection plants) to use for bioengineering abiotic stress tolerance in crop plants (Inan et al., 2004; Shi et al., 2013; Joshi et al., 2016; Bechtold, 2018). For example, the overexpression of the TF CpMYB10 from the resurrection plant Craterostigma plantagineum in A. thaliana led to desiccation and salt tolerance in transgenic lines (Villalobos et al., 2004).

CAM plants have evolved a specialized type of photosynthetic adaptation that allows them to live under conditions of severe water deficit and in semi-arid and arid regions of the world including deserts. These plants have shifted all or part of their primary CO₂ uptake and fixation to the nighttime, when evaporative water losses are minimal, and perform C₃ carboxylation reactions when stomata are closed during the daytime. This temporal separation of carbon fixation leads to the formation of the four-carbon organic acid malate, which is stored in the vacuole during the night and subsequently undergoes decarboxylation to release CO₂ for re-fixation during the day to produce carbohydrates (Borland et al., 2009). Because of this temporal separation of carbon fixation and inverted stomatal behavior, CAM plants can reduce water loss due to transpiration. These characteristics also allow CAM plants to fix net CO₂ 15% more efficiently than C₃ plants (Nobel, 1991) resulting in increased biomass of CAM plants while using less water than C₃ plants. Additionally, CAM plants can produce similar amounts of biomass using 80% less water in comparison to C₃ plants (Nobel, 1996; Borland et al., 2009). Thus, CAM plants have between 3- and 6-fold higher WUE than C₄ and C₃ plants, respectively (Garcia et al., 2014; Yang et al., 2015).

In addition to their higher WUE and associated drought tolerance (Yang et al., 2017), CAM plants can tolerate high temperature up to 70°C, whereas C₃ plants can tolerate only 50–55°C (Borland et al., 2009). CAM halophytes can also adapt to high salinity, as during the induction of CAM by salt stress in Mesembryanthemum crystallinum (Winter and Holtum, 2014). Moreover, CAM plants can better tolerate higher light intensities (>1000 μmol m^-2 s^-1) and UV-B irradiation levels than can agronomically important C₃ plants (Borland et al., 2009). Furthermore, CAM plants can increase daily net CO₂ uptake under increased atmospheric CO₂ concentrations, which might be advantageous in global climate change scenarios (Nobel, 1996). Some CAM plants such as Agave salmiana, Opuntia ficus-indica, and Stenocereus queretaroensis can also survive in subzero temperatures, and Agave utahensis Engelm. can tolerate temperatures as low as -18°C (Nobel, 1996).

The abiotic stress-adaptive characteristics of CAM plants will be particularly beneficial for adapting to the consequences of anthropogenic climate change, such as droughts and heatwaves, high soil salinity, temperature extremes, and high light or UV-B irradiation. Many of the stress-adaptive responses involve abscisic acid (ABA)-dependent and -independent response pathways (Song et al., 2016). ABA-dependent and independent signaling events likely participate in the stress-activation of CAM in M. crystallinum (Chu et al., 1990; Taybi and Cushman, 1999, 2002; Abdin et al., 2002; Cushman and Borland, 2002), suggesting that the CAM pathway likely has evolved in response to abiotic stress (Reyes-García and Andrade, 2009; Heyduk et al., 2018; Yin et al., 2018).

Crassulacean acid metabolism is thought to have evolved independently multiple times from ancestral C₃ plants (Silvera et al., 2010) because no unique metabolic pathways are required, although some CAM-specific variant enzymes apparently evolved in some instances (Ermolova et al., 2003; Gehrig et al., 2005; Vaasen et al., 2006). However, temporal changes in gene expression of the CAM enzyme variants likely occurred because of alterations in their regulation compared to their orthologs in C₃ plants (Hermans and Westhoff, 1990; Lepiniec et al., 1993; Cushman et al., 2008; Heyduk et al., 2018; Yin et al., 2018). Furthermore, the ABA-dependent stress response pathway is involved in CAM activation not only in M. crystallinum (Taybi and Cushman, 2002), but also in other CAM species (Taybi et al., 1995; Rodrigues et al., 2016; Yin et al., 2018). Although CAM is found in over 36 families of vascular plants (Silvera et al., 2010), we rely on only a few major CAM species such as pineapple (Ananas comosus), Agave, and Opuntia as agricultural crops to provide food, forage, fiber, and biofuels (Cushman et al., 2015). Well characterized CAM model species also provide abundant resources for the identification and selection of candidate TFs involved in abiotic stress adaptations (Hartwell et al., 2016). Hence, identification of candidate CAM pathway regulators (i.e., TFs) that are expressed or activated under water-deficit stress or CAM should be prioritized to exploit the molecular and regulatory machinery of abiotic stress adaptation in CAM plants as a vital resource for applications in C₃ crop species (Yang et al., 2015; Fernie, 2016; Yin et al., 2018).

Bioengineering a TF that can confer desirable traits such as increased drought tolerance (Villalobos et al., 2004) or increased biomass (Lim et al., 2018) into C₃ A. thaliana will be crucial as a proof of concept for the CAM abiotic stress-responsive TF-based approach to increase abiotic stress tolerance in C₃ plants. Fortunately, genetic resources (i.e., genome and transcriptome sequences) for CAM plants are now available for Agave (Abraham et al., 2016), Kalanchoe spp. (Yang et al., 2017), pineapple (Ming et al., 2015, 2016; Wai et al., 2017), and M. crystallinum (Chiang et al., 2016). Although genetic resources for CAM plants are becoming readily available, the underlying regulatory basis of CAM is still not completely understood. Many TFs of unknown function have been identified during recent genome and transcriptome sequencing efforts; thus, there are now many opportunities to analyze the functions of TFs involved in water-deficit-stress response or CAM function and to exploit the potential of bioengineering using CAM plant TFs to improve abiotic stress tolerance in crop plants. Indeed, candidate CAM TFs involved in C₃ to CAM transition in obligate CAM species of Agave (Heyduk et al., 2018; Huang et al., 2018; Yin et al., 2018), Kanlanchoe (Moseley et al., 2018), and Manfreda (Heyduk et al., 2018), and weak CAM species of Polianthes and Beschorneria (Heyduk et al., 2018), or the induction of CAM in Tralinum triangulare (Brilhaus et al., 2016) have been identified. Not surprisingly, a number of these candidate TFs are involved in the ABA stress responsive pathway (Heyduk et al., 2018; Yin et al., 2018). More importantly though, many of these candidate TFs are involved in the rewiring of the phase shift from C₃ to CAM transition in the evolution of CAM photosynthesis (Heyduk et al., 2018; Moseley et al., 2018; Yin et al., 2018). Although it would be interesting to attempt to reprogram a C₃ plant such that it becomes CAM performing, we are not suggesting to shift gene expression patterns of CAM pathway genes that might be present in extant C₃ plants, or to regulate the engineered CAM pathway in C₃ plants (Yang et al., 2015; Fernie, 2016; Heyduk et al., 2018; Yin et al., 2018), but rather identify and exploit the TFs involved in abiotic stress responses from obligate and inducible CAM plants to modulate the expression of appropriate genes in C₃ plants to improve their abiotic stress tolerance.

Eight most abundant candidate TFs under water-deficit stress diel and zeitgeber time have been identified that might regulate the CAM state, water-deficit stress response, or both in M. crystallinum (Garcia et al., 2014; Cushman, unpubl. data; Table 1). Mesembryanthemum crystallinum switches from C₃ to CAM when salt or water-deficit stressed (Cushman, 2001; Cushman and Borland, 2002; Taybi and Cushman, 2002). Kalanchoe fedtschenkoi orthologs of these top candidate TFs were also highly expressed during CAM induction in older leaf pairs of K. fedtschenkoi plants (Garcia et al., 2014; Cushman, unpubl. data). It is known that young leaves are C₃ performing, whereas mature leaves are CAM performing in K. fedtschenkoi (Cushman, 2001). These TFs share a base mean expression level of >100 FPKM and at least two-fold induction during a transition from C₃ to CAM or imposition of water-deficit stress in K. fedtschenkoi and M. crystallinum, respectively (Garcia et al., 2014; Cushman, unpubl. data). These candidate CAM TFs also belong to the families of TFs reported to be involved in abiotic stress responses (Rabara et al., 2014; Joshi et al., 2016; Roy, 2016; Wang et al., 2016). Orthologs of these candidate CAM TFs in A. thaliana also have several reported functions in plant development and abiotic stress tolerance (Zheng et al., 2009; Zhang et al., 2011; Yang et al., 2012; Chen et al., 2014; Swain et al., 2017; Zanetti et al., 2017). Intriguingly, recent functional characterization of two putative CAM regulators of water-deficit stress response or CAM activation via overexpression in A. thaliana strongly suggest that CAM TFs have high potential to increase tolerance to drought and other abiotic stresses in C₃ plants.

One of these candidate CAM TFs is a myeloblastosis (MYB59, closest ortholog in Arabidopsis) TF whose transcripts are 20-fold more abundant in CAM-performing older leaf pairs relative to C₃-performing younger leaves in K. fedtschenkoi (Hartwell et al. unpubl. data). Results from four, third-generation (T₃) transgenic lines carrying the KlMYB59 indicate increased rosette size and biomass at 4-week-old juvenile stage, and increased shoot length at 8-week-old mature stage in transgenic plants compared to WT (Table 2; Wone et al., unpubl. data; full results being presented in a separate publication). However, transgenic lines show delayed flowering in long-day photoperiod compared to WT plants (16 h light/8 h dark). In addition, these transgenic lines exhibit increased integrated WUE compared to WT plants. Furthermore, transgenic lines have longer primary roots despite exposure to 50 μM selenium compared to WT plants. In addition to MYB59, transcripts of the NAC83 TF (closest ortholog in Arabidopsis) were also more highly expressed in CAM-performing leaves of K. laxiflora and K. fedtschenkoi relative to C₃-performing leaves (Cushman et al., unpubl. data). The function of this K. fedtschenkoi NAC83 TF (KfNAC83) is not known in CAM- or C₃-performing CAM plants, but its A. thaliana ortholog suggests roles in abiotic stress responses and development (Table 1). Functional characterization of KfNAC83 shows enhanced water-deficit stress tolerance and increased integrated WUE in four independent transgenic T₃-generation A. thaliana lines compared to WT plants (Table 2; Wone et al., unpubl. data; full results being presented in a separate publication). Furthermore, KfNAC83-overexpressing lines show significantly increased rosette size, leaves in the mature rosette, shoot biomass, number of siliques, and lateral roots compared to WT. Interestingly, these transgenic lines also showed tolerance to 150 mM NaCl. Collectively, our characterization results strongly suggest that at least two of these candidate CAM TFs have functions in abiotic stress responses and CAM photosynthesis.

The abiotic stress-adaptive features of CAM plants provide a wealth of genetic resources, specifically TFs, that are now available for functional testing and possible improvement of WUE and abiotic stress responses in C₃ photosynthesis plants. Our recent findings strongly suggest that a bioengineering approach using CAM abiotic stress-responsive TFs has the potential to increase abiotic stress tolerance in A. thaliana and possibly in C₃ crop plants. Our results indicate that CAM abiotic stress-responsive gene expression can be modulated by the appropriate CAM TFs to generate stress-adaptive phenotypes in A. thaliana and likely other C₃ plants because these CAM abiotic stress-responsive genes are apparently conserved and present in C₃ plants (Heyduk et al., 2018). Furthermore, although K. fedtschenkoi is distantly related to A. thaliana, transgenic A. thaliana lines carrying the obligate CAM plants’ TFs showed favorable features for translational applications. We are optimistic that overexpressing TFs from the inducible CAM halophyte, M. crystallinum will have similar favorable responses in A. thaliana lines. Co-overexpression of only a small number of obligate and/or inducible CAM plant abiotic stress-responsive TFs with demonstrated abiotic stress-adaptive or -responsive functions would provide a facile approach for bioengineering desirable responses to abiotic stress (Song et al., 2016). Such an approach could open the door to potentially transformative applications to ensure long-term sustainable food, fiber, feed, and fuel production in a projected hotter and drier landscape in the 21st century and beyond.

BWMW conceived the CAM abiotic stress-responsive TF-based approach to improve C₃ crop plant abiotic stress responses. AA, KR, and BWMW wrote the manuscript. AA, KR, and BW conducted the experiments and provided the CAM plant TF overexpression data. WY, TG, and JC provided the CAM plant TF sequences. All authors reviewed the final 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.