Abstract
Reversible covalent attachment of SMALL UBIQUITIN-LIKE MODIFIERS (SUMOs) on target proteins regulate diverse cellular process across all eukaryotes. In Arabidopsis thaliana, most mutants with perturbed global SUMOylome display severe impairments in growth and adaptations to physiological stresses. Since SUMOs self-regulate activities of SUMOylation-associated proteins, existence of multiple isoforms introduces possibilities of their functional intersections which remain unexplored especially in plant systems. Using well-established defense responses elicited against virulent and avirulent Pseudomonas syringae pv. tomato strains, we investigated crosstalks in individual and combinatorial Arabidopsis sum mutants. Here we report that while SUM1 and SUM2 additively, but not equivalently suppress basal and TNL-specific immunity via down-regulation of salicylic acid (SA)-dependent responses, SUM3 promotes these defenses genetically downstream of SA. Remarkably, the expression of SUM3 is transcriptionally suppressed by SUMO1 or SUMO2. The loss of SUM3 not only lowers basal or post-bacterial challenge responsive enhancements of SUMO1/2-congugates but also reduces upregulation dynamics of defensive proteins and SUMOylation-associated transcripts. Combining a sum3 mutation partially attenuates heightened immunity of sum1 or sum2 mutants suggesting intricate functional impingements among these isoforms in optimizing immune amplitudes. Similar SUM1-SUM3 intersections also affect global SUMOylome responses to heat-shock affecting most notably the induction of selective heat-shock transcription factors. Overall, our investigations reveal novel insights into auto-regulatory mechanisms among SUMO isoforms in host SUMOylome maintenance and adjustments to environmental challenges.
Author Summary In plants, similar to animals, protein functions are regulated at multiple levels. One prevalent mode is to allow covalent linkage of small proteins to specific amino acids on targets thereby affecting its fate and function. One such kind of modification named as SUMOylation involves attachment of SUMO proteins. A plant maintains strict control over its pool of SUMOylated proteins (termed SUMOylome) which upon biotic or abiotic stresses are altered to facilitate appropriate responses, returning back to steady-state when the threat subsides. In mutants of the model plant Arabidopsis thaliana having disturbed steady-state SUMOylome, growth and developmental defects ensue. These mutants are auto-immune showing more resistance to infection by the bacterial pathogen Pseudomonas syringae. However, Arabidopsis SUMO-family are comprised of multiple members raising the question about their specificity or functional crosstalks. We discovered that two SUMO members function in coordination to suppress immunity including the repression of a third member which supports defenses. The expression of this third member during pathogen attack or heat-shock influences the responsive changes in the host SUMOylome likely suggesting SUMOs themselves play vital role in these adaptations. Overall, our work highlights novel intersections of SUMO members in mounting stress-specific responses.
Introduction
Post-translational modifications (PTMs) dynamically and reversibly modulate target proteins properties either by affecting their fate, location, or function and facilitate transitory adaptations often necessary for developmental passages and responding to biotic/abiotic cues. Of special importance in this category is the reversible attachment of the SMALL UBIQUITIN-LIKE MODIFIER (SUMO) through a process known as SUMOylation. This highly conserved mode of covalently modifying proteins is utilized primarily to regulate nuclear functions such as DNA replication, chromatin remodeling, transcription, and post-transcriptional processes in eukaryotes [1–3]. In plants, temperature perturbations, salinity changes and application of defensive hormones such as salicylic acid (SA) cause massive changes on global SUMOylome [4–6]. Interestingly, a vast majority of differentially SUMOylated proteins are nuclear, perhaps indicative of a rapid mechanism to modulate transcriptome in response to stimulus. Computational studies identify SUMOylation-annotated proteins as central relay players of protein-protein interaction webs, especially for transcriptional processes [7].
SUMOylation cascade recruits processed SUMOs with C-terminus di-glycine (GG) residues, a product of catalysis by specialized SUMO or Ubiquitin-like proteases (ULPs), to covalently attach via an isopeptide bond to ∊-amino group of lysine residues on target proteins. The modified lysine often is a part of a partially conserved motif, Ψ-K-X-D/E (Ψ = hydrophobic amino acid, X= any amino acid, D/E= aspartate/glutamate). A conjugatable SUMO, subsequently forms a thiol-ester bond with the SUMO E1 ACTIVATING ENZYME (SAE). A trans-esterification reaction further shuttles SUMO to the SUMO E2 CONJUGATING ENZYME 1 (SCE1) and then to target lysine on SUMOylation substrates. Enrichment of SUMOylated proteins upon heat shock reveal that SUMO1 can conjugate to itself and form poly-SUMO1 chains [5]. Although SCE1 is capable of catalyzing polySUMO-chain formation, members of SUMO E4 ligases the PROTEIN INHIBITOR OF ACTIVATED STAT LIKE protein (PIAL1/2 in Arabidopsis) enhance polySUMOylation [2,8,9]. Substrate specificities both for mono- or polySUMOylation are further regulated by SUMO E3 LIGASES such as HIGH PLOIDY2/METHYL METHANE SULFONATE21 (HPY2/MMS21) and SAP and MIZ1 (SIZ1) [10]. Fate of covalently-attached SUMOs may either include de-conjugation and recycling by SUMO proteases or targeted proteolysis of the substrate through ubiquitin-mediated pathway. Indeed, proteasome components are enriched in poly-SUMO1 pull-downs [5]. These interactions are non-covalent in nature facilitated by hydrophobic amino acid-rich SUMO-interaction motifs (SIMs) present in the recipient. A group of moderately conserved SUMO-targeted ubiquitin ligases (STUbLs) bind poly-SUMO chains via internal SIMs to ubiquitinylate the substrate [11]. Interestingly, SIMs are also only enriched in the SUMOylation-associated proteins such as SIZ1, SCE1, SAE2, SUMO-protease EARLY IN SHORT DAYS 4 (ESD4), PIAL1/2, STUbLs implying strong auto-regulatory mechanisms [8,12]. Not the least, increasing evidences that reciprocal influences of phosphorylation, ubiquitination, and acetylation that often compete with or modulate the efficacy of SUMOylation highlights the complexity of crosstalks among PTMs [13–15].
Unlike fruit fly, worm, or yeast, humans and plants express multiple SUMO isoforms [16]. In Arabidopsis thaliana, 4 SUMO isoforms are expressed namely SUMO1, SUMO2, SUMO3 and SUMO5 [4,17]. SUMO1 and SUMO2 share 89% sequence identity, whereas SUMO3 and SUMO5 are considerably diverged from SUMO1 (48% and 35% identical, respectively). The homolog pair SUMO1/2 are a result of genomic duplication event of a SUMO clade that preceded the evolution of monocots and eudicots from common ancient angiosperms [18]. Tandem organization of SUM2 and SUM3 genes in Arabidopsis are a result of gene duplication subsequently followed by diversification of only SUM3 sequences. At relative expression levels, SUM1/2 are more abundant than SUM3 or SUM5 [19]. Partial overlaps in tissue-specific expression patterns and biochemical properties taken together with embryonic lethality of sum1 sum2 double mutant indicate that plants require at least one functional copy of either of these redundant isoforms [20–22]. The Arabidopsis sum3 mutant is viable with mild late-flowering phenotype [21]. Acutely different SUMO3 however, unlike SUMO1/2, cannot form poly-SUMO chains in vitro, and shows little or no change to heat-shock treatment [4,17,19,21,23]. Not the least, SUMO1/2 but not SUMO3-modified targets, are efficiently deconjugated by SUMO proteases in vitro [17,24].
Host SUMOylome readjustments play a vital role in regulation of plant immunity and been comprehensively highlighted in several excellent reviews [25–27]. This was first evident in the Arabidopsis SUMO E3 ligase SIZ1 mutant (siz1-2) and subsequently in SUMO protease mutants of OVERLY TOLERANT TO SALT 1/2 (OTS1/2) and ESD4 [6,20,28,29]. Increased accumulation of defensive hormone salicylic acid (SA) and constitutive expression of PATHOGENESIS-RELATED (PR) proteins in these mutants conferred enhanced resistance when challenged with the bacterial pathogen Pseudomonas syringae pv. tomato strain DC3000 (PstDC3000). In this context, it is therefore not surprising that pathogens attempt to manipulate host SUMOylome to increase their colonization efficiencies [30–32]. Several bacterial phytopathogenic effectors interfere with host SUMOylation as a mode to suppress immunity [33,34]. XopD, a secreted effector from Xanthomonas campestris pv. vesicatoria (Xcv) de-conjugates SUMO from unknown targets in plants [35]. Mutations that disrupt SUMO-protease functions not only render the cognate strain deficient in virulence but also lower host defense induction.
Suggestively, both SUM1/2 jointly suppress SA-dependent defenses since plants null for SUM1 and expressing microRNA-silenced SUM2 (sum1-1 amiR-SUM2) have drastically reduced total SUMO1/2-conjugates compared to wild-type and display heightened anti-bacterial immunity [21]. The SUMO E3 ligase mutant siz1-2, also with lower levels of SUMO1/2 conjugates are similarly enhanced resistant to PstDC3000 [20,28]. On the contrary, esd4-1 or ots1ots2 plants with deficient SUMO protease ESD4 or OTS1/2 functions, respectively have increased SUMO1/2-conjugates yet display elevated SA-dependent defenses [1,6,29]. Constitutive activation of SA-dependent defenses also occur in SUM1/2-overexpressing transgenic plants regardless of whether the over-expressed SUMO isoforms are conjugation-proficient or deficient [21]. Thus, it is likely that perturbations rather than increase or decrease in SUMO1/2-conjugates per se regulate immune responses. Unlike SUM1/2, SUM3 is SA-inducible, upregulated in sum1-1 amiR-SUM2 plants, and upon over-expression enhance the resistance to PstDC3000 [21]. SUMO3-mediated SUMOylation of NPR1 (NONEXPRESSOR OF PATHOGENESIS-RELATED 1), a master transducer of SA-signaling is essential for defenses thus placing SUM3 as a bona fide positive immune regulator [36]. Whether increased SUMO3 activities during defenses intersect to alter SUMOylome outcomes of a host remains unexplored. Role of SUM3 in other stress responses similarly is also unknown. Evidently, perturbing host SUMOylome during biotic/abiotic stresses or through loss/gain of a specific isoform function initiates complex signaling more especially because the SUMOs themselves moderate SUMOylation proficiencies of targets. SAE2, SCE1, SIZ1, and EDS4 were identified as candidates whose SUMOylation footprints change upon stress exposure [5]. Animal studies also elegantly demonstrate ‘SUMO-preference/switching’ wherein the isoform choice for substrate conjugation is modulated not only by their relative levels but also by its influence on SUMOylation machinery [37–39]. Hence, it is beyond doubt that SUMO isoforms functionally intersect when the host SUMOylome-equilibrium is disturbed. However, such evidences from plant systems are completely lacking.
In this study, we utilized defense responses to various strains of PstDC3000 as a measure to test the contribution of individual and combinatorial Arabidopsis SUMO isoforms in immunity. Our data suggest that SUM1/2 function additively, but not equally, as negative regulators of SA-driven basal and TNL (Toll-Interleukin-1 receptor-like domain)-type immunity. In contrast, SUM3 has a more positive immune role potentiating auto-immunity that occur due to loss of SUM1 or SUM2. We demonstrate that accumulation kinetics of not only defense-associated markers but also of SUMOylation-associated genes are regulated by SUMO1-SUMO3 crosstalks. We further report that global change in a host SUMOylome in response to immune activation or heat-stress is also influenced by intersections of SUMO isoform activities. Overall, our results open newer avenues to unravel role of SUMO isoforms functions in maintenance and alterations of a host SUMOylome.
Results
SUM1/2 genetically function as negative immune regulators whereas SUM3 promotes defense responses
To study individual Arabidopsis SUMO isoform influences on defenses, we obtained previously characterized knockout mutant lines of sum1-1, sum2-1 and sum3-1 [20,21]. Wild-type (Col-0) and mutant plants were propagated either in short day (SD) or long day conditions (LD) as indicated for respective assays. Interestingly, in all propagation regimes and especially in SD conditions, we noticed clear growth defects in sum1-1, but not sum2-1 or sum3-1 plants. The sum1-1 mutant was developmentally dwarf with elongated leaves, reduced fresh tissue mass and increased trichomes density although with normal architecture compared to Col-0, sum2-1 or sum3-1 plants (Figs 1A,B; S1 Fig). Although identical sum1-1 mutant has been utilized previously the morphological defects we observed have not been reported earlier. The reason for this discrepancy is not clear and we speculate differences in soil compositions or growth variations as possible contributing factors. Nevertheless, the phenotypic attributes of sum1-1 were novel and we present evidences in subsequent sections that genetically link these defects solely to loss of SUM1.
We then investigated defense responses in the sum mutants utilizing the standard bacterial growth assays with virulent and avirulent PstDC3000 strains. In Col-0, PstDC3000 is virulent and triggers basal or PAMP-triggered immunity (PTI). In contrast, PstDC3000 harboring avrRps4, avrRpm1, or hopA1 effector is avirulent, triggering effector-triggered immunity (ETI), mediated by the cognate resistance gene RPS4, RPM1 or RPS6, respectively [40]. The virulent PstDC3000 strain we used harbors the plasmid backbone into which the avirulent effectors were cloned and hence named as PstDC3000 (empty vector; EV) in subsequent sections. Fully expanded leaves of 3-4 weeks-old SD-grown plants were infiltrated with the indicated PstDC3000 strains and the growth of the bacteria measured at 0- and 3-days post-infiltration (dpi). When compared to Col-0, total bacterial growth was reduced almost 10-fold in the sum1-1 and ~8-fold in sum2-1 plants whereas sum3-1 allowed more bacterial colonization (~8-fold higher) (Fig 1C). These results implied that SUM1/2 and SUM3 function as negative and positive regulators of PTI, respectively. When challenged with avirulent PstDC3000 expressing avrRps4 or hopA1, two TNL-specific ETI-eliciting effectors, lower bacterial accumulation than Col-0 persisted in the sum1-1 and sum2-1 plants (Fig 1D; S2A Fig). The sum3-1 plants were hyper-susceptible to these avirulent infections. Curiously, modest but consistent difference with less bacterial accumulation in sum1-1 than sum2-1 was observed to virulent or PstDC3000(hopA1) but not to PstDC3000(avrRps4) challenges. Further on, increased callose deposits, a well-established defensive phenomenon [41], were more pronounced in sum1-1 (~3-fold) and sum2 (~2-fold) than Col-0 in response to both virulent as well as avirulent (avrRps4) PstDC3000 (Fig 1E; S3 Fig). Although previous studies suggest redundant roles of SUM1/2 in defenses [21], as our results implicate that their degree of contribution however may slightly vary. Likely indicative of enhanced susceptibility to PstDC3000, lower levels of callose than Col-0 accumulated in sum3-1. Remarkably, ETI in all plants to avirulent PstDC3000 (avrRpm1) remained comparable suggesting that CNL-type responses are not affected by the loss of individual SUMO isoforms (S2B Fig). Indeed, as reported earlier avrRpm1-mediated HR is not affected in any sum mutants [21]. Overall, we identify a partial redundancy between SUM1/2 with a concomitant antagonism to SUM3 functions in immune responses to PstDC3000 strains.
Basal levels of SA and SA-responsive defense markers are upregulated in sum1-1 and sum2-1
Significant increases in salicylic acid, both free (SA) and glucose-conjugated (SAG), mediate signal responses against PstDC3000 [42]. A SUMOylation-deficient and enhanced resistant siz1-2 plants have elevated SA/SAG due to upregulated expression of SA-biosynthesis gene SID2/ICS1 (SALICYLIC ACID INDUCTION DEFICIENT2/ ISOCHORISMATE SYNTHASE1) [21,43]. To determine whether SA/SAG perturbations reflect the immune outcomes, we measured their relative levels in the different sum mutants. Remarkably, both sum1-1 and sum2-1 plants contained significantly elevated whereas the hyper-susceptible sum3-1 had lower SA/SAG, respectively than Col-0 (Figs 2A,B). We noted that SA/SAG levels in sum1-1 plants were slightly higher than sum2-1, perhaps indicative of its modestly higher degree of basal and TNL-type immunity. Transcripts of SID2/ICS1, defense-associated markers PR1 and PR2, and PR2 protein levels were upregulated in both sum1-1 and sum2-1 whereas in sum3-1 plants these were significantly lower than Col-0 (Figs 2C,D). Likewise, increased expressions of PTI markers FLG22-INDUCED RECEPTOR LIKE KINASE 1 (FRK1) and WRKY TRANSCRIPTION FACTOR 29 (WRKY29) [44] than Col-0 were detected in both sum1-1 and sum2-1 plants (Fig 2E). Relative to Col-0, FRK1 expression remained unaltered in sum3-1, whereas WRKY29 was drastically reduced suggesting that SUM3 promotes expression of only a subset of PTI-markers. Accumulation of a well-known SA-responsive TNL-type R protein SUPPRESSOR OF npr1-1 CONSTITUTIVE 1 (SNC1) is enhanced in several auto-immune mutants [45,46]. Remarkably, both sum1-1 and sum2-1 displayed upregulated SNC1 expressions whereas in sum3-1 the transcript levels remained comparable to Col-0 (Fig 2E). Since FRK1, WRKY29 and SNC1 are SA-inducible, upregulated SA-signaling sectors likely contribute to the enhanced basal and TNL-type immunity in sum1-1 and sum2-1 while these defenses are deficient in sum3-1. Also taking into account that SUM3 but not SUM1 or SUM2, is SA-inducible [21], we reasoned that enhanced resistance in sum1-1 or sum2-1 may be due to elevated SUM3 expression and its role as a positive immune regulator. Indeed, we detected ~3-and 2-fold higher SUM3 transcripts than Col-0 in sum1-1 and sum2-1, respectively (Fig 2F). Relative transcript levels of SUM1 in sum2-1, or SUM2 in sum1-1 however remain unaltered. Overall our results identify redundant but unequal roles of SUM1 and SUM2 in suppressing SA-dependent defenses that include expression of PTI markers, SNC1 and SUM3.
Expression of genomic copies of SUM1 or SUM3 complement immune response alterations in the respective sum mutants
Phenotypic abnormalities of sum1-1 plants necessitated us to genetically link the defects to loss of SUM1. Although our sum1-1 mutant is identical to earlier reports [20,21], to validate our observations, we utilized the transgenic His-H89R-SUM1 line that express a His6-tagged genomic SUM1 variant (H89R) from its native promoter in a sum1-1 sum2-1 background [5]. These plants as reported earlier not only complement the lethality of double homozygous sum1-1 sum2-1 mutant but also functionally mirror wild-type SUMO1 SUMOylome changes upon heat stress. From the His-H89R-SUM1 plants, the sum2-1 mutation was segregated out by generating F2 populations upon crossing with sum1-1. Henceforth, we termed these plants as sum1-1:His-H89R-SUM1. We constantly observed that growth defects we noted earlier were always linked to homozygous sum1-1 genotypes in absence of the His-H89R-SUM1 transgene. Thus, a functional SUM1 abolishes the growth defects apparent for sum1-1 plants (S4 Fig). We noted that the sum1-1:His-H89R-SUM1 plants had slightly elevated SUM1 (but not SUM2) transcripts compared to Col-0 (Fig 3A). Implicatively, as reported earlier for SUM1 over-expression [21], SA/SAG levels although considerably reduced in comparison to the sum1-1 parent, were still maintained at slightly higher levels than Col-0 (Fig 3B). Increased SA/SAG levels in sum1-1:His-H89R-SUM1 plants also resulted in upregulated SUM3 transcripts and higher PR1 proteins in comparison to Col-0 (Figs 3A,C). In pathogen growth assays using the virulent PstDC3000(EV), enhanced defenses noticeable for sum1-1 was abrogated to Col-0 levels in the sum1-1:His-H89R-SUM1 plants (Fig 3D). Interestingly, the avirulent PstDC3000(hopA1) challenge although diminished enhanced resistance relative to the sum1-1 parent, nevertheless these plants displayed stronger defenses than Col-0 (Fig 3D). This likely is attributed to higher than Col-0 levels of SA-regulated defense networks. These results unambiguously support the loss of SUM1 as the cause of enhanced resistance and phenotypic defects in sum1-1.
We also generated two independent complemented lines of sum3-1 to validate immune deficiencies due to the loss of SUM3. These transgenic plants express native promoter-driven His6-StrepII-tagged genomic fragment of SUM3 (sum3-1:HS-SUM3). Although owing to its low abundance [21] the expression of His6-StrepII-SUMO3 proteins remained undetectable in immunoblots, transcripts of SUM3 drastically reduced in the sum3-1, were restored to equivalent or slightly lower (~0.5X) than Col-0 levels in Line #1 and Line #2, respectively (Fig 3A). Relative abundance of SUM1 or SUM2 remained unaffected in both lines. For further assays, we therefore continued with sum3-1:HS-SUM3#1. We observed that deficiencies in SA/SAG or PR1 protein accumulations inherent to sum3-1 were reinstated to Col-0 levels in the sum3-1:HS-SUM3#1 plants suggesting functional complementation by the transgene (Figs 3B,C). Pathogen growth assays with either virulent PstDC3000(EV) or avirulent PstDC3000(hopA1) strains regained Col-0 levels of resistance in sum3-1:HS-SUM3#1 (Fig 3D). These data lead us to conclude that defensive deficiencies in sum3-1 is indeed due to loss of SUM3.
Constitutively active defenses in sum1-1 and sum2-1 caused rapid induction of immunity
To further establish that upregulated or dampened defense marker expressions in sum1-1, sum2-1 or sum3-1, respectively contribute to corresponding immune outcomes we challenged these plants with virulent PstDC3000(EV) or avirulent PstDC3000(avrRps4) strains. Leaf extracts harvested at 6-, 12-, and 24-hpi were used to compare accumulation kinetics of PR1/PR2 transcripts and proteins (Fig4; S5 Fig). Basal level (0-hpi) of PR1 in Col-0, sum2-1, or sum3-1 was below detection limit whereas in sum1-1 some enhancements were noticeable (Fig 4A). At 6- or 12-hpi although Col-0 or sum3-1 barely accumulated sufficient PR1/PR2 in response to both pathogen challenges, both sum1-1 and sum1-2 had markedly elevated levels of these proteins (Figs 4B,C). A similar trend continued at 24-hpi wherein in Col-0, PR2 but not PR1 proteins levels, almost matched sum1-1 or sum2-1. Most significantly, even at the 24-hpi sum3-1 plants were deficient in accumulating wild-type levels of PR1 or PR2. Real-time accumulation kinetics of PR1/PR2 transcripts to both bacterial infiltrations in general mirrored the corresponding protein levels although in some instances a direct correlation was not apparent (S5 Fig). Although the reason for this discrepancy is not clear, suggested role of SUMOylation in selective translation and/or its crosstalk with other PTMs that modulate protein synthesis or stability may likely be the cause [47,48]. Nevertheless, these data provide strong molecular evidence of constitutive SA-regulated defenses in sum1-1 and sum2-1 conferring their enhanced resistance to PstDC3000 strains. Contrastingly, delayed induction of defenses upon a pathogen challenge likely makes sum3-1 hypersusceptible.
Enhanced defenses in sum1-1 is SA-dependent
To genetically determine whether constitutive SA-signaling routes impart increased defenses to sum1-1, we generated sum1-1sid2-1 and sum1-1eds1-2 double mutants. A sid2-1 plant harbors a null mutation in SID2/ICS1 whereas an eds1-2 expresses a non-functional EDS1 (ENHANCED DISEASE SUSCEPTIBILITY1), a central player orchestrating SA-mediated defenses [43,49]. In segregating F2 populations, we noted that phenotypic defects always associated with homozygous sum1-1 allele provided at least one functional copy of EDS1 or SID2 were present (data not shown). Remarkably, the reduced leaf mass apparent in sum1-1 was improved in sum1-1eds1-2 and sum1-1sid2-1 plants (Fig 5A; S1 Fig). While sum1-1eds1-2 achieved wild-type mass, abolishing SID2/ICS1 in sum1-1 although ameliorated growth defects but these plants still retained sid2-1-characterstics including smaller and paler leaves than Col-0 (Fig 5A; S1 Fig). Interestingly, both sum1-1sid2-1 or sum1-1eds1-2 had wild-type trichome densities unlike the sum1-1 parent thus associating these defects to perturbations in SA-regulated networks (S6A Fig).
Loss of either EDS1 or SID2/ICS1 remarkably abolished accumulated PR2 proteins and upregulated PR1, PR2, FRK1 and WRKY29 transcripts in sum1-1 (Figs 4A, 5B; S6B Fig). Further on, pathogen growth assays using virulent and avirulent PstDC3000 strains demonstrated that sid2-1 or eds1-2 mutations were epistatic to sum1-1 abolishing not only the increased defenses but also conferring hyper-susceptibility to the respective double mutants towards virulent PstDC3000(EV) or avirulent (avrRps4- or hopA1-expressing) strains (Figs 5C,D; S6C Fig). Neither eds1-2 nor sid2-1 mutation affected resistance towards avirulent PstDC3000 (avrRpm1) which remained comparable to sum1-1 or Col-0 (S6D Fig). As expected, loss of EDS1 or SID2/ICS1 prevented accumulation of PR1 or PR2 proteins during PstDC3000 infections (Figs 4B,C). With these assays, we identify that enhanced resistance in sum1-1 is undoubtedly due to constitutive SA-routed defense signaling. A sum2-1eds1-2 double mutant, we also generated, mimicked sum1-1eds1-2 hyper-susceptible outcomes in disease assays (S7 Fig). Overall, our results provide several parallel lines of evidences suggesting SUM1/2 intersections on SA-mediated immune signaling.
sum3-1 alleviates enhanced defenses in sum1/2 mutants
Since our data suggested antagonistic involvement of SUM1/2 and SUM3 as negative or positive defense regulators, respectively we surmise a crosstalk between these isoforms. To investigate this, we generated sum1-1sum3-1 double mutant by genetic crossing. Unlike sum1-1eds1-2 or sum1-1sid2-1, only partial restoration of sum1-1 growth defects was observed in sum1-1sum3-1 plants (Fig 6A; S8A,B Figs). These observations hinted that SUM3 is mildly responsible for sum1-1 growth anomalies. Increased trichome densities noticed in sum1-1 however remained unaffected by the loss of SUM3 (S6A Fig). Because SUM3 affects basal SA accumulation (Figs 2A,B), we tested defense outputs are altered in sum1-1 sum3-1. Elevated SA levels in sum1-1 were marginally reduced in sum1-1 sum3-1 plants. Indeed, the expression of SID2/ICS1 remained comparable between sum1-1 and sum1-1 sum3-1 plants (Fig 2C). Additionally, PR1 or PR2 transcripts in sum1-1 sum3-1, were intermediate between sum1-1 and Col-0. These data imply SUM3 does not affect SA-biosynthesis but modulates positive feedback loop of SA-signaling wherein it intersects with SUM1 role as a transcriptional repressor of SID2/ICS1, PR1 or PR2. As is therefore expected, loss of SUM3 reduced PR1 or PR2 levels in sum1-1 (Fig 4A). Not the least, we also demonstrate that increased SNC1 accumulation in sum1-1 plants although reduced in sum1-1 sum3-1, are still maintained higher than Col-0 levels (Fig 6B).
In pathogen growth assays, either virulent PstDC3000(EV) or avirulent PstDC3000 (hopA1) strains accumulated to intermediate levels in sum1-1 sum3-1 leaves, particularly higher than sum1-1, but significantly lower than in Col-0 or sum3-1 (Fig 6C; S8C Fig). Interestingly, wild-type level of resistance was observed in sum1-1 sum3-1 to the avirulent effector avrRps4 (Fig 6D). Growth of avirulent PstDC3000 (avrRpm1) was not affected in sum1-1sum3-1 and remained comparable in all plants (S8D Fig). As investigated earlier, in response to virulent and avirulent PstDC3000(avrRps4) infections and unlike sum1-1, sum1-1 sum3-1 remained deficient in induction of PR1 or PR2 at 6-hpi (Figs 4B,C; S9A,B Figs). However, we observed that at later time points (12- and 24-hpi) loss of SUM3 did not affect the rapid induction of PR1/PR2 in sum1-1. This is a sharp contrast to sum3-1 plants which even at 24-hpi accumulated very less PR1/PR2 proteins. Thus, although SUM3 is deemed essential for upregulation of PR1/PR2 upon a pathogen attack, sum1-1 plants eventually overcome this requirement. Taken together, we principally support that SUM3 promotes PR1/PR2 transcription downstream of SA as suggested earlier [21]. We provide further molecular evidence that SUM1 suppression of defenses likely is upstream of SA and may involve transcriptional repression of SID2/ICS1 and its subsequent consequences on SA-responsive markers such as PR1, PR2, FRK1, WRKY29 and SNC1 among others.
Adjacent chromosomal arrangement of SUM2 and SUM3 in Col-0 genome suggest that they likely arose by a tandem duplication event [18]. Hence, unlike a sum1-1 sum3-1 double mutant, a sum2-1 sum3-1 is difficult to obtain by genetic crossing. To test whether SUM3 also intersects on SUM2 function as a negative immune regulator, we generated amiR-SUM2 sum3-1 plants by crossing the two parental lines. The amiR-SUM2 plants reported earlier contains specific knockdown of SUM2 and mimics the null mutant sum2-1 [21]. SUM1 or SUM3 transcripts remain unaffected in these lines. Plants homozygous for sum3-1 and containing amiR-SUM2 transgene have significantly downregulated SID2/ICS1, PR1 and PR2 levels in comparison to a sum2 mutant (Fig 2C). These transcripts however are still elevated than Col-0 plants implicating that SUM3 functions impinge partially on SUM2 role as a transcriptional repressor. This was also supported in pathogen challenges wherein growth of both virulent and avirulent PstDC3000 strains were intermediate between resistant Col-0 and hyper-resistant sum2 mutant (S7B,C Figs). Taken together, whether the functional antagonism in immune regulatory roles of SUM1/2 and SUM3 affect immune outcomes in plants.
SUMO3, but not SUMO1, can form non-covalent homo-dimers
Conjugation proficiencies in planta of SUMO isoforms are clearly different. Upon SA application, drastic increase in SUMO1/2 conjugates are observed whereas target modifications by SUMO3 is barely detectable even though SUM3 expression, unlike SUM1/2 is SA-inducible [6,21]. A distinct difference in the protein sequence of these isoforms is the presence of a predicted SIM motif in SUMO3 but not SUMO1/2 (S10 Fig). We utilized Bi-molecular Fluorescence Complementation (BiFC) assays to test homo- and hetero- non-covalent associations between SUMO1 and SUMO3. The BiFC vectors introduced via Agrobacterium-mediated transient transformation of N. benthamiana leaves expressed split YFP fusions of either SUMOylation-proficient (GG) or -deficient (AA) SUMOs wherein the C-terminal di-glycine (GG) residues were kept intact or mutated to di-alanine (AA), respectively (Fig 7A). Co-expression of only SUMO1GG/SUMO1GG, but not SUMO1GG/SUMO1AA or SUMO1AA/SUMO1AA showed reconstitution of the split YFP protein. Although positive BiFC indicate non-covalent interactions among protein partners, the lack of fluorescence in SUMO1AA/SUMO1AA suggests that SUMO1GG/SUMO1GG combinations and not homo- oligomers and likely reflect covalent polySUMO1-conjugates wherein the split fluorescent fusion proteins are in allowed proximity for reconstitution. Remarkably, all BiFC combinations of SUMO3 with itself, regardless of GG- or AA-forms, showed YFP fluorescence. Since SUMO3 lacks poly-SUMOylation properties [22], we reason that the observed associations are non-covalent SUMO3 oligomers. To test whether SUMO3 binds SUMO1 non-covalently, combinations of SUMO1GG or AA were co-expressed with SUMO3GG or AA. While SUMO1GG showed clear fluorescence when expressed with either SUMO3GG or AA, similar combinations of SUMO1AA did not. These observations are suggestive that only a conjugable-SUMO1 likely present on a target may achieve permissible molecular proximity in vivo to a SUMO3 that may be either bound to a SIM (as in SUMO1GG/SUMO3AA pair) or attached covalently to the same target at a separate SUMOylation motif (as in SUMO1GG/SUMO3GG pair). Evidences from human cell lines demonstrate that ortholog of Arabidopsis SUMO3, HsSUMO1 ‘cap’ poly-SUMO chains of the SUMO1 ortholog HsSUMO2 [50,51]. Whether Arabidopsis SUMO3 regulates poly-SUMO1/2 chain lengths remain a promising possibility to explore further. Since no interaction was detected with SUMO1AA/SUMO3AA, it is clear that SUMO3 and SUMO1 do not interact non-covalently. To validate further, we performed in vitro binding assays with tagged recombinant SUMO1AA/SUMO3AA proteins expressed in E. coli. Enrichment of His-SUMO1AA via Ni2+ beads failed to co-elute either Strep-SUMO1AA or Strep-SUMO3AA suggesting that SUMO1-SUMO1 homo- and SUMO1-SUMO3 hetero-dimers cannot form in vitro (Fig 7B). Homodimers of SUMO3AA however were detected by the presence of Strep-SUMO3AA in His-SUMO3AA-enriched eluates. These results support our hypothesis that BiFC interactions observed for SUMO1GG with SUMO3GG/AA may likely suggest close proximity, but not direct binding, of SUMO1 and SUMO3 in planta.
SUMO3 regulates SUMO1/2 conjugation efficiencies upon various stress exposures
With the above results, it is encouraging to speculate that SUMOylation efficiencies may be modulated by intersecting SUMO1-SUMO3 functions. Since SUM3 partially modulates sum1-1 defenses, we investigated dynamic changes in SUMO1/2-SUMOylome in Col-0, sum1-1, sum3-1, or sum1-1 sum3-1 upon a pathogen challenge. Leaf tissues either mock-treated or infiltrated with PstDC3000(EV), were collected at 24-hpi and processed for selective qPCRs or immunoblots for SUMO1/2-conjugates. The polyclonal anti-SUMO1 antibody we used cannot distinguish SUMO1/2 isoforms and the lack of anti-SUMO3 antibodies prevented us for analyzing SUMO3-conjugates. We observed that while a pathogen exposure caused remarkable increase in SUMO1/2-conjugates in Col-0, for sum1-1 these was considerably lower (Fig 8A).
These results suggested that SUMO1-rather than SUMO2-conjugates are not only more prevalent but also undergo massive increment upon pathogen treatment. Surprisingly, the conjugates are also distinctly lower in uninfected sum3-1 and are considerably less upregulated than Col-0 in response to the pathogen infection. As is expected, in sum1-1 sum3-1 extracts barely any SUMO-conjugates are detectable pre- or post-infection. With these results, we identify clear intersection of SUMO3 functions on SUMO1/2-conjugation efficiencies and its perturbations during a pathogen attack. We also observed that both sum1-1:His-H89R-SUM1 and sum3-1:HS-SUM3 plants achieved Col-0 levels of SUMO1/2-cojugates upon PstDC3000(EV) infection thus further validating the complete functional complementation of sum1-1 or sum3-1, respectively (S11 Fig).
To determine whether changes in conjugation efficiencies are due to differential expression of SUMO isoforms and/or SUMOylation-associated genes, we investigated their relative expressions in Col-0 and the sum mutants (Figs 8B,C; S11B Fig). We observed that SUM3, but not SUM1/2, was upregulated upon a PstDC3000(EV) challenge. This is in accordance with SUM3 being SA-inducible [21]. Additionally, we noticed that the enhanced SUM3 expression in sum1-1 was further boosted in pathogen-exposed samples. Curiously, for the investigated SUMOylation-associated genes although sum1-1 mutation did not significantly alter their basal expression levels, several of these (SAE2, SIZ1, HYP2, ESD4, ELS1, and OTS1) were down-regulated in sum3-1 plants. Thus, the modest reduction in global SUMOylome in sum3-1 we observe may be attributed to SUM3-dependent modulation of expression of these genes. Interestingly, sum1-1 sum3-1 plants showed slight upregulation of SAE2, SCE1, SIZ1, HPY2, ELS1 and OTS2 in mock-treatment supporting our earlier claim that sum1-1 plants with constitutively active defenses and elevated SA levels, overcome the requirement of SUM3 for the expression of these genes. We also noted that upregulation of SAE2, SIZ1, ELS1 or OTS1 that occurred only modestly in Col-0 upon PstDC3000(EV) infection were further aggravated in sum1-1, but intermediate in sum3-1 (Figs 8B,C). These enhancements are maintained in sum1-1 sum3-1 plants and are likely SUM3-independent. Remarkably, sum3-1 plants down-regulated the expression of SUM1 and SUM2 only upon pathogen-treatment highlighting that the positive immune function of SUM3 may involve transcriptional repression of negative defense regulators such as SUM1/2. Overall, our data lead us to speculate that immune responses recruit SUM1-SUM3 crosstalks to modulate the expression of these genes at a transcriptional level thereby influencing host SUMOylome changes.
SAE2, SCE1, SIZ1 or EDS4 are direct SUMO1 targets and both SUMO1/3 interact in planta with SCE1 and SIZ1 [5,52]. These led us to test whether SUMO1-SUMO3 intersections directly influence the efficiency of SUMO1/2-conjugate formation through modulation of SUMOylation-associated protein functions. We utilized the E. coli SIZ1-independent Arabidopsis SUMOylation reconstitution system [53]. Co-expression of His-SUMO1, and Strep-SUMO3 (GG or AA) along with SCE1, and SAE1/2 showed modest but clear increase in SUMO1-conjugates in comparison to extracts that lacked Strep-SUMO3 (S12 Fig). Remarkably, a more dramatic increase in SUMO3-conjugates when co-expressed with either Strep-SUMO1 GG or AA was also noted. From these results, it is undeniably clear that SUMO1 and SUMO3 have a mutually beneficial role in reciprocal SUMOylation process likely through modulation of SCE1 or SAE1/2 functions. The significant decrease in SUMO1/2 conjugates in absence of SUM3 (Fig 8A) provides reasonable endorsement to our hypothesis.
SUMO1/2-conjugates are enhanced in response to heat shock [20,21]. To test whether SUM3 role intersect on these responses, we subjected Col-0, sum1-1, sum3-1, or sum1-1 sum3-1 plants to heat shock and analyzed SUMO1/2-conjugates as well as relative expression levels of SUM1/2/3 and several previously characterized heat stress-responsive genes [54] (S13 Fig). In accordance with [20], we observe strong induction of SUMO1/2-conjugates in heat-treated Col-0 extracts (S13A Fig). These conjugates were low in heat-stressed sum1-1 plants reinstating that SUMO1-mediated SUMOylation is most predominant during heat stress. Surprisingly, in sum3-1 considerably reduced conjugates than Col-0 accumulated upon heat shock. We especially noted that unlike PstDC3000 challenges, SUM1/2 transcripts were significantly upregulated by heat-exposure, while SUM3 induction was almost negligible (S13B Fig). Further on, transcriptional upregulation of SUM1/2 was SUM3-independent. Therefore, reduced accumulation of SUMO1/2-conjugates in heat-treated sum3-1 suggests that SUM3 affects SUMO1-SUMOylation efficiencies at a post-transcriptional level. With these results, we infer that although different physiological stresses may cause apparently similar effects on global SUMO1/2-conjugates, the responsive routes to condition these are not only distinct but also stress-specific. Interestingly, increased expression of several heat stress-responsive markers such as HsfA2, HSP22.0-ER that are upregulated in Col-0 upon heat-treatment, were hyper-elevated in sum1-1 suggesting SUM1 suppresses their expression (S13C Fig). In sum3-1 plants, their fold-induction upon heat stress was relatively lower than Col-0 implying that SUM3 is partially responsible for their upregulation. Induction of other tested markers such as HSP18.2 or HSP23.5-P were unaffected in sum1-1 or sum3-1 plants (S13D Fig). Taken together, our results clearly identify SUM1-SUM3 crosstalks in responses to multiple stresses.
Discussion
Functional overlaps between Arabidopsis SUMO1 and SUMO2 isoforms, likely prevent noticeable phenotypic defects in the individual mutants under non-stressed conditions [17,20,21,23]. Interestingly, reduction in SUMO-conjugates both basally as well as upon heat-stress is more prominent in sum1-1 than sum2-1 suggesting a SUM1 predominance in these responses [20]. Our observations therefore of phenotypic defects in sum1-1 but not sum2-1 plants may reflect a similar developmental process preferentially regulated by SUM1. We convincingly establish that constitutive upregulation of SA-regulated networks is the primary cause of these defects in sum1-1 (S6A Fig). This is unlike siz1-2 where the associated growth abnormalities are EDS1- or SID2/ICS1-independent [55]. We deduce that reduction in global SUMO1/2-conjugates apparently similar between siz1-2 and sum1-1 affect downstream responses differently. At one instance, loss of SIZ1 may affect SUMO-conjugation for all isoforms, whereas the sum1-1 clearly restricts only SUMO1 functions. Previous studies propose SA antagonism or jasmonic acid (JA) promotion in increased trichome formation [56]. We reason that while direct application of SA or JA may affect trichome production as reported, the loss of SUM1 whose substrates include both JA and SA signaling regulators such as SIZ1, TPL, or JAZs among others may impact trichome density through a more complex SA-JA crosstalk [5,57]. Further investigations into relative SA-JA signaling routes perturbed in sum1-1 may unravel this mystery.
In context of defense responses, our data provide substantial molecular support not only to individual SUM1/2 or SUM3 but also to their intersecting contributions as negative or positive regulators, respectively of anti-bacterial basal and TNL-specific immunity. Firstly, bacterial accumulations are strongly reduced in both sum1-1 and sum2-1, with sum1-1 displaying stronger immunity than sum2-1, for both virulent as well as avirulent PstDC3000(hopA1) infections. Although comparable immunity between sum1-1 and sum2-1 is observed for avirulent PstDC3000(avrRps4) challenges, this likely is due to relatively weaker ETI responses elicited by AvrRps4 in comparison to HopA1-expresssing PstDC3000 (compare Fig 1D and S2A Fig). In contrast to sum1/2 mutants, immunity in sum3-1 is compromised implicating SUM3 is essential for these defenses. Secondly, we demonstrate that impairment of SA-signaling routes due to loss of EDS1 (eds1-2) or SID2/ICS1 (sid2-1) abolish enhanced resistance in sum1-1 and sum2-1 plants providing support to genetic placement of SUM1/2 as suppressors of these defenses [21] (Fig 8D). In this pathway, SUM3 partially delegates SA-defenses likely through positive feedback mechanisms. Indeed, both sum1-1 and sum2-1 plants accumulate increased SA than Col-0, while in sum3-1, these are relatively lower. Correspondingly, transcripts of SA-biosynthesis SID2/ICS1 and responsive markers such as FRK1, PR1, WRKY29, and SNC1 are upregulated in sum1-1 or sum2-1, and lower in sum3-1. Accumulation kinetics of PR1/PR2 proteins or transcripts (Fig 4; S5 Fig) further validate primed or deficient SA-mediated defenses in sum1-1 or sum2-1 and sum3-1, respectively. The modest difference we note between sum1-1 and sum2-1, in the upregulation of some of these SA-inducible markers, is suggestive of additive but unequal roles of SUM1 and SUM2 as negative immune regulators. Only ~17% common targets were identified between human SUMO1 and SUMO2 isoforms suggesting that they are partially redundant at best [58]. Whether SUMOylation-targets distinct between Arabidopsis SUMO1 and SUMO2 isoforms affect immune amplitudes differentially awaits further studies.
We provide the first in planta evidence of partial SUMO3 moderation on SUMO1/2 functions in biotic, abiotic and developmental responses. Introducing sum3-1 mutation not only attenuates reduced tissue mass but also subdues enhanced resistance of sum1-1 supporting SUM3 role in promoting defenses downstream of SA. Thus, elevated SA and PR1, PR2, or SNC1 expressions are partially dampened in either sum1-1 or in amiR-SUM2 plants when SUM3 is mutated (Fig 2). Intersections of SUM3 functions on SUMOylation-proficiencies, are also best noted on SUMO1/2-conjugate intensities in response to different stresses. Although, enhancement upon heat shock or SA-treatments have been previously reported [5,6], we first report their upregulation in PstDC3000 exposures. Induction of several SUMOylation-associated genes such as SAE2, SIZ1, ELS1, or OTS1 during PstDC3000 challenges are influenced by SUM1-SUM3 intersections and contribute to the corresponding defense outcomes (Fig 8). Recently, the activity of TPR1, a transcriptional co-repressor involved in suppression of negative defense regulators DND1/2 was reported to be regulated in a partial SA-dependent manner via SUMOylation by SIZ1 [59]. SUMOylation-deficient TPR1 not only is more enhanced in repressing DND1/2 expressions, but also its over-expression cause enhanced resistance than the SUMOylation-proficient TPR1. Overall, our results raised the possibility that increased SUMO1/2 conjugates may reflect negative feedback mechanisms to maintain immune responses as transitory thus regulating SUM3 functions in promoting defenses.
Functional intersections among SUMO isoforms are anticipated at multiple post-transcriptional events. Amino acid distinctions at key conserved positions among the SUMO isoforms influence their in vivo conjugation proficiencies [60] (Schematically summarized in S10 Fig). Aspartate63 (D63) in Arabidopsis SUMO1 (D75 in SUMO2) is replaced by Asparagine63 (N63) in SUMO3, reducing its relative interaction strength than SUMO1 with SCE1, thus lowering its poly-SUMO formation efficiencies. Other residues also diverged in SUMO3 weaken thioester-bond formation and interaction with SAE1. Whether SUM3 upregulation during defences improve these propensities and hence alters host SUMOylome outcomes although remains unknown, evidential support is obtained from several studies. Mutants with decreased SUMOylation including siz1-2 have increased SCE1 protein abundance [15,20,61]. And siz1-2 impairment in accumulating SUMO-conjugates during heat stress, is completely recovered in pial1 pial2 siz1-2 plants [8]. Considering PIAL1/2 improves SCE1 ability to form otherwise less efficient polySUMO3 chains, defense responses with upregulated SUM3 likely introduces competition between SUMO1 and SUMO3 for substrate polySUMOylation thereby altering their fates via respective isoform-specific SUMO-targeted E3 Ubiquitin ligases (STUbLs) [8].
Indeed, pial1/2 double mutants although less stress-tolerant, have upregulated levels of proteins related to biotic stress [8,15]. In sum1-1, the reduced global SUMO1/2-conjugates upon pathogen or heat-shock challenges taken together with elevated SUM3 transcripts is suggestive of this occurrence and need to be explored further.
As substrates of Arabidopsis SCE1 and ESD4, candidates predominantly involved in RNA-related processes such as nucleocytoplasmic transport, splicing, or turnover and chromatin-modification including transcriptional activation/repression-associated proteins were identified [62]. A majority of these candidates possessed predicted SUMOylation motifs and were covalently modified by both SUMO1 or SUMO3. In Arabidopsis, a SUMOylation-proficient SUMO1 is essential to covalently charge SCE1 at the catalytic site, promote its association with SIZ1, and form the ternary complex (SUMO-SCE1-SIZ1) [52]. Interestingly, the subcellular localization of this complex is determined by the specific SUMO isoform bound non-covalently at the SIM site distinct from the catalytic pocket of SCE1. Since SCE1 possess ability to distinguish substrates for SUMOylation [9], selection of substrates and formation of polySUMO chains therefore may be affected by fluctuations in relative SUMO1/3 levels, as has been previously suggested in animal studies [63,64]. Our data showing reciprocal improvements in SUMOylation efficiencies in vitro regardless of conjugation-proficiencies of the influencing SUMO isoform may suggest towards one such consequence of SUMO1-SUMO3 crosstalks (S12 Fig). Likewise, constitutive activation of defenses due to over-expression of either conjugation-proficient (GG) or -deficient (ΔGG) SUMO isoforms may at least be partially attributed to this phenomenon [21].
SUMOylation machineries such as SAE2, SIZ1 and ESD4 also bind SUMOs non-covalently owing to their intrinsic SIMs [12,65]. It is postulated that spatio-temporal regulations of SUMOylation/de-SUMOylation coordination primarily by SIZ1 and ESD4 regulate steady-state levels of host proteins conjugated to SUMOs which surprisingly are fewer in numbers [2]. Dramatic changes in this minimal SUMOylome across various stress conditions reveal that instead of newer targets undergoing covalent modifications by SUMOs, SUMOylation levels on prior-SUMOylated proteins pools are more altered [66]. Likely, functional inactivation of SUMO proteases especially noted post-stress in mammalian systems [67–69] and enhanced pool of SUMOylated pool SIZ1 [66] coordinate to achieve this. Taking this into account, a host SUMOylome output is likely to be influenced by relative changes in SUMO isoforms. Indeed, we demonstrate that transcripts of SIZ1, EDS4 or SAE2 are up-regulated basally in sum1-1, and in both Col-0 or sum1-1 upon pathogen challenge in a SUM3-dependent manner (Fig 8). Hence the requirement of SUM3 to maintain and adjust optimal level of SUMO1/2 SUMOylome is clearly evident. Non-covalent associations with SUMOs also affect chromatin architecture via their interaction with DNA methyltransferases and demethylases [11]. Expression of FLOWERING LOCUS C (FLC), a central transcription factor of flowering is dynamically regulated by DNA methylation status while the protein function is modulated by HPY2, SIZ1 and SUMO proteases [70,71]. FLC regulation presents a classical example of this versatility of SUMO-influences on a developmental process.
Stimulus-driven SUMO isoform switches and subsequent fate of substrates have been widely documented in animal systems [37–39]. The mammalian GTPase activating protein RanGAP1 although is equally modified by SUMO1/2/3 in vitro, conjugation in vivo to SUMO1, but not SUMO2, imparts more stability from isopeptidases thereby facilitating its association with Nup358 [39]. Similarly, HDAC1 is targeted for degradation upon SUMOylation by SUMO1, but not SUMO2, in cancerous cell lines thus potentiating invasive properties of the tumour [72]. Specific SUMO-proteases also regulate stimulus-dependent SUMO isoform switching [73].
Upon arsenic exposure, SUMO2 at Lys65 of PML (Promyelocytic leukemia protein) is replaced with SUMO1 to direct its ubiquitinylation. Clearly with the isoform selectivity regulated at multiple levels, a host SUMOylome undergoes dynamic changes in response to physiological perturbations. Evidences on SUMO isoform switches however are completely lacking from plant systems. Nonetheless several reports suggest the likelihood of this phenomenon. The effector XopD from Xcv is a SUMO3-specific isopeptidase [17]. The replication protein AL1 from Tomato Golden Mosaic Virus (TGMV) and the RNA-dependent RNA polymerase NIb from Turnip mosaic virus (TuMV) interact with and require SCE1 for replication [74,75].
While our results undoubtedly place SUMO3 as a strong contributor in regulating dynamics changes in a host SUMOylome, the loss of SUM3 in many Brassicaeae and other higher eudicots still remain an enigma [18]. Whether functions of SUM3 has been incorporated in SUM1/2 roles in these plants remains to be explored further. We foresee the immediate need of a plant system that would facilitate enrichment via distinct affinity tags on individual SUMO isoforms in order to obtain evidences of SUMO preference, switches and intersections. To summarize, our data lay the foundation on functional impingements of Arabidopsis SUMO isoforms and their adjustments on global SUMOylome in response to physiological changes.
Methods
Plant materials and growth conditions
Arabidopsis thaliana mutant lines sum1-1 (SAIL_296_C12), sum2-1 (SALK_029775C) and sum3-1 (SM_3_2707/SM_3_21645) were obtained from Arabidopsis stock centre (https://www.arabidopsis.org). Seeds of amiR-SUM2 sum3-1 were generated by Dr. Harrold van den Burg. All plants were grown at 22°C with 70% humidity under Long Days (LD; 16 h: 8 h, light: dark) or Short Days (SD; 8 h: 16 h, light : dark) having light intensity 100 μmol μm−2s−1 light. Specific growth conditions are indicated in respective legends. To generate combinatorial mutants, sum1-1 was genetically crossed with eds1-2, sid2-1 or sum3-1 mutant plants. The sum2-1 mutant was similarly crossed with eds1-2 to generate sum2-1 eds1-2. Double mutants (sum1-1 eds1-2, sum1-1 sid2-1 and sum1-1 sum3-1) were identified in F2 population by PCR based genotyping. To generate sum1-1:His-H89R-SUM1, His-H89R-SUM1 [5] was crossed with sum1-1. Segregating F2 populations were PCR-based genotyped for the presence of homozygous sum1-1, wild-type SUM2 and homozygous copies of His-H89R-SUM1 transgene. All primers used are listed in S1 Table.
Trichome visualisation and quantification
For trichome visualization, leaves from 4-weeks old SD-grown plants were harvested and washed in Acetic acid: Ethanol (1:3) solution for overnight to bleach all pigments then visualised under bright field in a fluorescence microscope. The images were captured as 4848 × 3648 pixels from 4.1 mm2 leaf area. The trichome numbers were counted in 10 randomly selected images. Experiment was repeated twice with similar observations.
Salicylic acid (SA) measurements
Free salicylic acid (SA) and glucose-conjugated SA (SAG) measurements were done using the Acinetobacter sp. ADPWH_lux biosensor system [76]. In brief, 100 mg leaf tissues were collected and frozen in liquid nitrogen. The tissue was homogenized in 250μl of acetate buffer (0.1 M, pH 5.6). Samples were centrifuged at 12000 rpm for 15 minutes. One aliquot (100μl) of the supernatant was used for free SA measurements, and another was incubated with 6U of β-glucosidase (Sigma-Aldrich) for 90 min at 37°C for total SA measurement. 20μl aliquot of each plant extract were added to 50μl of secondary culture of Acinetobacter sp. ADPWH_lux (OD600 of 0.4) with additional 60 μl of LB. Standard SA solutions (prepared in sid2-1 extract) were also taken to generate standard curve to calculate the amount of SA present in samples. Plate was incubated at 37°C for 1 h, and luminescence detected using the POLARStar Omega Luminometer (BMG Labtech). Data shown is representative of three biological replicates with SD.
In planta assays for bacterial growth and kinetics of defense-associated marker expressions
Bacterial growth assays were performed according to [46]. Briefly, PstDC3000 strains were infiltrated with a needleless syringe at a density of 5 × 104 cfu ml−1into fully expanded rosette leaves of 3-4-weeks-old SD grown plants. Leaf discs harvested from infiltrated area were macerated in 10mM MgCl2, serially diluted and plated onto selective medium plates. Bacterial growth was determined at 0 and 3-days post-infiltration. Bacterial infiltrations for kinetics of defensive markers were carried as above except a 5 × 106 cfu ml−1 bacterial inoculum was used. Infiltrated samples harvested at 6, 12 and 24 hpi were processed separately for total RNA extraction and qPCR analysis or for immunoblots with indicated antibodies.
Callose deposition assay and image analysis
The callose deposition assays were performed according to [77]. In brief, 4-week-old leaves of SD grown Col-0, sum1-1, sum2-1 and sum3-1 plants were infected with the indicated PstDC3000 strains. At 24 hpi, 3-4 random infiltrated leaves were harvested and washed first in Acetic acid: Ethanol (1:3) solution for overnight to bleach all pigments and then washed for 30 min in 150 mM K2HPO4 solution. The leaves were then incubated in dark for 2 hours in 150 mM K2HPO4 containing 0.01% aniline blue in a 16-well tray. Samples were then embedded in 50% glycerol and observed under a Nikon fluorescence microscope with DAPI filter. Images were analysed using IMARIS 8.0 software for quantifying the number of callose deposits/ mm2 leaf area.
RNA extraction and gene expression analysis by qRT-PCR
Total RNA extraction and cDNA synthesis was performed as described later. All qPCR primers used in this study are listed in S1 Table. qPCRs were performed in QuantStudio 6 Flex Real-Time PCR system (Applied Biosystems) with 5X HOT FIREPol® EvaGreen® qPCR Mix Plus (ROX) (Solis BioDyne) according to the manufacturer’s instructions. All qPCR experiments were replicated thrice with three biological and technical replicates (n=3). SAND (At2g28390) expression was used as internal control [46]. Relative expression was calculated according to the PCR efficiency^−∆∆Ct formula. Expression differences were normalised to Col-0 and plotted as fold change.
Protein extraction and western blotting
For immunoblotting, leaf tissues collected at indicated time points/treatments were homogenised in protein extraction buffer [50 mM Tris HCl (pH 8.0), 8M Urea, 50 mM NaCl, 1% v/v NP-40, 0.5% Sodium deoxycholate, 0.1% SDS and 1 mM EDTA] containing 20 mM N-ethylmaleimide (NEM), 1X plant protease inhibitors cocktail (Sigma Aldrich) and 2% w/v Polyvinylpyrrolidone (PVPP). The homogenates were clarified by centrifugation, mixed with 2X Laemmli buffer, proteins separated by SDS-PAGE and then transferred onto polyvinylidene fluoride (PVDF) membrane by wet transfer method. The membrane was blocked with 5% non-fat skim milk and western blots performed with indicated primary antibodies [anti-SNC1 (Abiocode), anti-PR1 or anti-PR2 (Agrisera), or anti-SUMO1 (Abcam), or anti-Actin C3 antibodies (Abiocode)] in 1X TBST at 4°C for overnight. Comparable protein loading was determined by Ponceau S staining of Rubisco subunit. Blots were washed thrice with 1X TBST and then incubated at RT for one hour with appropriate horse-radish peroxidase (HRP)-conjugated secondary antibodies. The blots developed using ECL™ Prime western blotting system (GE Healthcare) and visualised in ImageQuant™ LAS 4000 biomolecular imager (GE Healthcare).
SUMOylome changes in response to SA or heat shock treatments
For salicylic acid (SA)-induced SUMOylation changes, extracts were obtained from 2-weeks old seedlings treated with 2 mM SA or buffer alone for 1 hour [6]. For heat shock treatments, 2-weeks old seedlings were incubated at 37°C or at RT for 30 minutes. Total protein extracts were immunoblotted as described earlier.
Construction of Plasmid clones
To generate cDNA clones of Arabidopsis SUM1 and SUM3 genes, total RNA isolated from Col-0 plants (RNAiso Plus; Takara) was reverse transcribed (iScript™ cDNA Synthesis Kit; Bio-Rad) according to manufacturer’s instructions. Specific SUM1 and SUM3 (GG and AA forms) sequences were amplified (Phusion High-Fidelity DNA Polymerase; ThermoFisher) from the cDNA using the primers listed (S1 Table). PCR products were cloned into the Gateway entry vector pDONR201 and subsequently into pMDC-cCFP and pMDC-nVenus (BiFC destination vectors; [40] using the Clonase™ Recombination system (ThermoFisher). Confirmed BiFC clones were then electroporated into A. tumefaciens GV3101 strain.
For dimerization studies, cDNAs of SUM1AA or SUM3AA were cloned as EcoRI-SalI fragment into pASK-IBA16 vector (N-terminal Strep-tag II affinity tag; Neuromics). For in vitro isoform influence on SUMOylation assays SUM1GG or SUM3GG were cloned as EcoRI-SalI fragment into pASK-IBA16 vector. These clones were named accordingly as Strep-SUMO1GG, Strep-SUMO1AA, Strep-SUMO3GG, or Strep-SUMO3AA, respectively. For His-tagged versions, pCDFDuet vector harbouring SUM1GG (pKT-973), SUM1AA (pKT-1017), SUM3GG (pKT-975), or SUM3AA (pKT-1788) were used [53]. These plasmids were a kind gift from Prof. Katsunori Tanaka, Kwansei Gakuin University, Japan).
Cloning and generation of SUM3p-His-StrepII-SUM3g transgenic plants
Genomic DNA extracted from Col-0 plants were used for PCR of a ~1.3kbp genomic fragment with KpnI-SUM3p For/XbaI-SUM3-UTR Rev primers. Using the amplicon as a template, two independent PCRs were performed by using KpnI-SUM3p For/His-StrepII Rev or His-StrepII For/ XbaI-SUM3-UTR Rev primers combinations respectively. Individual PCR amplifications were used for overlap PCR with KpnI-SUM3p For/XbaI-SUM3-UTR Rev primers. The product was cloned into XbaI-KpnI site of the binary vector pBIB-Hyg [78]. Generated clones were sequenced and confirmed. The binary vector was introduced in Agrobacterium strain GV3101 via electroporation. Pool of sum3-1 plants were transformed via floral-dip method [79]. Transgenic sum3-1:HS-SUM3 plants were selected on Hygromycin containing medium, propagated through T3 generations to identify lines containing single locus but homozygous sum3-1:HS-SUM3 transgenes. Subsequently, the plants were used for assays as indicated.
Bimolecular Fluorescence Complementation (BiFC) assays
All BiFC assays were performed according to [40]. Briefly, Agrobacterium cells harbouring the indicated BiFC vectors were induced with 150 μM Acetosyringone for 4 hours, equal bacterial density suspensions of desired cCFP and nVenus BiFC combinations made and then infiltrated in leaves of 4-weeks old N. benthamiana plants. At 48-hpi, a small section of infiltrated area was imaged under a SP8 Leica confocal microscopy system using FITC filter (488-nm Argon Laser).
In vitro SUMO binding assays
SUMO-binding assays were performed with recombinant protein expressed in E. coli.
Expression was induced with 200 μg/L Anhydrotetracycline (for pASK-IBA16 clones) or with 0.5 mM IPTG (for pCDFDuet clones) in BL21 (DE3) for overnight at 25°C. Cell pellet was resuspended in lysis buffer (50 mM NaH2PO4, 300 mM NaCl, 10 mM Imidazole; pH 8.0) and sonicated. Equal volumes of His- and Strep-II tagged-SUMO combination supernatants were mixed and incubated at 4°C with rotation for 2-3 hours to allow binding. Ni2+-NTA beads (Qiagen) were added to the mix and incubated further for 2 hrs. His-tagged protein pull down was performed under native conditions according to manufacturer’s instructions. Immunoblots were performed with anti-His-HRP (Santa Cruz Biotech) or anti-StrepII-HRP (IBA Lifesciences) antibodies.
In vitro SUMOylation reconstitution assays
The His-tagged constructs used were according to [53]. Generation of Strep-tagged SUMO isoform clones have been described earlier. E. coli BL21 (DE3) cells containing both pKT-973 (SUMO1GG + SCE1) and pKT-978 (SAE2 + SAE1) was transformed with either empty vector (pASK-IBA16), Strep-SUMO3GG, or Strep-SUMO3AA plasmids. Similarly, BL21 (DE3) cells containing pKT-975 (SUMO3GG + SCE1) and pKT-978 (SAE2 + SAE1) was either transformed with either empty vector (pASK-IBA16), Strep-SUMO1GG, or Strep-SUMO1AA plasmids. As controls, conjugation-deficient SUMO1AA (pKT-1017) or SUMO3AA (pKT-1788) in combination with pKT-978 was used. Transformed cells were induced with 200 μg/L Anhydrotetracycline and 0.5 mM IPTG for overnight at 25°C. After harvesting, cell pellets were lysed and immunoblotted with anti-His or anti-Strep antibodies.
Statistical Analysis
For all gene expression experiments, Student’s t-test was performed to check significance and denoted by one, two and three asterisks (*) indicating p-value <0.05, <0.01, and 0.001, respectively. For growth curve and other assays, ANOVA was performed to check statistical significance in growth of bacteria among different genotypes and indicated by alphabets e.g. a, b, c, d, e etc. which depict statistical difference from each other at p-value <0.001.
Author Contributions
SB conceived the research. IKD, MK and SB designed the research. IKD and MK performed all the experiments. HvdB generated amiR-SUM2 sum3-1 line used in this study. SB helped in plasmid constructions and supervised the experiments. IKD, MK and SB analyzed the data and wrote the manuscript.
Supporting information
Additional supporting information may be found in the online version of this article.
S1 Fig: sum3-1 partially alleviates, whereas eds1-2 and sid2-1 abolishes growth deficiencies of sum1-1.
S2 Fig: sum1-1 and sum2-1 plants are enhanced resistant whereas sum3-1 is hypersusceptible to TNL-specific PstDC3000(hopA1) but not to CNL-specific PstDC3000(avrRpm1) avirulent strains.
S3 Fig: sum1-1 and sum2-1 leaves display increased whereas sum3-1 is deficient in callose deposition in response to virulent PstDC3000(EV) or avirulent PstDC3000(avrRps4) infections.
S4 Fig: Developmental phenotype of 4-week-old SD grown plants of indicated genotypes.
S5 Fig: sum1-1 and sum2-1 display rapid whereas sum3-1 is delayed than Col-0 in induction of PR1 and PR2 transcripts in response to virulent PstDC3000(EV) or avirulent PstDC3000(avrRps4) infections.
S6 Fig: Increased trichome density, elevated expression of PTI markers FRK1 or WRKY29, and enhanced resistance to avirulent PstDC3000(hopA1) in sum1-1 is SA-modulated.
S7 Fig: Enhanced basal immunity and ETI defences in sum2-1 to TNL-specific PstDC3000 strains is EDS1-dependent.
S8 Fig: Developmental defects and enhanced TNL-specific immunity in sum1-1 is partially SUM3 regulated.
S9 Fig: SUM3 buffers increased induction of PR1 in sum1-1 in response to PstDC3000 challenges.
S10 Fig: Schematic representation of key amino acid conservation and divergences among three Arabidopsis SUMO isoforms suggest their functional overlaps/distinctions.
S11 Fig: Complemented sum1-1 or sum3-1 lines have wild-type levels of SUMO1/2 conjugates upon PstDC3000 (EV) infection.
S12 Fig: SUMO1 and SUMO3 cause reciprocal enhancements of SUMO-conjugates in E. coli SUMOylation-reconstitution system.
S13 Fig: SUM3 mutation reduces SUMO1/2-conjugate enhancements during heat-shock.
S1 Table: List of primers used in this study.
Acknowledgements
IKD, MK and SB deeply acknowledge DST-SERB (Grant No: EMR/2016/001899) and Regional Centre for Biotechnology (RCB), Faridabad for providing financial support and central instrumental facilities. We express our gratitude to Prof. Walter Gassmann, Missouri University, USA, and Prof. Ashis Kumar Nandi, Jawaharlal Nehru University (JNU), New Delhi, India for providing PstDC3000 strains and sid2-1 lines and SA measurements, respectively used in this study. We also thank Dr. Souvik Bhattacharjee (JNU) for providing pASK-IBA16 plasmid and Prof. Katsunori Tanaka, (Kwansei Gakuin University, Japan) for providing E. coli SUMOylation reconstitution system constructs. IKD acknowledges Department of Biotechnology (DBT), Government of India for providing fellowships during his PhD tenure and KIIT University, Bhubaneswar, India for his PhD registration. MK thanks UGC for his PhD fellowship.
References
- 1.↵
- 2.↵
- 3.↵
- 4.↵
- 5.↵
- 6.↵
- 7.↵
- 8.↵
- 9.↵
- 10.↵
- 11.↵
- 12.↵
- 13.↵
- 14.
- 15.↵
- 16.↵
- 17.↵
- 18.↵
- 19.↵
- 20.↵
- 21.↵
- 22.↵
- 23.↵
- 24.↵
- 25.↵
- 26.
- 27.↵
- 28.↵
- 29.↵
- 30.↵
- 31.
- 32.↵
- 33.↵
- 34.↵
- 35.↵
- 36.↵
- 37.↵
- 38.
- 39.↵
- 40.↵
- 41.↵
- 42.↵
- 43.↵
- 44.↵
- 45.↵
- 46.↵
- 47.↵
- 48.↵
- 49.↵
- 50.↵
- 51.↵
- 52.↵
- 53.↵
- 54.↵
- 55.↵
- 56.↵
- 57.↵
- 58.↵
- 59.↵
- 60.↵
- 61.↵
- 62.↵
- 63.↵
- 64.↵
- 65.↵
- 66.↵
- 67.↵
- 68.
- 69.↵
- 70.↵
- 71.↵
- 72.↵
- 73.↵
- 74.↵
- 75.↵
- 76.↵
- 77.↵
- 78.↵
- 79.↵