Global SUMOylome Adjustments in Basal Defenses of Arabidopsis thaliana Involve Complex Interplay Between SMALL-UBIQUITIN LIKE MODIFIERs and the Negative Immune Regulator SUPPRESSOR OF rps4-RLD1

Steady-state SUMOylome of a plant is adjusted locally during developmental transitions and more globally during stress exposures. We recently reported that basal immunity in Arabidopsis thaliana against Pseudomonas syringae pv tomato strain DC3000 (PstDC3000) is associated with strong enhancements in the net SUMOylome. Transcriptional upregulations of SUMO conjugases, suppression of protease, and increased SUMO translations accounted for this enhanced SUMOylation. Antagonistic roles of SUMO1/2 and SUMO3 isoforms further fine-tuned the SUMOylome adjustments, thus impacting defense amplitudes and immune outcomes. Loss of function of SUPPRESSOR OF rps4-RLD1 (SRFR1), a previously reported negative regulator of basal defenses, also caused constitutive increments in global SUMO-conjugates through similar modes. These suggest that SRFR1 plays a pivotal role in maintenance of SUMOylation homeostasis and its dynamic changes during immune elicitations. Here, we demonstrate that SRFR1 degradation kinetically precedes and likely provides the salicylic acid (SA) elevations necessary for the SUMOylome increments in basal defenses. We show that SRFR1 not only is a SUMOylation substrate but also interacts in planta with both SUMO1 and SUMO3. In sum1 or sum3 mutants, SRFR1 stabilities are reduced albeit by different modes. Whereas a srfr1 sum1 combination is lethal, the srfr1 sum3 plants retain developmental defects and enhanced immunity of the srfr1 parent. Together with increasing evidence of SUMOs self-regulating biochemical efficiencies of SUMOylation-machinery, we present their impositions on SRFR1 expression that in turn counter-modulates the SUMOylome. Overall, our investigations reveal multifaceted dynamics of regulated SUMOylome changes via SRFR1 in defense-developmental balance.

Maintenance of SUMOylome homeostasis necessitates tight coordination of SUMO-conjugation/deconjugation cycles (Kurepa et al., 2003;Morrell and Sadanandom, 2019). SUMOylation requires the availability of processed SUMOs, with exposed diglycine (GG) residues at their C-terminus. SUMO proteases generate these conjugation-proficient SUMOs from precursors that contain extended C-terminus residues. The mature SUMOs through an ATP-dependent AMP-SUMO intermediate formation is linked via thiol-ester bond to the catalytic cysteine of SUMO E1 ACTIVATING ENZYME (SAE), a heterodimer of two subunits SAE1 and SAE2. SUMO E2 CONJUGATING ENZYME (SCE) then acquires the SUMO moiety from SAE via trans-esterification. Although SCE is capable of direct SUMOylation, its binding to SUMO E3 ligases, which also simultaneously interact with targeted substrates, augments the process and imparts specificity (Gareau and Lima, 2010). SUMOs are also self-SUMOylated at internal lysines forming poly-SUMO chains. These are catalyzed by SUMO E4 ligases that belong to the PIAL (protein inhibitor of activated stat-like) class of proteins (Tomanov et al., 2014). Some SUMO proteases remove SUMOs from SUMOylated substrates or disintegrate polySUMO chains, thus recycling free SUMOs (Chosed et al., 2006;Colby et al., 2006;Morrell and Sadanandom, 2019).
In broadly understood layers of plant defenses, conserved molecular signatures or pathogen-associated molecular patterns (PAMPs) present on microbes are sensed by extracellular transmembrane pattern-recognition receptors (PRRs) to transduce downstream induction of defense-associated genes (Jones and Dangl, 2006;Bentham et al., 2020). This route of immune signaling comprise the PAMP-triggered immunity (PTI). Intracellular perception of a pathogen attack is performed by strategically deployed resistance (R) proteins that directly or indirectly sense manipulations by the invader-secreted effectors. Defense responses elicited downstream of these perceptions are termed as effector-triggered immunity (ETI) and include heightened production of defensive hormone salicylic acid (SA), and prolonged and aggravated expression of PTI-responsive markers such as PATHOGENESIS-RELATED PROTEINS (PRs) and SA-biosynthesis SALICYLIC ACID-DEFICIENT 2/ISOCHORISMATE SYNTHASE 1 (SID2/ICS1) gene, among others. Basal and ETI mediated by the Toll-interleukin1 (TNL)-type of R proteins require ENHANCED DISEASE SUSCEPTIBILITY 1 (EDS1) to potentiate SA-based defenses (Aarts et al., 1998;Falk et al., 1999).
Disturbances in SUMO pathway affect immune responses often with developmental costs to the host (Lee et al., 2007;van den Burg et al., 2010;Bailey et al., 2016;Orosa et al., 2018;Verma et al., 2018). Studies overall suggest that SUMO1/2 suppresses whereas SUMO3 potentiates immunity primarily through modulation of SA-signaling networks (van den Burg et al., 2010;Saleh et al., 2015;Ingole et al., 2021b). However, adjustments of a host SUMOylome in response to pathogen attack or in SUMOylation-perturbed mutants present a more complicated involvement not only of SUMO isoforms but also of SUMO-machineries. For example, globally reduced SUMOylome in siz1-2 or enhanced SUMO1/2-conjugates in esd4-2 or ots1 ots2 both lead to elevated SA levels with constitutive activation of defenses (Villajuana-Bonequi et al., 2014;Bailey et al., 2016). ESD4 and SIZ1 are SUMOylated and non-covalently bind SUMOs, implying that their activities are self-regulated by SUMOylation (Miller et al., 2010;Mazur et al., 2017). Our studies recently identified both positive and negative immune regulators as differentially SUMOylated candidates upon a pathogen attack (Ingole et al., 2021a). Further, we also revealed that SUMOconjugation efficiencies are affected by the crosstalk between SUMO isoforms independent of their covalent-modification activities. Taken together, how a plant maintains SUMOylome homeostasis, performs response-appropriate modifications on substrates, and prevents fitness costs remains uncharacterized.
We recently showed that in the autoimmune mutant srfr1-4 constitutive and during PstDC3000 infections on wild-type plants progressive, increments of SUMO1/2 conjugations were observed, respectively (Ingole et al., 2021a). This implies a genetic link of SRFR1 to SUMOylome maintenance and responsive adjustments. Here, we reveal that SRFR1 potentiates PstDC3000-induced SUMOylome changes, undergoing transient instability that kinetically precedes SA accumulations and global SUMOylome enhancements. With known link in suppressing SID2/ICS1 expressions, SRFR1 reduction thus likely provides the SA stimulus previously shown to enhance SUMO-conjugations (Bailey et al., 2016). We deduce here that in addition to being a SUMOylation candidate, SUMOylome perturbation also reciprocally affects SRFR1 expressions at both transcriptional and post-transcriptional levels. Further, enhanced basal defenses or SUMO1/2-conjugation increments are negligibly SUM3-dependent in srfr1-4 plants. Overall, our investigations here present SRFR1 role in SUMOylome homeostasis and perturbations providing fine-tuning of immune amplitudes.

Salicylic Acid Measurements
Free and total SA (SA + glucose-conjugated, SAG) measurements were performed according to Defraia et al. (2008) using the Acinetobacter sp. ADPWH_lux biosensor system. Briefly, 100 mg of frozen tissue was homogenized in 250 µl of acetate buffer (0.1 M, pH 5.6) and clarified by centrifugation at 12,000 rpm for 15 min to remove cell debris. A 100 µl of the supernatant was kept for measuring free SA, while a similar volume was treated with 6 U of β-glucosidase (Sigma-Aldrich, St. Louis, MO, United States) for 90 min at 37 • C for total SA determination. A 20 µl aliquot of plant extract was mixed with 50 µl of Acinetobacter suspension (grown to OD 600 = 0.4) along with 60 µl of fresh Luria-Bertani (LB) broth. Standard curve was generated with sid2-1 tissue extracts spiked with known amounts of SA (Sigma-Aldrich, United States). After incubation for 1 h at 37 • C, luminescence was measured using a Luminometer (POLARStar Omega, BMG Labtech, Ortenberg, Germany). At least three biological replicates were used for each measurement, and data were reported as mean ± SD.  (Kim et al., 2010). Each experiment was repeated at least twice with three biological and technical replicates.
For immunoprecipitation assays between HA-SRFR1 and Myc-EDS1 or HA-PAD4 and Myc-EDS1, agro-infiltrated Nicotiana benthamiana leaves were used 48 h post-infiltration. Tissues were ground in chilled radioimmunoprecipitation assay (RIPA) buffer [5 mM of Tris-HCl pH 7.5, 150 mM of NaCl, 10 mM of MgCl 2 , 1 mM of EDTA, 1% NP-40, 1 mM of sodium deoxycholate, and 1 × protease inhibitor cocktail (Sigma-Aldrich, United States)] and clarified by centrifugation. Supernatant was precleared with IgG agarose beads (Sigma-Aldrich, United States) rotated at 4 • C for 1 h. The mix was then centrifuged to remove non-specific proteins bound on the IgG agarose beads, and supernatant was added to anti-HA-conjugated beads (Sigma, United States). After tumbling 3 h at 4 • C, the suspension was centrifuged, and agarose-bead pellet washed three times with RIPA buffer. Beads were then resuspended in loading dye and used for indicated immunoblots.
For MG132 treatments, the proteasome inhibitor at a dose of 100 µM in 10 mM of MgCl 2 , 10 mM of 2-(Nmorpholino)ethanesulfonic acid (MES) was infiltrated into plant tissues 12 h before harvesting.
Densitometric quantifications of protein bands in immunoblots were performed with the ImageJ software.

Construction of Clones for Escherichia coli SUMOylation Assay
SUM1 cDNA sequences in pCDFDuet TM1 vector were mutagenized via overlapping oligos at position 91 to incorporate arginine (R) replacing the threonine (T) residue. This resulted in SUMO1 T91R, which imparts smaller SUMOylation footprint of RGG on the modified lysine. For generating FLAG-SRFR1, cDNA sequence was directionally cloned as a NotI restriction fragment at the identical site of pFLAG-TEV vector (BioBharati LifeScience, India). For generating T7-EDS1, the cDNA sequence was cloned as a SalI-XhoI fragment into the pET28a(+) vector (Novagen, Madison, WI, United States). Details of primers used here are listed in Supplementary Table 1.

Sample Preparation and Liquid Chromatography-Tandem MS Analysis for Determination of SUMOylation Sites
Processing for in-gel trypsin digestion and MS analysis was according to Ingole et al. (2021a). Briefly, proteins from in vitro SUMOylation reaction were separated on 6% SDS-PAGE, stained with Coomassie brilliant blue (CBB R-250) solution, and destained; and gel slices were excised as 1-mm 3 pieces with sterile surgical blade. After processing as earlier, trypsin digestion (Promega, Madison, WI, United States) was performed; samples were desalted with C18 tips (Thermo Fisher Scientific, United States) according to manufacturer's instructions and dried in a speed-vac centrifuge. Tandem MS (MS/MS) analysis was done using TripleTOF R 5600+ (AB SCIEX, Redwood City, CA, United States) mass spectrometer instrument. Raw MS data files were searched for peptide sequences (MASCOT software, Matrix Science, Boston, MA, United States) against the A. thaliana protein database 1 .

Construction of Bimolecular
Fluorescence Complementation, GFP-Tagged SUM1 GG/AA Vectors, and in planta Assays Bimolecular fluorescence complementation (BiFC) vectors for SRFR1, EDS1, SUMO1 GG/AA, and SUMO3 GG/AA have been described earlier (Bhattacharjee et al., 2011;Ingole et al., 2021b). Clones were introduced into Agrobacterium tumefaciens GV3101 strain via electroporation. For in planta interaction assays, Agrobacterium strains harboring the indicated BiFC vectors were cultured overnight in LB broth, centrifuged, and resuspended in 10 mM of MgCl 2 , 10 mM of MES containing 150 mM of acetosyringone. After induction for 3-4 h, indicated combinations were mixed at equal bacterial density and infiltrated in fully expanded leaves of 4-week-old N. benthamiana. At 48 h post-infection (hpi), tissue sections from the infiltrated area were excised and imaged under a 40 × oil objective in SP8 Leica confocal microscope (Leica microsystems, Wetzlar, Germany) using the fluorescein isothiocyanate (FITC) filter (488-nm argon laser).
Construction of HA-SRFR1 K325R , HA-SRFR1 K427R , HA-SRFR1 K325R + K427R , GFP-SUMO1 GG/AA Forms, or SUMO3, Myc-EDS1 WT , or Myc-EDS1 K478R Expression Vectors and Generation of Transgenic Plants The construction of HA-SRFR1 expressing native promoterdriven HA-epitope tagged genomic clone of SRFR1 has been described earlier (Kim et al., 2010). With the use of this vector as a backbone, the indicated lysine residues were converted to arginine using overlapping primers (listed in Supplementary  Table 1). pDONR201 clone of SUMO1 GG or AA, SUMO3-GG described earlier (Ingole et al., 2021b), was subcloned via Gateway reaction into pSITE-2CA binary vector that carries an N-terminal GFP tag (Chakrabarty et al., 2007). The above clones were confirmed by sequencing and then electroporated into Agrobacterium GV3101 strain. Transformants obtained were cultured; and srfr1-4 or sum3-1, as appropriate, was transformed via floral-dip method (Clough and Bent, 1998). Transgenic plants were selected on kanamycin (for HA-SRFR1 or GFP-SUMO3 transformations) containing plant growth media and then propagated to T2 and T3 generations and genotyped. Simultaneously, expression of GFP-SUMO3 was also determined via anti-GFP immunoblots. GFP-SUM3 Line#1 was crossed to HA-SRFR1 expressing plants and F2/F3 population genotyped for sum3-1, srfr1-4, mutations, and HA-SRFR1 or GFP-SUM3 transgene homozygosity. The plants were then used for further analysis, as indicated.
For co-expression assays with HA-SRFR1, Agrobacterium strains with GFP alone (pSITE-2CA) or GFP-SUM1 GG or AA forms (in pSITE-2CA backbone) was co-infiltrated as above. Total protein from infiltrated tissues was isolated at 48 hpi and used for immunoblot with anti-HA antibodies.
Construction of Myc-EDS1 WT binary has been described earlier (Bhattacharjee et al., 2011). With the use of the pDONR201-EDS1 WT , the lysine 478 residue was mutated to arginine (EDS1 K478R ) by overlap PCR. This was then subcloned into pBA-Myc binary vector via Gateway-based cloning. Plant transformations with eds1-2 were done as earlier. Independent transgenic lines were selected on 1/2 MS + Basta (10 µg/ml) plates and grown to F3 for assays. Expression of recombinant protein was checked with anti-Myc immunoblots. Primers for various cloning are listed in Supplementary Table 1.

In planta Qualitative and Quantitative Bacterial Growth Assay
Bacterial growth assays were performed according to standardized protocol in Kim et al. (2010). Briefly, fully expanded leaves of 3-to 4-week-old SD-grown plants were infiltrated with a needleless syringe at a bacterial density of 10 6 cfu/ml (half-leaf infiltrated for qualitative disease symptom assay) or 5 × 10 4 cfu/ml (full leaf infiltrated for quantitative bacterial growth assays) with virulent PstDC3000 or avirulent PstDC3000 (avrRps4) strains. For disease symptom determinations, infiltrated leaves were detached 4 days post-infiltration (dpi) and imaged. For quantitative measurements on bacterial growth, leaf disks of defined area were harvested at 0 and 3 dpi, macerated in 10 mM of MgCl 2 , and plated with serial dilution on appropriate antibiotic containing media plates. Anti-SUMO1/2 immunoblot on Col-0, sum1-1, sum2-1, and sum3-1 post-PstDC3000 challenge (10 6 cfu/ml bacterial inoculum) was performed at 24 hpi.

SRFR1 Instability Precedes Global SUMO1/2ylation and Salicylic Acid Elevations During Basal Defenses
Recently, we demonstrated that SRFR1 regulates Arabidopsis SUMOylome maintenance (Ingole et al., 2021a). The srfr1-4 plants contained elevated levels of SUMO1/2-conjugates and displayed upregulated SUMO conjugation-promoting enzymes (SCE1, SAE1/2, SIZ1, and HPY2) and downregulated SUMO-protease (ESD4 and ELS1) expressions with preferential recruitment of SUM1 transcripts on polysomes. Though SUMOylome increments were also observed when wildtype (Col-0) plants were challenged with PstDC3000, SRFR1 functions were not directly attributed to these responses. To determine this, HA-SRFR1/srfr1-4 plants (referred hereafter as HA-SRFR1) expressing HA-epitope tagged SRFR1 protein from its native promoter were used. These plants were shown earlier to complement the growth and defense defects of srfr1-4 (Kim et al., 2010). HA-SRFR1 plants were challenged with PstDC3000, and tissue extracts from progressive time points were immuno-probed with anti-HA antibodies. Decrease in SRFR1 protein levels was noted as early at 1 hpi and persisted moderately at 3 and 6 hpi, after which gradual restoration to near wild-type levels was noted by 24 hpi (Figure 1A). Transcripts of SRFR1 showed progressive upregulation increasing from 1 to 6 hpi (∼2-to 2.5-fold) and restored to endogenous levels by 24 hpi (Figure 1B). Immunoblot with anti-SUMO1/2 antibodies showed gradual increments in SUMO1/2-conjugations from 1 to 24 hpi ( Figure 1C). Bailey et al. (2016) demonstrated that SA application enhances SUMO1/2-conjugations in Col-0. With previously known functions in suppressing SA-based defenses, to determine whether transient instability of SRFR1 preceded SA increase to cause SUMOylome changes, we measured free SA and its glucose-conjugate (SAG) levels in the above extracts. Significant increase in free and total (SA + SAG) levels (∼4and ∼8-fold, respectively) was first noted at 6 hpi, and by 24 hpi, drastic elevations were observed (∼15-and ∼55fold, respectively) ( Figure 1D). This moderately correlated to the upregulation in SID2/ICS1 expressions ( Figure 1E). Taken together, these data suggest that transient decrease in SRFR1 protein is linked to SUMOylome perturbation via SA-signaling routes. . Blots were probed with anti-HA or anti-SUMO1/2 antibodies, as indicated. Ponceau S-stained membrane or anti-actin immunoblot shows comparable protein loading across samples. Numbers above the HA-SRFR1 protein bands in panel (A) are densitometric quantification values relative to actin expression levels in the same extracts. Transcript abundance is relative to internal control MON1 gene expressions and presented as fold-change relative to uninfected (0 hpi) samples (n = 3). Statistical significance is by pairwise comparison with 0-hpi sample with Student's t-test (**p < 0.01, ****p < 0.0001, and ns = not significant).

SRFR1 Is a SUMOylation Candidate
SRFR1 harbors several intrinsically disordered regions (IDRs) especially between its multiple TPRs and the C-terminal domain (Gassmann and Bhattacharjee, 2012). Protein with IDR features often has various interacting partners adjusted according to the cellular signaling needs (Haynes et al., 2006;Liu et al., 2009). Most often, IDRs are also enriched for motifs that attract PTM changes (Gao and Xu, 2012;Narasumani and Harrison, 2018). In SRFR1, two sites that match canonical SUMOylation motifs (LK 325 EE and LK 427 QE) are predicted by the SUMOsp2.0 tool (Ren et al., 2009; Figure 3A). Plus, at least two each of high-scoring SIMs and non-consensus SUMOylation sites are also present in SRFR1. To test whether SRFR1 is SUMOylated by the Arabidopsis SUMO isoforms, we utilized the Escherichia coli SUMOylation reconstitution FIGURE 3 | SRFR1 is a SUMOylation and SUMO-binding candidate. (A) Prediction of putative SUMOylation and SIM motifs in SRFR1 and EDS1. In silico prediction was performed using SUMOsp2.0 tool. (B) SRFR1 is SUMOylated by SUMO1 T91R, SUMO2, or SUMO3 GG forms in Escherichia coli SUMOylation reconstitution system. Immunoblot was probed with anti-FLAG antibodies to detect FLAG-SRFR1. Migration position of molecular weight standards (in kDa), SUMOylated or non-SUMOylated SRFR1 (SRFR1 SUMO or SRFR1, respectively), is indicated. (C) Bimolecular fluorescence complementation (BiFC) panels of co-expressed nVenus-SRFR1 with cCFP-clones of SUMO1 GG/AA, SUMO3 GG/AA, or negative control GUS (β-glucuronidase). Images were acquired using a confocal microscope with GFP or bright-field filters. Merged images of YFP with bright field are shown. Scale bar = 50 µm. system, with minor modifications (Okada et al., 2009). The N-terminal 91st residue N-terminal to GG in SUMO1 was modified from threonine to arginine (T91R) to facilitate smaller SUMO footprints on targets and improved identification via MS. SUMO3 contains a relatively proximal arginine at the 88th position to the diglycine G 92 G 93 ends. MYB30, a previously known target tested by Okada et al. (2009), was SUMOylated by SUMO1 T91R variant (Supplementary Figure 2). SRFR1 cDNA sequences were co-expressed with a FLAG-epitope tag (FLAG-SRFR1) in the presence of SUMO1 T91R, SUMO2, or SUMO3 SUMOylation-competent (containing diglycine, GG) or SUMOylation-deficient (diglycine replaced by dialanine, AA) isoforms. Induced E. coli extracts probed with anti-FLAG antibodies showed a slower migrating band than FLAG-SRFR1 in the presence of only GG, but not AA forms of the coexpressed SUMOs (Figure 3B). This band was more intense when SUMO3 was co-expressed than with SUMO1/2. These results suggested that SRFR1 is a potential SUMOylation candidate of multiple SUMO isoforms.
We subjected the slower migrating SRFR1 protein bands to trypsin digests and liquid chromatography (LC)-MS/MS analysis. SUMO1, but not SUMO2 or SUMO3, footprints on selective lysine residues of SRFR1 were detected in the identified peptides (Supplementary Figure 3). Remarkably, K 325 in the predicted LK 325 EE SUMOylation motif was identified with SUMO1-modifications. In addition, SRFR1 K 229 was also covalently modified by SUMO1, though conserved features of a SUMOylation motif were absent for this lysine (CK 229 PC). These results demonstrated SRFR1 as a SUMOylation candidate, at least in the E. coli system. To test the functional relevance of in silico predicted SRFR1 SUMOylation sites, we generated native promoter-driven HA-SRFR1 expressing transgenic plants containing K325R, K427R, or K325R + K427R mutations (named as HA-SRFR1 K325R , HA-SRFR1 K427R , or HA-SRFR1 K325R + K427R , accordingly), in the srfr1-4 background. Functional complementation of the expressed SRFR1 variant is expected to abolish growth defects of srfr1-4. Transgenic plants expressing HA-SRFR1 K325R , HA-SRFR1 K427R , or HA-SRFR1 K325R + K427R versions of SRFR1 restored wild-type growth in srfr1-4 (Supplementary Figure 4). These results implied that R substitutions of at least these K residues do not affect SRFR1 function. With one additional lysine residue (K 229 ) identified in SRFR1 as a potential SUMO-acceptor, similar mutational and complementation studies are needed to decipher its functional relevance. Detection in planta of SUMO1-modifications on K 721 of TPR1, which is not a predicted SUMOylation site, encourages possibility of a non-conventional motif in SUMO-influences on SRFR1 (Niu et al., 2019).
To evaluate non-covalent binding of SUMOs to SRFR1, we utilized the BiFC assays. N. benthamiana leaves transiently expressing SRFR1 showed restoration of YFP signal with SUMO1 or SUMO3 co-expressed as GG (SUMOylationcompetent) forms ( Figure 3C). When SUMOs were expressed as AA (SUMOylation-deficient) forms, only SUMO3 but not SUMO1 bound SRFR1. As controls, neither SRFR1 nor SUMOs interacted with the negative control beta-glucuronidase (GUS) protein or the corresponding empty vector (Figure 3C and Supplementary Figure 5). These results suggest that SRFR1 is a likely SUMOylation target as well as SUMO-binding protein.

EDS1 K 478 , a Predicted SUMOylated Residue, Is Essential for Interaction With SRFR1
To test for possible functional relevance of SRFR1 SIMs, we utilized its previously known interaction with EDS1 (Bhattacharjee et al., 2011). As a negative immune regulator, SRFR1 likely sequesters EDS1 from activation of defenses. A high-scoring LK 478 NE matching the consensus SUMOylation motif is identified in the EDS1 protein sequence (Figure 3A). EDS1 K 478 is located in the EP domain and is bracketed by residues R 475 and D 481 that form salt bridges with the corresponding loop residues in its partner SAG101 (Wagner et al., 2013). When tested in the E. coli-SUMOylation system, EDS1 was not SUMOylated by any of the SUMO isoforms (Supplementary Figure 6). Interestingly, in BiFC assays, coexpression of EDS1 and SUMO1 GG but not SUMO1 AA or SUMO3 GG/AA allowed reconstitution of YFP fluorescence, hinting that SUMO1 is a direct modifier of EDS1 ( Figure 4A). To test the relevance of EDS1 K 478 in plant defenses, we generated binary vectors that express CaMV 35S promoterdriven Myc-epitope tagged wild-type (Myc-EDS1 WT ) or K478R (lysine replace by arginine; Myc-EDS1 K478R ) versions of EDS1 cDNA. The binary vectors were then used to generate transgenic plants in the EDS1-null (eds1-2) background (Cui et al., 2017).
To detect for SUMOylated EDS1 in planta, we obtained F1 plants that co-expressed EDS1-YFP and His-H89R-SUM1. The crossed parents expressed EDS1-YFP or His-H89R-SUM1 that functionally complement the loss of respective endogenous EDS1 or SUM1 (Garcia et al., 2010;Miller et al., 2010). Expression of His 6 -SUM1-H89R facilitates improved affinity-based enrichment of SUMO1-SUMOylated proteins. EDS1-YFP (control) or EDS1-YFP/His-H89R-SUM1 plants were sprayed with SA to mimic defense responses and then enriched for His-SUMO1-conjugated proteins with Ni 2+ -affinity chromatography, as earlier (Ingole et al., 2021a). Eluates showed enrichment of SUMO1-conjugates and when probed with anti-GFP antibodies identified a protein band at expected migration positions for SUMOylated EDS1-YFP (EDS1 SUMO1 ) only in EDS1-YFP/His 6 -SUM1-H89R but not EDS1-YFP samples (Supplementary Figure 7). Input extracts from both samples displayed similar levels of EDS1-YFP expression. Because of their low amounts, we were unable to perform MS analysis on the hypothesized EDS1 SUMO1 protein bands. In a parallel approach, we transiently co-expressed Myc-EDS1 WT or Myc-EDS1 K478R with GFP-SUMO1 or His-StrepII-SUM3g (Ingole et al., 2021b) in N. benthamiana leaves and probed whether SUMO1/3 co-expression caused reduced migration of EDS1 indicative of its SUMOylation. Immunoblot, however, showed similar migration of Myc-EDS1 WT across all SUMO combinations (Supplementary Figure 8). Thus, despite our attempts, reasonable doubt whether in planta EDS1 is SUMOylated (on K 478 ) persists, and further investigations to elucidate this are therefore warranted.

sum1-1 or sum3-1 Affect Endogenous Levels and Dynamics of SRFR1 During Basal Defenses
With indications of SRFR1 as a potential SUMOylation and SIM-harboring candidate, we tested whether enhanced or compromised defenses, respectively, in sum1-1 or sum3-1 (Ingole et al., 2021b), are the result of changes in SRFR1 levels. Toward this, we generated HA-SRFR1/sum1-1 plants by crossing sum1-1 to the HA-SRFR1 line. Plants with homozygous sum1-1 mutation displayed reduced HA-SRFR1 protein in comparison with the HA-SRFR1 parent (Figure 5A). To determine whether the lower protein levels were due to transcriptional or post-transcriptional effects, we measured SRFR1 transcript abundance. In sum1-1 plants, SRFR1 transcripts displayed only a slight reduction (∼0.8fold), suggesting that reduced SRFR1 protein is mostly due to post-transcriptional effects ( Figure 5B). Addition of the 26S proteasome inhibitor MG132 slightly improved HA-SRFR1 levels in HA-SRFR1/sum1-1 plants. Furthermore, GFP-SUMO1 GG or AA variant co-expressed with HA-SRFR1 in N. benthamiana leaves caused an increase in HA-SRFR1 protein abundance in comparison with the GFP alone control (Figure 5C). Taken together, our results imply that SUM1 role as a negative immune regulator may be via maintaining the steady-state levels of SRFR1.
Unlike what was noted for HA-SRFR1/sum3-1, MG132 treatment stabilized HA-SRFR1 protein in the GFP-SUM3 overexpressing plants, suggesting that as in sum1-1, the proteasome pathway was responsible for lower SRFR1 accumulations. Together with known SUM3 overexpression enhancing basal defenses, our results connect these responses again to consequences of reduced SRFR1 levels (van den Burg et al., 2010).
To further substantiate this, we challenged HA-SRFR1/sum1-1 or HA-SRFR1/sum3-1 plants with virulent PstDC3000 and evaluated SRFR1 protein changes ( Figure 5E). Interestingly, lack of SUM1 delayed SRFR1 restoration, and even at 24 hpi, HA-SRFR1 levels were barely detectable. Contrastingly, HA-SRFR1 recovery was rapid in sum3-1 reaching native levels by 6 hpi. Accumulation of SUMO1/2-conjugates with PstDC3000 challenge was dramatically lower in sum3-1 and in agreement with our earlier observations (Ingole et al., 2021b; Figure 5F). Together, these data demonstrated intricate interplay between SUMO isoforms in basal defenses with SUM1 essential for SRFR1 restitution at post-transcriptional level and SUM3 modulating transcriptional efficiency as a part of feedback loop mechanism.
Several TFs are SUMOylation targets (Miller et al., 2010). SUMOylome disturbances cause expression differences of several SUMOylation-associated genes (Ingole et al., 2021b). To evaluate this in esd4-2 combination mutants, we performed quantitative real-time PCRs (qRT-PCRs). Expressions especially for SIZ1 and HPY2, the two E3 ligases, were higher and SUM3-independent in esd4-2 (Supplementary Figure 10C). SUM1 transcripts remain unaffected in all plants tested (Supplementary Figure 10B). Taken together, these data reflected consequences of deficient ESD4 activity rather than increased SUMOylation efficiencies by upregulated SIZ1/HPY2 or due to SUM3 involvement in their expressions as the primary cause of elevated SUMO1/2conjugates in esd4-2 plants.
Enhanced Basal Defenses in srfr1-4 Are Partially SUM3-Dependent Our results here and earlier suggested mutual expression influences between SRFR1 and SUMOylation-associated genes (Ingole et al., 2021a). To evaluate this interplay in the context of enhanced immunity in srfr1-4, we attempted to generate srfr1-4 sum1-1 or srfr1-4 sum3-1 double mutants. Considering SUM1 role as a negative immune regulator and SUM3 as a positive immune regulator, we tested whether pathogenesis outcomes of srfr1-4 are altered. From the F2/F3 segregating population of sum1-1 crossed with HA-SRFR1 plants used earlier, we screened for srfr1-4 sum1-1 double mutants. Even after extensive screening, srfr1-4 sum1-1 plants were not obtained, suggesting their embryonic lethality. It remains a possibility that developmental consequences that occur in srfr1-4 are compounded by the loss of SUM1. Observed growth and immune enhancements that we reported recently in sum1-1 support our speculation (Ingole et al., 2021b).
SA targets SUMO proteases OTS1/2 for proteasomemediated degradation, thus enhancing SUMO1/2-conjugate levels (Bailey et al., 2016). We show here that SRFR1 undergoes transient instability when basal defenses are induced upon PstDC3000 exposure. The reduction in SRFR1 correlates with subsequent increase in SA levels and SUMO1/2-conjugate enhancements. Considering SA degrades OTS1/2, SRFR1 changes via downstream SA increase may drive the condemned fates of OTS1/2, presenting another example of its host SUMOylome adjustment role at the post-transcriptional level (Bailey et al., 2016). Recuperation of SRFR1 levels at latter time points of basal defense elicitation perhaps is indicative of fine-tuning of immune amplitudes, preventing overshoots or pleiotropic consequences like in srfr1-4 or in the SUMOylationperturbed mutants such as siz1-2 or esd4-2 (Lee et al., 2007;Kim et al., 2010;Villajuana-Bonequi et al., 2014).
SUMOylome changes in turn regulate SRFR1 expressions. Firstly, in sum1-1 plants, unchanged SRFR1 transcripts but reduced protein levels are stabilized by the proteasome inhibitor MG132. Secondly, increased SRFR1 protein accumulation is detected by transient overexpression of SUM1 in N. benthamiana leaves. Thirdly, SRFR1 restoration is deficient in sum1-1 plants when challenged with PstDC3000. Last but not the least, gradual restoration of SRFR1 protein levels parallels progressive increase in SUMO1/2-conjugates during PstDC3000 challenges. Overall, these results suggest that SUM1 affects SRFR1 at a posttranscriptional level, possibly regulating protein turnover. Unlike sum1-1, reduced SRFR1 protein in sum3-1 matches its lower transcript levels. Taking into consideration our recent data that SUM3 potentiates SUMO1/2-conjugation efficiencies and in sum3-1 plants endogenous as well as defense-induced increase in SUMO1/2-conjugates are deficient, lower SRFR1 levels in this mutant likely also incorporate additive effects from this intersection (Ingole et al., 2021b). Thus, sum3-1 partially mimics sum1-1 consequences in destabilizing SRFR1. This notion is supported by our observation that MG132 treatment improves SRFR1 protein stability to a modest extent in sum3-1. Immune responses, however, remain deficient in sum3-1 likely because of functional inadequacies in the SA-signaling sector through NPR1, a known exclusive SUMO3-substrate (Saleh et al., 2015).
SUM3 expression is induced transiently at early time points of SA application (van den Burg et al., 2010). Increased SA levels and concomitant upregulation of SRFR1 transcripts during basal defenses noted in our data imply transcriptional contributions of SUM3 as we suggested above. However, SRFR1 instability also noted at same time points leads us to speculate that SUMO3, antagonistic to SUMO1, may negatively affect SRFR1 protein accumulation upon SUMOylation. Our data that SRFR1 is unstable in SUM3 overexpressing transgenic lines are in accordance with this hypothesis. Presence of multiple predicted SUMOylation motifs and in vitro SUMOylation by SUMO3 more prominent than for SUMO1, indeed, present SRFR1 as a candidate whose functions/stabilities may in turn be affected by host SUMOylome/SUMO isoform ratio changes. Stimulus-driven SUMO isoform switches that affect the stability of a substrate have been widely documented in animal systems (Meulmeester et al., 2008;Zhu et al., 2008Zhu et al., , 2009). 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 . Similarly, HDAC1 is targeted for degradation upon SUMOylation by SUMO1, but not SUMO2 (Citro et al., 2013). Reduced SRFR1 levels in sum1-1 plants can therefore be also attributed to increased SUM3 expressions (Ingole et al., 2021a). However, that the same is not noted for esd4-2 plants with elevated SUM3 levels indicates that ESD4 functions are necessary for this process. Indeed, involvement of a SUMO-protease has been reported recently for stimulus-dependent SUMO paralog switching and its impact on the substrate stability (Fasci et al., 2015).
Bioinformatics predictions of SIMs in SRFR1 taken together with non-covalent binding of multiple SUMO isoforms also present another mode by which a change in SUMO homeostasis may impact SRFR1 activities. Our observation that SRFR1 interaction with EDS1 K478R , a speculated SUMOylation-deficient version, is abolished indicates biological implication of SRFR1 SIMs in interaction with positive defense regulators. The placement of a predicted SIMs (I 39 LDIC 43 ) in the first TPR motif of SRFR1 (residues 39-72, TPRpred, toolkit.tuebingen.mpg.de), a well-known platform for protein-protein interaction, introduces encouraging direction to pursue further. Precise mutagenesis of this and other predicted SIMs followed by interaction analysis with EDS1 or other known interactors (such as RPS4/6, or SNC1) is needed to be tested in vivo (Kim et al., 2010;Bhattacharjee et al., 2011). With our assays, here, we are, however, unable to convincingly demonstrate EDS1 as an in vivo SUMOylation candidate. Since only a minor pool of EDS1 interacts with FIGURE 8 | Simplified schematic representation of crosstalk between SRFR1 and SUMOylome maintenance in defense-developmental balance. Under homeostasis, SRFR1 prevents SUMOylome perturbations via its suppression of SID2/ICS1 expression that drive SA responses. This mode of negative regulation incorporates transcriptional suppression of SUMO-conjugases (such as SIZ1 and SCE1, among others), SUM3, and protein stabilities of SUMO proteases OTS1/2. Basal SUMOylome adjustments in turn reciprocates on SRFR1 function at multiple levels. As SUMOylation and/or SUMO-binding candidates, intermolecular SUMO-SIM-type associations of SRFR1 with SNC1, TCPs, and EDS1 prevent mis-primed immune activations. Similar nature of association occurring between these immune players and their downstream targets (TPRs and HDA19) regulates expression of defense-responsive genes such as DND1/2 and PR1/2, among others. SUMO1/2 or SUMO3 isoforms with their respective roles as negative or positive immune modulators balance the association stoichiometry through SRFR1 expression/stability. During a pathogen threat, transient reduction in SRFR1 initiates SUMOylome perturbations and activates defense players to execute defense signaling. A feedback signaling loop also initiated restoration of SRFR1 levels to avoid development penalties of constitutive immunity. SRFR1, in planta identification of SUMOylated EDS1 may prove challenging (Bhattacharjee et al., 2011). SNC1 or TCP8/14/15 also contains multiple predicted SUMOylation motifs. Although SUMOylated SNC1 or TCPs has been detected in vivo or in vitro, respectively, their functional relevance remains untested (Gou et al., 2017;Mazur et al., 2017). Consolidated SUMO1/3 SUMOylome changes have strong potential to interfere with SRFR1 properties at multiple levels and modulate immune signaling. A mechanistic implication into this is presented in the schematic (Figure 8).
In summary, our results here provide a unique molecular basis into SRFR1 role in the SUMO-immune balances. With regulatory influences on expression of several SUMOylationassociated genes countered by SRFR1's own stability determined by SUMO isoform crosstalk, we highlight its mediation of immune response amplitudes in plants (Ingole et al., 2021a). In addition to positive and negative immune regulators that are directly affected, SUMOylation efficiencies are also modulated by SUMO changes. This self-regulation is evident from the observation that SIZ1, SCE1, and ESD4 are themselves substrates for SUMOylation, and depending on the SUMO isoform they associate with, their localization or specificities are affected (Miller et al., 2010;Mazur et al., 2017). For example, SUMO1 bound to SCE1 SIM localizes the ternary complex (SUMO-SCE1-SIZ1) to nuclear bodies, whereas SUMO3 binding partitions it as nucleocytoplasmic. Overall, defense amplitudes therefore rely on strict harmony between SUMOylome modulators such as SRFR1, (de)SUMOylation efficiencies, localization of SUMOylation-machineries, and selection of substrates (Ingole et al., 2021a). Comprehensive elucidation of these events requires not only qualitative (i.e., which SUMO isoformmodified) but also quantitative (ratio of SUMOylated versus non-SUMOylated) evaluation of a host SUMOylome adaptation during immunity. While SUMO1/2-modified protein list is everincreasing, SUMO3-targets remain grossly underrepresented. Development of an efficient SUMO3-enrichment system, on the similar theme to Miller et al. (2010), is therefore a necessity. Equally important is the characterization of proteinprotein interactome changes that are defined by the SUMO-SIM nature. Lastly, functional intersection among SUMOmachineries, especially the SUMO-ligases/proteases, recently identified DeSIs (de-SUMOylating isopeptidases) (Orosa et al., 2018), STUbLs (SUMO-targeted ubiquitin ligases; Elrouby et al., 2013), and PIALs (E4-type SUMO ligases) (Tomanov et al., 2014), and their cognate substrates are warranted to decipher the net impact on immune signaling. Toward this endeavor, we present SRFR1 role in SUMOylome regulations and the srfr1-4 mutant as a promising system to pursue these investigations.

DATA AVAILABILITY STATEMENT
The datasets presented in this study can be found in online repositories. The names of the repository/repositories and accession number(s) can be found below: PRIDE Archive, accession no: PXD026117.

AUTHOR CONTRIBUTIONS
SB and WG conceived the research. MK, KI, SR, and SB designed the research. SR generated T91R SUMO1 clone. MK and KI generated other clones and plants used here and performed the experiments. MK, KI, and SB analyzed the data. MK, WG, and SB wrote the manuscript. All authors contributed to the article and approved the submitted version.