The yeast two-hybrid constructs received from Hybrigenics Services (Evry-Courcouronne, France) were used to switch the prey and the bait clones by using SfiI restriction sites to validate the MORC1-MIP interaction. Genomic MED9 amplified from BAC T7N22 was digested with XhoI and ligated into the pBKS plasmid to generate pBKS-gMED9. A Flag-tag was introduced into pBKS-gMED9 using Cas9 Nuclease from New England Biolabs (NEB, Ipswhich, MA, USA) to create pBSK-gMED9-Flag. The gMED9-Flag fragment was then excised using XhoI and ligated into the pPZP-RCS2-hptII vector using SalI to generate pPZP-gMED9-Flag. The pER8-Myc-MORC1 construct was described previously (Kang et al., 2012). The coding sequence of the MIP genes were cloned into pER8 tagged at the 3′ end with triple HA (pET-HA) by using AscI sites and transformed into Agrobacterium tumefaciens strain GV2260 using electroporation.

pB27 bait and pP6 prey plasmids were transformed into Y187 and L40 yeast strains, respectively. This yeast two-hybrid system uses the HIS3 reporter gene, which encodes imidazoleglycerol-phosphate dehydratase, an enzyme required for the synthesis of the amino acid histidine (His). In the yeast two-hybrid system, the interaction between the bait and prey proteins leads to the reconstitution of a functional transcription factor, which activates the expression of the HIS3 reporter gene. Minimal media containing His was used to demonstrate growth, while media lacking His was used to test the interaction between the prey and bait proteins. A competitive inhibitor of His synthesis, 3-amino-1,2,4-triazole (3-AT), was used to evaluate the strength of the interaction. A stronger interaction leads to higher HIS3 expression, which can outcompete 3-AT. A concentration of 0.1 mM 3-AT was used to create a stringent condition for selecting strong interactions.

The growing of the Agrobacterium culture and its infiltration process were performed as described (Menke et al., 2005). pBA-HcPro was used to suppress gene silencing to maximize the transient expression. A solution of 30 μM estradiol in 0.1% Tween-20 was sprayed on the leaves to induce the expression of pER8-based constructs at 1-day post-infiltration. These plants were then grown for an additional 2 days at room temperature.

Immunoblotting analysis was performed as described using an enhanced chemiluminescence method (Kang and Klessig, 2005). Tissue samples in 2 ml tubes were ground with a ceramic bead with 4× sodium dodecyl sulfate (SDS) sample buffer containing dithiothreitol (DTT) at 60 mg/ml using a homogenizer. Once homogenized, these samples were boiled for 5 min and centrifuged at 21,100 ×g for 1 min. Five microliters of supernatant was loaded into sodium dodecyl sulfate polyacrylamide gel electrophoresis and transferred onto a polyvinylidene difluoride membrane (Immobilon, Millipore, Billerica, MA, USA) using a standard transfer buffer. The rest of the immunoblotting analysis was performed as previously described (Kang and Klessig, 2005) with the following differences: ECL2 solution (Thermo Fisher Scientific, Waltham, MA, USA) was used to detect the antibodies, and the membrane was imaged using an Azure Biosystems C600 imager (Dublin, CA, USA). Antibodies used include HRP-conjugated anti-Myc (1:5,000, Santa Cruz Biotechnology, Santa Cruz, CA, USA), HRP-conjugated anti-FLAG (1:5,000, Sigma, St. Louis, MO, USA), and HRP-conjugated anti-HA (1:10,000, Sigma).

Leaves of Arabidopsis or N. benthamiana were homogenized with a mortar and pestle in liquid nitrogen. Samples were ground again in 750 μl of extraction buffer (GTEN buffer: 150 mM NaCl, 1 mM EDTA, 25 mM Tris pH 7.5, and 10% glycerol) with 2% polyvinylpolypyrrolidone, 1× protease inhibitor cocktail (Sigma), and 0.1% Triton X-100. These protein extracts were collected and centrifuged at 5,000 ×g at 4°C for 5 min, and 0.5 ml of the supernatant was then subjected to size exclusion chromatography using Illustra NAP-5 Sephadex G-25 DNA grade columns (GE Healthcare, Pittsburgh, PA, USA) to remove salts and perform buffer exchange. Eluates were collected in 1.5 ml tubes, mixed with 20 μl of anti-mouse IgG agarose beads (Sigma), and incubated at 4°C for 2 h with gentle rotation. After removing the anti-mouse IgG agarose beads, the remaining solutions were mixed with 20 μl of agarose beads conjugated with a target antibody and incubated at 4°C overnight. These solutions were washed a minimum of six times with 1 ml of IP buffer (GTEN buffer with 0.15% NP-40 and 5 mM DTT) and mixed with 30 μl of 4× SDS sample buffer before proceeding to immunoblotting analysis.

Three point five-week-old Arabidopsis plants were syringe-infiltrated with an indicated inoculum of Pseudomonas syringae pv. tomato DC3000 (Pst) grown for 2 days at 28°C in King’s B medium with appropriate antibiotics. Imaging-based resistance assays were performed as previously described (Pujara et al., 2021) except that Pst carrying the luxCDABE operon was used.

For a conventional leaf-disc-based resistance assay in Arabidopsis, three leaf disks were collected into a single 2 ml tube and homogenized using a homogenizer for 1 min in 500 μl of 0.01% Triton X-100. The samples were incubated on a mild rotating shaker for 10 min and then homogenized again in the shaker for 1 min. Samples were vortexed briefly and serially diluted. Twenty microliters of all dilutions were then plated on Luria-Bertani, and colonies were counted after 36 h.

Three point five weeks after germination, two leaves from each plant were infected with Pst carrying AvrRpt2 at 1 × 10⁵ cfu/ml every 2 days for 3 weeks. Upon completion of the treatment, plants were imaged to check the growth and development characteristics.

Two leaves were generally collected in a single tube and homogenized in a shaker for 1 min using a bead. Their RNA was extracted using the Purelink RNA Mini kit 250 (Thermo Fisher Scientific). For quantification, RNA was converted to cDNA using SuperScript RT (Invitrogen, Carlsbad, CA, USA). The primers used in reverse transcription PCR are as follows: PR1-For (AAAACTTAGCCTGGGGTAGCGG), PR1-Rev (CCACCATTGTTACACCTCACTTT), PR2-For (ATCAAGGAGCTTAGCCTCAC), PR2-Rev (TGTAAAGAGCCACAACGTCC), PR5-For (CTCTTCCTCGTGTTCACTCAC), PR5-Rev (GAAGCACCTGGAGTCAATTC), Tip41-Like-For (GCGATTTTGGCTGAGAGTTGAT), and Tip41-Like-Rev (GGATACCCTTTCGCAGATAGAGAC). Maxima SYBR Green qPCR Master Mix (Thermo Fisher Scientific) was used for quantitative reverse transcription PCR (qRT-PCR). PCR was performed with the following temperature/time profile: 50°C for 2 min and 95°C for 10 min, followed by 40 cycles of 95°C for 25 s and 60°C for 1 min.

Generation of Myc-gCRT1 transgenic line is as described (Kang et al., 2012). med9 was tranformed with pPZP-gMED9-Flag and subsequentely crossed with Myc-gCRT1 line to generate Myc-gCRT1 gMED9-Flag in morc1/2 med9. The following mutant lines were obtained from The Arabidopsis Information Resource (TAIR): cpl3 (c-terminal phosphatase-like3, med4), SALK 051322; GNARLED (med5), SALK 014298; smc1 (structural maintenance of chromosomes1, mip8), CS87452; a trihelix transcription factor (mip9), SALK 024424; bzip29 (basic leucine zipper29, mip10), SALK 018426; a mobile domain protein (med12), SALK 066042; med9 (med13), SALK 029120.

Plants were grown in soil in a growth chamber under long-day photoperiod (16 h light, 12k lux of cool white fluorescence bulbs) at 23°C, 60% relative humidity.

A conventional yeast two-hybrid (Y2H) screening for MORC1 as a bait was performed by Hybrigenics using 87 million prey clones. The commercial service provided 14 confirmed clones that interacted with the bait, which are listed in Table 1. To validate the physical interaction between MORC1 and its interacting proteins, MIPs, we generated a targeted Y2H assay with the bait and prey vectors for MORC1 and all 14 MIPs, respectively (Fig. 1). In addition, we also generated constructs that switched the bait and the prey. MORC1, when used as bait, interacted with all the MIP clones identified in the screen, confirming the outcome provided by Hybrigenics, while MIP6 displayed autoactivation. MORC1 showed a weak interaction with itself, suggesting homodimerization. When MORC1 was switched to prey (pP6-MORC1), it interacted with a subset of the MIPs: MIP2, MIP3, MIP7, MIP11, and MIP12 (Fig. 1). When a transcription-activating domain is present in a bait vector, it generally activates without a prey vector. This transactivation was observed with MIP6 and MIP13, and to a lesser degree, with MIP7, suggesting that these MIP proteins may carry a transactivation domain. As previously stated, MIP13 (MED9) is part of the MED complex that interfaces a transcription factor with RNA polymerase II (Allen and Taatjes, 2015; Buendía-Monreal and Gillmor, 2016; Kidd et al., 2011). Thus, when used as bait, MIP13 appears to function as a coactivator and brings in a factor(s) with transactivation activity. There were several interactions shown across MIPs. MIP11, a sorting nexin (Phan et al., 2008), interacted with the highest number of MIPs, including MIP2, MIP11, MIP12, MIP13, and MIP14 (Fig. 1). MIP3 is a MORC1 homolog, MORC6. The interaction between MORC1 and MORC6 was previously reported (Liu et al., 2014), and their heterodimer was shown to maintain heterochromatin compaction (Moissiard et al., 2012). Interestingly, MORC6 did not interact with any of the MIPs, suggesting that the MIPs are likely specific to MORC1.

The MIP proteins were further examined to determine if the interaction occurs with MORC1 in planta. To this end, we transiently co-expressed MIPs and MORC1 tagged with HA and Myc, respectively, in N. benthamiana. The expressed proteins were subjected to a co-immunoprecipitation (co-IP) experiment in which an αHA antibody immunoprecipitated MIP proteins, and its co-IP of MORC1 was monitored by western analysis with an αMyc antibody. MIP4, MIP6, MIP7, MIP8, and MIP12 were not detectable via HA immunoblotting analysis (Fig. 2). MIP3 and MIP13 co-immunoprecipitated significantly more MORC1 than other weak interactors such as MIP9, MIP10, and MIP11 (Fig. 2). Note that MORC1 is known to be expressed as doublets, with the higher band suspected to result from post-translational modifications (Kang et al., 2012).

Eight T-DNA insertion mutants for selected MIPs with annotated nuclear localization, similar to MORC1, were obtained from TAIR. We used an image-based resistance assay involving Pst carrying the luminescent luxCDABE reporter gene (Fan et al., 2008) to analyze the disease resistance phenotypes in a high-throughput manner. We used virulent Pst (VirPst) for infection, which provided the resistance phenotypes for basal resistance. mip3, mip10, and mip13 were susceptible against VirPst, while mip4 and mip9 were resistant (Fig. 3). mip8, due to a homozygous lethality, was not further characterized.

We showed that a knock-out mutant of MED9 displayed compromised resistance to VirPst (Fig. 3). As we have used an imaging-based assay, we suspected the magnitude of altered resistance could have been underestimated. Thus, we performed a conventional leaf-disc method to assess bacterial resistance in med9, morc1/2, and their higher-order mutant (morc1/2 med9).

The single mutants, med9 and morc1/2, were susceptible to VirPst and avirulent Pst (AvrPst), suggesting that these components are involved in basal resistance and ETI. Unusually, morc1/2 med9 regained WT-level resistance through an unknown mechanism (Fig. 4). Restoring resistance by combining two susceptible mutants is counterintuitive, which indicates that the MORC1 and MED9 interaction is likely multifaceted, perhaps involving various differential gene regulations over time. For instance, these factors may be essential in inducing defense genes early, while more engaged in curbing the induced genes later to minimize unnecessary fitness costs.

We further characterized MED9 by tracking the expression of defense genes in med9 and morc1/2 med9 together with morc1/2 in response to AvrPst and VirPst infection. We chose three well-characterized defense genes for our targeted transcriptional analysis: PR1, PR2, and PR5 (Shah et al., 1997). Transcription dynamics of defense genes are also known to be an important predictor for successful resistance. For instance, PR1 induction was delayed in morc1/2 when examined at multiple time points (Bordiya et al., 2016) which explains the susceptible resistance phenotype against AvrPst and VirPst. Thus, we decided to use the same multiple time points for our transcriptional analysis: 0, 1, 6, 24, and 48 hpi (hours post-infection).

qRT-PCR analysis was carried out for quantitative transcriptional analysis of PR1, PR2, and PR5 in response to infection with AvrPst and VirPst; mock infection was used as a negative control. PR1 was induced similarly in response to AvrPst among WT, med9, and morc1/2 (Fig. 5A). However, the induction of PR1 was slow (see 24 hpi with AvrPst in Fig. 5A) and trended upward even at 48 hpi in morc1/2 med9. This upward trend, while to a lesser degree, was seen with PR2 and PR5 at 48 hpi in med9 and morc1/2. While it was less obvious than those responding to AvrPst, the induction rate of all the defense genes from 24 hpi to 48 hpi was steepest in morc1/2 med9 responding to VirPst. Induction of PR2 and PR5 in med9 and morc1/2 was less robust and rapid relative to WT (Fig. 5B and C), consistent with their susceptible phenotypes to AvrPst and VirPst.

Physical MORC1-MED9 interaction was established in the Y2H assay (Fig. 1). As MORC1 displays nucleus-translocation in response to biotic stress (Kang et al., 2012), it is feasible that the MORC1-MED9 interaction depends on stress. To test this possibility, a transgenic Arabidopsis line carrying Myc-MORC1 and MED9-Flag under their cognate promoters was subjected to a co-IP experiment, which confirmed physical interaction of the two proteins (Fig. 2). The MORC1-MED9 interaction was further investigated by monitoring levels of PR1, PR2, and PR5 at 6 and 24 hpi with AvrPst treatments (Fig. 6) using co-IP, in which MORC1 was precipitated and MED9 was examined for co-IP. The results demonstrated that AvrPst significantly increased the MORC1-MED9 interaction at 24 hpi. In response to AvrPst, over 99% of the inducible defense genes in Arabidopsis peak at around 6 hpi and begin curbing this induced expression at 24 hpi (Mine et al., 2018). Note that three PR genes that display a 24-h peak are rather the exceptions than the rule in an overall group of genes showing a dynamic induction pattern (Mine et al., 2018). Given the kinetics of the defense gene induction, it is unlikely that MED9 interacts with MORC1 to enhance transcription. Instead, it is conceivable that MED9 interacts with MORC1 to restrain the overexpression of defense genes. Since MED14 interacts with a co-repressor LUG to facilitate target gene repression in Arabidopsis (Gonzalez et al., 2007), MED9 could also coordinate with MORC1 to repress defense genes at later time points.

We observed that morc1/2 and med9 displayed compromised resistance to AvrPst and VirPst, while morc1/2 med9 regained WT-level resistance (Fig. 4). We further investigated this seemingly counterintuitive observation by tracking three commonly used defense genes (PR1, PR2, and PR5) to obtain a snapshot of defense signaling; this investigation revealed the slow but sustained expression of defense genes in morc1/2 med9. While regaining bacterial resistance to the level of WT through extended expression is likely beneficial to the host, it is frequently observed that excessive resistance incurs fitness-costs (Huot et al., 2014). Therefore, we hypothesized that regaining resistance in morc1/2 med9 likely incurs a fitness-penalty.

We observed little difference in the morphology of morc1/2, med9, and morc1/2 med9 at 3.5 weeks after germination relative to WT, as shown in the upper panel of Fig. 7A. These plants were then repeatedly infected with AvrPst at 1 × 10⁵ cfu/ml every other day for 3 weeks and evaluated for their growth characteristics (Fig. 7B and C). First, the total weight was significantly reduced for morc1/2 med9 when repeatedly challenged with AvrPst (Fig. 7B). We further characterized their growth and fertility characteristics: the weight and number of leaves, the weight and number of inflorescences, and the number of siliques (the long and narrow seedpods in Arabidopsis). Remarkably, only morc1/2 med9 plants showed a significant reduction in all five measurements in response to AvrPst infection relative to the mock control (Fig. 7C), indicating a severe fitness cost incurred due to recurrent infection. For instance, the total leaf weight of morc1/2 med9 was reduced by over 25% in response to AvrPst infection, while those of WT and the single mutants showed negligible changes. In particular, siliques of morc1/2 med9 were significantly reduced in number, suggesting that sexual reproduction was also affected as a part of the fitness penalty.