Understanding SUMO-mediated adaptive responses in plants to improve crop productivity

Abstract The response to abiotic and biotic stresses in plants and crops is considered a multifaceted process. Due to their sessile nature, plants have evolved unique mechanisms to ensure that developmental plasticity remains during their life cycle. Among these mechanisms, post-translational modifications (PTMs) are crucial components of adaptive responses in plants and transduce environmental stimuli into cellular signalling through the modulation of proteins. SUMOylation is an emerging PTM that has received recent attention due to its dynamic role in protein modification and has quickly been considered a significant component of adaptive mechanisms in plants during stress with great potential for agricultural improvement programs. In the present review, we outline the concept that small ubiquitin-like modifier (SUMO)-mediated response in plants and crops to abiotic and biotic stresses is a multifaceted process with each component of the SUMO cycle facilitating tolerance to several different environmental stresses. We also highlight the clear increase in SUMO genes in crops when compared with Arabidopsis thaliana. The SUMO system is understudied in crops, given the importance of SUMO for stress responses, and for some SUMO genes, the apparent expansion provides new avenues to discover SUMO-conjugated targets that could regulate beneficial agronomical traits.

Multiple studies have demonstrated the importance of the different SUMO components in all eukaryotes from mammals and plants to single-cell yeasts over the past two decades. Deletion of the first identified SUMO homolog suppressor of mif two3 (SMT3) in the budding yeast, Saccharomyces cerevisiae, has shown to result in a loss of cell viability [27]. Similarly, the introduction of single Atsae1, Atsae2 and Atsce1 mutants and double Atsumo1 Atsumo2 and Atsiz1 Atmms21 mutants in A. thaliana plants is embryo lethal [28]. Aside from development, the adaptive response of SUMO proteins has also shown to be beneficial during their response to stress due to their ability to be both rapid and reversible upon the introduction of stress stimuli. This is evident in A. thaliana where Kurepa et al. [24] observed that heat induction initiates the accumulation of AtSUMO1/2 conjugates as early as 2 min, following a temperature shift and upon returning the plants to normal growth conditions: the increase in SUMO conjugation decreased and the pool of free SUMO increased to prestressed levels.
The process of SUMOylation in A. thaliana requires the attachment of SUMO proteins to a lysine residue in a target substrate using several enzymatic reactions. First, free SUMO is translated as a precursor that undergoes maturation by SUMO proteases. Currently, all SUMO proteases that have been identified in the SUMO system are cysteine proteases and are members of either the AtULP or AtDeSI gene families; however, only AtULP proteases have been shown to play a role in maturation [2]. During maturation, SUMO AtULP proteases recognize a C-terminal diglycine motif and cleave off approximately ten amino acids-exposing the diglycine motif [21]. The SUMO protein is then catalyzed into activation by the SUMO-activating enzyme (SUMO E1), which consists of a heterodimer of a regulatory subunit AtSAE1a/AtSAE1b and a catalytic subunit AtSAE2 [29]. For this reaction to take place an ATP molecule is hydrolyzed, forming a high-energy thioester bond between the cysteine residue in the catalytic subunit AtSAE2 and the exposed glycine residue in the SUMO protein [30]. The activated SUMO protein is then transferred in a transesterification reaction from AtSAE2 to a cysteine residue in the E2 conjugation enzyme-AtSCE1 to form a SUMO-SCE1 thioester complex [31,32]. The SUMO-SCE1 complex then catalyzes the process of SUMOylation using an isopeptide bond between the C-terminal diglycine residue of the SUMO protein and a lysine residue in the target protein [33].
Although the process of SUMO conjugation can be directly catalyzed by the E2 SUMO-SCE1 complex, this interaction alone is often insufficient; therefore, E3 ligases are required to aid in the transfer [26]. A further step also identified involves SUMO E4 proteins, which are specialized E3 ligases that exhibit elongase activity and promote the formation of SUMO chains in an AtSCE1-dependent manner onto target proteins [33,35]. In the final step, SUMO proteases cleave SUMO from target proteins to create pools of free SUMO, therefore, making SUMOylation a reversible process [2]- Figure 1.

The role of SUMO in adaptive responses against abiotic and biotic stresses in A. thaliana and major crops
The first implications of the role of SUMO in response to abiotic and biotic stresses were discovered by detecting an increased abundance of SUMO conjugates in protein extracts of stressed plants [28]. In plants, the SUMO-conjugate accumulation is not only influenced by the induction of a stress stimulus but also by the severity and duration of the stress and more crucially differential regulation of one or more of the components that comprise the SUMO cycle [36][37][38][39].

E1-activating enzymes and stress responses
In A. thaliana, the SUMO E1-activating enzyme is made up of three subunits-AtSAE1a, AtSAE1b, and AtSAE2. It is suggested that the E1-activating enzyme plays one of the most essential roles in the SUMO cycle due to its location within the first step of the conjugation pathway. The majority of the E1 function is located within the catalytic AtSAE2 subunit, which has been shown to play a critical role in early plant development where Atsae2 mutant plants are embryo lethal [14,31]. There is limited information on the role of AtSAE2 in response to plant stress, but this is likely due to Atsae2 mutants been embryo lethal.
Initially, it was first speculated that AtSAE1 may be redundant and unessential-with initial results indicating that T-DNA Atsae1a mutants are viable. However, in recent studies, this has since been disproven and a role for AtSAE1 in the regulation of SUMO conjugation during abiotic stress has been proposed. Mutant Atsae1a A. thaliana plants, when exposed to heat and drought stresses, displayed phenotypic defects and reduced SUMO-conjugate accumulation. These results suggest that AtSAE1 plays a role in the response to drought and heat stresses, whereas AtSAE2 is essential for plant development [40,41]. Currently, it is not possible to obtain Atsae1b null T-DNA insertion plants; therefore, its role remains unknown [31].
Currently, there are several known genes encoding for E1 enzymes in common crops; however, there is no indication of the presence of E1 genes in S. tuberosum but this is likely due to the SUMO machinery not yet being completely described in this species- Figure 2. While no data have currently been reported on the role of E1 in response to stress in the majority of the major crops, some data have been recorded in G. max. The results showed that levels of GmSAE1a were slightly increased following 6 h of salinity treatment-with levels returning to normal after 24 h. A similar pattern of expression was also observed under heat-shock treatment [42]. Although these results are The SUMO cycle begins with free precursor SUMO undergoing maturation via a SUMO ULP protease-cleaving off the C-terminus exposing a diglycine motif. Mature SUMO is then activated by a hydrolyzed ATP molecule and a SUMO E1 enzyme-a heterodimer of AtSAE1a/b and AtSAE2 [21]. The activated SUMO is transferred from AtSAE2 to AtSCE1, an E2 conjugation enzyme-forming a SUMO-AtSCE1 thioester complex, which catalyzes the process of SUMOylation onto a target protein [29]. SUMO E3 ligases aid in the transfer of SUMO proteins from AtSCE1 onto the lysine residue of target proteins [25]. SUMO E4 is a further step in the SUMO cycle and promotes the formation of SUMO chains [34,35]. Finally, SUMO proteases cleave SUMO from target proteins via a process called deSUMOylation to create pools of free SUMO, therefore, making the process of SUMOylation reversible [2].
suggestive that GmSAE1a may play a role in stress response, it is unclear as to what phenotypic advantage it offers. This is a common theme among E1 enzymes in crops; therefore, more research is required to understand their role during stress and if the addition of more E1 enzymes constitutes an evolutionary advantage.
While searching the literature, we also found that low copy numbers of SAE genes in A. thaliana and the major crops, excluding S. tuberosum, can be ascribed to single copies of these genes in common ancestors of dicots and monocots [43]. In addition, E1 enzymes are closely related among the six species, which is demonstrated in Figure  3, where branches are much shorter compared with other SUMO components. Another notable feature is that the E1 subunits observed in A. thaliana are identical with those observed in S. lycopersicum. It could be speculated that perhaps the E1 subunits in these two species operate in a similar way; however, there is currently no literature ULP and DeSI genes are SUMO proteases that function to either mature or deSUMOylate SUMOylated proteins; however, only ULP proteases play a role in maturation [2]. E1 and E2 genes are essential in the SUMO cycle as they encode for proteins that are essential to catalyzing the activation of matured SUMO proteins and facilitating their transfer onto a lysine residue in the target protein through a process called SUMO conjugation [29][30][31][32]. Although this process can occur directly through E2, E3 ligase genes are also essential as they encode for proteins that further aid in the transfer of the SUMO protein on the target lysine residue-making the process of SUMO conjugation more efficient [25]. Proteins encoded by E4 genes provide a further step in the SUMO cycle that promotes the formation of SUMO chains [33,35]. All data displayed in this figure were gathered from Ghimire et al. [12] and Li et al. [42]. The colors used in this figure each correlate to a separate plant species and display the number of each of the SUMO machinery components that have been found in the specific species so far. NA is used where currently no data is available for this component in the relevant species.
demonstrating the role of E1 in S. lycopersicum. A final observation to make is that the species G. max, similar to A. thaliana, also has two SAE1 subunits-GmSAE1a and GmSAE1b; however, the GmSAE2 subunit in G. max, as opposed to A. thaliana, is divided further into two subunits-GmSAE2a and GmSAE2b- Table 1. As knockouts of AtSAE2 are embryo lethal, could the division of GmSAE2 be an evolutionary development to enhance stress survival or could it be that more E1 genes are required in crop species due to their need for another layer of regulation?

E2-conjugating enzymes
AtSCE1 is the only SUMO-conjugating E2 enzyme described in A. thaliana. This enzyme is critical within the SUMO cycle due to its role in SUMO conjugation and AtSCE1 mutants being embryonic lethal-arresting early during embryo development [41]. As AtSCE1 is required during early development, the literature is limited regarding the role  Table 1 were aligned by MUSCLE, constructed, and visualized via a phylogenetic tree by MEGA 11, using the Neighbor Joining method with 1000 bootstraps. Red dots and green dots represent protein sequences from A. thaliana and yeast, respectively. All displayed data were gathered from Ghimire et al. [12] and Li et al. [42].
of AtSCE1 during stress; however, multiplication of E2-conjugating enzymes and their role within stress in crops is slightly better established.
Compared with A. thaliana, in crops, there are several genes encoding SCE1- Table 1 and Figure 2. This expansion of SCE genes may constitute an evolutionary advantage to produce higher stress and disease-tolerant crops and is likely due to tandem duplication in some crops and the tetraploid nature of S. tuberosum [44,45]. This is evident in the literature where overexpression of OsSCE3 has shown to enhance drought tolerance with a stimulated growth recovery of 90-96% observed after 2 days of drought compared with WT with a recovery rate of 40-64% [46]. G. max transcript levels of GmSCE1a and GmSCE1d have also shown to be up-regulated during stress with GmSCE1d transcript levels significantly increased as early as 1 h, following salinity stress and GmSCE1a exhibiting a faster response during heat-shock treatment compared with GmSCE1d. A role for GmSCE1a in immunity was also observed with transcript levels up-regulated, following Phytophthora sojae infection [42,47]. The S. tuberosum SCES-StSCE1, StSCE5, StSCE6, and StSCE7 transcript levels all increased during both salinity and PEG-induced water stress; however, StSCE9 was up-regulated during salinity stress and down-regulated during PEG-induced water stress [48]. Solannum peruvianum, a wild species of tomato, also showed that silencing of SpSCE1 leads to an increase in disease vulnerability to Clavibacter michiganensis ssp. michiganensis [49]. Although informative, the majority of E2 data in crops lacks phenotypic observations. Initial results in Z. mays indicated that ZmSCE1e transcript levels increased during both salinity and PEG-induced water stress. Using these initial findings, Wang et al. [50] went on to experimentally overexpress ZmSCE1 in three different transgenic tobacco lines and conferred tolerance to drought. After 15 days in the absence of water, WT tobacco plants demonstrated severe wilting, whereas the ZmSCE1e transgenic tobacco plants showed only moderate signs of stress with most of the upper leaves remaining green. After rewatering, nearly 50% of all WT plants died, whereas all transgenic lines started to grow again. More recently, Wang et al. [51] also overexpressed ZmSCE1d in four transgenic A. thaliana lines and were able to again confer drought tolerance and observed similar results. Following 17 days without water, the majority of WT plants experienced severe wiliting or lethality, whereas transgenic ZmSCE1d lines showed only moderate water deficit with 50% of upper leaves remaining green and fully expanded. These initial results suggest that SCEs may have a conserved function in stress protection. We speculate that SCEs may offer a SUMO-dependent mechanism for crop protection under stress. Phylogenetic analysis of the SUMO machinery in both A. thaliana and crops reveals the conservation of protein sequences within each SUMO component among the six species- Figure 3.

E3 ligases
Unlike Atsae2 and Atsce1 knockouts, single mutants of the SUMO E3 ligases are viable and in fact, have shown to be effective in several stress responses [53,54]. This is apparent where following exposure to 45 • C for 1 h, AtSIZ1 WT 5-day-old seedlings developed severe chlorotic whereas, under the same conditions, Atsiz1 mutant seedlings remained healthy [55]. AtSIZ1 has also been found to negatively regulate immunity in A. thaliana where increased resistance to P. syringae pv. Tomato DC3000 (Pst DC3000) was observed in Atsiz1 mutants due to an elevation in the levels of SA and the up-regulation of immune response genes [56].
In A. thaliana, the E3 ligase AtMMS21/AtHPY2 has been shown to negatively regulate drought tolerance. Zhang et al. [56] observed that following 17 days of drought conditions Atmms21 knockout A. thaliana plants showed no drought stress symptoms, whereas WT plants showed weak drought symptoms and AtMMS21/AtHPY2 overexpressed plants exhibited severe drought symptoms. The present study suggests that like AtSIZ1, AtMMS21/AtHPY2 negatively regulates stress in A. thaliana; however, this pattern of negative regulation is somewhat contradictory in Atmms21 plants as although mutants are viable and exhibit a level of drought resistance, this is accompanied by severe dwarfism, stunted development in root growth and defective meristem [57]. The main reason for this is due to the role of AtMMS21/AtHPY2 in repressing the endocycle onset in A. thaliana meristem with Atmms21/Athpy2 mutants displaying defective meristems accompanied by severe dwarfism due to early transitioning from the mitotic cycle to the endocycle [58]. Although knockout of the E3 ligase AtMMS21/AtHPY2 is considered viable and beneficial in response to drought stress it is accompanied by undesirable severe defects that are not worth the trade-off. This pattern of negative regulation, however, is not observed in S. lycopersicum, Z. mays, and G. max where overexpression of SiSIZ1 in S. lycopersicum has shown to boost tolerance during heat stress by reducing the accumulation of reactive oxygen species (ROS) and up-regulating heat-shock transcription factors (HSFs) [11]. In Z. mays, up-regulation of ZmSIZ1a, ZmSIZ1b, and ZmSIZ1c transcript levels were observed in response to drought and salinity stress, and in G. max GmSIZ1a and GmSIZ1b expression was stimulated in response to heat and dehydration-promoting GmSUMO1 conjugation [59,60].
The SUMO E3 ligase SIZ1 has also shown to exhibit conserved function similar to E2. By overexpressing OsSIZ1 in transgenic A. thaliana lines, it is possible to increase thermotolerance and tolerance to both drought and salinity stress-a response previously negatively regulated in A. thaliana [61]. Similarly, overexpression of OsSIZ1 in cotton plants has also been shown to increase drought tolerance, improve growth, and increase fibre yield [62]. These results further support the notion that the conserved function of the SUMO machinery may be utilized to confer stress tolerance universally across different crops and plants. Phylogenetic analysis of the E3 ligases SIZ1 and HPY2, unlike other clusters, is placed into two separate groups signifying unique sequences and possible unique functions with null mutants in AtSIZ1 in A. thaliana plants, showing a severe pleiotropic phenotype [32,63]- Figure 3.

E4 ligases
E4 SUMO ligases function by promoting SCE1-dependent SUMO chain formation and are currently not as vastly identified in crops in comparison with A. thaliana as observed in other SUMO components- Table 1. In A. thaliana, Tomanov et al. [35] observed that Atpial1 and Atpial2 mutants displayed better growth compared with WT when under both salinity and osmotic stress as well as exhibiting altered sulfur metabolism. The literature is again limited regarding the role of crop E4 ligases in stress response.

SUMO proteases
SUMO proteases are essential within the SUMO cycle due to their role in both SUMO maturation and deSUMOylation. So far seven DeSI and eight ULP proteases have been identified in A.thaliana [24]- Table 1. Two ULPs that have been shown to confer tolerance to several stresses are AtOTS1 (overly tolerant to salt1) and AtOTS2. These ULPs were initially identified in response to salinity stress where it was observed that during high-salt growth conditions, AtOTS1/2 were degraded to regulate salt stress response, whereas Atots1 Atots2 double mutants displayed extreme sensitivity to salt exposure [64]. Similarly, AtOTS1/2 are also responsible for the deSUMOylation of auxin response factor 7 (ARF7) to provide roots with hydropatterning during wet environments by forming lateral roots in the direction of water-regulating osmotic stress. [65]. Aside from abiotic stresses, AtOTS1/2 have also been shown to play a role in immunity, where Bailey et al. [66] found that the double mutant Atots1 Atots2 displayed elevated levels of conjugated SUMO, SA, and increased resistance to Pst DC3000 compared with WT plants. In their WT form, the SUMO proteases AtOTS1/2 limit the production of SA by suppressing the expression of isochorismate synthase1 (ICS1) and as a feedback loop are degraded upon SA treatment to regulate SA signalling.
Similarly to AtOTS1/2, the ULP protease Atesd4-1 mutant is also influenced by the of ICS1 with overexpression leading to an increase in the accumulation of SA [25,36]. The ULPs AtSPF1 and AtSPF2 also play a role in immunity by mediateing the SUMOylation of WRKY33 during flg22 treatment and infection from Botrytis cinerea. SUMOylation mediates the phosphorylation of WRKY33 facilitated by the interaction of mitogen-activated protein kinases (MAPKs), MAPK3 and MAPK6 [13].There is limited information known regarding the role of DeSI proteases in the model plant; however, the AtDesi3a protease has been shown to play a critical role in FLS2-mediated immunity. Orosa-Puente et al. [67] found that in A. thaliana, after induction of flagellin, AtDesi3a was degraded that promoted the SUMOylation of FLS2 to promote the dissociation and release of BIK1 from the FLS2 complex and thus activated PTI intracellular immune signalling.
When it comes to the role of SUMO proteases in the response to stress in crops very little is known; however, recent literature has shown an expansion of ULPs in O. sativa where a role for the SUMO proteases OsOTS1 and OsOTS2 has been observed during drought and salinity stresses. Transgenic-rice plants overexpressing OsOTS1 have been shown to enhance salt tolerance with an increase in the levels of salinity shown to trigger degradation of OsOTS1-indicating that SUMO conjugation in rice plants during salinity stress is due to down-regulation of OsOTS1/2 activity [68]. OsOTS1 has also been shown to mediate drought tolerance in O. sativa plants. This is evident where knockout of OsOTS1 in transgenic plants leads to an increase in the accumulation of ABA and more productive agronomic traits during drought stress, whereas OsOTS1 overexpressing lines displayed increased drought-sensitivity and ABA insensitivity [69]. In current literature, there is no identification of DeSI proteases in any crops due to their recent characterization in A. thaliana [2,70].

The implication for SUMO in future-proofing crops against climate change
There is relatively less information available about the role of SUMO in crops; therefore, in the present review, we provided an outline of the SUMO components in crops and how these provide a mechanism of defence against biotic and abioitic stresses in comparison with the model organism A. thaliana. SUMO is an essential PTM that is a multifaceted process that plays an essential role in the regulation of several stress responses in plants and major crops [25]. It is evident that the number of SUMO components, especially E2, E3, and ULP genes, have increased in major crops in comparison with A. thaliana [11,42]- Table 1 and Figure 2. This increase is likely ascribed to an evolutionary advantage in crops to enable their survival as it has shown that polyploidy offers benefits to some physiological characteristics by providing higher stress and disease tolerance; however, this could also be due to the requirement of extra SUMO genes in crop species to provide an extra layer of regulation that is not needed in A. thaliana [50,51,71]. Phylogenetic analysis of E2 SCE genes in crop species shows that there has been an independent duplication of E2 SCE genes, suggesting that crops have a greater reliance on the SUMO system for yield stability under environmental stress- Figure 3.
Multiple studies have successfully demonstrated the conserved function of SUMO in stress tolerance across different species. As E2 genes are not closely related among the different crops and A. thaliana on the phylogenetic tree it suggests that perhaps it is not required that species must be closely related to take advantage of this conserved function. Therefore, core stress response pathways in different species can be revealed by studying their SUMO-dependent mechanisms and this knowledge can readily be transferred to different crops to provide the key tools needed to produce stress-resistant crops to maintain global security.
A final point to make is regarding the limited information available on the role of SUMO proteases in crops. In A. thaliana, the previous literature has shown that SUMO proteases play a role in multiple abiotic and biotic stresses; however, the protease enzymes in crops are not well-characterized and their function remains unknown aside from the role of some ULPs in rice and maize [70][71][72]. Currently, it is known that there are 12-22 ULP proteases in rice; however, only seven OsULPs have been tested for protease function [69,[71][72][73][74]. It has been speculated that the expansion of ULPs in crops could have resulted from domestication and could give rise to target specificity during deSUMOylation [11,72,73,75]. As for DeSI proteases, to our knowledge, there is currently none identified in crops due to the recent characterization in A. thaliana; however, with more importance being placed on SUMO and its potential use in future-proofing crops, this is likely to change [2,70].

Summary
• SUMO plays a critical role in stress responses in the model plant A. thaliana and major crops.
• Major crops have an increase in SUMO genes compared with A. thaliana.
• The SUMO reveals conserved core stress response pathways in model plants and crops.
• Identifying SUMOylated targets regulating these core responses could pave the way to generating stress-tolerant crops.